23886 lines
3.5 MiB
23886 lines
3.5 MiB
"project_id","number","text","issue_type"
|
|
"6013","1000","newaxis incompatible with array indexing Trac 402).
|
|
Original ticket http projects.scipy.org numpy ticket 402 on 2006 12 16 by brickZA, assigned to unknown.
|
|
Hi.
|
|
I get the following import numpy as N dofs = N.arange 5) 1 physVals = N.ones 5,4,3)) ....
|
|
In [35] physVals[0 3] dofs[[4,3,2], N.newaxis, N.newaxis] exceptions.TypeError Traceback most recent call last) home brick akademie NewCode working <ipython console> TypeError an integer is required I can get the intended result like this In [36] physVals[0 3] dofs[[4,3,2]][ , N.newaxis, N.newaxis] Out[36] array [[[ 5., 5., 5.
|
|
], [ 5., 5., 5.
|
|
], [ 5., 5., 5.
|
|
], [ 5., 5., 5.
|
|
]], [[ 4., 4., 4.
|
|
], [ 4., 4., 4.
|
|
], [ 4., 4., 4.
|
|
], [ 4., 4., 4.
|
|
]], [[ 3., 3., 3.
|
|
], [ 3., 3., 3.
|
|
], [ 3., 3., 3.
|
|
], [ 3., 3., 3.]]])","enhancement"
|
|
"6013","1001","speedup for copyswap ) Trac 403).
|
|
Original ticket http projects.scipy.org numpy ticket 403 on 2006 12 20 by astraw, assigned to unknown.
|
|
This patch, created by Francesc Altet, elicits a small but significant speedup from the copyswap ) functions.
|
|
I post it here, to the tracker, so it won t be forgotten.
|
|
See http projects.scipy.org pipermail numpy discussion 2006 December 025120.html","enhancement"
|
|
"6013","1002","numpy.fromfile on file like object Trac 404).
|
|
Original ticket http projects.scipy.org numpy ticket 404 on 2006 12 20 by trac user eisele, assigned to unknown.
|
|
I would have expected that the following steps would allow to read an array from a compressed file from os import popen import numpy f = popen zcat movieID.gz ) a = numpy.fromfile f, int32 , 1) However, I get an array of length 0 and the message zcat stdout Broken pipe Is it reasonable to assume that this could be done?
|
|
Thanks and best regards, Andreas","bug"
|
|
"6013","1003","nan functions do not preserve matrix subclass Trac 405).
|
|
Original ticket http projects.scipy.org numpy ticket 405 on 2006 12 21 by trac user svetosch, assigned to unknown.
|
|
Keith reported on the mailing list sorry don t have latest numpy installed currently to test myself but Keith is usually reliable ...) that the functions nanmin, nanmax, nanargmin, nanargmax, nansum return an array when given a matrix subclass, as opposed to virtually all other numpy functions.
|
|
due to first line being y=array a)) sven","bug"
|
|
"6013","1004","Missing docstring for nan to num Trac 406).
|
|
Original ticket http projects.scipy.org numpy ticket 406 on 2006 12 21 by nilswagner01, assigned to unknown.
|
|
help nan to num) Help on function nan to num in module numpy.lib.type check nan to num x)","bug"
|
|
"6013","1005","apply along axis fails for ndim>2 arrays Trac 407).
|
|
Original ticket http projects.scipy.org numpy ticket 407 on 2006 12 22 by trac user user 0901, assigned to unknown.
|
|
Sorry, I don t know how to submit patches and am in a time crunch, or I would....
|
|
I first noticed this when trying to use stats.stats.nanmedian which is broken because of this bug).
|
|
It fails when the input function returns a scalar for arrays larger than 2dimensions >>> a = numpy.arange 27).reshape 3,3,3)) >>> numpy.apply along axis numpy.sum,0,a) array [[45, 45, 45], [48, 48, 48], [51, 51, 51]]) Here s a fix for this, though.
|
|
In numpy lib shape base.py, the lines outarr[ind] = res and outarr[ind] = res should have the ind replaced with tuple ind) outarr[tuple ind)] = res and outarr[tuple ind)] = res Cheers!","bug"
|
|
"6013","1006","problem using chararray with argsort ) Trac 408).
|
|
Original ticket http projects.scipy.org numpy ticket 408 on 2006 12 22 by chanley, assigned to unknown.
|
|
The following works with numarray but not numpy.
|
|
NUMARRAY Example >>> from numarray import strings as chararray >>> a = chararray.array [ a , b , c ]) >>> index=a.argsort ) >>> print index [0 1 2] >>> NUMPY Example >>> from numpy import char >>> a = char.array [ a , b , c ]) >>> index =a.argsort ) Traceback most recent call last) File <stdin> , line 1, in <module> File data sparty1 dev site packages lib python numpy core defchararray.py , line 43, in array finalize raise ValueError, Can only create a chararray from string data.
|
|
ValueError Can only create a chararray from string data.
|
|
>>> >>> import numpy >>> numpy.
|
|
version 1.0.2.dev3487 >>>","bug"
|
|
"6013","1007","svn trunk build doesn t pass numpy.test ) Trac 409).
|
|
Original ticket http projects.scipy.org numpy ticket 409 on 2006 12 24 by trac user darren.weber, assigned to unknown.
|
|
.............................
|
|
File usr lib python2.4 site packages numpy lib tests test ufunclike.py , line 20, in test ufunclike Failed example U.isposinf a) Expected array [True, False, False, False, False, False], dtype=bool) Got array [ True, False, False, False, False, False], dtype=bool) It looks like all the array [True],dtype=bool) values are returning True instead of True values.","bug"
|
|
"6013","1008","setmember1d problem Trac 410).
|
|
Original ticket http projects.scipy.org numpy ticket 410 on 2007 01 02 by trac user curiousjan, assigned to unknown.
|
|
Please see the attached scripts.
|
|
I am simply comparing two arrays, storing items that appear in both lists.
|
|
I do this by a) storing all items from list 1 that are also in list 2 b) storing all items from list 2 that are also in list 1 The two resulting arrays should be identical, but they are not.
|
|
A simple fix is svn diff numpy lib Index numpy lib arraysetops.py =================================================================== numpy lib arraysetops.py revision 3491) numpy lib arraysetops.py working copy) 126,7 126,7 zlike = nm.zeros like ar = nm.concatenate ar1, ar2 ) ) tt = nm.concatenate zlike ar1 ), zlike ar2 ) 1) ) perm = ar.argsort ) perm = ar.argsort kind= mergesort ) aux = ar[perm] aux2 = tt[perm] flag = ediff1d aux, 1 ) == 0","bug"
|
|
"6013","1009","indices function too slow Trac 411).
|
|
Original ticket http projects.scipy.org numpy ticket 411 on 2007 01 07 by trac user auger, assigned to unknown.
|
|
The indices function is quite slow 12s to create indices for a 4000x4000 array on my machine).
|
|
Here is a version that is between 4 and 8 times faster depending on the array sizes).
|
|
def indices shape,dtype=int) lshape = [] oshape = [len shape)] for i in shape lshape.append i) oshape.append i) out = empty tuple oshape)) dims = len lshape) for i in range dims) v = lshape.pop i) lshape.append 1) subshape = tuple lshape) coords = arange shape[i]) tiled = tile coords,subshape) ndims = tiled.ndim tindx = arange ndims) tindx[tindx>i] = 1 tindx[i] = 1 lshape.pop ) lshape.insert i,v) out[i] = tiled.transpose tindx) return out.astype dtype)","enhancement"
|
|
"6013","1010","segault in numpy.float64 z) for complex z Trac 412).
|
|
Original ticket http projects.scipy.org numpy ticket 412 on 2007 01 08 by timleslie, assigned to unknown.
|
|
Passing a complex value to N.float or N.int raises a type error as expected.
|
|
In [1] import numpy as N In [2] N.float 2.5 2.5j) exceptions.TypeError Traceback most recent call last) home timl src scipy Lib sparse sparsetools <ipython console> TypeError can t convert complex to float; use abs z) Passing to any of int32, int64, float32 or float64 I havn t tested other types) results in a segfault.
|
|
In [6] N.float64 2.5 2.5j) Segmentation fault core dumped) This is with In [2] numpy.
|
|
version Out[2] 1.0.2.dev3497 And crashes on my macbook and also my 64bit linux box.","bug"
|
|
"6013","1011","Reduction with noncommutative ufuncs is broken Trac 413).
|
|
Original ticket http projects.scipy.org numpy ticket 413 on 2007 01 08 by rkern, assigned to teoliphant.
|
|
In [21] import numpy In [22] numpy.subtract.reduce numpy.arange 5)) Out[22] 2 In [23] import Numeric In [24] Numeric.subtract.reduce Numeric.arange 5)) Out[24] 10 In [25] print numpy.
|
|
version 1.0.2.dev3493 In [26] print Numeric.
|
|
version 24.2","bug"
|
|
"6013","1012","bug in reduce for numpy.array Trac 414).
|
|
Original ticket http projects.scipy.org numpy ticket 414 on 2007 01 08 by trac user lbolla, assigned to unknown.
|
|
[[BR]] Wrong results are given by subtract.reduce and divide.reduce, with respect to the standard function reduce.
|
|
Take a look at this piece of code from numpy import x = arange 4) x = array [0,1,2,3]) def myadd x,y) re define the binary sum function return x y print reduce myadd, x) 6, as expected print add.reduce x) 6, as expected def mysub x,y) re define the binary diff function return x y print reduce mysub, x) 6, as expected print subtract.reduce x) 2 > WRONG!
|
|
It probably depends on the wrong order of the operands in the binary operation subtract or divide.
|
|
Any non commutative operation can lead to this problem and it s worth checking!).
|
|
This code works correctly if the Numeric package is imported istead of numpy first row of the snippet).","bug"
|
|
"6013","1013","update numpy testing docstring Trac 415).
|
|
Original ticket http projects.scipy.org numpy ticket 415 on 2007 01 08 by jarrodmillman, assigned to unknown.
|
|
The docstring in `numpy testing info.py` still calls itself the Scipy Test tools.
|
|
All occurrences of Scipy should be replaced with Numpy.","bug"
|
|
"6013","1014","The docstrings of functions which take an out= parameter should mention that fact Trac 416).
|
|
Original ticket http projects.scipy.org numpy ticket 416 on 2007 01 09 by rkern, assigned to rkern.
|
|
Most of the functions that take an `out` optional argument fail to mention that they do so.
|
|
Each of them ought to do so.","bug"
|
|
"6013","1015","Numpy 1.0.1 compilation fails on IRIX 6.5 Trac 417).
|
|
Original ticket http projects.scipy.org numpy ticket 417 on 2007 01 10 by trac user goddard, assigned to cournape.
|
|
Compiling numpy 1.0.1 on SGI IRIX 6.5.24m fails with the error shown below.
|
|
The trouble is that it is compiling with the n32 new 32 ABI) but it is also explicitly adding L usr lib to the link line and that system directory contains incompatible o32 libraries.
|
|
The usr lib directory was added because it found BLAS libraries in that directory.
|
|
After working around that problem by changing default lib dirs to [] in numpy 1.0.1 numpy distutils system info.py I get the C code errors shown in the second error message below undefined symbols expm1l, rintf, ..., non constant initializers, ...).
|
|
These problems are using SGI compilers not gcc).
|
|
Tom python2.4 setup.py install ... C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc n32 configtest.o L usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib L usr local lib L usr lib o configtest ld32 FATAL 12 Expecting n32 objects usr lib crt1.o is o32.
|
|
ld32 FATAL 12 Expecting n32 objects usr lib crt1.o is o32.
|
|
failure.
|
|
removing configtest.c configtest.o Traceback most recent call last) File setup.py , line 89, in ?
|
|
setup package ) File setup.py , line 82, in setup package configuration=configuration ) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils core.py , line 174, in setup return old setup new attr) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils core.py , line 149, in setup dist.run commands ) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils dist.py , line 946, in run commands self.run command cmd) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils command install.py , line 16, in run r = old install.run self) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils command install.py , line 506, in run self.run command build ) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils command build.py , line 112, in run self.run command cmd name) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils command build src.py , line 87, in run self.build sources ) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils command build src.py , line 106, in build sources self.build extension sources ext) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils command build src.py , line 212, in build extension sources sources = self.generate sources sources, ext) File usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils command build src.py , line 270, in generate sources source = func extension, build dir) File numpy core setup.py , line 50, in generate config h raise ERROR Failed to test configuration ERROR Failed to test configuration After setting default lib dirs = [] get several C code errors python2.4 setup.py install Running from numpy source directory.
|
|
Appending numpy.distutils.command configuration to numpy.distutils Ignoring attempt to set name from numpy.distutils to numpy.distutils.command ) Appending numpy.distutils.fcompiler configuration to numpy.distutils Ignoring attempt to set name from numpy.distutils to numpy.distutils.fcompiler ) Appending numpy.distutils configuration to numpy Ignoring attempt to set name from numpy to numpy.distutils ) Appending numpy.testing configuration to numpy Ignoring attempt to set name from numpy to numpy.testing ) Appending numpy.f2py.lib.parser configuration to numpy.f2py.lib Ignoring attempt to set name from numpy.f2py.lib to numpy.f2py.lib.parser ) Appending numpy.f2py.lib configuration to numpy.f2py Ignoring attempt to set name from numpy.f2py to numpy.f2py.lib ) F2PY Version 2 3473 Appending numpy.f2py configuration to numpy Ignoring attempt to set name from numpy to numpy.f2py ) Ignoring attempt to set version from None to 2 3473 ) blas opt info blas mkl info libraries mkl,vml,guide not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE atlas blas threads info Setting PTATLAS=ATLAS libraries lapack,blas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE atlas blas info libraries lapack,blas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1302 UserWarning Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) blas info libraries blas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1311 UserWarning Blas http www.netlib.org blas ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [blas]) or by setting the BLAS environment variable.
|
|
warnings.warn BlasNotFoundError.
|
|
doc ) blas src info NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1314 UserWarning Blas http www.netlib.org blas ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [blas src]) or by setting the BLAS SRC environment variable.
|
|
warnings.warn BlasSrcNotFoundError.
|
|
doc ) NOT AVAILABLE Appending numpy.core configuration to numpy Ignoring attempt to set name from numpy to numpy.core ) Appending numpy.lib configuration to numpy Ignoring attempt to set name from numpy to numpy.lib ) Appending numpy.oldnumeric configuration to numpy Ignoring attempt to set name from numpy to numpy.oldnumeric ) Appending numpy.numarray configuration to numpy Ignoring attempt to set name from numpy to numpy.numarray ) Appending numpy.fft configuration to numpy Ignoring attempt to set name from numpy to numpy.fft ) lapack opt info lapack mkl info mkl info libraries mkl,vml,guide not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE NOT AVAILABLE atlas threads info Setting PTATLAS=ATLAS libraries lapack,blas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib libraries lapack atlas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib numpy.distutils.system info.atlas threads info NOT AVAILABLE atlas info libraries lapack,blas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib libraries lapack atlas not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib numpy.distutils.system info.atlas info NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1211 UserWarning Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) lapack info libraries lapack not found in usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1222 UserWarning Lapack http www.netlib.org lapack ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [lapack]) or by setting the LAPACK environment variable.
|
|
warnings.warn LapackNotFoundError.
|
|
doc ) lapack src info NOT AVAILABLE usr local src staff goddard chimera foreign numpy numpy 1.0.1 numpy distutils system info.py 1225 UserWarning Lapack http www.netlib.org lapack ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [lapack src]) or by setting the LAPACK SRC environment variable.
|
|
warnings.warn LapackSrcNotFoundError.
|
|
doc ) NOT AVAILABLE Appending numpy.linalg configuration to numpy Ignoring attempt to set name from numpy to numpy.linalg ) Appending numpy.random configuration to numpy Ignoring attempt to set name from numpy to numpy.random ) Appending numpy configuration to Ignoring attempt to set name from to numpy ) running install running build running config fc running build src building py modules sources building extension numpy.core.multiarray sources Generating build src.irix64 6.5 2.4 numpy core config.h customize MipsFCompiler customize MipsFCompiler customize MipsFCompiler using config C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc n32 configtest.o L usr local src staff chimera build IRIX X11 foreign Python 2.4.3 lib o configtest configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc n32 configtest.o o configtest configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
expl; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
expf; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
log1p; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
expm1; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
asinh; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier atanhf is undefined.
|
|
atanhf; 1 error detected in the compilation of configtest.c .
|
|
cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier atanhf is undefined.
|
|
atanhf; 1 error detected in the compilation of configtest.c .
|
|
failure.
|
|
removing configtest.c configtest.o C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier atanhl is undefined.
|
|
atanhl; 1 error detected in the compilation of configtest.c .
|
|
cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier atanhl is undefined.
|
|
atanhl; 1 error detected in the compilation of configtest.c .
|
|
failure.
|
|
removing configtest.c configtest.o C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
isnan; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier isinf is undefined.
|
|
isinf; 1 error detected in the compilation of configtest.c .
|
|
cc 1020 cc ERROR File = configtest.c, Line = 4 The identifier isinf is undefined.
|
|
isinf; 1 error detected in the compilation of configtest.c .
|
|
failure.
|
|
removing configtest.c configtest.o C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1171 cc WARNING File = configtest.c, Line = 4 The indicated expression has no effect.
|
|
rint; cc n32 configtest.o o configtest success!
|
|
removing configtest.c configtest.o configtest adding build src.irix64 6.5 2.4 numpy core config.h to sources.
|
|
executing numpy core code generators generate array api.py adding build src.irix64 6.5 2.4 numpy core multiarray api.h to sources.
|
|
creating build src.irix64 6.5 2.4 numpy core src conv template > build src.irix64 6.5 2.4 numpy core src scalartypes.inc adding build src.irix64 6.5 2.4 numpy core src to include dirs.
|
|
conv template > build src.irix64 6.5 2.4 numpy core src arraytypes.inc numpy.core nothing done with h files= [ build src.irix64 6.5 2.4 numpy core src scalartypes.inc , build src.irix64 6.5 2.4 numpy core src arraytypes.inc , build src.irix64 6.5 2.4 numpy core config.h , build src.irix64 6.5 2.4 numpy core multiarray api.h ] building extension numpy.core.umath sources adding build src.irix64 6.5 2.4 numpy core config.h to sources.
|
|
executing numpy core code generators generate ufunc api.py adding build src.irix64 6.5 2.4 numpy core ufunc api.h to sources.
|
|
conv template > build src.irix64 6.5 2.4 numpy core src umathmodule.c adding build src.irix64 6.5 2.4 numpy core src to include dirs.
|
|
numpy.core nothing done with h files= [ build src.irix64 6.5 2.4 numpy core src scalartypes.inc , build src.irix64 6.5 2.4 numpy core src arraytypes.inc , build src.irix64 6.5 2.4 numpy core config.h , build src.irix64 6.5 2.4 numpy core ufunc api.h ] building extension numpy.core.
|
|
sort sources adding build src.irix64 6.5 2.4 numpy core config.h to sources.
|
|
executing numpy core code generators generate array api.py adding build src.irix64 6.5 2.4 numpy core multiarray api.h to sources.
|
|
conv template > build src.irix64 6.5 2.4 numpy core src sortmodule.c numpy.core nothing done with h files= [ build src.irix64 6.5 2.4 numpy core config.h , build src.irix64 6.5 2.4 numpy core multiarray api.h ] building extension numpy.core.scalarmath sources adding build src.irix64 6.5 2.4 numpy core config.h to sources.
|
|
executing numpy core code generators generate array api.py adding build src.irix64 6.5 2.4 numpy core multiarray api.h to sources.
|
|
executing numpy core code generators generate ufunc api.py adding build src.irix64 6.5 2.4 numpy core ufunc api.h to sources.
|
|
conv template > build src.irix64 6.5 2.4 numpy core src scalarmathmodule.c numpy.core nothing done with h files= [ build src.irix64 6.5 2.4 numpy core config.h , build src.irix64 6.5 2.4 numpy core multiarray api.h , build src.irix64 6.5 2.4 numpy core ufunc api.h ] building extension numpy.core.
|
|
dotblas sources building extension numpy.lib.
|
|
compiled base sources building extension numpy.numarray.
|
|
capi sources building extension numpy.fft.fftpack lite sources building extension numpy.linalg.lapack lite sources creating build src.irix64 6.5 2.4 numpy linalg Warning Using unoptimized lapack adding numpy linalg lapack litemodule.c to sources.
|
|
adding numpy linalg zlapack lite.c to sources.
|
|
adding numpy linalg dlapack lite.c to sources.
|
|
adding numpy linalg blas lite.c to sources.
|
|
adding numpy linalg dlamch.c to sources.
|
|
adding numpy linalg f2c lite.c to sources.
|
|
building extension numpy.random.mtrand sources creating build src.irix64 6.5 2.4 numpy random C compiler cc n32 OPT Olimit=0 DNDEBUG O compile options Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc configtest.c cc 1035 cc WARNING File = configtest.c, Line = 7 error directive No WIN32 error No WIN32 cc n32 configtest.o o configtest configtest failure.
|
|
removing configtest.c configtest.o configtest building data files sources running build py creating build lib.irix64 6.5 2.4 creating build lib.irix64 6.5 2.4 numpy copying numpy init .py > build lib.irix64 6.5 2.4 numpy copying numpy import tools.py > build lib.irix64 6.5 2.4 numpy copying numpy add newdocs.py > build lib.irix64 6.5 2.4 numpy copying numpy ctypeslib.py > build lib.irix64 6.5 2.4 numpy copying numpy dual.py > build lib.irix64 6.5 2.4 numpy copying numpy matlib.py > build lib.irix64 6.5 2.4 numpy copying numpy setup.py > build lib.irix64 6.5 2.4 numpy copying numpy version.py > build lib.irix64 6.5 2.4 numpy copying build src.irix64 6.5 2.4 numpy config .py > build lib.irix64 6.5 2.4 numpy creating build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils init .py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils version .py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils ccompiler.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils conv template.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils core.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils cpuinfo.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils exec command.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils extension.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils from template.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils info.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils intelccompiler.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils interactive.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils log.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils lib2def.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils line endings.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils mingw32ccompiler.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils misc util.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils setup.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils system info.py > build lib.irix64 6.5 2.4 numpy distutils copying numpy distutils unixccompiler.py > build lib.irix64 6.5 2.4 numpy distutils copying build src.irix64 6.5 2.4 numpy distutils config .py > build lib.irix64 6.5 2.4 numpy distutils creating build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command init .py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command bdist rpm.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build clib.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build ext.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build py.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build scripts.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command build src.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command config.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command config compiler.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command egg info.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command install.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command install data.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command install headers.py > build lib.irix64 6.5 2.4 numpy distutils command copying numpy distutils command sdist.py > build lib.irix64 6.5 2.4 numpy distutils command creating build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler init .py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler absoft.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler compaq.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler g95.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler gnu.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler hpux.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler ibm.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler intel.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler lahey.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler mips.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler nag.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler none.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler pg.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler sun.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler copying numpy distutils fcompiler vast.py > build lib.irix64 6.5 2.4 numpy distutils fcompiler creating build lib.irix64 6.5 2.4 numpy testing copying numpy testing init .py > build lib.irix64 6.5 2.4 numpy testing copying numpy testing info.py > build lib.irix64 6.5 2.4 numpy testing copying numpy testing numpytest.py > build lib.irix64 6.5 2.4 numpy testing copying numpy testing setup.py > build lib.irix64 6.5 2.4 numpy testing copying numpy testing utils.py > build lib.irix64 6.5 2.4 numpy testing creating build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py init .py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py svn version .py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py version .py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py auxfuncs.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py capi maps.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py cb rules.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py cfuncs.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py common rules.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py crackfortran.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py diagnose.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py f2py2e.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py f2py testing.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py f90mod rules.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py func2subr.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py info.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py rules.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py setup.py > build lib.irix64 6.5 2.4 numpy f2py copying numpy f2py use rules.py > build lib.irix64 6.5 2.4 numpy f2py creating build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib init .py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib api.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib main.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib py wrap.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib py wrap subprogram.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib py wrap type.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib setup.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib test derived scalar.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib test module module.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib test module scalar.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib test scalar function in.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib test scalar in out.py > build lib.irix64 6.5 2.4 numpy f2py lib copying numpy f2py lib wrapper base.py > build lib.irix64 6.5 2.4 numpy f2py lib creating build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser Fortran2003.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser init .py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser api.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser base classes.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser block statements.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser parsefortran.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser pattern tools.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser readfortran.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser sourceinfo.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser splitline.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser statements.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser test Fortran2003.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser test parser.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser utils.py > build lib.irix64 6.5 2.4 numpy f2py lib parser copying numpy f2py lib parser typedecl statements.py > build lib.irix64 6.5 2.4 numpy f2py lib parser creating build lib.irix64 6.5 2.4 numpy core copying numpy core init .py > build lib.irix64 6.5 2.4 numpy core copying numpy core svn version .py > build lib.irix64 6.5 2.4 numpy core copying numpy core internal.py > build lib.irix64 6.5 2.4 numpy core copying numpy core arrayprint.py > build lib.irix64 6.5 2.4 numpy core copying numpy core defchararray.py > build lib.irix64 6.5 2.4 numpy core copying numpy core defmatrix.py > build lib.irix64 6.5 2.4 numpy core copying numpy core fromnumeric.py > build lib.irix64 6.5 2.4 numpy core copying numpy core info.py > build lib.irix64 6.5 2.4 numpy core copying numpy core ma.py > build lib.irix64 6.5 2.4 numpy core copying numpy core memmap.py > build lib.irix64 6.5 2.4 numpy core copying numpy core numeric.py > build lib.irix64 6.5 2.4 numpy core copying numpy core numerictypes.py > build lib.irix64 6.5 2.4 numpy core copying numpy core records.py > build lib.irix64 6.5 2.4 numpy core copying numpy core setup.py > build lib.irix64 6.5 2.4 numpy core copying numpy core code generators generate array api.py > build lib.irix64 6.5 2.4 numpy core creating build lib.irix64 6.5 2.4 numpy lib copying numpy lib init .py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib arraysetops.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib convdtype.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib function base.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib getlimits.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib index tricks.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib info.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib machar.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib polynomial.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib scimath.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib setup.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib shape base.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib twodim base.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib type check.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib ufunclike.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib user array.py > build lib.irix64 6.5 2.4 numpy lib copying numpy lib utils.py > build lib.irix64 6.5 2.4 numpy lib creating build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric init .py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric alter code1.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric alter code2.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric array printer.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric compat.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric fft.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric fix default axis.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric functions.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric linear algebra.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric ma.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric matrix.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric misc.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric mlab.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric precision.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric random array.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric rng.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric rng stats.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric setup.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric typeconv.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric ufuncs.py > build lib.irix64 6.5 2.4 numpy oldnumeric copying numpy oldnumeric user array.py > build lib.irix64 6.5 2.4 numpy oldnumeric creating build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray init .py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray alter code1.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray alter code2.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray compat.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray convolve.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray fft.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray functions.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray image.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray linear algebra.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray ma.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray matrix.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray mlab.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray nd image.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray numerictypes.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray random array.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray session.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray setup.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray ufuncs.py > build lib.irix64 6.5 2.4 numpy numarray copying numpy numarray util.py > build lib.irix64 6.5 2.4 numpy numarray creating build lib.irix64 6.5 2.4 numpy fft copying numpy fft init .py > build lib.irix64 6.5 2.4 numpy fft copying numpy fft fftpack.py > build lib.irix64 6.5 2.4 numpy fft copying numpy fft helper.py > build lib.irix64 6.5 2.4 numpy fft copying numpy fft info.py > build lib.irix64 6.5 2.4 numpy fft copying numpy fft setup.py > build lib.irix64 6.5 2.4 numpy fft creating build lib.irix64 6.5 2.4 numpy linalg copying numpy linalg init .py > build lib.irix64 6.5 2.4 numpy linalg copying numpy linalg info.py > build lib.irix64 6.5 2.4 numpy linalg copying numpy linalg linalg.py > build lib.irix64 6.5 2.4 numpy linalg copying numpy linalg setup.py > build lib.irix64 6.5 2.4 numpy linalg creating build lib.irix64 6.5 2.4 numpy random copying numpy random init .py > build lib.irix64 6.5 2.4 numpy random copying numpy random info.py > build lib.irix64 6.5 2.4 numpy random copying numpy random setup.py > build lib.irix64 6.5 2.4 numpy random running build ext customize UnixCCompiler customize UnixCCompiler using build ext building numpy.core.multiarray extension compiling C sources C compiler cc n32 OPT Olimit=0 DNDEBUG O creating build temp.irix64 6.5 2.4 creating build temp.irix64 6.5 2.4 numpy creating build temp.irix64 6.5 2.4 numpy core creating build temp.irix64 6.5 2.4 numpy core src compile options Ibuild src.irix64 6.5 2.4 numpy core src Inumpy core include Ibuild src.irix64 6.5 2.4 numpy core Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc numpy core src multiarraymodule.c cc 1552 cc WARNING File = numpy core src scalartypes.inc.src, Line = 1123 The variable numbytes is set but never used.
|
|
int numbytes; cc 1178 cc WARNING File = numpy core src arraytypes.inc.src, Line = 885 Argument is incompatible with the corresponding format string conversion.
|
|
num = fscanf fp, LONGLONG FMT, ip); cc 1178 cc WARNING File = numpy core src arraytypes.inc.src, Line = 885 Argument is incompatible with the corresponding format string conversion.
|
|
num = fscanf fp, ULONGLONG FMT, ip); cc 1185 cc WARNING File = numpy core src arrayobject.c, Line = 1279 An enumerated type is mixed with another type.
|
|
fortran = PyArray ISFORTRAN m1); cc 1185 cc WARNING File = numpy core src arrayobject.c, Line = 1853 An enumerated type is mixed with another type.
|
|
order = PyArray ISFORTRAN self); cc 1515 cc WARNING File = numpy core src arrayobject.c, Line = 4433 A value of type void cannot be assigned to an entity of type int ) void , void , int, int) .
|
|
cmpfunc = func; cc 1185 cc WARNING File = numpy core src arrayobject.c, Line = 4792 An enumerated type is mixed with another type.
|
|
temp1 = PyArray Ravel arr,0); cc 1185 cc WARNING File = numpy core src arraymethods.c, Line = 364 An enumerated type is mixed with another type.
|
|
if ret = PyArrayObject )PyArray NewCopy self, 1)) == NULL) cc 1185 cc WARNING File = numpy core src arraymethods.c, Line = 654 An enumerated type is mixed with another type.
|
|
obj = ARET PyArray NewCopy self,0)); cc 1515 cc WARNING File = numpy core src arrayobject.c, Line = 5416 A value of type void cannot be assigned to an entity of type PyArray FinalizeFunc .
|
|
cfunc = PyCObject AsVoidPtr func); cc 1515 cc WARNING File = numpy core src arrayobject.c, Line = 7446 A value of type void cannot be assigned to an entity of type PyArray VectorUnaryFunc .
|
|
castfunc = PyCObject AsVoidPtr cobj); cc 1552 cc WARNING File = numpy core src arrayobject.c, Line = 9267 The variable typecode is set but never used.
|
|
PyArray Descr typecode; cc 1185 cc WARNING File = numpy core src arrayobject.c, Line = 9526 An enumerated type is mixed with another type.
|
|
return PyArray Flatten it >ao, 0); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 221 An enumerated type is mixed with another type.
|
|
fortran = PyArray ISFORTRAN a); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 381 An enumerated type is mixed with another type.
|
|
order = PyArray ISFORTRAN a); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 632 An enumerated type is mixed with another type.
|
|
fortran = PyArray ISFORTRAN self); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 1156 An enumerated type is mixed with another type.
|
|
new = PyArray NewCopy self, 1); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 1542 An enumerated type is mixed with another type.
|
|
otmp = PyArray Ravel mps[i],0); cc 1185 cc WARNING File = numpy core src multiarraymodule.c, Line = 1873 An enumerated type is mixed with another type.
|
|
retval = descr >f >scalarkind arr ?
|
|
arr NULL)); ld n32 shared all build temp.irix64 6.5 2.4 numpy core src multiarraymodule.o o build lib.irix64 6.5 2.4 numpy core multiarray.so ld32 WARNING 47 This module build temp.irix64 6.5 2.4 numpy core src multiarraymodule.o .text) contains branch instruction s) that might degrade performance on an older version rev.
|
|
2.2) R4000 processor.
|
|
building numpy.core.umath extension compiling C sources C compiler cc n32 OPT Olimit=0 DNDEBUG O creating build temp.irix64 6.5 2.4 build creating build temp.irix64 6.5 2.4 build src.irix64 6.5 2.4 creating build temp.irix64 6.5 2.4 build src.irix64 6.5 2.4 numpy creating build temp.irix64 6.5 2.4 build src.irix64 6.5 2.4 numpy core creating build temp.irix64 6.5 2.4 build src.irix64 6.5 2.4 numpy core src compile options Ibuild src.irix64 6.5 2.4 numpy core src Inumpy core include Ibuild src.irix64 6.5 2.4 numpy core Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c cc build src.irix64 6.5 2.4 numpy core src umathmodule.c cc 1020 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 66 The identifier expm1l is undefined.
|
|
static void expm1 data[] = { void )expm1f, void )expm1, void )expm1l, void )nc expm1f, void )nc expm1, void )nc expm1l, void ) expm1 }; cc 1041 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 66 The indicated expression must have arithmetic or pointer type.
|
|
static void expm1 data[] = { void )expm1f, void )expm1, void )expm1l, void )nc expm1f, void )nc expm1, void )nc expm1l, void ) expm1 }; cc 1020 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 165 The identifier rintf is undefined.
|
|
static void rint data[] = { void )rintf, void )rint, void )rintl, void )nc rintf, void )nc rint, void )nc rintl, void ) rint }; cc 1041 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 165 The indicated expression must have arithmetic or pointer type.
|
|
static void rint data[] = { void )rintf, void )rint, void )rintl, void )nc rintf, void )nc rint, void )nc rintl, void ) rint }; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; 7 errors detected in the compilation of build src.irix64 6.5 2.4 numpy core src umathmodule.c .
|
|
cc 1020 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 66 The identifier expm1l is undefined.
|
|
static void expm1 data[] = { void )expm1f, void )expm1, void )expm1l, void )nc expm1f, void )nc expm1, void )nc expm1l, void ) expm1 }; cc 1041 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 66 The indicated expression must have arithmetic or pointer type.
|
|
static void expm1 data[] = { void )expm1f, void )expm1, void )expm1l, void )nc expm1f, void )nc expm1, void )nc expm1l, void ) expm1 }; cc 1020 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 165 The identifier rintf is undefined.
|
|
static void rint data[] = { void )rintf, void )rint, void )rintl, void )nc rintf, void )nc rint, void )nc rintl, void ) rint }; cc 1041 cc ERROR File = build src.irix64 6.5 2.4 numpy core umath generated.c, Line = 165 The indicated expression must have arithmetic or pointer type.
|
|
static void rint data[] = { void )rintf, void )rint, void )rintl, void )nc rintf, void )nc rint, void )nc rintl, void ) rint }; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; cc 1028 cc ERROR File = numpy core src ufuncobject.c, Line = 2071 The expression used must have a constant value.
|
|
int arg types[3] = {otype, otype, otype}; 7 errors detected in the compilation of build src.irix64 6.5 2.4 numpy core src umathmodule.c .
|
|
error Command cc n32 OPT Olimit=0 DNDEBUG O Ibuild src.irix64 6.5 2.4 numpy core src Inumpy core include Ibuild src.irix64 6.5 2.4 numpy core Inumpy core src Inumpy core include I usr local src staff chimera build IRIX X11 foreign Python 2.4.3 include python2.4 c build src.irix64 6.5 2.4 numpy core src umathmodule.c o build temp.irix64 6.5 2.4 build src.irix64 6.5 2.4 numpy core src umathmodule.o failed with exit status 2","bug"
|
|
"6013","1016","Segfault dotblas.c Trac 418).
|
|
Original ticket http projects.scipy.org numpy ticket 418 on 2007 01 11 by nilswagner01, assigned to unknown.
|
|
Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 46912509653888 LWP 30502)] dotblas matrixproduct dummy=<value optimized out>, args=<value optimized out>) at dotblas.c 233 233 Py DECREF ap1);","bug"
|
|
"6013","1017","Make web page links to NumPy bug report ticket) system Trac 419).
|
|
Original ticket http projects.scipy.org numpy ticket 419 on 2007 01 11 by trac user goddard, assigned to jarrodmillman.
|
|
The [http projects.scipy.org scipy numpy newticket NumPy bug reporting system] is hard to find on the web and hard to use.
|
|
This ticket requests some changes to simplify bug reporting for new users.
|
|
Could you add a link to the [http numpy.scipy.org NumPy project web page] that goes to the bug reporting page [http projects.scipy.org scipy numpy developer project page])?
|
|
Since the Trac ticket system will be unfamiliar to most users it would be helpful to make a separate bug reporting page that says a bug report is called a ticket .
|
|
It should have a search form for searching tickets, a link to view a list of active tickets, and an explanation saying you must login to create a new ticket bug report and have a link to register.
|
|
Currently you don t even see the New Ticket tab unless you have logged in.
|
|
The [http sourceforge.net projects numpy sourceforge NumPy page] has bug tracking for numarray, no numarray downloads, and no numpy bug tracking.
|
|
This is needlessly confusing.
|
|
Could numarray bug tracking be moved to a separate numarray sourceforge project?
|
|
Could the sourceforge bug tracker summary page have a sentence saying bug tracking is handled at scipy.org and have a link to that bug tracking page?","bug"
|
|
"6013","1018","Add numpy.floatc for C float type Trac 420).
|
|
Original ticket http projects.scipy.org numpy ticket 420 on 2007 01 11 by trac user goddard, assigned to unknown.
|
|
I d like numpy.floatc added as a synonym for the existing numpy.single.
|
|
This type represents the native C floating point type for the machine.
|
|
The name single reduces code readability as it is not even clear that it is a data type.
|
|
The proposed name floatc parallels the existing numpy.intc which is the C int type.
|
|
C or C code modules that operate on numpy arrays often require C float arrays.
|
|
So when working with these modules it is often required to create numpy arrays of type floatc in Python.
|
|
Currently we use from numpy import single as floatc and then use this floatc name in creating the arrays in Python.
|
|
Direct use of the name single makes the code far less readable.","enhancement"
|
|
"6013","1019","{{{ScipyTest}}} should be deprecated.
|
|
Trac 421).
|
|
Original ticket http projects.scipy.org numpy ticket 421 on 2007 01 12 by jarrodmillman, assigned to timleslie.
|
|
`ScipyTest` should be deprecated.
|
|
Currently, `ScipyTest = NumpyTest` in `numpy testing numpytest.py`.
|
|
Instead we should have `ScipyTest` be a small wrapper, which raises a deprecation warning and then calls `NumpyTest`.
|
|
Something like class ScipyTest NumpyTest) def init self, package=None) warnings.warn ScipyTest is now called NumpyTest; please update your code , DeprecationWarning) NumpyTest.
|
|
init self, package)","bug"
|
|
"6013","1020","getitem method of MaskedArray returns ndarray in some cases Trac 422).
|
|
Original ticket http projects.scipy.org numpy ticket 422 on 2007 01 12 by trac user reggie, assigned to unknown.
|
|
When using extended slicing on a MaskedArray with its mask set to nomask, the ` getitem ` method called because of the extended slicing) will return only the data portion of the MaskedArray thereby causing the result of your slicing operation to be an ndarray instead of a MaskedArray.
|
|
I m attaching a patch that seems to fix this problem and also eliminates the ` getslice ` and ` setslice ` methods as I believe they re only needed in Python versions prior to 2.0 which aren t supported by numpy, if I m not mistaken.","bug"
|
|
"6013","1021","clip segfaults when using out argument with non expected datatype Trac 423).
|
|
Original ticket http projects.scipy.org numpy ticket 423 on 2007 01 15 by cournape, assigned to unknown.
|
|
The clip function segfaults when using a out argument which does not have the expected type import numpy as N a = N.linspace 0, 10, 100).astype numpy.int32) m = N.int32 0) M = N.int32 1) works a.clip m, M) b = N.zeros a.shape, dtype = N.int32) works a.clip m, M, out = b) b = N.zeros a.shape, dtype = N.int64) always segfaults a.clip m, M, out = b) This is with a recent SVN numpy 1.0.2.dev3510 ).","bug"
|
|
"6013","1022","stable argsort ) and sort ) behave differently on 2 column uint64 arrays.
|
|
possibly a bug in argsort ) or sort ) .
|
|
Trac 424).
|
|
Original ticket http projects.scipy.org numpy ticket 424 on 2007 01 16 by trac user ciro, assigned to unknown.
|
|
numpy 1.0.1 fails to pass the following in a reproducible way.
|
|
!
|
|
usr bin env python2.4 from numpy import N = 500000 X = zeros N,2), uint64) for i in range N) X[i,0] = random.randint 2 20) << 43 if we set highest bit problems X[i,0] = random.randint 2 20) << 42 this works, instead X2 = X[ ,0].copy ) X2.sort kind= mergesort ) X index = X[ ,0].argsort kind= mergesort ) X[ ,0] = take X[ ,0], X index) print alltrue X2 == X[ ,0]) Conversely, the following always works N = 500000 X = zeros N, uint64) for i in range N) X[i] = random.randint 2 20) << 43 if we set highest bit, ok in this case X[i] = random.randint 2 20) << 42 this works X2 = X.copy ) X2.sort kind= mergesort ) X index = X.argsort kind= mergesort ) X = take X, X index) print alltrue X2 == X) behavior seems inconsistent across architectures and numpy versions.","bug"
|
|
"6013","1023","Faster clip for scalar clipping values Trac 425).
|
|
Original ticket http projects.scipy.org numpy ticket 425 on 2007 01 18 by cournape, assigned to teoliphant.
|
|
Here is a patch against numpy svn 3511 for faster clip.
|
|
It adds a function fastclip to numpy.core a function PyArray FastClip to the numpy C Api which has the exact same semantics than PyArray Clip.
|
|
various tests to check the results are exactly the same than current clip implementation.
|
|
I was not sure where to add the new functions in the code, so I kepts most change in separate files when possible.
|
|
It still uses the current implementation for some cases I hope to implement soon.
|
|
The improvements are significant for the cases where scalar clipping values min max) are used 5 to 30 times faster); it can uses the optional out argument, including for in place.
|
|
It also works in some cases where the current clip does not work or segfaults see ticket 422).
|
|
If the patch is considered acceptable, the C function can be a drop in replacement for the current PyArray Clip so that no change to the numpy C api or python API are necessary.
|
|
Any suggestions are welcomed of course, David","enhancement"
|
|
"6013","1024","give ctypes.data as a reference to the array Trac 426).
|
|
Original ticket http projects.scipy.org numpy ticket 426 on 2007 01 18 by trac user martin wiechert, assigned to unknown.
|
|
Is there a way to let a ctypes.POINTER have a reference to a python object?
|
|
If not, could the ctypes maintainer be talked into providing this facility?)
|
|
If so, the pointer returned by ndarray.ctypes.data as ctypes.POINTER ctypes.c <type>)) should have a reference to the array.
|
|
This would help avoid bugs like data pointer = a b).ctypes.data as ctypes.POINTER ctypes.c double)) c library function data pointer) which passes a stale pointer to c library function.
|
|
If this is not feasible, at least a warning in the numpy documentation might be helpful.","enhancement"
|
|
"6013","1025","numpy.random.permutation cannot handle numpy integer scalars Trac 427).
|
|
Original ticket http projects.scipy.org numpy ticket 427 on 2007 01 18 by rkern, assigned to rkern.
|
|
As reported by Emanuele Olivetti Look at this bug.py import numpy a=numpy.array [1,2]) b=a.sum ) print type b) c=numpy.random.permutation b) If I run it Python 2.5, numpy 1.0.1 on a Linux box) I get > python tmp bug.py <type numpy.int32 > Traceback most recent call last) File tmp bug.py , line 5, in <module> c=numpy.random.permutation b) File mtrand.pyx , line 1227, in mtrand.RandomState.permutation File mtrand.pyx , line 1211, in mtrand.RandomState.shuffle TypeError len ) of unsized object We are currently using `type x) is int` to distinguish integers from arrays.
|
|
We should test for `isinstance x, int) or isinstance numpy.integer)`, I believe.
|
|
Possibly, we should also check for `hasattr x, index )`.","bug"
|
|
"6013","1026","remove old SciPy references Trac 428).
|
|
Original ticket http projects.scipy.org numpy ticket 428 on 2007 01 21 by jarrodmillman, assigned to unknown.
|
|
There are still some old references to !SciPy, which need to be updated to !NumPy 1.
|
|
`numpy testing numpytest.py` `class SciPyTextTestResult unittest.
|
|
TextTestResult) ` `class SciPyTextTestRunner unittest.TextTestRunner) ` ` return SciPyTextTestResult self.stream, self.descriptions, self.verbosity)` 1.
|
|
`numpy testing setup.py` ` setup maintainer = SciPy Developers ,` ` description = SciPy test module ,` ` license = SciPy License BSD Style) ,` 1.
|
|
`numpy distutils cpuinfo.py` `terms of the SciPy BSD style) license.
|
|
See LICENSE.txt that came with` 1.
|
|
`numpy distutils system info.py ` `terms of the SciPy BSD style) license.
|
|
See LICENSE.txt that came with`","bug"
|
|
"6013","1027","typo in doc DISTUTILS.txt Trac 429).
|
|
Original ticket http projects.scipy.org numpy ticket 429 on 2007 01 21 by jarrodmillman, assigned to unknown.
|
|
.
|
|
should be replaced with .
|
|
at the end of the sentence, which reads Note that all classes that are inherited from ``TestCase`` class, are picked up by the test runner when using ``testoob``.","bug"
|
|
"6013","1028","numpy fails with PyWin build 210 PyWin32 210.win32 py2.5.exe) Trac 430).
|
|
Original ticket http projects.scipy.org numpy ticket 430 on 2007 01 22 by trac user rstockford, assigned to unknown.
|
|
I have installed Py 2.5 for Windows with PyWin as above.
|
|
Everything works as before, except that numpy causes PyWin to crash PyWin32 has encountered a problem and needs to close... ).
|
|
The details are module umath.pyd offset 0001d310 code 0xc0000091 address 0x000000006485d310 numpy was installed with no errors and works fine with Idle it can be imported and used.
|
|
Any sugestions?
|
|
Richard Stockford","bug"
|
|
"6013","1029","default x11 lib dirs should contain 64 bit library paths Trac 431).
|
|
Original ticket http projects.scipy.org numpy ticket 431 on 2007 01 22 by lebedov, assigned to dmcooke.
|
|
The list of X11 library directories in `numpy distutils system info.py` default x11 lib dirs) ought to include those directories unique to 64 bit systems e.g., ` usr lib64`, ` usr X11R6 lib64`).","bug"
|
|
"6013","1030","cov segfaults when input is oriented incorrectly Trac 432).
|
|
Original ticket http projects.scipy.org numpy ticket 432 on 2007 01 23 by huard, assigned to unknown.
|
|
r = random.randn 1000,2) cov r) Crashes the shell.
|
|
Ubuntu Dapper, 1.0.1.dev3423.
|
|
cov r.T) works fine.","bug"
|
|
"6013","1031","longdouble on G5 OS X operates incorrectly Trac 433).
|
|
Original ticket http projects.scipy.org numpy ticket 433 on 2007 01 24 by rkern, assigned to unknown.
|
|
As reported by Vincent Nijs Mac OS X 10.4.8 G5 PPC Numpy svn 2602 Target powerpc apple darwin8 Configured with private var tmp gcc gcc 5367.obj 1 src configure disable checking enable werror prefix= usr mandir= share man enable languages=c,objc,c ,obj c program transform name= [cg][ . ]
|
|
s 4.0 with gxx include dir= include c 4.0.0 with slibdir= usr lib build=powerpc apple darwin8 host=powerpc apple darwin8 target=powerpc apple darwin8 Thread model posix gcc version 4.0.1 Apple Computer, Inc. build 5367) ====================================================================== FAIL check large types numpy.core.tests.test scalarmath.test power) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packag es numpy core tests test scalarmath.py , line 46, in check large types assert b == 6765201, error with r got r t,b) AssertionError error with <type numpy.float128 > got 0.0 ====================================================================== FAIL check types numpy.core.tests.test scalarmath.test types) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packag es numpy core tests test scalarmath.py , line 19, in check types assert a == 1, error with r got r atype,a) AssertionError error with <type numpy.float128 > got 1.0 Ran 526 tests in 1.133s FAILED failures=2)","bug"
|
|
"6013","1032","Incomplete docstring linalg.eigh Trac 434).
|
|
Original ticket http projects.scipy.org numpy ticket 434 on 2007 01 24 by nilswagner01, assigned to unknown.
|
|
Help on function eigh in module numpy.linalg.linalg eigh a, UPLO= L ) Compute eigenvalues for a Hermitian symmetric matrix.
|
|
The eigenvectors are missing.
|
|
There is no docstring for eigvalsh.","enhancement"
|
|
"6013","1033","numpy.mean ) accumulator default type should not be single precision Trac 435).
|
|
Original ticket http projects.scipy.org numpy ticket 435 on 2007 01 24 by chanley, assigned to unknown.
|
|
The accumulator used in the mean algorithm should not be single precision by default.
|
|
This default can cause unexpected results.
|
|
Please see the following example In [5] a.dtype Out[5] dtype >f4 ) In [6] print a [[ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]
|
|
[ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]
|
|
[ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]
|
|
..., [ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]
|
|
[ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]
|
|
[ 132.
|
|
132.
|
|
132.
|
|
..., 132.
|
|
132.
|
|
132.]]
|
|
In [7] a.min ) Out[7] 132.0 In [8] a.max ) Out[8] 389.0 In [9] a.mean ) Out[9] 129.742439153 However, if you recast the array as float64 you get the correct result In [11] a.astype numpy.float64).mean ) Out[11] 132.062805059 I believe that double precision would be a more appropriate default type for the accumulator.","bug"
|
|
"6013","1034","a[ 1].tofile f) creates junk Trac 436).
|
|
Original ticket http projects.scipy.org numpy ticket 436 on 2007 01 27 by sebhaase, assigned to unknown.
|
|
Hi!
|
|
`arr.tofile f)` creates garbage in `f`.
|
|
`f` in my case is an open file object at position 1024 that s the length of our file format s header) N. version 1.0.1 It cannot determine what exactly goes wrong the file size seems to be varying even if `arr` always has the same size my `arr` has `dtype=uint16`) Thanks, Sebastian Haase","bug"
|
|
"6013","1035","Updates to f2py callbacks desired Trac 437).
|
|
Original ticket http projects.scipy.org numpy ticket 437 on 2007 01 29 by trac user mesmith, assigned to pearu.
|
|
I have a couple of changes I have made to f2py related to call backs that I would like to have included.
|
|
I have been using these changes for over a year with no adverse side affects.
|
|
Two attached patch files have the changes.","enhancement"
|
|
"6013","1036","extension modules on x86 64 fail, march is set to i686 Trac 438).
|
|
Original ticket http projects.scipy.org numpy ticket 438 on 2007 01 29 by tovrstra, assigned to dmcooke.
|
|
When trying to compile this fortran example module into a python module MODULE test IMPLICIT NONE CONTAINS FUNCTION add a,b) REAL 8),INTENT IN) a,b REAL 8) add add = a b END FUNCTION END MODULE with f2py f2py c test.f90 on my x86 64 box, I get the following output toon dropklop f2py c test.f90 running build running config fc running build src building extension untitled sources f2py options [] f2py > tmp tmpus7JnZ src.linux x86 64 2.4 untitledmodule.c creating tmp tmpus7JnZ creating tmp tmpus7JnZ src.linux x86 64 2.4 Reading fortran codes... Reading file test.f90 format free) Post processing... Block untitled Block test Block add Post processing stage 2)... Block untitled Block unknown interface Block test Block add Building modules... Building module untitled ... Constructing F90 module support for test ...
|
|
Creating wrapper for Fortran function add add )... Constructing wrapper function test.add ... add = add a,b) Wrote C API module untitled to file tmp tmpus7JnZ src.linux x86 64 2.4 untitledmodule.c Fortran 90 wrappers are saved to tmp tmpus7JnZ src.linux x86 64 2.4 untitled f2pywrappers2.f90 adding tmp tmpus7JnZ src.linux x86 64 2.4 fortranobject.c to sources.
|
|
adding tmp tmpus7JnZ src.linux x86 64 2.4 to include dirs.
|
|
copying usr lib python2.4 site packages numpy f2py src fortranobject.c > tmp tmpus7JnZ src.linux x86 64 2.4 copying usr lib python2.4 site packages numpy f2py src fortranobject.h > tmp tmpus7JnZ src.linux x86 64 2.4 adding tmp tmpus7JnZ src.linux x86 64 2.4 untitled f2pywrappers2.f90 to sources.
|
|
running build ext customize UnixCCompiler customize UnixCCompiler using build ext Could not locate executable g77 Could not locate executable f77 customize GnuFCompiler Could not locate executable ifort Could not locate executable ifc Could not locate executable ifort Could not locate executable efort Could not locate executable efc Could not locate executable ifort Could not locate executable efort Could not locate executable efc customize IntelFCompiler customize LaheyFCompiler customize PGroupFCompiler customize AbsoftFCompiler customize NAGFCompiler customize VastFCompiler customize GnuFCompiler customize CompaqFCompiler customize IntelItaniumFCompiler customize IntelEM64TFCompiler customize Gnu95FCompiler customize Gnu95FCompiler customize Gnu95FCompiler using build ext building untitled extension compiling C sources C compiler gcc pthread fno strict aliasing DNDEBUG g O2 Wall Wstrict prototypes fPIC creating tmp tmpus7JnZ tmp creating tmp tmpus7JnZ tmp tmpus7JnZ creating tmp tmpus7JnZ tmp tmpus7JnZ src.linux x86 64 2.4 compile options I tmp tmpus7JnZ src.linux x86 64 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c gcc tmp tmpus7JnZ src.linux x86 64 2.4 untitledmodule.c gcc tmp tmpus7JnZ src.linux x86 64 2.4 fortranobject.c compiling Fortran 90 module sources Fortran f77 compiler usr bin gfortran Wall ffixed form fno second underscore fPIC O3 funroll loops march=i686 mmmx msse2 msse fomit frame pointer Fortran f90 compiler usr bin gfortran Wall fno second underscore fPIC O3 funroll loops march=i686 mmmx msse2 msse fomit frame pointer Fortran fix compiler usr bin gfortran Wall ffixed form fno second underscore Wall fno second underscore fPIC O3 funroll loops march=i686 mmmx msse2 msse fomit frame pointer compile options I tmp tmpus7JnZ src.linux x86 64 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c extra options J tmp tmpus7JnZ I tmp tmpus7JnZ gfortran f90 test.f90 test.f90 0 error CPU you selected does not support x86 64 instruction set test.f90 0 error CPU you selected does not support x86 64 instruction set test.f90 0 error CPU you selected does not support x86 64 instruction set test.f90 0 error CPU you selected does not support x86 64 instruction set error Command usr bin gfortran Wall fno second underscore fPIC O3 funroll loops march=i686 mmmx msse2 msse fomit frame pointer I tmp tmpus7JnZ src.linux x86 64 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c c test.f90 o tmp tmpus7JnZ test.o J tmp tmpus7JnZ I tmp tmpus7JnZ failed with exit status 1 It seems that f2py is not correctly recognizing the cpu architecture.
|
|
i.e.
|
|
the compiler option march=i686 is wrong.
|
|
As a workaround, one can simply disable the architecture optimization FARCH= f2py c test.f90","bug"
|
|
"6013","1037","poly1d objects can not be pickled Trac 439).
|
|
Original ticket http projects.scipy.org numpy ticket 439 on 2007 01 30 by tovrstra, assigned to unknown.
|
|
Due to a bug in the ` getattr ` method of the `poly1d` class, instances of this class can not be pickled.
|
|
The current version of the method is like this in file `numpy lib polynomial.py`) def getattr self, key) if key in [ r , roots ] return roots self.coeffs) elif key in [ c , coef , coefficients ] return self.coeffs elif key in [ o ] return self.order else return self.
|
|
dict [key] while it should be something like this def getattr self, key) if key in [ r , roots ] return roots self.coeffs) elif key in [ c , coef , coefficients ] return self.coeffs elif key in [ o ] return self.order elif key not in self.
|
|
dict raise AttributeError s has not attribute s self.
|
|
class , key)) else return self.
|
|
dict [key] In the original version a `KeyError` is raised when one tries to access a non existing attribute, while an `AttributeError` should be raised.
|
|
This confuses the pickle process since it tries to access the method ` getstate ` as an attribute and it expects an `AttributeError` to be raised when this method does not exist.
|
|
I have tested the new version and it works.","bug"
|
|
"6013","1038","Inconsistent behaviour for empty arrays Trac 440).
|
|
Original ticket http projects.scipy.org numpy ticket 440 on 2007 01 30 by stefanv, assigned to stefanv.
|
|
Functions returning floating point values based on array input, like mean, average and median should behave in a consistent fashion.
|
|
Current situation mean []) > nan average []) > ZeroDivisionError median []) > IndexError","bug"
|
|
"6013","1039","numpy from iter taking shape rather than count Trac 441).
|
|
Original ticket http projects.scipy.org numpy ticket 441 on 2007 02 01 by trac user tdennist, assigned to unknown.
|
|
It would be great if the fromiter function in numpy took a shape argument rather than a count argument and interpretted its input as n dimensional iterator of iterators of iterators etc) of values instead of a 1 dimensional iterator of values.
|
|
Right now to add two dimension values, for instance you need to use a custom dtype which gives you a recarray like structure, which is not always what you want.
|
|
I ve looked at the code.
|
|
It doesn t look too involved but would require someone more adept with reference counting than myself.
|
|
Tom","enhancement"
|
|
"6013","1040","Dirichlet random generator Trac 442).
|
|
Original ticket http projects.scipy.org numpy ticket 442 on 2007 02 01 by cournape, assigned to unknown.
|
|
I attached a patch for a Dirichlet random generator http en.wikipedia.org wiki Dirichlet distribution","enhancement"
|
|
"6013","1041","ERROR check complex bad numpy.lib.tests.test type check.test nan to num) Trac 443).
|
|
Original ticket http projects.scipy.org numpy ticket 443 on 2007 02 02 by nilswagner01, assigned to unknown.
|
|
====================================================================== ERROR check complex bad numpy.lib.tests.test type check.test nan to num) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test type check.py , line 245, in check complex bad vals = nan to num v) File usr lib64 python2.4 site packages numpy lib type check.py , line 132, in nan to num are inf = isposinf y) File usr lib64 python2.4 site packages numpy lib ufunclike.py , line 33, in isposinf umath.logical and isinf x), signbit x), y) TypeError function not supported for these types, and can t coerce safely to supported types ====================================================================== ERROR check complex bad2 numpy.lib.tests.test type check.test nan to num) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test type check.py , line 253, in check complex bad2 vals = nan to num v) File usr lib64 python2.4 site packages numpy lib type check.py , line 132, in nan to num are inf = isposinf y) File usr lib64 python2.4 site packages numpy lib ufunclike.py , line 33, in isposinf umath.logical and isinf x), signbit x), y) TypeError function not supported for these types, and can t coerce safely to supported types ====================================================================== ERROR check complex good numpy.lib.tests.test type check.test nan to num) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test type check.py , line 238, in check complex good vals = nan to num 1 1j) File usr lib64 python2.4 site packages numpy lib type check.py , line 132, in nan to num are inf = isposinf y) File usr lib64 python2.4 site packages numpy lib ufunclike.py , line 33, in isposinf umath.logical and isinf x), signbit x), y) TypeError function not supported for these types, and can t coerce safely to supported types Ran 526 tests in 0.518s FAILED errors=3) <unittest.TextTestRunner object at 0x2aaaae34ffd0> >>> numpy.
|
|
version 1.0.2.dev3534","bug"
|
|
"6013","1042","from numpy.numarray import fails with AttributeError module object has no attribute NewAxis Trac 444).
|
|
Original ticket http projects.scipy.org numpy ticket 444 on 2007 02 03 by trac user jachaves, assigned to unknown.
|
|
from numpy.numarray import fails with AttributeError module object has no attribute NewAxis because file numpy numarray init .py is missing the following line from compat import","bug"
|
|
"6013","1043","normed histogram returns incorrect results Trac 445).
|
|
Original ticket http projects.scipy.org numpy ticket 445 on 2007 02 05 by trac user cvwright, assigned to rkern.
|
|
The histogram ) function, when called with normed=True, returns incorrect results for the bins frequencies.
|
|
The frequencies should sum to 1.0 in order to be a real probability distribution, but usually with the current code they do not.
|
|
Looking at the current version of the code in trac, I believe the problem is on lines 105 and 106 of function base.py.
|
|
On line 105, we calculate the width of a bin, and on line 106, we divide the counts for each bin by the size of the input times the bin width.
|
|
I think the problem could be fixed by simply removing the bin width calculation and instead dividing the bin counts by only the number of samples in the input.","bug"
|
|
"6013","1044","numpy.linalg.qr bug on 64 bit platforms Trac 446).
|
|
Original ticket http projects.scipy.org numpy ticket 446 on 2007 02 07 by fperez, assigned to unknown.
|
|
Hi all, I recently got a report of a bug triggered only on 64 bit hardware, and on a machine in case it s relevant) that runs python 2.5 In [1] import numpy In [2] numpy.
|
|
version Out[2] 1.0.2.dev3540 In [3] a = numpy.array [[1.0,2],[3,4]]) In [4] numpy.linalg.qr a) On entry to DGEQRF parameter number 2 had an illegal value sage[ ]> dumped back at system prompt On my 2.4 using, 32 bit desktop, this code runs just fine.
|
|
A quick grep exercise in the sources shows this as the probable source in numpy linalg dlapack lite.c, this code is the one that seems to matter around line 4920) Function Body info = 0; nb = ilaenv c 1, DGEQRF , , m, n, c n1, c n1, ftnlen)6, ftnlen) 1); lwkopt = n nb; work[1] = doublereal) lwkopt; lquery = lwork == 1; if m < 0) { info = 1; } else if n < 0) { info = 2; } else if lda < max 1, m)) { info = 4; } else if lwork < max 1, n) !
|
|
lquery)) { info = 7; } if info != 0) { i 1 = info); xerbla DGEQRF , i 1); return 0; } else if lquery) { return 0; } I m not sure why n should have come out negative while m is fine in this case though.","bug"
|
|
"6013","1045","Missing docstrings in linalg.lapack lite Trac 447).
|
|
Original ticket http projects.scipy.org numpy ticket 447 on 2007 02 07 by nilswagner01, assigned to unknown.
|
|
help linalg.lapack lite.dgeqrf) yields Help on built in function dgeqrf in module numpy.linalg.lapack lite dgeqrf ...) which is uninformative.","enhancement"
|
|
"6013","1046","memory error in polymul Trac 448).
|
|
Original ticket http projects.scipy.org numpy ticket 448 on 2007 02 09 by stefanv, assigned to unknown.
|
|
Reproduce by N.polymul [],[1.])
|
|
Fixed in r3543.","bug"
|
|
"6013","1047","Swaping byteorder in recarrays with multidimensional fields doesn t work Trac 449).
|
|
Original ticket http projects.scipy.org numpy ticket 449 on 2007 02 12 by FrancescAlted, assigned to unknown.
|
|
When swaping the byteorder of a recarray with multidimensional fields, those are not swapped at all as can be seen in In [45] r=numpy.array [ 1, 0,1,2))], dtype= i2,3i2 ) In [46] r.byteswap ) Out[46] array [ 256, [0, 1, 2])], dtype=[ f0 , <i2 ), f1 , <i2 , 3)]) Curiously, this seems to work fine with nested recarrays without multidimensional columns, of course) In [47] mydescriptor = numpy.dtype [ x , f4 ), y , f4 ), nested , [ i , i2 ), j , i2 )])]) In [48] nra=numpy.array [ 1.0, 2.0, 1,2)), 2.1, 3.2, 3,2))], dtype=mydescriptor) In [49] nra.byteswap ) Out[49] array [ 4.6006029882248069e 41, 8.9683101716788293e 44, 256, 512)), 2.7156489580261977e 23, 428443648.0, 768, 512))], dtype=[ x , <f4 ), y , <f4 ), nested , [ i , <i2 ), j , <i2 )])])","bug"
|
|
"6013","1048","Make a.min ) not copy data Trac 450).
|
|
Original ticket http projects.scipy.org numpy ticket 450 on 2007 02 14 by trac user goddard, assigned to unknown.
|
|
I would like the min max array methods not to copy the array.
|
|
This limits the ability to handle very large data sets our case is 3 D electron microscopy).
|
|
Here s an example >>> import numpy as n >>> a = n.zeros 1024, 1024, 1400), n.int8) >>> a.min ) 0 >>> b = a[1 ,1 ,1 ] >>> b.min ) Traceback most recent call last) File <pyshell 509> , line 1, in ?
|
|
b.min ) MemoryError","bug"
|
|
"6013","1049","easy no autodetect for numpy distutils Trac 451).
|
|
Original ticket http projects.scipy.org numpy ticket 451 on 2007 02 14 by trac user bicatali, assigned to dmcooke.
|
|
numpy distutils automatically find cpu, installed compilers and libraries on the system.
|
|
Although nice for users, packagers have and like their own system for options dependencies and choosing the right options for various architectures.
|
|
Supporting all possible configurations could rapidly become complex for numpy devs.
|
|
Looking around at packaging numpy scipy in the nix land, a lot of distributions actually patch system info.py or make use of a complex set of both environment, patches and setup.cfg file.
|
|
[[BR]] Ex the system has atlas, acml, netlib lapack, ifort and gfortran, threaded or not.
|
|
Configuring numpy with netlib lapack and ifort even though others are installed can be tricky to package.
|
|
[[BR]] A simple no autodetect option called with setup could be available, letting the nix package system autodetect via dependencies.
|
|
A setup.cfg file with all default values explained would be helpful for packagers.
|
|
As far as I can see on open tickets, a lot of problems could also be avoided with such a simple feature.","enhancement"
|
|
"6013","1050",".tolist ) for void scalar type is returning a tuple Trac 452).
|
|
Original ticket http projects.scipy.org numpy ticket 452 on 2007 02 14 by FrancescAlted, assigned to unknown.
|
|
This exposes the problem In [37] a=numpy.array 1,2, asa ), i2,i2,a3 ) In [38] a.tolist ) Out[38] 1, 2, asa ) In [39] a.item ) Out[39] 1, 2, asa ) In [40] type a.tolist )) Out[40] <type tuple > In [41] type a.item )) Out[41] <type tuple > .tolist ) for void scalar type is returning a tuple, and provided that .item ) already returns a tuple, perhaps it would nice if .tolist ) could return a list instead.","bug"
|
|
"6013","1051","coercion rules for boolean arrays are broken Trac 453).
|
|
Original ticket http projects.scipy.org numpy ticket 453 on 2007 02 16 by chanley, assigned to unknown.
|
|
In Python boolean multiplication results in the following In [1] True 4096 Out[1] 4096 NUMARRAY is consistent with what you get from Python In [2] import numarray In [3] a = numarray.ones 10,type=numarray.Bool) In [4] print a [1 1 1 1 1 1 1 1 1 1] In [5] a 4096 Out[5] array [4096, 4096, 4096, 4096, 4096, 4096, 4096, 4096, 4096, 4096]) However, numpy gives the following results In [6] import numpy In [7] b = numpy.ones 10,dtype=numpy.bool ) In [8] print b [ True True True True True True True True True True] In [9] b 4096 Out[9] array [0, 0, 0, 0, 0, 0, 0, 0, 0, 0], dtype=int8) The NUMPY coercion rules are not even self consistent.
|
|
If you extract an element from array b to get a rank zero array you get the same behavior as Python In [10] print b[0] True In [11] b[0] 4096 Out[11] 4096","bug"
|
|
"6013","1052","Importing numpy prevents decrementing references for local variables Trac 454).
|
|
Original ticket http projects.scipy.org numpy ticket 454 on 2007 02 16 by trac user goddard, assigned to unknown.
|
|
If the first import of numpy is within a function it prevents that function s local variables and all locals in higher call frames that led to the import from being deleted.
|
|
This can prevent freeing up large numpy arrays.
|
|
The problem is that numpy init .py imports numpy import tools.py and creates a PackageLoader object that saves the parent frame self.parent frame = frame = sys.
|
|
getframe 1) Saving the parent frame results in saving the entire call stack even after all those functions have returned.
|
|
I ve attached Python code illustrating the problem.","bug"
|
|
"6013","1053","histogram2d crashes if all points are outliers Trac 455).
|
|
Original ticket http projects.scipy.org numpy ticket 455 on 2007 02 17 by trac user npadmana, assigned to unknown.
|
|
histogram2d crashes if all the points are outliers the issue appears to be that the xy array is empty, resulting in bincount returning a very long array which exceeds the dimensions of hist.
|
|
I think simply protecting that piece of code with an if xy.size > 0 should solve this problem.","bug"
|
|
"6013","1054","Inconsistent behaviour of representation of numpy scalar strings when NULLs exists Trac 456).
|
|
Original ticket http projects.scipy.org numpy ticket 456 on 2007 02 22 by FrancescAlted, assigned to unknown.
|
|
Hi, I think the next behaviour is somewhat inconsistent >>> numpy.array a\x00\x0b\x0c\x00 ) array a , dtype= |S5 ) i.e.
|
|
the representation of scalar strings truncates when the first NULL is encountered.
|
|
This is not the default behaviour in Python, that accepts NULLs as part of the representation >>> s= a\x00b\x0c\x00 >>> s a\x00b\x0c\x00 Furthermore, it seems that some conversion methods works well in this case >>> numpy.array a\x00b\x0c\x00 ).tostring ) a\x00b\x0c\x00 But others seems to be broken >>> numpy.array a\x00b\x0c\x00 ).item ) a","bug"
|
|
"6013","1055","searchsorted docstring patch Trac 457).
|
|
Original ticket http projects.scipy.org numpy ticket 457 on 2007 02 28 by astraw, assigned to unknown.
|
|
The attached patch fixes a docstring bug.
|
|
The following demonstrates that the output array is the shape of v, not a as described in the docstring.
|
|
>>> import numpy >>> >>> a=numpy.arange 20) >>> v=numpy.array [2,3,6,3,5]) >>> >>> result = numpy.searchsorted a,v) >>> print result.shape 5,)","bug"
|
|
"6013","1056","Compilation fails on Cygwin Trac 458).
|
|
Original ticket http projects.scipy.org numpy ticket 458 on 2007 03 01 by mdehoon, assigned to teoliphant.
|
|
Compiling NumPy 1.0.1 with Python 2.5 failed on Cygwin with the following error ... gcc shared Wl, enable auto image base build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o L usr local lib python2.5 config lpython2.5 o build lib.cyg win 1.5.23 i686 2.5 numpy numarray capi.dll build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA checkF PErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA clearF PErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA checkA ndReportFPErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept collect2 ld returned 1 exit status build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA checkF PErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA clearF PErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept build temp.cygwin 1.5.23 i686 2.5 numpy numarray capi.o In function `NA checkA ndReportFPErrors home mdehoon numpy 1.0.1 numpy numarray capi.c 2947 undefined reference to ` fetestexcept home mdehoon numpy 1.0.1 numpy numarray capi.c 2954 undefined reference to ` feclearexcept collect2 ld returned 1 exit status error Command gcc shared Wl, enable auto image base build temp.cygwin 1.5.2 3 i686 2.5 numpy numarray capi.o L usr local lib python2.5 config lpython2.5 o build lib.cygwin 1.5.23 i686 2.5 numpy numarray capi.dll failed with exit s tatus 1","bug"
|
|
"6013","1057","select does not preserve dtype for 1 element case Trac 459).
|
|
Original ticket http projects.scipy.org numpy ticket 459 on 2007 03 03 by trac user batripler, assigned to unknown.
|
|
in lib function base.py def select 421 if type S) in ScalarType or max asarray S).shape)==1 422 pfac = asarray 1) 423 for k in range n2 1) 424 pfac = pfac asarray choicelist[k]) 425 S = S ones asarray pfac).shape) the last line should preserve dtype 425 S = S ones asarray pfac).shape, S.dtype) An example that works is n=5; numpy.select [numpy.ones n)>200],[numpy.ones n)]) An example that does not currently work, and should n=1; numpy.select [numpy.ones n)>200],[numpy.ones n)]) The patch is attached.","bug"
|
|
"6013","1058","better integration of matrices using properties Trac 460).
|
|
Original ticket http projects.scipy.org numpy ticket 460 on 2007 03 04 by trac user batripler, assigned to unknown.
|
|
I think that adding the following properties on arrays matrices would greatly simplify code [[BR]] ndarray.M to return asmatrix self)[[BR]] ndarray.A to return self[[BR]] matrix.M to return self[[BR]] Thus, typing foo.M or foo.A will always work.
|
|
At the very least, adding merely the first property would be a great help to shorten code, and improve clarity.","enhancement"
|
|
"6013","1059","std stdp Trac 461).
|
|
Original ticket http projects.scipy.org numpy ticket 461 on 2007 03 04 by trac user batripler, assigned to unknown.
|
|
The current definition of .std is non standard it returns the population standard deviation dividing by n), rather than the unbiased sample estimate dividing by n 1).
|
|
Moreover, it is an inconsistent naming versus other packages [[BR]] Matlab std X) vs. std X,1)[[BR]] Excel STDEV vs. STDEVP [[BR]] Mathematica StandardDeviation[[BR]] Could the current .std be renamed .stdp and a new .std method do the right thing ?
|
|
Or, could an optional parameter be added?","enhancement"
|
|
"6013","1060","numpy.rec parser problem Trac 462).
|
|
Original ticket http projects.scipy.org numpy ticket 462 on 2007 03 06 by chanley, assigned to unknown.
|
|
The format parser in the records module in numpy.core cannot handle a repeat count of zero.
|
|
This feature is actually important in pyfits so that a user can define a field that has zero length.
|
|
The difference between the numpy and numarray behavior is illustrated below In [10] import numarray.records as r In [11] x = r.array None,formats= a8,3f8,0f8,i4,f4,a1,f4,3f4,a1,f4,3f4 , shape=28) In [12] from numpy import rec In [13] y = rec.array None,formats= a8,3f8,0f8,i4,f4,a1,f4,3f4,a1,f4,3f4 , shape=28) <type exceptions.TypeError > Traceback most recent call last) data sparty1 dev devCode <ipython console> in <module> ) data sparty1 dev site packages lib python numpy core records.py in array obj, dtype, shape, offset, strides, formats, names, titles, aligned, byteorder, copy) 516 elif formats is not None 517 dtype = format parser formats, names, titles, > 518 aligned, byteorder).
|
|
descr 519 else 520 kwds = { formats formats, data sparty1 dev site packages lib python numpy core records.py in init self, formats, names, titles, aligned, byteorder) 46 self.
|
|
parseFormats formats, aligned) 47 self.
|
|
setfieldnames names, titles) > 48 self.
|
|
createdescr byteorder) 49 50 def parseFormats self, formats, aligned=0) data sparty1 dev site packages lib python numpy core records.py in createdescr self, byteorder) 107 formats self.
|
|
f formats, 108 offsets self.
|
|
offsets, > 109 titles self.
|
|
titles}) 110 if byteorder is not None) 111 byteorder = byteorderconv[byteorder[0]] <type exceptions.TypeError > data type not understood","enhancement"
|
|
"6013","1061","hex ), oct ) fails on int64 array Trac 463).
|
|
Original ticket http projects.scipy.org numpy ticket 463 on 2007 03 06 by trac user gregsmithto, assigned to unknown.
|
|
This probably only fails on 32 bit platforms >>> a = array [0,257<<31]) >>> a array [ 0, 551903297536], dtype=int64) >>> a[1] 551903297536 >>> hex a[1]) 0x80000000 >>> hex a.tolist )[1]) 0x8080000000L It looks like the value is being squeezed through an int32 pipe before the conversion.
|
|
oct ) is likewise.","bug"
|
|
"6013","1062","division inconsistent with Python; a b inconsistent with a b, scary scalar behaviour Trac 464).
|
|
Original ticket http projects.scipy.org numpy ticket 464 on 2007 03 06 by trac user gregsmithto, assigned to unknown.
|
|
With integer `n d` and `n d` , if `n <0` and `d>0` , python always rounds `n d` to inf and gives `0 <= n d < d` ; whereas numpy gives you whatever the C division does, which is usually rounding towards zero for `n d` .
|
|
I brought this up for Numeric a few years ago, and was informed that this had been discussed on the group and it was done this way for performance reasons.
|
|
I would have voted strongly the other way, but that is what it is, and it wasn t such a big deal for Numeric.
|
|
However, with numpy, this behaviour is inherited by integers extracted from arrays, which otherwise behave like Python ints.
|
|
These objects could travel an arbitrary distance into other parts of a Python application which are completely unaware of numpy, so there is the potential for very strange failures.
|
|
Also, it seems that numpy unlike Numeric) now gives the python result for ` `, but still gives the C result for ` ` and ` `).
|
|
So we have ` n d) d n d) != n` because ` ` and ` ` are inconsistent ` ` operator which is called floor division internally), does not do floor division very surprising behavior with scalars, as shown below.
|
|
>>> from numpy import array >>> a,b= 7,10 >>> a,b 7, 10) >>> a b, a b 1, 3) >>> divmod a,b) 1, 3) >>> pair=array a,b)) >>> pair array [ 7, 10]) >>> a,b=pair >>> a,b 7, 10) >>> a b, a b 0, 3) >>> divmod a,b) 0, 3) At least it is now possible to get a future proof,portable floor division without needing conditionals ` n n d)) d `, since the ` ` works.
|
|
I recognize the problems associated with changing this behaviour.
|
|
But it s already been changed I don t know when) for ` `; and having it propogated into the scalars creates a serious new problem which wasn t there in Numeric.
|
|
Also at least in principle the current behaviour is platform dependent, so apps using numpy may encounter the same issue simply by moving to a different C compiler platform.","bug"
|
|
"6013","1063","ndarray s mean method should be computed using double precision Trac 465).
|
|
Original ticket http projects.scipy.org numpy ticket 465 on 2007 03 07 by chanley, assigned to unknown.
|
|
The default data type for the accumulator variable in the mean method should be double precision.
|
|
The problem can best be illustrated with the following example Python 2.4.3 2, Dec 7 2006, 11 01 45) [GCC 4.0.1 Apple Computer, Inc. build 5367)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as n >>> n. version 1.0.2.dev3571 >>> a = n.ones 1000,1000),dtype=n.float32) 132.00005 >>> print a [[ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578] [ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578] [ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578] ..., [ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578] [ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578] [ 132.00004578 132.00004578 132.00004578 ..., 132.00004578 132.00004578 132.00004578]] >>> a.min ) 132.000045776 >>> a.max ) 132.000045776 >>> a.mean ) 133.96639999999999 Having the mean be greater than the maximum is a tad odd.
|
|
The calculation of the mean is occurring with a single precision accumulator variable.
|
|
A user can force a double precision calculation with the following command and receive a correct result >>> a.mean dtype=n.float64) 132.00004577636719 >>> However, this is not going to be obvious to the casual user and will appear to be an error.
|
|
I realize that one reason for not doing all calculations as double precision is performance.
|
|
However, it is probably better to always receive the correct answer than to quickly arrive at the wrong one.
|
|
The current default behavior needs to be changed.
|
|
All calculations should be done in double precision.
|
|
If performance is needed the expert user can go back and start setting data types after having shown that their application arrives at a correct result.
|
|
Not having to worry about overflow problems in the accumulator variable would also make numpy consistent with numarray s behavior.","bug"
|
|
"6013","1064","Segmentation fault when reversing an array enough times Trac 466).
|
|
Original ticket http projects.scipy.org numpy ticket 466 on 2007 03 08 by FrancescAlted, assigned to dmcooke.
|
|
Hi, I m getting a consistent seg fault by running this code import numpy print numpy version > , numpy.
|
|
version N = 1000 1000 a = numpy.array [1,2]) for i in xrange N) a = a[ 1] and the output for my machine numpy version > 1.0.2.dev3546 Violacixc3xb3 de segment I need 1 million of iterations for getting the segfault, but your mileage may vary.","bug"
|
|
"6013","1065","inconsistent mgrid results Trac 467).
|
|
Original ticket http projects.scipy.org numpy ticket 467 on 2007 03 14 by trac user acorrigan, assigned to unknown.
|
|
>>> print mgrid[2.45 2.6 0.05, 0 5 1] [[[ 2.45 2.45 2.45 2.45 2.45] [ 2.5 2.5 2.5 2.5 2.5 ]] [[ 0.
|
|
1.
|
|
2.
|
|
3.
|
|
4. ]
|
|
[ 0.
|
|
1.
|
|
2.
|
|
3.
|
|
4. ]]]
|
|
>>> print mgrid[2.45 2.6 0.05] [ 2.45 2.5 2.55] I m using 2.45 2.6 0.05 to specify the grid in the first dimension, but get different results in each case.","bug"
|
|
"6013","1066","Use ellipses to auto complete indices for slice assignments Trac 468).
|
|
Original ticket http projects.scipy.org numpy ticket 468 on 2007 03 14 by trac user jlagarde, assigned to unknown.
|
|
I have not used numpy fro that long and I apologize if this is a completely stupid idea, but it seems that the slice assignment a[4 4 len b)] = b is a common use case for me anyway).
|
|
It would be a nice convenience if the ellipses notation could be used to mean as many as required in this context as well, i.e.
|
|
a[4 ...] = b, or a![...
|
|
10] = b There would be no confusion with the other use of ellipses because in this case they would be immediately next to the colon.
|
|
I have not spent much time thinking of the multidimentional case, but think it would conceptually work also, i.e.
|
|
a[4 ...,... 3] = b Of course, a shape mismatch error would still occur if there is just not enough room in a to fit b at the requested indices.","enhancement"
|
|
"6013","1067","Memory error string concatenation Trac 469).
|
|
Original ticket http projects.scipy.org numpy ticket 469 on 2007 03 17 by stefanv, assigned to teoliphant.
|
|
With r3580, the following code produces a memory error import numpy as N x = N.array []) N.append x, asdasd\tasdasd ) Valgrind says ==9184== Invalid write of size 1 ==9184== at 0x4022CE2 memcpy mc replace strmem.c 405) ==9184== by 0x463D5E4 PyArray Concatenate multiarraymodule.c 1596) ==9184== by 0x463D71E array concatenate multiarraymodule.c 6277) ==9184== by 0x805A0B6 PyObject Call in usr bin python2.5) ==9184== by 0x80BE4E9 PyEval EvalFrameEx in usr bin python2.5) ==9184== by 0x80C37F4 PyEval EvalCodeEx in usr bin python2.5) ==9184== by 0x80C183D PyEval EvalFrameEx in usr bin python2.5) ==9184== by 0x80C37F4 PyEval EvalCodeEx in usr bin python2.5) ==9184== by 0x80C3866 PyEval EvalCode in usr bin python2.5) ==9184== by 0x80E2EEB PyRun FileExFlags in usr bin python2.5) ==9184== Address 0x43FCFE1 is 1 bytes after a block of size 8 alloc d ==9184== at 0x4021396 malloc vg replace malloc.c 149) ==9184== by 0x4629CD4 PyArray NewFromDescr arrayobject.c 5386) ==9184== by 0x463D56B PyArray Concatenate multiarraymodule.c 1584) ==9184== by 0x463D71E array concatenate multiarraymodule.c 6277) ==9184== by 0x805A0B6 PyObject Call in usr bin python2.5) ==9184== by 0x80BE4E9 PyEval EvalFrameEx in usr bin python2.5) ==9184== by 0x80C37F4 PyEval EvalCodeEx in usr bin python2.5) ==9184== by 0x80C183D PyEval EvalFrameEx in usr bin python2.5) ==9184== by 0x80C37F4 PyEval EvalCodeEx in usr bin python2.5) ==9184== by 0x80C3866 PyEval EvalCode in usr bin python2.5)","bug"
|
|
"6013","1068","numpy uses mixture of tabs and spaces in source code Trac 470).
|
|
Original ticket http projects.scipy.org numpy ticket 470 on 2007 03 19 by trac user bhoel, assigned to unknown.
|
|
Building a module of mine usin numpy.distutils I get an error message related to the inconsistent use of tabs and spaces in indentation when invoking python with the tt command line switch.
|
|
Traceback most recent call last) File setup.py , line 24, in ?
|
|
from numpy.distutils.core import setup File C \Python24\lib\site packages\numpy\distutils\ init .py , line 5, in ?
|
|
import ccompiler File C \Python24\lib\site packages\numpy\distutils\ccompiler.py , line 11, in ?
|
|
import log File C \Python24\lib\site packages\numpy\distutils\log.py , line 7, in ?
|
|
from misc util import red text, yellow text, cyan text, is sequence, is string File C \Python24\lib\site packages\numpy\distutils\misc util.py , line 975 import numpy.numarray.util as nnu TabError inconsistent use of tabs and spaces in indentation Component is set to `numpy.core`, but problem might apply to other components as well.","bug"
|
|
"6013","1069","Recent changes in mtrand.pyx Trac 471).
|
|
Original ticket http projects.scipy.org numpy ticket 471 on 2007 03 20 by nilswagner01, assigned to unknown.
|
|
scipy.test 1) results in several errors, e.g.
|
|
File mtrand.pyx , line 1234, in mtrand.RandomState.binomial NameError PyInt AsLong File mtrand.pyx , line 946, in mtrand.RandomState.standard t File mtrand.pyx , line 149, in mtrand.cont1 array sc AttributeError module object has no attribute Float64 File mtrand.pyx , line 1206, in mtrand.RandomState.triangular File mtrand.pyx , line 248, in mtrand.cont3 array sc AttributeError module File mtrand.pyx , line 1323, in mtrand.RandomState.zipf File mtrand.pyx , line 420, in mtrand.discd array sc AttributeError module object has no attribute Int","bug"
|
|
"6013","1070","pyrex example doesn t compile.
|
|
Trac 472).
|
|
Original ticket http projects.scipy.org numpy ticket 472 on 2007 03 20 by trac user royd, assigned to unknown.
|
|
After make clean the pyrex example doesn t compile any more make python setup.py build ext inplace usr lib python2.4 site packages numpy lib utils.py 83 DeprecationWarning get numpy include is deprecated, use get include DeprecationWarning) running build ext building numpyx extension home royd python pyrex c numpy.pxd 99 22 Array element cannot be a Python object make [all] Error 1 rpm q numpy numpy 1.0.1 3.fc6 pyrexc v Pyrex version 0.9.4 Both numpy and pyrex are installed using std.
|
|
rpms provided for Fedora 6.
|
|
Regards, r.","bug"
|
|
"6013","1071","Matrix multiplication returns wrong result Trac 473).
|
|
Original ticket http projects.scipy.org numpy ticket 473 on 2007 03 22 by charris, assigned to unknown.
|
|
A row vector is returned when the result should be a column vector.
|
|
In [2] I = matrix eye 2)) In [3] x = array [1,2]) In [4] I x Out[4] matrix [[ 1., 2.]])","bug"
|
|
"6013","1072","Record type inconsistency between rank 0 recarrays and rank > 0 recarrays Trac 474).
|
|
Original ticket http projects.scipy.org numpy ticket 474 on 2007 03 23 by trac user SuperElectric, assigned to unknown.
|
|
= Short description = The records of rank N record arrays are usually returned as rank N ndarrays.
|
|
However, if N is 0, the records are returned as scalars.
|
|
This makes it problematic to write general vectorized code that operates on arbitrary rank record arrays, as expressions like recarr.x[...] = 2 fail if rank = 0.
|
|
= Example = >>> import numpy as N >>> dt = N.dtype [ x , f8 ), y , f8 )]) >>> rarr = N.zeros ), dtype = dt).view N.recarray) >>> rarr recarray 0.0, 0.0), dtype=[ x , <f8 ), y , <f8 )]) >>> oddly, rarr.x is not an array >>> rarr.x 0.0 >>> This makes it impossible to fill values in the usual manner >>> rarr.x[...] = 2.0 TypeError object does not support item assignment >>> A workaround is to reshape to non zero rank >>> rarr.shape = [1] >>> Now rarr.x returns a ndarray of the same rank as rarr >>> rarr.x array [ 0.])
|
|
>>> Value setting now works as expected.
|
|
>>> rarr.x[...] = 2.0 >>> rarr recarray [ 2.0, 0.0)], dtype=[ x , <f8 ), y , <f8 )]) = Workaround = The fallout of this bug can be illustrated by its workaround.
|
|
You need to insert special checks for ndim==0, and temporarily reshape the record array to rank > 0 before performing array operations on its records.
|
|
In other words, you need to replace this rarr.x[...] = blah return rarr With this if rarr.ndim == 0 rarr = output.reshape [1]) rarrWas0dim = True else rarrWas0dim = False rarr.x[...] = blah if rarrWas0dim rarr = rarr.reshape []) return rarr This code bloat quickly gets nasty when dealing with arithmetic between many record arrays, all of which may be of rank 0.","bug"
|
|
"6013","1073","Right Left multiplication of matrix by sequence is inconsistent.
|
|
Trac 475).
|
|
Original ticket http projects.scipy.org numpy ticket 475 on 2007 03 24 by charris, assigned to unknown.
|
|
In [12] i = matrix eye 2)) In [13] [1,2] i Out[13] matrix [[ 1., 2.]])
|
|
works In [14] i [[1],[2]] exceptions.TypeError Traceback most recent call last) home charris <ipython console> TypeError can t multiply sequence by non int doesn t.","bug"
|
|
"6013","1074","Strings with leading nulls don t compare properly Trac 476).
|
|
Original ticket http projects.scipy.org numpy ticket 476 on 2007 03 24 by trac user kreymborg, assigned to unknown.
|
|
Comparison of equal length numpy strings appears to cease at the first null character.
|
|
I used MSVC 7.1 compiler on 1.0.2.dev3593.
|
|
Example s1 = \000\001 [[BR]] s2 = \000\002 [[BR]] print s1 == s2 this evaluates as True[[BR]] print array s1) == array s2) this evaluates as False[[BR]]","bug"
|
|
"6013","1075","isscalar give strange results Trac 477).
|
|
Original ticket http projects.scipy.org numpy ticket 477 on 2007 03 25 by charris, assigned to unknown.
|
|
This seems to be expected behaviour.","bug"
|
|
"6013","1076","numpy.rec.array is inconsistent on objects Trac 478).
|
|
Original ticket http projects.scipy.org numpy ticket 478 on 2007 03 26 by FrancescAlted, assigned to unknown.
|
|
Hi, I find this to be incosistent In [134] 1.
|
|
== numpy.array 1.)
|
|
Out[134] True However...
|
|
In [135] numpy.rec.array [[1.
|
|
]], dtype=[ var1 , f8 )]) Out[135] recarray [ 1.0,)], dtype=[ var1 , <f8 )]) In [136] numpy.rec.array [[numpy.array 1.
|
|
)]], dtype=[ var1 , f8 )]) Out[136] recarray [[ 1.0,)]], dtype=[ var1 , <f8 )]) out[135] has a different level of nesting than out[136], but they should be equal.
|
|
In the same way In [138] numpy.rec.array [[1.
|
|
]], dtype=[ var1 , f8 , 2,))]) Out[138] recarray [ array [ 1., 1.
|
|
]),)], dtype=[ var1 , <f8 , 2,))]) In [139] numpy.rec.array [numpy.array [1.
|
|
])], dtype=[ var1 , f8 , 2,))]) Out[139] recarray [1.0, 1.0],), dtype=[ var1 , <f8 , 2,))]) Why the difference?","bug"
|
|
"6013","1077","MemoryError creating recarrays from medium sized python objects Trac 479).
|
|
Original ticket http projects.scipy.org numpy ticket 479 on 2007 03 26 by FrancescAlted, assigned to unknown.
|
|
The next exposes the weird) behaviour In [207] N=8670 In [208] numpy.rec.array [[range N)]], dtype=[ var1 , f8 , N,))]) Out[208] recarray [ array [ 0.00000000e 00, 1.00000000e 00, 2.00000000e 00, ..., 8.66700000e 03, 8.66800000e 03, 8.66900000e 03]),)], dtype=[ var1 , <f8 , 8670,))]) In [209] N=8680 In [210] numpy.rec.array [[range N)]], dtype=[ var1 , f8 , N,))]) exceptions.MemoryError Traceback most recent call last) home faltet <ipython console> usr lib python2.4 site packages numpy core records.py in array obj, dtype, shape, offset, strides, formats, names, titles, aligned, byteorder, copy) 531 elif isinstance obj, list, tuple)) 532 if isinstance obj[0], tuple, list)) > 533 return fromrecords obj, dtype=dtype, shape=shape, kwds) 534 else 535 return fromarrays obj, dtype=dtype, shape=shape, kwds) usr lib python2.4 site packages numpy core records.py in fromrecords recList, dtype, shape, formats, names, titles, aligned, byteorder) 384 385 try > 386 retval = sb.array recList, dtype = descr) 387 except TypeError list of lists instead of list of tuples 388 if shape is None or shape == 0) MemoryError The memory of my machine is 512 MB, and the recarray python object should fit in far less than 1 MB.
|
|
Curiously enough, you can achieve the desired recarray by using a regular numpy array In [211] numpy.rec.array numpy.arange N 1.
|
|
), dtype=[ var1 , f8 , N,))]) Out[211] recarray [ array [ 0.00000000e 00, 1.00000000e 00, 2.00000000e 00, ..., 8.67700000e 03, 8.67800000e 03, 8.67900000e 03]),)], dtype=[ var1 , <f8 , 8680,))]) But note the strange nesting of the resulting recarray see ticket 1076).","bug"
|
|
"6013","1078","Recent changes in defmatrix.py Trac 480).
|
|
Original ticket http projects.scipy.org numpy ticket 480 on 2007 03 26 by nilswagner01, assigned to unknown.
|
|
scipy.test 1) ====================================================================== ERROR check matmat scipy.sparse.tests.test sparse.test csc) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 142, in check matmat assert array almost equal a bsp).todense ), a b) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check rmatvec scipy.sparse.tests.test sparse.test csc) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 107, in check rmatvec assert array almost equal row M, row M.todense )) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check matmat scipy.sparse.tests.test sparse.test csr) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 142, in check matmat assert array almost equal a bsp).todense ), a b) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check rmatvec scipy.sparse.tests.test sparse.test csr) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 107, in check rmatvec assert array almost equal row M, row M.todense )) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check matmat scipy.sparse.tests.test sparse.test dok) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 142, in check matmat assert array almost equal a bsp).todense ), a b) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR Does the matrix s mean ,axis=0) method work?
|
|
Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 54, in check mean assert array equal self.dat.mean axis=0), self.datsp.mean axis=0)) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 423, in mean mean = self.sum 0) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 402, in sum return o self File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check rmatvec scipy.sparse.tests.test sparse.test dok) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 107, in check rmatvec assert array almost equal row M, row M.todense )) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR Does the matrix s sum ,axis=0) method work?
|
|
Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 46, in check sum assert array equal self.dat.sum axis=0), self.datsp.sum axis=0)) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 402, in sum return o self File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check matmat scipy.sparse.tests.test sparse.test lil) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 142, in check matmat assert array almost equal a bsp).todense ), a b) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR Does the matrix s mean ,axis=0) method work?
|
|
Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 54, in check mean assert array equal self.dat.mean axis=0), self.datsp.mean axis=0)) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 423, in mean mean = self.sum 0) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 402, in sum return o self File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR check rmatvec scipy.sparse.tests.test sparse.test lil) Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 107, in check rmatvec assert array almost equal row M, row M.todense )) File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned ====================================================================== ERROR Does the matrix s sum ,axis=0) method work?
|
|
Traceback most recent call last) File usr local lib64 python2.5 site packages scipy sparse tests test sparse.py , line 46, in check sum assert array equal self.dat.sum axis=0), self.datsp.sum axis=0)) File usr local lib64 python2.5 site packages scipy sparse sparse.py , line 402, in sum return o self File usr local lib64 python2.5 site packages numpy core defmatrix.py , line 162, in mul return N.dot self, other) ValueError objects are not aligned","bug"
|
|
"6013","1079","Inconsistency between plot like commands and the getp command Trac 481).
|
|
Original ticket http projects.scipy.org numpy ticket 481 on 2007 03 26 by trac user jbmichel, assigned to unknown.
|
|
The following piece of code does not work p = plot whatever) getp p, anyProperty) The reason for this is that the getp function is tailored to take as an input an instance of, say, Line2D, but the command plot produces a list of such objects.
|
|
2 solutions 1.
|
|
Take this into account, and run the command p = plot whatever) getp p[0], anyProperty) 2.
|
|
Modify the getp code at line 416 and use func = getattr o[0], get name) instead of func = getattr o, get name) JB","bug"
|
|
"6013","1080","crash on fancy indexing Trac 482).
|
|
Original ticket http projects.scipy.org numpy ticket 482 on 2007 03 26 by pv, assigned to unknown.
|
|
The following code crashes, but not always on numpy 1.0.1) import numpy as N A = N.zeros 800, 7, 8)) B = N.zeros 6,6)) 1 print N. version , N. file for i in xrange 800) A[N.ix [i], [0,1,2,3,5,6], [0,1,2,3,4,5])] = B crash here print no crash It s probably a non local error, since puzzling things happen echo x {PYTHONPATH}x xx python2.4 xxx.py 1.0.1 usr lib python2.4 site packages numpy init .pyc no crash PYTHONPATH=this path doesnt exist python2.4 xxx.py 1.0.1 usr lib python2.4 site packages numpy init .pyc Segmentation fault core dumped) PYTHONPATH= python2.4 xxx.py 1.0.1 usr lib python2.4 site packages numpy init .pyc Segmentation fault core dumped) There are no lurking numpy installations or source dirs, only the one in usr.
|
|
I have been able to reproduce this on different machines different gcc, processors, and build options).
|
|
Also, although this crashes on python2.4, there is no crash on python2.5.
|
|
Stack trace some debug info attached.","bug"
|
|
"6013","1081","Junk in the representation of string fields of recarrays Trac 483).
|
|
Original ticket http projects.scipy.org numpy ticket 483 on 2007 03 28 by FrancescAlted, assigned to unknown.
|
|
The next exposes the problem In [27] r=numpy.array [[ abc ]], dtype=[ var1 , |S20 )]) In [28] r[ var1 ] Out[28] array [[ abc\xb7\x08\xc1 \x08\xc0>\xde\xb7 \xec\x95\xb7`\xec\x95\xb7 ]], dtype= |S20 ) In [30] r[ var1 ][0] Out[30] array [ abc\xb7\x08\xc1 \x08\xc0>\xde\xb7 \xec\x95\xb7`\xec\x95\xb7 ], dtype= |S20 ) In [31] r[ var1 ][0][0] Out[31] abc\xb7\x08\xc1 \x08\xc0>\xde\xb7 \xec\x95\xb7`\xec\x95\xb7 In [32] str r[ var1 ][0][0]) Out[32] abc\xb7\x08\xc1 \x08\xc0>\xde\xb7 \xec\x95\xb7`\xec\x95\xb7 This seems to be a problem related only with string fields on recarrays, because I m not able to reproduce it on plain string arrays In [34] s2=numpy.array [[ abc ]], dtype= S20 ) In [35] s2 Out[35] array [[ abc ]], dtype= |S20 ) In [36] s2[0][0] Out[36] abc In [37] str s2[0][0]) Out[37] abc Using 1.0.2.dev3546 here.
|
|
Thanks!","bug"
|
|
"6013","1082","random state not portable between 32 bit and 64 bit architectures Trac 484).
|
|
Original ticket http projects.scipy.org numpy ticket 484 on 2007 03 28 by trac user berkes, assigned to unknown.
|
|
The random state in the numpy.random module contains an array of int32 or int64 depending on the architecture, making it not portable.
|
|
This makes simulations not reproducible on different machines.
|
|
For example, try the following on two machines with different architecures Python 2.4.4 2, Jan 13 2007, 17 50 26) [GCC 4.1.2 20061115 prerelease) Debian 4.1.1 21)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import platform >>> import pickle >>> >>> platform.architecture ) 32bit , ) >>> >>> state = numpy.random.get state ) >>> pickle.dump state, file tmp random state.pic , w )) >>> numpy.random.permutation numpy.arange 10)) array [8, 6, 3, 1, 0, 5, 7, 4, 2, 9]) >>> Python 2.5 r25 51908, Mar 19 2007, 13 41 07) [GCC 4.1.0 SUSE Linux)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import platform >>> import pickle >>> >>> platform.architecture ) 64bit , ELF ) >>> >>> state = pickle.load file tmp random state.pic , r )) >>> numpy.random.set state state) >>> numpy.random.permutation numpy.arange 10)) array [3, 9, 4, 8, 1, 6, 2, 0, 5, 7]) >>>","bug"
|
|
"6013","1083","Problem with numpy.rec.fromarrays when using titles argument Trac 485).
|
|
Original ticket http projects.scipy.org numpy ticket 485 on 2007 03 29 by trac user obi, assigned to unknown.
|
|
The error is documented by the following ipython log 1 import numpy 2 a = numpy.array [1,2,3]) 3 b = numpy.array [0.1,0.2,0.3]) 4 print numpy.rec.fromarrays [a,b],names= A,B ) > 5 print numpy.rec.fromarrays [a,b],names= A,B ,titles= Spalte 1,Spalte 2 ) usr lib python2.4 site packages numpy 1.0.1 py2.4 linux i686.egg numpy core records.py in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 317 Determine shape from data type.
|
|
318 if len descr) != len arrayList) > 319 raise ValueError, mismatch between the number of fields \ 320 and the number of arrays 321 ValueError mismatch between the number of fields and the number of arrays WARNING Failure executing file <test fromarrays.py> [ 1, 0.10000000000000001) 2, 0.20000000000000001) 3, 0.29999999999999999)] Note, that the problem does not occur in version 0.9.6.","bug"
|
|
"6013","1084","FAIL check basic numpy.core.tests.test multiarray.test clip) on Solaris 8 system Trac 486).
|
|
Original ticket http projects.scipy.org numpy ticket 486 on 2007 04 01 by chanley, assigned to unknown.
|
|
Numpy 1.0.2 rev 3632 fails on a Solaris 8 system running Python 2.4.3.
|
|
====================================================================== FAIL check basic numpy.core.tests.test multiarray.test clip) Traceback most recent call last) File data basil5 site packages lib python numpy core tests test multiarray.py , line 388, in check basic self.
|
|
clip type float ,1024, 12.8,100.2, inplace=inplace) File data basil5 site packages lib python numpy core tests test multiarray.py , line 382, in clip type assert equal byteorder,x.dtype.byteorder) File data basil5 site packages lib python numpy testing utils.py , line 143, in assert equal assert desired == actual, msg AssertionError Items are not equal ACTUAL < DESIRED = Ran 562 tests in 19.376s FAILED failures=1)","bug"
|
|
"6013","1085","Segfault with recarrays and r3632 Trac 487).
|
|
Original ticket http projects.scipy.org numpy ticket 487 on 2007 04 01 by FrancescAlted, assigned to unknown.
|
|
It seems that a problem has slipped in lately because the next code import numpy dtype = [ var1 , S4 )] for i in xrange 10) recarr = numpy.empty shape=1, dtype=dtype) print > , recarr consistently causes a segfault at least with r3632.
|
|
This seems to work well using r3546.","bug"
|
|
"6013","1086","seeded randint returns different values on 32 and 64bit architectures Trac 488).
|
|
Original ticket http projects.scipy.org numpy ticket 488 on 2007 04 02 by trac user benstone, assigned to unknown.
|
|
This returns different values on 32 and 64bit Ubuntu feisty) using Numpy1.0.1 import numpy.random as r r.seed 7777) r.randint 10)","bug"
|
|
"6013","1087","Filling polygons with Basemap Trac 489).
|
|
Original ticket http projects.scipy.org numpy ticket 489 on 2007 04 02 by trac user jbmichel, assigned to unknown.
|
|
Basemap had no simple function for filling polygons wich can be very useful if one wants to represent countries with different colors, corresponding to some social or economic parameter for instance.
|
|
One can copy paste Basemap s plot function and only replace the word plot by the word fill .
|
|
This yields good results.
|
|
The code to add is exactly def fill self, args, kwargs) Draw lines and or markers on the map see pylab.plot documentation).
|
|
extra keyword ax can be used to override the default axis instance.
|
|
if not kwargs.has key ax ) and self.ax is None try ax = pylab.gca ) except import pylab ax = pylab.gca ) elif not kwargs.has key ax ) and self.ax is not None ax = self.ax else ax = popd kwargs, ax ) allow callers to override the hold state by passing hold=True|False b = ax.ishold ) h = popd kwargs, hold , None) if h is not None ax.hold h) try ret = ax.fill args, kwargs) try pylab.draw if interactive ) except pass except ax.hold b) raise ax.hold b) set axes limits to fit map region.
|
|
self.set axes limits ax=ax) make sure axis ticks are turned off.
|
|
if self.noticks ax.set xticks []) ax.set yticks []) return ret","enhancement"
|
|
"6013","1088","problem with simple matrix addition Trac 490).
|
|
Original ticket http projects.scipy.org numpy ticket 490 on 2007 04 03 by trac user gauthampdas, assigned to unknown.
|
|
when we declare an row vector as a matrix, and doing addition i noticed a problem.
|
|
>>> a=mat [1,2,3]) >>> a transpose a) matrix [[2, 3, 4], [3, 4, 5], [4, 5, 6]]) >>> a a matrix [[2, 4, 6]]) In my application i have to handle different matrices and vectors.
|
|
Just noticed it.
|
|
This is a simple mistake but can make big errors","bug"
|
|
"6013","1089","Build cares about being in an SVN checkout Trac 491).
|
|
Original ticket http projects.scipy.org numpy ticket 491 on 2007 04 03 by rkern, assigned to rkern.
|
|
Currently, the build process for numpy requires one to be in an SVN checkout or have the appropriate files generated from an sdist done from an SVN checkout.
|
|
This prevents people from working from an SVN export or with distributed version control tools like SVK or any of the DVCSs that can be supported by tailor.
|
|
This patch removes the restriction.
|
|
Builds done from outside an SVN checkout and are not releases will still have a ` .dev ` appended to the version number.
|
|
F2PY will have a version of ` 2 ?
|
|
` as seems to have been the default when ` svn version .py` could not be found.
|
|
If no one complains, I will check this in shortly.
|
|
Pearu, I m CCing you about the F2PY version number.
|
|
If you have another preference, please suggest it.","bug"
|
|
"6013","1090","integer overflow on plus and sum at least) Trac 492).
|
|
Original ticket http projects.scipy.org numpy ticket 492 on 2007 04 05 by davidsocha, assigned to unknown.
|
|
There are several places at least) where integer operations overflow, instead of upgrading the size of the result to accommodate larger values.
|
|
Here are some examples.
|
|
C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> a = array [127,127], dtype= int8 ) >>> a a array [ 2, 2], dtype=int8) <== overflow bad!)
|
|
>>> a.sum ) 254 <== correct value >>> a.sum ).dtype dtype int32 ) <== correctly upgraded to larger type >>> b = array [1250000000,1250000000], dtype= int32 ) >>> b b array [ 1794967296, 1794967296]) <== overflow bad!)
|
|
>>> b.sum ) 1794967296 <== overflow >>> b.sum ).dtype dtype int32 ) <== did not upgrade type to available int64 type >>> >>> from numpy import int64 <== int64 exists >>> c = array [1250000000,1250000000], dtype= int64 ) >>> c array [1250000000, 1250000000], dtype=int64) >>> c.sum ) 2500000000 <== correct value >>> c c array [2500000000, 2500000000], dtype=int64) <== correct value >>> >>> f = array [8750000000000000000, 8750000000000000000], dtype=int64) >>> f f array [ 946744073709551616, 946744073709551616], dtype=int64) <== overflow bad too?)
|
|
>>> f.sum ) 946744073709551616 <== overflow bad too?)
|
|
>>> David Socha Daniel Terhorst, UrbanSim Project, http www.urbansim.org","bug"
|
|
"6013","1091","maximum sctype returns a smaller type than is actually used by the data Trac 493).
|
|
Original ticket http projects.scipy.org numpy ticket 493 on 2007 04 05 by davidsocha, assigned to unknown.
|
|
Here s an example C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> from numpy import maximum sctype >>> g = array [1], dtype= int64 ) >>> maximum sctype g) <type numpy.int32 > <== Why is this not at least int64?
|
|
>>> David Socha Daniel Terhorst, UrbanSim Project, http www.urbansim.org","bug"
|
|
"6013","1092","Arrays of type float96 or float128 do not display correctly Trac 494).
|
|
Original ticket http projects.scipy.org numpy ticket 494 on 2007 04 05 by davidsocha, assigned to unknown.
|
|
Here is an example C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> f = array [1,2,3], dtype= float96 ) >>> f array [0.0, 0.0, 2.0], dtype=float96) <== float96 prints incorrectly >>> f.astype float32 ) array [ 1., 2., 3.
|
|
], dtype=float32) <== Yet data seems intact >>> f.sum ) 2.0 <== prints incorrectly >>> f.sum ).astype float32 ) 6.0 <== prints correctly if force to float32 >>> We have the same symptoms with float128 on a Mac with an Intel 2GHz Core Duo).
|
|
David Socha Daniel Terhorst, UrbanSim Project, http www.urbansim.org","bug"
|
|
"6013","1093","maximum sctype returns incompatible types on different computers Trac 495).
|
|
Original ticket http projects.scipy.org numpy ticket 495 on 2007 04 05 by davidsocha, assigned to unknown.
|
|
On our Windows XP Pro computer we get C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> from numpy import maximum sctype >>> g = array [1], dtype= float32 ) >>> maximum sctype g) <type numpy.float96 > >>> from numpy import float128 Traceback most recent call last) File <stdin> , line 1, in ?
|
|
ImportError cannot import name float128 >>> Note that the maximum type for float32 is float96, and we cannot import float128.
|
|
On our Macintosh with an Intel 2GHz Core Duo, maximum sctype returns float128 for float32, and float96 cannot be imported.
|
|
This makes it problematic to share results between the two computers, since files written as float96 will not be readable on the other computers.
|
|
We d expect both computers to return the same type from maximum sctype.
|
|
David Socha Daniel Terhorst, UrbanSim Project, http www.urbansim.org","bug"
|
|
"6013","1094","ones and zeros return floats by default not ints) Trac 496).
|
|
Original ticket http projects.scipy.org numpy ticket 496 on 2007 04 05 by davidsocha, assigned to teoliphant.
|
|
Despite what the documentation says, ones and zeros return floats by default, not ints.
|
|
Here s an example C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import ones >>> ones 3) array [ 1., 1., 1.])
|
|
>>> from numpy import zeros >>> zeros 3) array [ 0., 0., 0.])
|
|
>>> Yet on page 88 of the manual, it says zeros shape=, dtype=int, order=xe2x80x99Cxe2x80x99) ... ones shape=, dtype=int, order=xe2x80x99Cxe2x80x99) which indicates that the default type is int.
|
|
David Socha Daniel Terhorst, UrbanSim Project, http www.urbansim.org","bug"
|
|
"6013","1095","The dot and inner products fail for large arrays on 64 bit linux using the ATLAS libraries.
|
|
Trac 497).
|
|
Original ticket http projects.scipy.org numpy ticket 497 on 2007 04 10 by charris, assigned to unknown.
|
|
The problem is easy to exhibit In [2] dot eye 1000),eye 1000)) Illegal instruction The products work for smaller arrays, so I suspect either the output array or a work array is being created with the wrong size, probably a 64bit integer thing in the cblas interface.
|
|
The example works on 32 bit linux.
|
|
This ticket replaces 1030","bug"
|
|
"6013","1096","array.var method does not have docstring Trac 498).
|
|
Original ticket http projects.scipy.org numpy ticket 498 on 2007 04 11 by trac user eric, assigned to unknown.
|
|
the var ) function does, but the array method doesn t.","enhancement"
|
|
"6013","1097","array.diagonal method doesn t have a doc string Trac 499).
|
|
Original ticket http projects.scipy.org numpy ticket 499 on 2007 04 11 by trac user eric, assigned to unknown.","enhancement"
|
|
"6013","1098","Numpy distutils relies on obsoleted gfortran version output Trac 500).
|
|
Original ticket http projects.scipy.org numpy ticket 500 on 2007 04 17 by trac user jarodwilson, assigned to dmcooke.
|
|
Numpy s distutils relies on the output of gcc s gfortran version to determine the presence of fortran 95 support.
|
|
Upstream gcc has removed 95 from the version string, because gcc fortran actually supports f95, f77 and now part of f2003 as well.
|
|
At present, numpy still builds against Fedora s development tree, but scipy, which uses numpy s distutils, crashes and burns.
|
|
References http gcc.gnu.org ml gcc patches 2007 03 msg00320.html http gcc.gnu.org bugzilla show bug.cgi?id=31050 https bugzilla.redhat.com bugzilla show bug.cgi?id=236670","bug"
|
|
"6013","1099","print arr fails with nan when seterr all= raise ) [test included] Trac 501).
|
|
Original ticket http projects.scipy.org numpy ticket 501 on 2007 04 17 by astraw, assigned to unknown.
|
|
The included testcase a patch to go in numpy s unittests) triggers what I consider to be a bug.
|
|
One should be able to print an array no matter what the seterr state is.
|
|
If necessary, I suggest array str pushes the errors state when setting to ignore , does the conversion to string, and pops the error state again.","bug"
|
|
"6013","1100","Unbiased estimates of variance std deviation Trac 502).
|
|
Original ticket http projects.scipy.org numpy ticket 502 on 2007 04 18 by pierregm, assigned to unknown.
|
|
Unbiased estimates of the variance and standard deviation are used far more often than their biased counterparts.
|
|
Currently, the var std methods functions return biased estimates.
|
|
Unbiased estimates can be obtained simply by multiplying the variance by n float n 1) where n is the size of the array along a particular axis).
|
|
This extra step and the test on n it implies) becomes quickly tedious when used repeatedly.
|
|
I suggest the introduction of 2 new methods and the corresponding functions), varu and stdu, that would give direct access to the unbiased estimates.","enhancement"
|
|
"6013","1101","non native endian does not work with frompyfunc Trac 503).
|
|
Original ticket http projects.scipy.org numpy ticket 503 on 2007 04 20 by trac user hugo, assigned to unknown.
|
|
>>> uradians = numpy.frompyfunc math.radians, 1, 1) >>> big = numpy.array [83.4, 83.5], dtype= >f8 ) >>> uradians big) array [ 2.6919063283e 182, 1.26773292198e 318], dtype=object) >>> little = numpy.array [83.4, 83.5], dtype= <f8 ) >>> uradians little) array [1.45560459616, 1.45734992542], dtype=object) This particular example can better be done by multiplying by pi 180, it is just a simple prove of concept.","bug"
|
|
"6013","1102","Building with Intel Math Kernel Library in windows Trac 504).
|
|
Original ticket http projects.scipy.org numpy ticket 504 on 2007 04 24 by trac user mspacek, assigned to dmcooke.
|
|
this was posted to numpy discussion) Here s a patch that allows numpy to build using Intel s MKL library in windows.
|
|
Builds with MSVC7.1 with no unittest errors.
|
|
I ve also included a .numpy site.cfg file that works with this.","enhancement"
|
|
"6013","1103","fromfile does not work with files opened with gzip Trac 505).
|
|
Original ticket http projects.scipy.org numpy ticket 505 on 2007 04 24 by trac user mesmith, assigned to unknown.
|
|
I would like to use fromfile with a file opened with the gzip module fp = gzip.open filename, r ) arraybuffer = fromfile fp, sep= , ) but this gives an error IOError first argument must be an open file I found a similar problem at http projects.scipy.org astropy pyfits ticket 7 Has this problem been addressed in numpy?
|
|
I am using version 1.0rc1.","bug"
|
|
"6013","1104","fft is unusable of moderately large arrays with a size which is a prime number Trac 506).
|
|
Original ticket http projects.scipy.org numpy ticket 506 on 2007 04 27 by cournape, assigned to unknown.
|
|
Currently, using fft on arrays of size which has a prime number is extremely slow up to 15 minutes for eg an array of size 100003).
|
|
The major problem though is that the code being uninterruptible, this effectively kills the current python session.
|
|
My understanding is that the fft in numpy is using fftpack converted in C, which uses a O N 2) algorithm for prime numbers ?
|
|
Would it be possible at least to make the code interruptible ?
|
|
{{ !python import numpy as N from scipy import fftpack prime is a prime number prime = 100003 a = N.random.randn prime) print scipy computing... fftpack.fft a) print numpy computing... N.fft.fft a) }}","bug"
|
|
"6013","1105","invalid math result in numpy.sqrt ) and numpy.square ) when using 2 32 and 2 32) 2 Trac 507).
|
|
Original ticket http projects.scipy.org numpy ticket 507 on 2007 04 27 by trac user Xiph1980, assigned to unknown.
|
|
I encountered this error a little while ago, and I couldn t find it already mentioned in other tickets.
|
|
[[BR]] I apologise if I totally misplaced this bug report or made any other mistakes, but this is the best place I ve found so far to report this ) I m not a programmer so I don t know where or how the error occurs, but with numpy.square weird stuff starts to happen in the area of 2 32 where I first encountered it, for example, numpy.square 2 32) 1) gives the negative number 8589934591, and numpy.square 2 32) 1) gives 2 32 2 1 as answer.
|
|
Obviously this isn t correct...[[BR]] I filed this as critical since numpy.square ) just gives a false result, and not an error to attend the programmer that something s going pretty wrong.
|
|
Following is a direct copy from the interpreter window where the error how I encountered it, shows in lines 12 17.
|
|
Also can be seen by the test in line 18 19 with the sqrt function in math that that version of sqrt doesn t bug out.
|
|
Python 2.5 r25 51908, Sep 19 2006, 09 52 17) [MSC v.1310 32 bit Intel)] on win 32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import math >>> print numpy.
|
|
version 1.0.2 >>> 2 32 4294967296L >>> 2 32) 2 18446744073709551616L >>> numpy.square 2 32) 0 >>> numpy.sqrt 2 32) 2) Traceback most recent call last) File <stdin> , line 1, in <module> AttributeError long object has no attribute sqrt >>> math.sqrt 2 32) 2) 4294967296.0 After a bit of fiddling, it seems that 9223372036854775807 is the largest number for with numpy.sqrt ) works before it passing on the abovementioned AttributeError.
|
|
Systemspecs [[BR]] AMD Athlon 64 3200 [[BR]] 2 512MB DDR[[BR]] Windows XP 32bit[[BR]] Python 2.5[[BR]] numpy 1.02[[BR]]","bug"
|
|
"6013","1106","astype modifies fortran ordering Trac 508).
|
|
Original ticket http projects.scipy.org numpy ticket 508 on 2007 05 01 by stefanv, assigned to unknown.
|
|
In [5] z = N.array [[1,2],[3,4]],order= F ).astype N.float64) In [6] z Out[6] array [[ 1., 2.
|
|
], [ 3., 4.]])
|
|
In [7] z.flags Out[7] C CONTIGUOUS True F CONTIGUOUS False OWNDATA True WRITEABLE True ALIGNED True UPDATEIFCOPY False In [8] x = N.array [[1,2],[3,4]],order= F ) In [9] x.flags Out[9] C CONTIGUOUS False F CONTIGUOUS True OWNDATA True WRITEABLE True ALIGNED True UPDATEIFCOPY False In [10] x.astype N.float64).flags Out[10] C CONTIGUOUS True F CONTIGUOUS False OWNDATA True WRITEABLE True ALIGNED True UPDATEIFCOPY False","bug"
|
|
"6013","1107","histogramdd fails on indentical input arrays Trac 509).
|
|
Original ticket http projects.scipy.org numpy ticket 509 on 2007 05 02 by huard, assigned to unknown.
|
|
Emanuele Olivetti reported the following bug on the numpy discussion list.
|
|
import numpy x = numpy.array [0,0]) y = numpy.array [0,1]) numpy.histogram2d x,y,bins=[2,2]) Warning divide by zero encountered in log10 exceptions.OverflowError Traceback most recent call last) The patch fixes this and adds the relevant test.
|
|
The fix makes the behavior identical to histogram, namely that if min ) == max ), the range is chosen as [min ) .5, max ) .5].","bug"
|
|
"6013","1108","Difference in the number of elements in a fromfile ) between Windows and Linux Trac 510).
|
|
Original ticket http projects.scipy.org numpy ticket 510 on 2007 05 04 by trac user Matt, assigned to cournape.
|
|
Let 0.0 inf 13.9040914426 14.7406669444 inf 4.41783247603 inf inf 6.05071515635 inf inf inf 15.6925185021 inf inf inf inf inf inf inf be the content of a text file data.txt If data.txt is loaded with data = numpy.fromfile data.txt , sep= , dtype = numpy.float) , the behaviour is different under Linux and Windows Linux the array contains 20 elements.
|
|
Windows the array contains only 6 elements.
|
|
Some thoughts float inf ) in the Python interpreter is not recognised under Windows as the infinite value, it is for Linux.","bug"
|
|
"6013","1109","array does not consistently handle list with strings and Nones Trac 511).
|
|
Original ticket http projects.scipy.org numpy ticket 511 on 2007 05 04 by davidsocha, assigned to unknown.
|
|
When creating an array from a list that has a combination of strings and Nones, it fails if None is not the last element, but succeeds if None is the last element.
|
|
Microsoft Windows XP [Version 5.1.2600] C) Copyright 1985 2001 Microsoft Corp. Z \>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> array [ a ,None]) array [a, None], dtype=object) >>> array [ a ,None, b ]) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError expected a readable buffer object >>> array [ a ,None, b , c ,None]) array [a, None, b, c, None], dtype=object) >>> David Socha Daniel Terhorst [http www.urbansim.org UrbanSim]","bug"
|
|
"6013","1110","Incorrect conversion of None into a string for string arrays of size less than 4.
|
|
Trac 512).
|
|
Original ticket http projects.scipy.org numpy ticket 512 on 2007 05 04 by davidsocha, assigned to unknown.
|
|
I would have expected each of the following to have None in them, and a type of |S4 Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> array [ a ,None, c ], dtype= S ) array [ a , N , c ], dtype= |S1 ) >>> array [ abd ,None, cde ], dtype= S ) array [ abd , Non , cde ], dtype= |S3 ) >>> array [ abcd ,None, cdef ], dtype= S ) array [ abcd , None , cdef ], dtype= |S4 ) >>> David Socha Daniel Terhorst [http www.urbansim.org UrbanSim])","bug"
|
|
"6013","1111","Cygwin build problem Multi include protection needed for fenv.c Trac 513).
|
|
Original ticket http projects.scipy.org numpy ticket 513 on 2007 05 07 by trac user jenisys, assigned to unknown.
|
|
VERSION 1.0.2 PLATFORM Cygwin 1.5.24 2 PYTHON 2.5.1 file numpy core include numpy fenv fenv.c seems to be included multiple times in some source files.
|
|
Compilation fails because symbol fe dfl env exists twice in same source object file.
|
|
QUICKFIX Add header like multi include protection to source file.","bug"
|
|
"6013","1112","memory corruption crash with array of strings Trac 514).
|
|
Original ticket http projects.scipy.org numpy ticket 514 on 2007 05 09 by bthyreau, assigned to unknown.
|
|
Hi, The following code seems to trigger a crash for me.
|
|
from numpy import hstack, version print version s = aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa t = [] print hstack t, s )) in some case, try a second time print hstack t, s )) 1.0.3.dev3736 glibc detected python malloc ) memory corruption 0x08c70a98 Tested on Python 2.4.1, 2.4.3 and 2.5, with numpy 1.0.1.dev3460 and 1.0.3.dev3736 Note that i m aware that the code itself is silly ) Thanks","bug"
|
|
"6013","1113","Wrong march flags selected for modern Xeon processors Trac 515).
|
|
Original ticket http projects.scipy.org numpy ticket 515 on 2007 05 10 by trac user jitter, assigned to dmcooke.
|
|
Executive summary A wrong march flag for gfortran causes the scipy build to break.
|
|
cat proc cpuinfo contains processor 3 vendor id GenuineIntel cpu family 6 model 15 model name Intel R) Xeon R) CPU 5160 3.00GHz stepping 6 cpu MHz 3000.149 cache size 4096 KB physical id 3 siblings 2 core id 1 cpu cores 2 fpu yes fpu exception yes cpuid level 10 wp yes flags fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall lm constant tsc pni monitor ds cpl vmx est tm2 cx16 xtpr lahf lm bogomips 6000.72 clflush size 64 cache alignment 64 address sizes 36 bits physical, 48 bits virtual power management ... and it s correctly detected as Xeon CPU by cpuinfo.py.
|
|
So far so good.
|
|
numpy distutils fcompiler gnu.py fails to list Xeon processors, though.
|
|
Therefore the default fallback march=i686) is chosen.
|
|
This doesn t work because march=i686 doesn t support 64bit instructions.
|
|
The bug isn t triggered inside numpy, but is triggered when building scipy Lib fftpack dfftpack).
|
|
My current solution is to apply the following patch Index numpy distutils fcompiler gnu.py =================================================================== numpy distutils fcompiler gnu.py revision 3737) numpy distutils fcompiler gnu.py working copy) 225,6 225,8 march opt = march=nocona elif cpu.is Core2 ) march opt = march=nocona elif cpu.is Xeon ) march opt = march=nocona elif cpu.is Prescott ) march opt = march=prescott elif cpu.is PentiumIV ) You might have to add an additional condition that it s a 64bit Xeon.
|
|
I unfortunately don t have access to a 32bit Xeon to test that.
|
|
No idea whether nocona cares about 32 64bit instructions.","bug"
|
|
"6013","1114","Error transposing arrays with many dimensions [fix included] Trac 516).
|
|
Original ticket http projects.scipy.org numpy ticket 516 on 2007 05 10 by trac user gcross, assigned to unknown.
|
|
== Symptom == Numpy incorrectly reports the error dimensions too large when transposing arrays with many dimensions.
|
|
For example, the following code produces an error from numpy.random import rand arr1 = rand 2,) 16) arr2 = arr1.transpose range 16)) Clearly the above should cause no problems, since not only should arr2 not be too big given we could already create arr1, but also in this particular case we are not even changing the indices of arr2!
|
|
== Cause == The problem can be traced to an erroneous assumption made in the C routine PyArray Transpose.
|
|
In line 1875 of this routine, a call is made to PyArray NewFromDescr to allocate space for a new array descriptor ret = PyArrayObject )\ PyArray NewFromDescr ap >ob type, ap >descr, n, permutation, NULL, ap >data, ap >flags, PyObject )ap); To see the problem, look at the fourth argument of the call permutation .
|
|
This passes in an array which contains the permuted indices, but in an argument which meant to specify the new dimensions of the array!
|
|
For arrays with, say, 16 dimensions, this confuses numpy into thinking that we are creating an array of size 16!=20922789888000, which is clearly not what we want.
|
|
As far as I can tell, this was done to save a little memory.
|
|
The programmer preferred not to create intermediate arrays with the correct values for the dimensions and strides to pass into PyArray NewFromDescr, so instead he figured that he could just pass any dummy array with the correct size into the dims argument, and then later write the correct dimensions and strides directly into the fields of the newly allocated descriptor.
|
|
However, when doing this he forgot that the dims argument is used to check whether the array being allocated is too big, so that you cannot pass just anything into it and expect the call to work.
|
|
== Solution == I am not sufficiently familiar with the code to know whether passing a dummy array into this field is even a good idea, but assuming that it is a better choice would be the dimensions field of the old array ap , i.e.
|
|
ap >dimensions , so that the call becomes ret = PyArrayObject )\ PyArray NewFromDescr ap >ob type, ap >descr, n, ap >dimensions, NULL, ap >data, ap >flags, PyObject )ap); This seems to fix the problem for me.
|
|
== DIFF == numpy core src multiarraymodule.c 1878c1878 < n, permutation, > n, ap >dimensions,","bug"
|
|
"6013","1115","Mean, median, and std dev need optimization in numpy Trac 517).
|
|
Original ticket http projects.scipy.org numpy ticket 517 on 2007 05 10 by trac user cwarner62, assigned to unknown.
|
|
I have been using numpy scipy while writing software to deal with sets of large 2048x2048) astronomical images and have discovered that the built in mean, median, and std methods for numpy are very slow much slower than calculating those quantities by hand.
|
|
There is no reason for this and I am not sure why they were written in such a way as to be so slow but for instance m = reduce add,z) z.shape[0] is nearly twice as fast as m = z.mean 0).
|
|
See below for full results from some time tests I ran >>> z.shape 10, 2048, 2048) >>> z array [[[ 186., 154., 134., ..., 385., 359., 251.
|
|
], [ 232., 208., 210., ..., 344., 333., 220.
|
|
], [ 236., 208., 226., ..., 330., 324., 226.
|
|
], ..., [ 428., 400., 475., ..., 334., 304., 371.
|
|
], [ 447., 488., 455., ..., 296., 285., 340.
|
|
], [ 495., 480., 533., ..., 236., 226., 275.
|
|
]], [[ 138., 119., 111., ..., 309., 278., 189.
|
|
], [ 185., 168., 183., ..., 289., 269., 176.
|
|
], [ 212., 175., 188., ..., 280., 282., 188.
|
|
], ..., [ 360., 346., 398., ..., 277., 246., 275.
|
|
], [ 362., 407., 378., ..., 258., 233., 241.
|
|
], [ 382., 381., 424., ..., 176., 175., 209.
|
|
]], [[ 136., 117., 106., ..., 304., 268., 182.
|
|
], [ 174., 169., 180., ..., 285., 275., 169.
|
|
], [ 199., 167., 183., ..., 280., 278., 188.
|
|
], ..., [ 352., 349., 398., ..., 279., 245., 267.
|
|
], [ 367., 410., 378., ..., 250., 227., 251.
|
|
], [ 379., 373., 418., ..., 188., 186., 200.
|
|
]], ..., [[ 144., 115., 101., ..., 304., 274., 193.
|
|
], [ 183., 168., 173., ..., 285., 274., 171.
|
|
], [ 209., 169., 184., ..., 273., 278., 182.
|
|
], ..., [ 356., 346., 395., ..., 272., 244., 264.
|
|
], [ 361., 412., 377., ..., 244., 235., 243.
|
|
], [ 384., 378., 428., ..., 194., 186., 208.
|
|
]], [[ 133., 117., 107., ..., 295., 273., 183.
|
|
], [ 180., 167., 185., ..., 286., 273., 174.
|
|
], [ 199., 168., 189., ..., 277., 285., 189.
|
|
], ..., [ 362., 344., 399., ..., 283., 243., 262.
|
|
], [ 362., 408., 383., ..., 249., 227., 242.
|
|
], [ 377., 376., 421., ..., 181., 175., 204.
|
|
]], [[ 139., 121., 105., ..., 309., 273., 189.
|
|
], [ 177., 174., 181., ..., 289., 280., 177.
|
|
], [ 209., 177., 185., ..., 273., 273., 183.
|
|
], ..., [ 353., 339., 396., ..., 264., 248., 266.
|
|
], [ 367., 414., 386., ..., 246., 228., 249.
|
|
], [ 387., 373., 423., ..., 183., 178., 202.]]])
|
|
1.
|
|
Mean is slower than reduce add) >>> t = time.time ); m = z.mean 0); print time.time ) t 1.68424701691 >>> m array [[ 142.4, 121.9, 108.8, ..., 311.9, 279.9, 193.2], [ 184.4, 173.9, 181.9, ..., 293.1, 280.4, 179.8], [ 206.5, 174.1, 190.9, ..., 284.2, 284.7, 190.7], ..., [ 365.2, 349.9, 406.3, ..., 279.7, 255.3, 276.4], [ 373. , 418.6, 389.1, ..., 254.1, 237.7, 252.9], [ 392.3, 385.4, 433.4, ..., 190.3, 183.3, 209.8]]) >>> t = time.time ); m = reduce add,z) z.shape[0]; print time.time ) t 1.00891780853 >>> m array [[ 142.4, 121.9, 108.8, ..., 311.9, 279.9, 193.2], [ 184.4, 173.9, 181.9, ..., 293.1, 280.4, 179.8], [ 206.5, 174.1, 190.9, ..., 284.2, 284.7, 190.7], ..., [ 365.2, 349.9, 406.3, ..., 279.7, 255.3, 276.4], [ 373. , 418.6, 389.1, ..., 254.1, 237.7, 252.9], [ 392.3, 385.4, 433.4, ..., 190.3, 183.3, 209.8]]) 2.
|
|
Std Dev is much slower than calculating by hand >>> t = time.time ); s = z.std 0); print time.time ) t 16.9463210106 >>> s array [[ 14.90771612, 11.42322196, 8.87468309, ..., 24.65542537, 26.60996054, 19.55914109], [ 16.13815355, 11.81058847, 10.22203502, ..., 17.3634674 , 18.25486237, 13.99142595], [ 10.92931837, 11.86970935, 12.16100325, ..., 16.22837022, 13.55027675, 12.09173271], ..., [ 21.27345764, 16.87275911, 23.2467202 , ..., 18.83640093, 17.51028269, 31.8031445 ], [ 24.95195383, 23.26886332, 22.26858774, ..., 14.54269576, 16.81695573, 29.48711583], [ 34.40363353, 31.65185619, 33.29324256, ..., 16.14341971, 15.36912489, 22.18017132]]) >>> t = time.time ); m = reduce add,z) z.shape[0]; s = sqrt reduce add,z z) 1. z.shape[0]) m m); print time.time ) t 2.93472313881 >>> s array [[ 14.90771612, 11.42322196, 8.87468309, ..., 24.65542537, 26.60996054, 19.55914109], [ 16.13815355, 11.81058847, 10.22203502, ..., 17.3634674 , 18.25486237, 13.99142595], [ 10.92931837, 11.86970935, 12.16100325, ..., 16.22837022, 13.55027675, 12.09173271], ..., [ 21.27345764, 16.87275911, 23.2467202 , ..., 18.83640093, 17.51028269, 31.8031445 ], [ 24.95195383, 23.26886332, 22.26858774, ..., 14.54269576, 16.81695573, 29.48711583], [ 34.40363353, 31.65185619, 33.29324256, ..., 16.14341971, 15.36912489, 22.18017132]]) 3.
|
|
Median slower than sorting and taking middle value >>> t = time.time ); m = median z); print time.time ) t 18.8376431465 >>> m array [[ 138. , 120. , 106.5, ..., 304. , 273. , 188.
|
|
], [ 179. , 169. , 180.5, ..., 287. , 274.5, 176.5], [ 202. , 169.5, 188. , ..., 280. , 281. , 188.
|
|
], ..., [ 360.5, 344. , 398.5, ..., 274.5, 248. , 265.
|
|
], [ 366.5, 410.5, 381.5, ..., 249.5, 233.5, 242.5], [ 382. , 374.5, 422.5, ..., 187. , 179. , 203. ]])
|
|
>>> from arraymedian import >>> t = time.time ); m = arraymedian z,axis= Y ); print time.time ) t 11.0808348656 >>> m array [[ 138. , 120. , 106.5, ..., 304. , 273. , 188.
|
|
], [ 179. , 169. , 180.5, ..., 287. , 274.5, 176.5], [ 202. , 169.5, 188. , ..., 280. , 281. , 188.
|
|
], ..., [ 360.5, 344. , 398.5, ..., 274.5, 248. , 265.
|
|
], [ 366.5, 410.5, 381.5, ..., 249.5, 233.5, 242.5], [ 382. , 374.5, 422.5, ..., 187. , 179. , 203. ]])
|
|
My arraymedian method simply sorts and takes the middle value.
|
|
I can send it to you if you wish.
|
|
I have also noticed that doing this is still a factor of 2.5 slower than IDL s median method.
|
|
Perhaps the built in sort is not optimized to do a quick sort?
|
|
Thanks.
|
|
Craig warner astro.ufl.edu","bug"
|
|
"6013","1116","a.var ), a.std ), and a.mean ) return incorrect type when no axis specified Trac 518).
|
|
Original ticket http projects.scipy.org numpy ticket 518 on 2007 05 12 by charris, assigned to unknown.
|
|
The dtype keyword has no effect when the {variance,std,mean} is computed using the default axis In [28] eye 2).var dtype=float32).dtype Out[28] dtype float64 ) It does work when the axis is specified.
|
|
In [29] eye 2).var axis=0,dtype=float32).dtype Out[29] dtype float32 ) Integer types are always ignored In [35] eye 2).var axis=0,dtype=int).dtype Out[35] dtype float64 ) The question here is whether int specifies the accumulator type, which should probably raise an error, or the return type for values computed using float64.","bug"
|
|
"6013","1117","allclose fails with inf Trac 519).
|
|
Original ticket http projects.scipy.org numpy ticket 519 on 2007 05 14 by trac user qwerty, assigned to unknown.
|
|
allclose x,y, rtol, atol) | x y | < atol rtol | y | This formula is good so long as x and y do not contain inf inf) at the same index.
|
|
In such a case, the difference is not well defined.
|
|
The problem then, is that allclose fails and returns False when it should be returning True.
|
|
>>> a = array [log 0)]) >>> b = array [log 0)]) >>> a == b array [True], dtype=bool) >>> allclose a,b) False alltrue a==b) will work if you have an array of inf...but this is definitely undesirable when you comparing floats.
|
|
Somehow, it would be nice if the following returned True instead of False.
|
|
>>> a = array [log 0), log .5)]) >>> b = array [log 0), log .5000000001)]) >>> allclose a,b) False Eck.
|
|
Here is an uninformed solution pruned x = [] pruned y = [] for xx, yy in zip x,y) if xx != yy pruned x.append xx) pruned y.append yy) return old allclose array pruned x), array pruned y)) This means we d need an additional run through the values, but this certainly makes results more predictable...and prevents the operator from having to check for cases like these.","bug"
|
|
"6013","1118","intel compiler package uses arch specific flags in get flags linker so that can t be overridden Trac 520).
|
|
Original ticket http projects.scipy.org numpy ticket 520 on 2007 05 14 by mbeachy, assigned to unknown.
|
|
I ve been trying to set up configuration files to standardize a local 1.02 numpy installation and have run into a problem with the intel compiler package.
|
|
If I try `python setup.py config fc fcompiler=intel build ext` from the top level numpy 1.0.2 directory) I get the following failures with ifort 8.1) ifort Command line warning extension M not supported ignored in option x ifort Command line error Unrecognized keyword SSE2 for option arch ifort Command line warning extension M not supported ignored in option x ifort Command line error Unrecognized keyword SSE2 for option arch I assumed that the noarch option to the config fc command would get me around this, but it does not.
|
|
This seems to be because the get flags linker so method in the numpy.distutils.fcompiler.intel.IntelFCompiler class always extends the options with the result of self.get flags arch ).
|
|
Removing that call fixes the problem.
|
|
I don t think that having architecture specific flags in the shared linker options even matters, does it?
|
|
It would be nice if we could just get rid of the self.get flags arch ) call in get flags linker so for the 1.03 release.","bug"
|
|
"6013","1119","add stacklevel=2 to ScipyTest DeprecationWarning Trac 521).
|
|
Original ticket http projects.scipy.org numpy ticket 521 on 2007 05 14 by mbeachy, assigned to unknown.
|
|
Please add stacklevel=2 to the !ScipyTest !DeprecationWarning in numpy testing numpytest.py so the deprecation warning is for the user of !ScipyTest, not the numpytest module.
|
|
This should be safe and easy and could presumably target the 1.03 release.","bug"
|
|
"6013","1120","Compiling extension modules libraries should use individual compilers Trac 522).
|
|
Original ticket http projects.scipy.org numpy ticket 522 on 2007 05 15 by pearu, assigned to pearu.
|
|
When a setup.py defines two extension modules, one is pure C and the other one is C , then C compiler is used to compile both extension modules.
|
|
This feature is harmless in the world of C C ) and inherited from Python std distutils.
|
|
When dealing with extensions with Fortran sources then this feature should be changed compiling and linking extension modules should use minimal compiler combination.
|
|
Note that if there exists an extension module which uses Fortran codes then also all other extension modules, no matter if they contain only C or C sources, are linked with a Fortran compiler.
|
|
This will cause problems when setup.py defines C based and Fortran based extension modules at the same time.
|
|
Similar problems may appear when there is a choice to made between F77 or F90 compilers.
|
|
So, I am planning to fix this by using minimal compiler set for each extension module based on the language keyword.","bug"
|
|
"6013","1121","Tiny bug causes compilation on Cygwin to fail Trac 523).
|
|
Original ticket http projects.scipy.org numpy ticket 523 on 2007 05 17 by mdehoon, assigned to unknown.
|
|
In numpy numarray capi.c, for a number of platforms fenv.h or fenv.c is included twice.
|
|
Lines 227 232 in current source if defined GLIBC ) || defined APPLE ) || defined CYGWIN ) || defined MINGW32 ) if defined GLIBC ) || defined APPLE ) || defined MINGW32 ) include <fenv.h> elif defined CYGWIN ) include numpy fenv fenv.c endif and further down in lines 2940 2945 elif defined GLIBC ) || defined APPLE ) || defined CYGWIN ) || defined MINGW32 ) if defined GLIBC ) || defined darwin) || defined MINGW32 ) include <fenv.h> elif defined CYGWIN ) include numpy fenv fenv.h endif In numpy release 1.0.2, the second include for CYGWIN actually reads include numpy fenv fenv.c so it includes fenv.c instead of fenv.h twice.
|
|
This causes the somewhat cryptic error message In file included from numpy numarray capi.c 2945 numpy core include numpy fenv fenv.c 32 error redefinition of fe dfl env numpy core include numpy fenv fenv.c 32 error previous definition of fe dfl env was here In the latest revision of numpy numarray capi.c, the second include reads fenv.h instead of fenv.c.
|
|
This appears to be a lucky typo, which actually allows the compilation to proceed.
|
|
Nevertheless, the second set of include s seem to be unnecessary.","bug"
|
|
"6013","1122","Scalar arithmetic with arrays produces values that masquerade as a particular type but are not Trac 524).
|
|
Original ticket http projects.scipy.org numpy ticket 524 on 2007 05 21 by davidsocha, assigned to unknown.
|
|
Multiplication and addition of scalars with arrays produces values that say they are of a particular type but are not actually of that type.
|
|
Here s an example Microsoft Windows XP [Version 5.1.2600] C) Copyright 1985 2001 Microsoft Corp. C \Documents and Settings\socha>python Python 2.4.3 Enthought Edition 1.0.0 69, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.0.2 >>> from numpy import array >>> a = array [1000]) >>> b = 1000 array [1]) >>> type a[0]) == type b[0]) False >>> type a[0]) <type numpy.int32 > >>> type b[0]) <type numpy.int32 > >>> from numpy import int32 >>> type a[0]) == int32 True >>> type b[0]) == int32 False >>> b = array [1000]) >>> type b[0]) == int32 True >>> type a[0]) == type b[0]) True >>> c = 1000 array [1]) also happens for >>> c array [1001]) >>> type c[0]) == int32 False >>> Daniel Terhorst and David Socha [http www.urbansim.org UrbanSim])","bug"
|
|
"6013","1123","Inconsistency in the length of buffers of unicode scalars Trac 525).
|
|
Original ticket http projects.scipy.org numpy ticket 525 on 2007 05 22 by FrancescAlted, assigned to unknown.
|
|
I think there is an inconsistency here >>> import sys >>> sys.maxunicode 65535 using python with UCS2 here >>> u=numpy.unicode popo ) >>> u.data <read only buffer for 0x82276e0, size 16, offset 0 at 0xb7dc3780> >>> len u.data) 8 should be 16!
|
|
>>> u=numpy.array u popo ) >>> u.data <read write buffer for 0x82454d0, size 16, offset 0 at 0xb7dc3780> >>> len u.data) 16 This works fine for 0 dim arrays This prevents thinks like >>> numpy.ndarray buffer=numpy.unicode popo ), dtype= uint32 , shape=4) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError buffer is too small for requested array to work correctly.
|
|
However, the above works well in UCS4 Python interpreters >>> import sys >>> sys.maxunicode 1114111 >>> import numpy >>> u=numpy.unicode popo ) >>> u.data <read only buffer for 0x8203b60, size 16, offset 0 at 0xb7283640> >>> len u.data) 16 >>> numpy.ndarray buffer=u, dtype= uint32 , shape=4) array [112, 111, 112, 111], dtype=uint32) Works fine Thanks","bug"
|
|
"6013","1124","g95 support is broken Trac 526).
|
|
Original ticket http projects.scipy.org numpy ticket 526 on 2007 05 24 by trac user bgranger, assigned to pearu.
|
|
When using g95 as the fortran compiler, numpy won t build.
|
|
This is because the g95 fortran compiler doesn t set the linker exe key of the executables dict.
|
|
Thus you get a KeyError running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize G95FCompiler Traceback most recent call last) File setup.py , line 89, in <module> setup package ) File setup.py , line 82, in setup package configuration=configuration ) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils core.py , line 174, in setup return old setup new attr) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils core.py , line 151, in setup dist.run commands ) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils command install.py , line 16, in run r = old install.run self) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils command install.py , line 506, in run self.run command build ) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils command build.py , line 112, in run self.run command cmd name) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File home2 user work txp txpython 0.2 linux local lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils command build ext.py , line 113, in run self.fcompiler.customize self.distribution) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils fcompiler init .py , line 351, in customize linker exe flags = self.
|
|
get flags self.get flags linker exe, LDFLAGS ) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils fcompiler init .py , line 533, in get flags var = command ) File home2 user work txp txpython 0.2 linux spkg build numpy 1.0.3.dev3673 numpy distutils fcompiler init .py , line 203, in get flags linker exe if self.executables[ linker exe ] KeyError linker exe The current syntax in `fcompiler init .py`, in `get flags linker exe` does not protect against this error.
|
|
My guess is that `linker exe` should simply be set properly in the g95 fcompiler class.","bug"
|
|
"6013","1125","fortran linking flag option... Trac 527).
|
|
Original ticket http projects.scipy.org numpy ticket 527 on 2007 05 25 by trac user fred, assigned to pearu.
|
|
Hi, Could you add some option to f2py cmdline such as f90ldflags= ... to take in account some linking parameters ?
|
|
Thanks in advance.
|
|
Cheers,","enhancement"
|
|
"6013","1126","select changes choicelist as side effect Trac 528).
|
|
Original ticket http projects.scipy.org numpy ticket 528 on 2007 05 28 by trac user markmat, assigned to unknown.
|
|
select ) changes choicelist as side effect.
|
|
Example >>> condlist = [zeros 4,6), int) for i in range 3)] >>> choicelist = [zeros 4,6)) for i in range 3)] >>> print len condlist), len choicelist) 3 3 >>> result = numpy.select condlist, choicelist) >>> print len condlist), len choicelist) 3 4 I running Windows XP, Python 2.4.3, Numpy 1.0.2","bug"
|
|
"6013","1127","Changeset r3830 ended up in the wrong place Trac 529).
|
|
Original ticket http projects.scipy.org numpy ticket 529 on 2007 05 29 by alberts, assigned to unknown.
|
|
Changeset r3830 was checking against tags 1.0.3 instead of against trunk.","enhancement"
|
|
"6013","1128","SciPy Windows build with Intel Visual Fortran doesn t work Trac 530).
|
|
Original ticket http projects.scipy.org numpy ticket 530 on 2007 05 31 by alberts, assigned to dmcooke.
|
|
!NumPy build on Windows with MSVC is working now.
|
|
Next up, !SciPy build with with Intel Visual Fortran.
|
|
Command I m using to build python setup.py v config compiler=msvc fcompiler=intelv build clib compiler=msvc fcompiler=intelv build ext compiler=msvc fcompiler=intelv bdist wininst site.cfg contents in !NumPy root directory when I built that [mkl] include dirs = C \Program Files\Intel\MKL\9.0\include library dirs = C \Program Files\Intel\MKL\9.0\ia32\lib mkl libs = mkl c, libguide40 lapack libs = mkl lapack Build output when compiling SciPy attached.
|
|
Looks like it s struggling with a path name that contains a space.","bug"
|
|
"6013","1129","SciPy Windows build with Intel Visual Fortran doesn t work Trac 531).
|
|
Original ticket http projects.scipy.org numpy ticket 531 on 2007 05 31 by alberts, assigned to dmcooke.
|
|
!NumPy build on Windows with MSVC is working now.
|
|
Next up, !SciPy build with with Intel Visual Fortran.
|
|
Command I m using to build python setup.py v config compiler=msvc fcompiler=intelv build clib compiler=msvc fcompiler=intelv build ext compiler=msvc fcompiler=intelv bdist wininst site.cfg contents in !NumPy root directory when I built that [mkl] include dirs = C \Program Files\Intel\MKL\9.0\include library dirs = C \Program Files\Intel\MKL\9.0\ia32\lib mkl libs = mkl c, libguide40 lapack libs = mkl lapack Build output when compiling SciPy attached.
|
|
Looks like it s struggling with a path name that contains a space.","bug"
|
|
"6013","1130","Invalid numpy.trace docstring Trac 532).
|
|
Original ticket http projects.scipy.org numpy ticket 532 on 2007 06 01 by pv, assigned to unknown.
|
|
The numpy.trace docstring is invalid, trace actually sums over the diagonal in the first two axes by default import numpy as N print N. version N.info N.trace) x = N.zeros 4,4,5,6)) print N.trace x).shape x = N.zeros 5,6,4,4)) print N.trace x).shape outputs 1.0.3 trace a, offset=0, axis1=0, axis2=1, dtype=None, out=None) trace a,offset=0, axis1=0, axis2=1) returns the sum along diagonals defined by the last two dimenions) of the array.
|
|
5, 6) 4, 4)","bug"
|
|
"6013","1131","numpy.trace is slow for small dimensions Trac 533).
|
|
Original ticket http projects.scipy.org numpy ticket 533 on 2007 06 02 by pv, assigned to pv.
|
|
numpy.trace is slow for tracing over small dimensions import numpy as N import timeit print N. version print timeit.Timer x.trace axis1=1, axis2=2) , import numpy as N; x = N.zeros 1000, 4, 4)) ).timeit 500) print timeit.Timer x[...,0,0] x[...,1,1] x[...,2,2] x[...,3,3] , import numpy as N; x = N.zeros 1000, 4, 4)) ).timeit 500) outputs 1.0.3 3.793227911 0.0330259799957 It d be nice if the performance could be improved there seems to be room for this, as the difference is 100x.","enhancement"
|
|
"6013","1132","build problem on Intel Mac running OSX 10.4 Trac 534).
|
|
Original ticket http projects.scipy.org numpy ticket 534 on 2007 06 03 by chanley, assigned to unknown.
|
|
I cannot build numpy r3857 on my Intel MacBook running OS X 10.4.
|
|
It appears that distutils is looking for a FORTRAN compiler which should not be required to build numpy.
|
|
Build log is attached to the ticket.","bug"
|
|
"6013","1133","distutils places extra files in installation directory Trac 535).
|
|
Original ticket http projects.scipy.org numpy ticket 535 on 2007 06 05 by chanley, assigned to dmcooke.
|
|
When installing numpy, the COMPATIBILITY , scipy compatibility , and site.cfg.example files are placed in site packages or whatever the installation target might be).
|
|
[familyfun site packages lib python] chanley ls al total 40 drwxr xr x 7 chanley chanley 238 Jun 5 09 14 .
|
|
drwxr xr x 3 chanley chanley 102 May 2 15 24 .. rw r r 1 chanley chanley 1620 Apr 22 15 51 COMPATIBILITY drwxr xr x 36 chanley chanley 1224 Jun 5 09 14 numpy rw r r 1 chanley chanley 1619 Jun 5 09 14 numpy 1.0.4.dev3859 py2.5.egg info rw r r 1 chanley chanley 123 Jan 11 10 32 scipy compatibility rw r r 1 chanley chanley 4613 Apr 16 10 33 site.cfg.example [familyfun site packages lib python] chanley This is true for all platforms I have tested Solaris 10, Mac OSX 10.4.9, RHE3).
|
|
I don t think this is intended.
|
|
Chris","bug"
|
|
"6013","1134","importing numpy crashes python if python is built with COUNT ALLOCS Trac 536).
|
|
Original ticket http projects.scipy.org numpy ticket 536 on 2007 06 12 by cournape, assigned to unknown.
|
|
I tried to build numpy with a special built python using COUNT ALLOCS).
|
|
Unfortunately, importing numpy crashes the interpreter.
|
|
It looks like the C function import array is where the crashs happens, when importing umath.so","bug"
|
|
"6013","1135","Compatibility issues with old Numeric Trac 537).
|
|
Original ticket http projects.scipy.org numpy ticket 537 on 2007 06 13 by trac user multijon, assigned to unknown.
|
|
Hi, I ve been using the Python Gnuplot interface a few months now, and have noticed there is a strange discrepancy between the numbers I provide it with and the results displayed Needed background Gnuplot.py hasn t been updated for the past couple of years, and is using old Numeric in order to convert input data into a standard input type).
|
|
Today, after diving into its source code, I found out that Gnuplot is trying to cast a float first to float32, and if it doesn t succeed, it then casts the input data into float64.
|
|
My guess is that old Numeric would have raised an Exception when trying to cast a number larger than 32bits into a 32 bit data type, a feature that isn t available in the current version of Numeric.
|
|
This, of course, should be fixed for compatibility s sake.
|
|
Jon","bug"
|
|
"6013","1136","numpy r3871 build problem Trac 538).
|
|
Original ticket http projects.scipy.org numpy ticket 538 on 2007 06 15 by chanley, assigned to dmcooke.
|
|
The latest numpy build fails on my RHE3 machine using the default gnu compilers.
|
|
The build log is attached.","bug"
|
|
"6013","1137","config parameters are not passed on to build Trac 539).
|
|
Original ticket http projects.scipy.org numpy ticket 539 on 2007 06 26 by pv, assigned to cournape.
|
|
numpy 1.0.3 and dev.r3875), Python 2.5 on Linux) Propagation of some command line parameters given to the `config` and `config fc` commands to `build` appear not to work as intended in `setup.py config help`) `.
|
|
setup.py config I somepath build` ` I` is not passed to the C compiler gcc).
|
|
Ditto for ` l` and ` L` `.
|
|
setup.py config Dfoo=bar build` results to error `error error in command line command config has no such option define ` Ditto for ` Ufoo` For `build ext` the flags ` I` and ` L` appear to work, but flags ` D`, ` U`, ` l`, ` R` are not recognized although they appear in the help text.
|
|
Instead, `error option l not recognized` Also, `build ext` accepts the switches of the commands `build src` and `config` which is slightly confusing.
|
|
For `build clib`, there appears to be no way to pass the parameters ` I` and ` L` the parameters from `config` are ignored, and `build clib` doesn t accept them.
|
|
For `config fc fcompiler=gnu95`, the parameters ` noopt`, ` noarch` and ` debug` appear to have no effect on the resulting compiler flags used, I always get ` O3 funroll loops mmmx m3dnow msse` The above occurs on the following setup.py, which AFAIK is correct !
|
|
usr bin env python from numpy.distutils.misc util import Configuration def configuration parent package= , top path=None) config = Configuration xxx , parent package, top path, package path= xxx , ) config.add library a , sources=[ xxx a.c ]) config.add extension b , sources=[ xxx b.c ], libraries=[ a ]) config.add library c , sources=[ xxx c.f90 ]) return config if name == main from numpy.distutils.core import setup setup configuration top path= ).todict )) I think I had also some problems with the compiler parameter on Windows needed to specify it multiple times both for config, build clib and build ext), but I ll need to verify this later.","bug"
|
|
"6013","1138","Merge sort doesn t work correctly for string arrays.
|
|
Trac 540).
|
|
Original ticket http projects.scipy.org numpy ticket 540 on 2007 06 27 by charris, assigned to unknown.
|
|
Merge sort doesn t do a stable sort on string arrays.
|
|
In [3] charArr = numpy.array [ a , a , a , b , b , b , c , d ]) In [12] numpy.argsort charArr, kind= m ) Out[12] array [2, 1, 0, 5, 4, 3, 6, 7]) The result should be array [0, 1, 2, 3, 4, 5, 6, 7])","bug"
|
|
"6013","1139","int division by zero give wrong results Trac 541).
|
|
Original ticket http projects.scipy.org numpy ticket 541 on 2007 06 28 by trac user chipschips, assigned to unknown.
|
|
integer division by zero returns 0 rather then inf or some other specially defined integer infinity constant).
|
|
This behavior is not consistent with floats and is mathematically wrong.
|
|
Example >>> 1 int 0) 0 >>> 1 float 0) inf Is this a bug or there is some good reason why ints behave this way?
|
|
I m using numpy 1.0.1 and scipy 0.5.2","bug"
|
|
"6013","1140","int division by zero gives wrong results Trac 542).
|
|
Original ticket http projects.scipy.org numpy ticket 542 on 2007 06 28 by trac user chipschips, assigned to unknown.
|
|
integer division by zero returns 0 rather then inf or some other specially defined integer infinity constant).
|
|
This behavior is not consistent with floats and is mathematically wrong.
|
|
Example >>> 1 int 0) 0 >>> 1 float 0) inf Is this a bug or there is some good reason why ints behave this way?
|
|
I m using numpy 1.0.1 and scipy 0.5.2","bug"
|
|
"6013","1141","documentation string for cumproduct is wrong Trac 543).
|
|
Original ticket http projects.scipy.org numpy ticket 543 on 2007 07 01 by tovrstra, assigned to unknown.
|
|
Hi, This is just an obvious type oh, nothing to be really worried about Help on function cumproduct in module numpy.core.fromnumeric cumproduct x, axis=None, dtype=None, out=None) Sum the array over the given axis.
|
|
>>> numpy.
|
|
version 1.0.2 thanks for all the nice work so far!
|
|
Toon","bug"
|
|
"6013","1142","typo in arrayprint.py added in r3877 Trac 544).
|
|
Original ticket http projects.scipy.org numpy ticket 544 on 2007 07 02 by chanley, assigned to chanley.
|
|
A typo in arrayprint r3877 causes the unittests to fail ====================================================================== ERROR Test of conversions and indexing Traceback most recent call last) File data sparty1 dev site packages lib python numpy core tests test ma.py , line 200, in check testCI junk, garbage = str x2), repr x2) File data sparty1 dev site packages lib python numpy core ma.py , line 761, in repr data str self.filled )), File data sparty1 dev site packages lib python numpy core numeric.py , line 473, in array str return array2string a, max line width, precision, suppress small, , , str) File data sparty1 dev site packages lib python numpy core arrayprint.py , line 240, in array2string separator, prefix) File data sparty1 dev site packages lib python numpy core arrayprint.py , line 153, in array2string len str mininum.reduce data)))) NameError global name mininum is not defined","bug"
|
|
"6013","1143","Array of non native numbers don t work correctly Trac 545).
|
|
Original ticket http projects.scipy.org numpy ticket 545 on 2007 07 03 by trac user usovalx, assigned to unknown.
|
|
Consider the following example a = arange 3, dtype= >f ) print a[a.argmax )] == a.max )) For the machines I have tested FC6 32 and 64 bit) this gives False.","bug"
|
|
"6013","1144","Array of non native numbers don t work correctly Trac 546).
|
|
Original ticket http projects.scipy.org numpy ticket 546 on 2007 07 03 by trac user usovalx, assigned to unknown.
|
|
Consider the following example a = arange 3, dtype= >f ) print a[a.argmax )] == a.max ) For the machines I have tested FC6 32 and 64 bit, numpy 1.0.3) this gives False.","bug"
|
|
"6013","1145","numpy distutils system info.py should summarize missing failed dependencies when build fails Trac 547).
|
|
Original ticket http projects.scipy.org numpy ticket 547 on 2007 07 08 by johannct, assigned to dmcooke.
|
|
currently, the script writes out many lines and the user has to scroll back to check what was available not available","enhancement"
|
|
"6013","1146","fromfile crashes due to failed allocation Trac 548).
|
|
Original ticket http projects.scipy.org numpy ticket 548 on 2007 07 10 by alberts, assigned to unknown.
|
|
I m seeing a crash that appears to be due to fromfile or something it calls ignoring a failed allocation.
|
|
Code to reproduce import numpy as N def main ) dtype = N.dtype f4 ) iocount = 6 x = N.zeros 33554432,), dtype=dtype) assert x.nbytes == 128 1024 1024 count = len x) print writing fp = open matrix.out , wb ) for i in xrange iocount) x.tofile fp) fp.close ) del x fp = open matrix.out , rb ) toomuch = [] for i in xrange 100) print i fp.seek 0) y = N.fromfile fp, dtype=dtype, count=5 count) print len y) toomuch.append y) fp.close ) if name == main main ) Output on my Windows machine, which can t handle more than one 640 MB allocation even though it has 2 GB RAM heap fragmentation rocks) writing 0 167772160 1 167772160 items requested but only 0 read Crash dialog python.exe Application Error The instruction at 0x10022439 referenced memory at 0x0000001c .
|
|
The memory could not be read .
|
|
Click on OK to terminate the program Click on CANCEL to debug the program OK Cancel","bug"
|
|
"6013","1147","int ) of numpy scalar fails silently Trac 549).
|
|
Original ticket http projects.scipy.org numpy ticket 549 on 2007 07 10 by trac user zouave, assigned to unknown.
|
|
[posted on numpy discussion on 2007 07 09] The conversion from a numpy scalar to a python int is not consistent with python s native conversion or numarray s) if the scalar is out of bounds for an int, python and numarray automatically create a long while numpy still creates an int... with the wrong value.
|
|
e.g.
|
|
using numpy 1.0.3) Python 2.4.3 2, Apr 27 2006, 14 43 58) [GCC 4.0.3 Ubuntu 4.0.3 1ubuntu5)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> l= [1e3, 1e9, 1e15, 1e3, 1e9, 1e15] >>> a= array l) >>> map int, l) [1000, 1000000000, 1000000000000000L, 1000, 1000000000, 1000000000000000L] >>> map int, a) [1000, 1000000000, 2147483648, 1000, 1000000000, 2147483648] >>> map long, a) [1000L, 1000000000L, 1000000000000000L, 1000L, 1000000000L, 1000000000000000L] >>> IMHO, numpy s conversions to int should behave like Python s float int or long int functions see PYTHON SRC DIR Objects floatobject.c, PYTHON SRC DIR Objects longobject.c) if it doesn t fit in an int, return a long.
|
|
For now svn), it seems that numpy is always using PyInt FromLong after an implicit C cast to long which silently fails; see NUMPY SRC DIR numpy core src scalarmathmodule.c.src)","bug"
|
|
"6013","1148","tofile ) broken on 64 bit linux Trac 550).
|
|
Original ticket http projects.scipy.org numpy ticket 550 on 2007 07 11 by trac user myrvoll, assigned to unknown.
|
|
Numpy compiled on X86 64 Fedora Core 3 SMP system under Python 2.5 and GCC3.4.2 The following breaks numpy 1.0.3 >>> import numpy >>> G = numpy.zeros [30000,30000]) >>> fh = open foo , w ) >>> G.tofile fh) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError 900000000 requested and 0 written","bug"
|
|
"6013","1149","numpy.ndarray messed up after unpickling Trac 551).
|
|
Original ticket http projects.scipy.org numpy ticket 551 on 2007 07 16 by trac user cotackst, assigned to unknown.
|
|
After unpickling a pickled numpy.ndarray matrix using cPickle), I get a segmentation fault when trying to multiply the matrix by a vector.
|
|
However I can multiply a copy of the matrix with the same vector.
|
|
The code to generate the actual matrix is quite large, so I only attach the pickled matrix mat.bug ).
|
|
Reproducing code [[BR]] from cPickle import Unpickler[[BR]] from numpy import ones, dot[[BR]] bug mat = Unpickler file mat.bug , r )).load )[[BR]] ok mat = bug mat.copy )[[BR]] v = ones 1,bug mat.shape[0]))[[BR]] dot v,ok mat) this works[[BR]] dot v,bug mat) this segfault[[BR]]","bug"
|
|
"6013","1150","Bad import in oldnumeric random array.py Trac 552).
|
|
Original ticket http projects.scipy.org numpy ticket 552 on 2007 07 17 by trac user ColdWind, assigned to unknown.
|
|
oldnumeric random array.py has the following import import numpy as Numeric which seems that should be import numpy.oldnumeric as Numeric If not, it complains about ArrayType.
|
|
To reproduce execute python c from numpy.oldnumeric.random array import ; randint 0, 50, [2,3])","bug"
|
|
"6013","1151","poly1d division error Trac 553).
|
|
Original ticket http projects.scipy.org numpy ticket 553 on 2007 07 23 by trac user DrIdiot, assigned to unknown.
|
|
Using ipython pylabs, I get the error when dividing two polynomials In [5] polydiv numpy.poly1d [1,2,3]), numpy.poly1d [1,2,3,4,5])) exceptions.NameError Traceback most recent call last) home harrison <console> NameError name polydiv is not defined I am running Gentoo Linux and numpy version 1.0.3","bug"
|
|
"6013","1152","poly1d equate error Trac 554).
|
|
Original ticket http projects.scipy.org numpy ticket 554 on 2007 07 24 by trac user DrIdiot, assigned to unknown.
|
|
When I try to see if two polynomials are equal, I get an error In [13] numpy.poly1d [1, 1, 1, 1,0,1]) == numpy.poly1d [1,0]) exceptions.AttributeError Traceback most recent call last) home harrison <console> usr lib python2.4 site packages numpy lib polynomial.py in eq self, other) 601 602 def eq self, other) > 603 return self.coeffs == other.coeffs).all ) 604 605 def ne self, other) I am running Gentoo with numpy 1.0.3","bug"
|
|
"6013","1153","setting random seed does not work with a numpy.int64 Trac 555).
|
|
Original ticket http projects.scipy.org numpy ticket 555 on 2007 07 24 by trac user scottzelak, assigned to unknown.
|
|
The following code fails for loop num in arange 4) random.seed loop num) Traceback most recent call last) File <stdin> , line 2, in <module> File mtrand.pyx , line 515, in mtrand.RandomState.seed ValueError object of too small depth for desired array but will work if I replace the command in the loop with [[BR]] random.seed int loop num))","bug"
|
|
"6013","1154","polydiv function gets the wrong answer!!
|
|
Trac 556).
|
|
Original ticket http projects.scipy.org numpy ticket 556 on 2007 07 24 by trac user subheight640, assigned to unknown.
|
|
the numpy.polydiv function does not seem to work for all cases.
|
|
For example, >>> numpy.polydiv [7,6,5,4,3,2,1],[6,5,4,3,2,1]) array [ 1.16666667, 0.
|
|
]), array [1])) The answer I got from two other polynomial division programs, my own function, and by working the problem out by hand is >>> poly divider [7,6,5,4,3,2,1],[6,5,4,3,2,1]) [1.1666666666666667, 0.027777777777777679], [0.194444444444, 0.388888888889, 0.583333333333, 0.777777777778, 0.972222222222]) polydiv seems to suffer from rounding errors.","bug"
|
|
"6013","1155","pinv sefgaults with empty arrays Trac 557).
|
|
Original ticket http projects.scipy.org numpy ticket 557 on 2007 07 26 by cournape, assigned to unknown.
|
|
This is a copy of ticket 381 from scipy When I create an empty matrix and ask for the .I attribute, scipy crashes.
|
|
I think this operation should raise an exception not crash scipy.
|
|
Details are below.
|
|
Python 2.5 r25 51908, Sep 19 2006, 09 52 17) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import scipy >>> scipy.version.version 0.5.2 >>> scipy.version.release True >>> x = scipy.matrix []) >>> x.I On entry to DGESDD parameter number 10 had an illegal value","bug"
|
|
"6013","1156","axis support for numpy.median ) Trac 558).
|
|
Original ticket http projects.scipy.org numpy ticket 558 on 2007 07 26 by cournape, assigned to unknown.
|
|
median does not support axis argument.
|
|
Maybe we can simply use the scipy version ?","enhancement"
|
|
"6013","1157","var ) on a size=1 array should raise exception.
|
|
Trac 559).
|
|
Original ticket http projects.scipy.org numpy ticket 559 on 2007 07 30 by trac user gpk, assigned to unknown.
|
|
Properly, a single value does not have a variance.
|
|
Whoa!, you might say I can compute the sum of the squares of the distance from the mean, even when there is only one value.
|
|
It s zero.
|
|
That s true, but misleading.
|
|
Since you call it var , you are implying that it is a variance, and variance is undefined unless you have two or more values.
|
|
More importantly, even the biassed estimator of variance sum x xbar) 2) N ) does not have a well defined value when N==1.
|
|
All estimates of the variance are equally good or equally bad when you have only one value.
|
|
Follow the logic to compute the minimum variance estimator of variance for a single sample, and you ll see all kinds of absurdities, including characteristic functions for the probability distribution that do not have a limiting value.
|
|
From a practical point of view, anyone who takes the variance of a single value is probably in deep trouble.
|
|
Variances tend to get used for F tests, which are undefined for zero degrees of freedom.
|
|
The get used to build confidence intervals for t tests, but the t distribution is undefined for zero degrees of freedom.
|
|
Some people will construct confidence intervals via [mean 3 sqrt var), mean 3 sqrt var)] or some similar approximation this will fail badly and lead to tears.
|
|
And, the same logic applies to the std ), of course.
|
|
>>> import numpy >>> x = numpy.array [1.0]) >>> x array [ 1.])
|
|
>>> x.var ) 0.0 >>>","bug"
|
|
"6013","1158","asfarray applied to a matrix returns a matrix instead of an array Trac 560).
|
|
Original ticket http projects.scipy.org numpy ticket 560 on 2007 08 01 by nilswagner01, assigned to unknown.
|
|
>>> from numpy import >>> asfarray matrix [[0.3]])) matrix [[ 0.3]]) >>> array matrix [[0.3]])) array [[ 0.3]]) >>> numpy.
|
|
version 1.0.4.dev3937","bug"
|
|
"6013","1159","problem compiling, cannot find multiarray Trac 561).
|
|
Original ticket http projects.scipy.org numpy ticket 561 on 2007 08 03 by johannct, assigned to unknown.
|
|
I just tried to compile SVN version of numpy, and I get [cohen localhost numpy svn] python setup.py build Traceback most recent call last) File setup.py , line 90, in <module> setup package ) File setup.py , line 62, in setup package from numpy.distutils.core import setup File data1 sources python numpy svn numpy init .py , line 39, in <module> import core File data1 sources python numpy svn numpy core init .py , line 8, in <module> import numerictypes as nt File data1 sources python numpy svn numpy core numerictypes.py , line 83, in <module> from numpy.core.multiarray import typeinfo, ndarray, array, empty, dtype ImportError No module named multiarray I have a hard time believing that this is a bug, but rather that something is wrong in my setup.... though I can t find what right now.
|
|
thanks","bug"
|
|
"6013","1160","Reference leaks Trac 562).
|
|
Original ticket http projects.scipy.org numpy ticket 562 on 2007 08 03 by alberts, assigned to unknown.
|
|
Lisandro Dalcin reported some possible reference leaks http projects.scipy.org pipermail numpy discussion 2007 August 028749.html http projects.scipy.org pipermail numpy discussion 2007 August 028757.html","bug"
|
|
"6013","1161","Python 2.6 DeprecationWarnings Trac 563).
|
|
Original ticket http projects.scipy.org numpy ticket 563 on 2007 08 04 by alberts, assigned to jarrodmillman.
|
|
The following !DeprecationWarnings are emitted when building !NumPy with Python 2.6 from trunk numpy distutils misc util.py 1279 !DeprecationWarning os.popen4 is deprecated.
|
|
Use the subprocess module.
|
|
numpy core code generators genapi.py 9 !DeprecationWarning the md5 module is deprecated; use hashlib instead","bug"
|
|
"6013","1162","poly1d subtraction yields coefficient of 0 Trac 564).
|
|
Original ticket http projects.scipy.org numpy ticket 564 on 2007 08 06 by trac user DrIdiot, assigned to unknown.
|
|
Using ipython pylabs, explicitly setting the leading coefficient to 0 does this In [3] a = numpy.poly1d [1,2,3]) In [4] a[2] = 0 In [5] print a 2 x 3 In [6] a.coeffs Out[6] array [0, 2, 3]) Likewise, In [10] a = numpy.poly1d [1,2,3]) In [11] a[2] = 0 In [12] a[1] = 0 In [13] print a 3 In [14] a.coeffs Out[14] array [0, 0, 3]) Likewise, In [15] a = numpy.poly1d [1,2,3]) In [16] a[2] = 0 In [17] a[1] = 0 In [18] a[0] = 0 In [19] print a In [20] a.coeffs Out[20] array [0, 0, 0]) The correct behavior is ) In [22] a = numpy.poly1d [0]) In [23] print a 0 I am running Gentoo Linux and numpy version 1.0.3","bug"
|
|
"6013","1163","numpy cannot be built when scipy is installed Trac 565).
|
|
Original ticket http projects.scipy.org numpy ticket 565 on 2007 08 06 by johannct, assigned to unknown.
|
|
hi, each time I update numpy from svn and try to build it, it gives an error Traceback most recent call last) File setup.py , line 90, in <module> setup package ) File setup.py , line 62, in setup package from numpy.distutils.core import setup File data1 sources python numpy svn numpy init .py , line 39, in <module> import core File data1 sources python numpy svn numpy core init .py , line 8, in <module> import numerictypes as nt File data1 sources python numpy svn numpy core numerictypes.py , line 83, in <module> from numpy.core.multiarray import typeinfo, ndarray, array, empty, dtype ImportError No module named multiarray This error disappears, and the build stage proceeds smoothly, when I remove the installed scipy from the python site package install area..... Is there a way to avoid this impractical situation?
|
|
Am I doing smthg wrong?","enhancement"
|
|
"6013","1164","lexsort fails to completely sort Trac 566).
|
|
Original ticket http projects.scipy.org numpy ticket 566 on 2007 08 10 by trac user hattawayd, assigned to unknown.
|
|
I am sorting a large array with lexsort, and while it seems to sort, the sort is not complete, I still have to sort it again to be finished.
|
|
I will try to attach my test case once I get permission to release the data.","bug"
|
|
"6013","1165","Aligned allocator for numpy Trac 567).
|
|
Original ticket http projects.scipy.org numpy ticket 567 on 2007 08 11 by cournape, assigned to unknown.
|
|
Here is a first patch to have aligned allocator in numpy this does not give yet a high interface to create arrays with a given alignment, only aligned allocators in the array interface.
|
|
The patch defines aligned allocators Make PyDataMem macro to point to the aligned allocator with a default alignment fixed to 16 bytes) Define new macro PyDataAligned to align buffer with arbitrary alignment","enhancement"
|
|
"6013","1166","Aligned allocator for numpy Trac 568).
|
|
Original ticket http projects.scipy.org numpy ticket 568 on 2007 08 11 by cournape, assigned to unknown.
|
|
Here is a first patch to have aligned allocator in numpy this does not give yet a high interface to create arrays with a given alignment, only aligned allocators in the array interface.
|
|
The patch defines aligned allocators Make PyDataMem macro to point to the aligned allocator with a default alignment fixed to 16 bytes) Define new macro PyDataAligned to align buffer with arbitrary alignment","enhancement"
|
|
"6013","1167",".transpose ) of memmap array fails to close ) Trac 569).
|
|
Original ticket http projects.scipy.org numpy ticket 569 on 2007 08 13 by alberts, assigned to unknown.
|
|
http projects.scipy.org pipermail numpy discussion 2007 August 028851.html","bug"
|
|
"6013","1168","Vectorize leaks Trac 570).
|
|
Original ticket http projects.scipy.org numpy ticket 570 on 2007 08 13 by alberts, assigned to unknown.
|
|
http projects.scipy.org pipermail numpy discussion 2007 August 028860.html","bug"
|
|
"6013","1169","Support universal gfortran compiler on OS X Trac 571).
|
|
Original ticket http projects.scipy.org numpy ticket 571 on 2007 08 16 by trac user jmarshall, assigned to dmcooke.
|
|
Modify numpy distutils to support the universal gfortran found here http r.research.att.com tools This will allow an easy universal SciPy build.
|
|
See email here http projects.scipy.org pipermail numpy discussion 2007 August 028866.html A sample patch is attached.
|
|
Note this will break non universal gfortran on OS X only), so it is just a proof of concept.
|
|
We will need a method to detect whether a universal compiler is available.","enhancement"
|
|
"6013","1170","Memory leak in insert when using rank 0 input Trac 572).
|
|
Original ticket http projects.scipy.org numpy ticket 572 on 2007 08 24 by cournape, assigned to unknown.
|
|
The function insert has a memory leak, and this shows when using for example cdf functions of distributions classes in scipy.stats see scipy 1074).
|
|
The fix is trivial just call Py DECREF on ainput in rank 0 case, as in general case the attached patch avoids separate code path for memory cleaning instead).","bug"
|
|
"6013","1171","numpy does not build when LDFLAGS are set Trac 573).
|
|
Original ticket http projects.scipy.org numpy ticket 573 on 2007 08 24 by trac user bicatali, assigned to unknown.
|
|
On Linux, [[BR]] LDFLAGS= Wl, s python setup.py build will not link properly, at least the cblas libraries.
|
|
This happen with whatever LDFLAGS.
|
|
I ve seen similar posts in lists and tickets, but never a proper answer or resolution.
|
|
The only work around is to build the lite libraries, i.e.
|
|
unset BLAS and LAPACK env variables.
|
|
scipy suffers from it even more.
|
|
[[BR]] Do you plan to add user LDFLAGS for linking?","bug"
|
|
"6013","1172","tri n, dtype=bool) return None Trac 574).
|
|
Original ticket http projects.scipy.org numpy ticket 574 on 2007 08 24 by trac user zz2liu, assigned to stefanv.
|
|
def tri N, M=None, k=0, dtype=float) returns a N by M array where all the diagonals starting from lower left corner up to the k th are all ones.
|
|
if M is None M = N m = greater equal subtract.outer arange N), arange M)), k) if m.dtype != dtype return m.astype dtype) else return m tri 3, dtype=bool) will return None maybe should fix it by adding the block commented to the end of the code.","bug"
|
|
"6013","1173","fast putmask implementation does not work on big endian systems Trac 575).
|
|
Original ticket http projects.scipy.org numpy ticket 575 on 2007 08 25 by chanley, assigned to stefanv.
|
|
The fast putmask implementation introduced in r3981 does not work on big endian machines.
|
|
The record array test added in r3982 confirms this fact.
|
|
====================================================================== FAIL test record array numpy.core.tests.test multiarray.test putmask) Traceback most recent call last) File data basil5 site packages lib python numpy core tests test multiarray.p y , line 450, in test record array assert array equal rec[ x ],[10,5]) File data basil5 site packages lib python numpy testing utils.py , line 223, in assert array equal verbose=verbose, header= Arrays are not equal ) File data basil5 site packages lib python numpy testing utils.py , line 215, in assert array compare assert cond, msg AssertionError Arrays are not equal mismatch 50.0 ) x array [ 4.58492919e 320, 5.00000000e 000]) y array [10, 5]) Ran 673 tests in 13.436s FAILED failures=1) <unittest.TextTestRunner object at 0x565f70> >>> numpy.
|
|
version 1.0.4.dev4011 >>> This test does pass on our Redhat Enterprise systems which our little endian.","bug"
|
|
"6013","1174","numpy installation doesn t detect gfortran or g95 Trac 576).
|
|
Original ticket http projects.scipy.org numpy ticket 576 on 2007 09 06 by trac user gruel, assigned to unknown.
|
|
I m using the last svn version of numpy .
|
|
I tried the solution found on the mail list python setup.py config fc fcompiler=gnu95 build the two interesting part at least for me) in the building process are customize GnuFCompiler Found executable usr bin g77 gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found customize GnuFCompiler gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found customize GnuFCompiler using config and later customize G95FCompiler Found executable usr local bin g95 but I still have the same answer when I m trying to do f2py c m hello hello.f90 customize GnuFCompiler Found executable usr bin g77 gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found customize GnuFCompiler gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found and so error f90 not supported by GnuFCompiler needed for hello.f90 I can t compile my fortran module because of this.
|
|
I ll attach the file obtain with python setup.py v config fc fcompiler=gnu95 build thank you for numpy and f2py.","bug"
|
|
"6013","1175","Return value of sum of masked array is inconsistent Trac 577).
|
|
Original ticket http projects.scipy.org numpy ticket 577 on 2007 09 12 by trac user reggie, assigned to stefanv.
|
|
When calling the sum method of a masked array with no arguments, the result can be either a scalar or a 1 element array depending on whether the mask is nomask or not.
|
|
I m attaching a patch for ma.py below that should fix this problem.","bug"
|
|
"6013","1176","Indexing of multi D arrays is counterintuitive Trac 578).
|
|
Original ticket http projects.scipy.org numpy ticket 578 on 2007 09 13 by trac user gic888, assigned to unknown.
|
|
numpy 1.0.4.dev3869 with Python 2.5 on Mac OS 10.4.9 Intel) Slicing 1 dimension of a 2D array does not have the same effect as using the same slice on a 1D array.
|
|
Also, calling array on the slice generates an array different than the slice, and some slice indexes behave differently than others.
|
|
I will clarify with examples ipy > z=reshape arange 30), 15,2)) ipy > z[3 ,0]=z[ 3,0] ipy > z out array [[ 0, 1], [ 2, 3], [ 4, 5], [ 0, 7], [ 2, 9], [ 4, 11], [ 0, 13], [ 2, 15], [ 4, 17], [ 0, 19], [ 2, 21], [ 4, 23], [ 0, 25], [ 2, 27], [ 4, 29]]) but ipy > z=reshape arange 30), 15,2)) ipy > z[ 3,0]=z[3 ,0] ipy > z out array [[ 6, 1], [ 8, 3], [10, 5], [12, 7], [14, 9], [16, 11], [18, 13], [20, 15], [22, 17], [24, 19], [26, 21], [28, 23], [24, 25], [26, 27], [28, 29]]) and ipy > z=reshape arange 30), 15,2)) ipy > z[3 ,0]=array z[ 3,0]) ipy > z out array [[ 0, 1], [ 2, 3], [ 4, 5], [ 0, 7], [ 2, 9], [ 4, 11], [ 6, 13], [ 8, 15], [10, 17], [12, 19], [14, 21], [16, 23], [18, 25], [20, 27], [22, 29]])","bug"
|
|
"6013","1177","Implement bluestein algorithm for prime size FFT Trac 579).
|
|
Original ticket http projects.scipy.org numpy ticket 579 on 2007 09 14 by cournape, assigned to unknown.
|
|
See 1104 and http en.wikipedia.org wiki Bluestein s FFT algorithm","bug"
|
|
"6013","1178","ExtGen uses Python 2.4 features Trac 580).
|
|
Original ticket http projects.scipy.org numpy ticket 580 on 2007 09 14 by rkern, assigned to jarrodmillman.
|
|
I noticed in a buildbot log that property was being used by !ExtGen.
|
|
As these are Python 2.4 features and we are still targeting Python 2.3, these need to be changed to the ` foo = property foo) ` form.","bug"
|
|
"6013","1179","random.set state does not reset state of random.standard normal Trac 581).
|
|
Original ticket http projects.scipy.org numpy ticket 581 on 2007 09 17 by trac user Yung Chin Oei, assigned to rkern.
|
|
When calling random.set state the function random.standard normal may return an internally retained old number before starting to use the new state of the generator.
|
|
I think this happens because the Box Muller algorithm generates normally distributed numbers in pairs.
|
|
If an uneven number of calls to the function was made before using set state, the second number of such a pair is returned before new numbers are generated.
|
|
Maybe useful for fixing this problem using random.seed 0) does seem to properly reset the internal state of random.standard normal the next call to standard normal always returns the expected new random number.","bug"
|
|
"6013","1180","Crash when embedded using Borland Builder 6 Trac 582).
|
|
Original ticket http projects.scipy.org numpy ticket 582 on 2007 09 20 by trac user thomas, assigned to unknown.
|
|
Hi, when used in an embedded scenario with some non g compilers, every call of import numpy leads to a crash Floating Point Overflow) because of the following function in umathmodule.c static double pinf init void) { double mul = 1e10; double tmp = 0.0; double pinf; pinf = mul; for ;;) { pinf = mul; if pinf == tmp) break; tmp = pinf; } return pinf; } When pinf supercedes 1E300, g will turn it to infinity, whereas other compilers e.g.
|
|
Borland C Builder) will throw an overflow exception.
|
|
My recommendation is to directly generate the double value for infinity by IEEE 754 specification char inf string[9] = \x00\x00\x00\x00\x00\x00\xF0\x7F ; double pinf = double )inf string)[0]; on little endian architectures, and char inf string[9] = \x7F\xF0\x00\x00\x00\x00\x00\x00 ; double pinf = double )inf string)[0]; on big endian architectures.
|
|
Thank you!","bug"
|
|
"6013","1181","fromiter memory error on invalid dtype Trac 583).
|
|
Original ticket http projects.scipy.org numpy ticket 583 on 2007 09 25 by stefanv, assigned to unknown.
|
|
Triggered by import numpy as N x = [[ 12 , ],[ 13 , ]] z = N.fromiter x,str)","bug"
|
|
"6013","1182","new max min functions that returns tuple max, idx) Trac 584).
|
|
Original ticket http projects.scipy.org numpy ticket 584 on 2007 10 05 by trac user steve geo, assigned to unknown.
|
|
I request that new max min functions of some suitable name)be created that returns a tuple of the max value and the index that max is located.
|
|
In my data sets the location of the max is often more important that the value of max itself, while I have created a python function to iterate through the data set and return what I need.
|
|
I think it would be trivial and faster for numpy to support this request.","enhancement"
|
|
"6013","1183","Patch which detects max < min for histogram X, range) Trac 585).
|
|
Original ticket http projects.scipy.org numpy ticket 585 on 2007 10 06 by trac user brorson, assigned to unknown.","enhancement"
|
|
"6013","1184","Retry) Patch to detect max < min in histogram X, range) Trac 586).
|
|
Original ticket http projects.scipy.org numpy ticket 586 on 2007 10 06 by trac user brorson, assigned to jarrodmillman.
|
|
As discussed on numpy discuss on 10.5.2007.","enhancement"
|
|
"6013","1185","f2py fails on INTEGER 8) return statement in f90 code Trac 587).
|
|
Original ticket http projects.scipy.org numpy ticket 587 on 2007 10 07 by tovrstra, assigned to pearu.
|
|
This is the example code MODULE fc IMPLICIT NONE CONTAINS SUBROUTINE color stat valsg, valsp, alphasg, alphasp, result, n) !f2py intent hide) n INTEGER 1),DIMENSION n),INTENT in) valsg, valsp, alphasg, alphasp INTEGER 8),INTENT out),DIMENSION 256,3) result !
|
|
local vars INTEGER i, n, weight result , ) = 0 DO i=1,n weight = alphasg n) alphasp n) IF weight > 0) THEN result valsp n),1) = result valsp n),1) weight result valsp n),2) = result valsp n),2) valsg n) weight result valsp n),3) = result valsp n),3) valsg n) valsg n) weight END IF END DO END SUBROUTINE END MODULE When I try to make a module with f2py, I get this toon poony tmp trevi f2py m helpers c helpers.f90 running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src building extension helpers sources f2py options [] f2py > tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c creating tmp tmpR6Rx7w creating tmp tmpR6Rx7w src.linux i686 2.4 Reading fortran codes... Reading file helpers.f90 format free) Post processing... Block helpers Block fsim Block color stat Post processing stage 2)... Block helpers Block unknown interface Block fsim Block color stat Building modules... Building module helpers ... Constructing F90 module support for fsim ... Constructing wrapper function fsim.color stat ... result = color stat valsg,valsp,alphasg,alphasp) Wrote C API module helpers to file tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c Fortran 90 wrappers are saved to tmp tmpR6Rx7w src.linux i686 2.4 helpers f2pywrappers2.f90 adding tmp tmpR6Rx7w src.linux i686 2.4 fortranobject.c to sources.
|
|
adding tmp tmpR6Rx7w src.linux i686 2.4 to include dirs.
|
|
copying usr lib python2.4 site packages numpy f2py src fortranobject.c > tmp tmpR6Rx7w src.linux i686 2.4 copying usr lib python2.4 site packages numpy f2py src fortranobject.h > tmp tmpR6Rx7w src.linux i686 2.4 adding tmp tmpR6Rx7w src.linux i686 2.4 helpers f2pywrappers2.f90 to sources.
|
|
running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 Could not locate executable ifort Could not locate executable ifc Could not locate executable ifort Could not locate executable efort Could not locate executable efc Could not locate executable ifort Could not locate executable efort Could not locate executable efc customize IntelFCompiler customize LaheyFCompiler customize PGroupFCompiler customize AbsoftFCompiler customize NAGFCompiler customize VastFCompiler customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 customize CompaqFCompiler customize IntelItaniumFCompiler customize IntelEM64TFCompiler customize Gnu95FCompiler Could not locate executable f95 customize Gnu95FCompiler Could not locate executable f95 customize Gnu95FCompiler using build ext building helpers extension compiling C sources C compiler i686 pc linux gnu gcc pthread fno strict aliasing DNDEBUG fPIC creating tmp tmpR6Rx7w tmp creating tmp tmpR6Rx7w tmp tmpR6Rx7w creating tmp tmpR6Rx7w tmp tmpR6Rx7w src.linux i686 2.4 compile options I tmp tmpR6Rx7w src.linux i686 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c i686 pc linux gnu gcc tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 136 error expected declaration specifiers or xe2x80x98...xe2x80x99 before xe2x80x98long longxe2x80x99 tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c In function xe2x80x98f2py rout helpers fsim color statxe2x80x99 tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error xe2x80x98long longxe2x80x99 undeclared first use in this function) tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error Each undeclared identifier is reported only once tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error for each function it appears in.)
|
|
tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error xe2x80x98resultxe2x80x99 undeclared first use in this function) tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 190 error expected expression before xe2x80x98)xe2x80x99 token tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 240 error too many arguments to function xe2x80x98f2py funcxe2x80x99 tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 136 error expected declaration specifiers or xe2x80x98...xe2x80x99 before xe2x80x98long longxe2x80x99 tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c In function xe2x80x98f2py rout helpers fsim color statxe2x80x99 tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error xe2x80x98long longxe2x80x99 undeclared first use in this function) tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error Each undeclared identifier is reported only once tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error for each function it appears in.)
|
|
tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 165 error xe2x80x98resultxe2x80x99 undeclared first use in this function) tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 190 error expected expression before xe2x80x98)xe2x80x99 token tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c 240 error too many arguments to function xe2x80x98f2py funcxe2x80x99 error Command i686 pc linux gnu gcc pthread fno strict aliasing DNDEBUG fPIC I tmp tmpR6Rx7w src.linux i686 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.c o tmp tmpR6Rx7w tmp tmpR6Rx7w src.linux i686 2.4 helpersmodule.o failed with exit status 1 When I replace INTEGER 8) by INTEGER 4) in the example code, f2py works fine.
|
|
Some system info numpy 1.0.3.1 compiled with gcc version 4.1.2 Gentoo 4.1.2) P.S.
|
|
I can not select the right numpy version from the new ticket form.","bug"
|
|
"6013","1186","numpy.dot ) doesn t work correctly on non contingous arrays Trac 588).
|
|
Original ticket http projects.scipy.org numpy ticket 588 on 2007 10 08 by robince, assigned to teoliphant.
|
|
Using dot ) with non contigous arrays doesn t work correctly.
|
|
Example In [1] from numpy import In [2] x=array [[1.,5.,25.,125.,625.]])
|
|
In [3] y=array [[20.],[160.],[640.],[1280.],[1024.]])
|
|
In [4] z=array [[1024.],[1280.],[640.],[160.],[20.]])
|
|
In [5] y2 = y[ 1] In [6] y2 z Out[6] array [[ 0.
|
|
], [ 0.
|
|
], [ 0.
|
|
], [ 0.
|
|
], [ 0.]])
|
|
In [8] dot x,z) Out[8] array [[ 55924.]])
|
|
In [9] dot x,y2) Out[9] array [[ 640000.]])
|
|
In [10] dot x,array y2)) Out[10] array [[ 55924.]])
|
|
In [11] version Out[11] 1.0.4.dev4155 Perhaps a test could be added to try and catch this type of problem testing core functions with non contigous arrays).","bug"
|
|
"6013","1187","problem installing with latest version of Intel MKL Trac 589).
|
|
Original ticket http projects.scipy.org numpy ticket 589 on 2007 10 09 by trac user DavidPowell, assigned to dmcooke.
|
|
There is a problem I think with disutils) in finding the libraries for version 9.1.021 of the Intel MKL.
|
|
The installation looks for libraries mkl lapack32 and mkl lapack64, which do not exist.
|
|
I was able to get around it by creating symbolic links from these files to mkl lapack.
|
|
Following the wiki instructions on installation with ifort and mkl did not help.
|
|
In addition, I found that it was necessary to add the SELinux context textrel shlib t to the shared libraries libguide.so, libirc.so, libmkl.so, libvml.so and libmkl lapack.so, though I don t really know what this actually did.","bug"
|
|
"6013","1188","arr.max ) does not always return maximum Trac 590).
|
|
Original ticket http projects.scipy.org numpy ticket 590 on 2007 10 10 by stefanv, assigned to unknown.
|
|
See attachment for data) import numpy as N z = N.fromfile z.dat ,sep= ) print z.max ), z[z > z.max )] 0.149478116585 [ 0.15997875 0.17071139 0.16408428 0.17560516]","bug"
|
|
"6013","1189","tab completion on record array fields Trac 591).
|
|
Original ticket http projects.scipy.org numpy ticket 591 on 2007 10 11 by trac user jdh2358, assigned to unknown.
|
|
I posted some time ago on the numpy mailing list about the desirability of supporting tab completion, eg in ipython, on the fields of numpy record arrays.
|
|
Travis had this to say > Not hard, in fact somebody suggested a patch that does exactly that.
|
|
> > The only question is what impact that might have on other things.
|
|
For > example, I think we would have to make sure that the proper order for > fields that conflict with object attributes would be I d have to look > to remember what the current order is).
|
|
I would like to again make a pitch to have this feature added to numpy record arrays, as it would be really useful.
|
|
Thanks, John Hunter","enhancement"
|
|
"6013","1190","numpy 1.0.4.dev3947 segfaults on |= for largish array sizes Trac 592).
|
|
Original ticket http projects.scipy.org numpy ticket 592 on 2007 10 16 by trac user AaronParsons, assigned to teoliphant.
|
|
I m on Fedora Core 7, running Python 2.5 with numpy 1.0.4.dev3947.
|
|
The following >>> import numpy >>> a = numpy.zeros 484,256)) >>> a |= numpy.ones 484,256)) Segfaults Expected behavior TypeError unsupported operand type s) for | float and float Seems to be array size dependent.","bug"
|
|
"6013","1191","Avoiding polluting public namespace in numpy headers Trac 593).
|
|
Original ticket http projects.scipy.org numpy ticket 593 on 2007 10 18 by cournape, assigned to unknown.
|
|
This patch splits the config.h header into two distinct files one numpyconfig.h, which can be used in public headers one config.h, which can NOT be used in public headers but can be included in c files).
|
|
All symbols in numpyconfig.h are NPY prepended this avoids polluting the C namespace with symbols such as SIZEOF which are often used most projects using autotools will define those too, for example).
|
|
This makes the numpy headers a bit cleaner, a bit easier to integrate in a autotools based project, and is more standard now, the config.h and numpyconfig.h follow more stricty autoheader convention).
|
|
This does not change in any way the C API, or the behaviour in numpy it is a bug otherwise).
|
|
Within all the symbols declared in config.h, here is the list of the symbols which are part of the numpy API e.g.
|
|
which should stay public) NPY NOSIGNAL NPY ALLOW THREADS CHAR BITS All other are kept private not visible from numpy headers) SIZEOF HAVE MATHLIB DISTUTILS USE SDK PyOS ascii strtod For SIZEOF symbols, I prepended them by NPY in numpyconfig.h this was already the case for most of them, though I just sanitize the header so that only the prepended symbols are used).
|
|
To make the review easier, I implemented this enhancement in a branch cleanconfig).","enhancement"
|
|
"6013","1192","Major slowdown when mixing array with Python s list Trac 594).
|
|
Original ticket http projects.scipy.org numpy ticket 594 on 2007 10 20 by trac user Boris, assigned to unknown.
|
|
I encountered a major slowdown every time I use a simple Python list with an array function.
|
|
Here is an example which occurs quite often in my code t1 = Timer setup= from numpy import ; x=arange 1,100); k=array [97,43,3,6,23,0]) , stmt= x[k] ) t1.timeit 10000) =0.029493949996 t2 = Timer setup= from numpy import ; x=arange 1,100); k=[97,43,3,6,23,0] , stmt= x[k] ) t2.timeit 10000) =0.177627355842 I would expect that t2 is as fast as t1.
|
|
Using an array instead of the list is also no option since I create this index lists very often and k=[97,43,3,6,23,0] is a lot faster than k=array [97,43,3,6,23,0]) This slowdown also occurs with other functions, e.g.
|
|
sum, sort.","bug"
|
|
"6013","1193","sum function slow Trac 595).
|
|
Original ticket http projects.scipy.org numpy ticket 595 on 2007 10 21 by trac user Boris, assigned to unknown.
|
|
Numpy s sum function is very slow for 1 D vectors at least on my pc), it is a lot faster to use e.g.
|
|
the dot product with a ones vector which returns the same result) t2 = Timer setup= from numpy import ; c=1.
|
|
arange 1,1001) , stmt= sum c) ) t3 = Timer setup= from numpy import ; c=1.
|
|
arange 1,1001) , stmt= c.sum ) ) t4 = Timer setup= from numpy import ; c=1.
|
|
arange 1,1001); d=ones 1000) , stmt= dot c,d) ) t2.timeit 10000) =0.204298445144 t3.timeit 10000) =0.164432998837 t4.timeit 10000) =0.0447709820354 My configuration Windows XP Pentium D), Numpy 1.0.3","enhancement"
|
|
"6013","1194","setuptools develop mode broken Trac 596).
|
|
Original ticket http projects.scipy.org numpy ticket 596 on 2007 10 21 by astraw, assigned to dmcooke.
|
|
As of svn 4271, I could not get the numpy.scons branch to install using the develop mode of setuptools Steps to reproduce 0.
|
|
`svn co http svn.scipy.org svn numpy branches numpy.scons` 1.
|
|
`python setup.py build` may not be neccessary, but I did it) 2.
|
|
Attempt to install using setuptools develop mode set `sys.argv[0], name , main ` to their expected values in case setup.py introspect these) `python c import numpy.distutils, setuptools, sys;f= setup.py ;sys.argv[0]=f;execfile f,{ file f, name main }) develop` 3.
|
|
`python setupegg.py develop` also breaks, and there s a fairly load warning about numpy.distutils not being imported.
|
|
The error I get Traceback most recent call last) File <string> , line 1, in <module> File setup.py , line 90, in <module> setup package ) File setup.py , line 83, in setup package configuration=configuration ) File home astraw other peoples src numpy.svn.scons numpy distutils core.py , line 182, in setup return old setup new attr) File distutils core.py , line 151, in setup File distutils dist.py , line 974, in run commands File distutils dist.py , line 994, in run command File usr lib python2.5 site packages setuptools command develop.py , line 26, in run self.install for development ) File usr lib python2.5 site packages setuptools command develop.py , line 81, in install for development self.process distribution None, self.dist, not self.no deps) File usr lib python2.5 site packages setuptools command easy install.py , line 497, in process distribution self.install egg scripts dist) File usr lib python2.5 site packages setuptools command develop.py , line 111, in install egg scripts script path = os.path.abspath convert path script name)) File usr lib python2.5 posixpath.py , line 402, in abspath if not isabs path) File usr lib python2.5 posixpath.py , line 49, in isabs return s.startswith ) AttributeError function object has no attribute startswith Thanks so far for your work on the scons branch I hope I can get it to work soon!","bug"
|
|
"6013","1195","assigning negative to a uint64 array element gives wrong answer on 32 bit machine Trac 597).
|
|
Original ticket http projects.scipy.org numpy ticket 597 on 2007 10 22 by trac user gregsmith to, assigned to unknown.
|
|
!python a = numpy.zeros 2,), numpy.uint64) a[0] = 1234567 a array [18446744073709551615, 0], dtype=uint64) The right answer is 18446744073708317049 to be consistent with the uint32 behaviour anyway).
|
|
The value you get 184...615) is the largest possible uint64.
|
|
Failure does not occur on a 64 bit machine.
|
|
Sorry for posting against an older version; I could not find any tickets for this issue so it may remain.","bug"
|
|
"6013","1196","incorrect behaviour with d a[i]) on uint64 Trac 598).
|
|
Original ticket http projects.scipy.org numpy ticket 598 on 2007 10 22 by trac user gregsmith to, assigned to unknown.
|
|
!python >>> a array [18446744073709551615, 18446744073709551615], dtype=uint64) >>> d a[0] 1 >>> u a[0] 1 >>> x a[0] 1 This occurs on 64 and 32 bit machines.
|
|
Sorry for posting against an older version, can t find a ticket for this issue though.
|
|
It used to be that hex a[0]) would give you 0x1 here, but that was fixed by 1.02 whilst this remains.
|
|
Workaround d long a[0]) or s a[0]","bug"
|
|
"6013","1197","Possible error in fromnumeric.py wrapit for nonarray types Trac 599).
|
|
Original ticket http projects.scipy.org numpy ticket 599 on 2007 10 24 by trac user dharland, assigned to unknown.
|
|
By chance I noticed this odd looking snippet of code in wrapit.
|
|
if wrap and isinstance result, mu.ndarray) if not isinstance result, mu.ndarray) SURELY THIS CAN NEVER BE TRUE?!
|
|
result = asarray result) result = wrap result) Should this instead be?
|
|
if wrap if not isinstance result, mu.ndarray) result = asarray result) result = wrap result)","bug"
|
|
"6013","1198","Cannot load numpy.unicode dumped by cPickle Trac 600).
|
|
Original ticket http projects.scipy.org numpy ticket 600 on 2007 10 26 by trac user Kaloo, assigned to unknown.
|
|
The following code is crashing dumping a unicode field of a ndArray, then loading it) import numpy import cPickle titi = numpy.array [ DROND , DROND1 ], dtype= <U6 , order= FORTRAN ) toto = titi[1] f = open c toto.txt , w ) cPickle.dump toto, f) f.close ) f = open c toto.txt , r ) t = cPickle.load f) Error is File <interactive input> , line 1, in <module> ValueError initialization string is too small , <built in function scalar>, dtype <U6 ), D\x00R\x00O\x00N\x00D\x001\x00 )) With Python 2.5, numpy 1.0.3.1, on Windows XP","bug"
|
|
"6013","1199","function for computing powers of a matrix Trac 601).
|
|
Original ticket http projects.scipy.org numpy ticket 601 on 2007 10 28 by lebedov, assigned to unknown.
|
|
It would be nice to add the following function to numpy to allow for the raising of matricies to arbitrary exponents and perhaps modify `matrix.
|
|
pow )` to use this approach when one attempts to raise a matrix to a noninteger exponent) from numpy import diag,dot,shape,eye from numpy.linalg import eig,inv def mpower x,y) Compute x raised to the power y when x is a square matrix and y is a scalar.
|
|
s = shape x) if len s) != 2 or s[0] != s[1] raise ValueError matrix must be square ) if y == 0 return eye s[0]) [e,v] = eig x) d = diag e) return dot dot v,d y),inv v))","enhancement"
|
|
"6013","1200","Memory leak in ufunc update use defaults Trac 602).
|
|
Original ticket http projects.scipy.org numpy ticket 602 on 2007 10 28 by trac user langner, assigned to unknown.
|
|
A missing Py DECREF errobj) in ufunc update use defaults ufuncobject.c) causes references to cumulate and memory to leak.
|
|
The function is called by ufunc seterr, which is called for instance when formatting float arrays from the method FloatFormat.fillFormat).
|
|
Resident process memory grows rapidly due to this, for instance in the loop while True s = numpy.array [1.0]).
|
|
str ) Patch for r4210 of trunk numpy core src ufuncobject.c 3206a3207 > Py DECREF errobj); This has been discussed on the list http projects.scipy.org pipermail numpy discussion 2007 October 029667.html http projects.scipy.org pipermail numpy discussion 2007 October 029669.html","bug"
|
|
"6013","1201","numpy 1.0.3.1, python crash when indexing an array with a int32 ndarray see example) Trac 603).
|
|
Original ticket http projects.scipy.org numpy ticket 603 on 2007 10 29 by trac user djkawa, assigned to jarrodmillman.
|
|
>>> import numpy >>> from numpy import array >>> array = array [0.0], dtype = numpy.float64) >>> index = array [0.0], dtype = numpy.int32) >>> cum var[index] array [ 0.])
|
|
>>> index = array 0., dtype = numpy.int32) >>> cum var[index] > CRASH","bug"
|
|
"6013","1202","array creation does not work with generators Trac 604).
|
|
Original ticket http projects.scipy.org numpy ticket 604 on 2007 11 03 by trac user chtito, assigned to unknown.
|
|
I think that `array n 2 for n in range 10))` should create an array with ten elements.
|
|
Right now it creates an array with only one generator object which is not consistent with the behaviour for lists.
|
|
Indeed, `array [n 2 for n in range 10)])` creates an array with ten elements instead of an array of one list object.","bug"
|
|
"6013","1203","Incorrect behaviour of numpy.histogram Trac 605).
|
|
Original ticket http projects.scipy.org numpy ticket 605 on 2007 11 04 by trac user Elby, assigned to unknown.
|
|
The behavior of numpy.histogram is not consistent with its doc string doc string says that, with a range argument, values outside of this range are allocated to the closest bin.
|
|
in fact values below the range are simply ignored.
|
|
There was a discussion on this subject on the scipy.user mailing list http groups.google.com group scipy user browse frm thread 3b3166e2200f846b d6040fb6b659c6dd?hl=fr lnk=gst q=histogram d6040fb6b659c6dd IMHO, the current behavior of numpy.histogram, that is assuming that values below the range are outliers, is not what a neophyte is waiting for, and should be clearly stated.
|
|
Besides, the user should have the possibility to choose what to do with values outside the range just ignoring them is not a good idea in most of the case I ve seen.","bug"
|
|
"6013","1204","ImportError No module named distutils.ccompiler Trac 606).
|
|
Original ticket http projects.scipy.org numpy ticket 606 on 2007 11 05 by nilswagner01, assigned to dmcooke.
|
|
I cannot install the latest svn version of numpy.
|
|
I am using python2.3 on x86 64. python setup.py install prefix= HOME local Running from numpy source directory.
|
|
Traceback most recent call last) File setup.py , line 90, in ?
|
|
setup package ) File setup.py , line 62, in setup package from numpy.distutils.core import setup File data home nwagner svn numpy numpy distutils init .py , line 6, in ?
|
|
import ccompiler File data home nwagner svn numpy numpy distutils ccompiler.py , line 6, in ?
|
|
from distutils.ccompiler import ImportError No module named distutils.ccompiler","bug"
|
|
"6013","1205","segfault from particular indexing of empty array Trac 607).
|
|
Original ticket http projects.scipy.org numpy ticket 607 on 2007 11 06 by efiring, assigned to unknown.
|
|
The following causes a segfault upon execution as a script import numpy u = numpy.array []) g = numpy.array [True, True]) u[g] = 0 See the thread starting with http projects.scipy.org pipermail numpy discussion 2007 November 029801.html","bug"
|
|
"6013","1206","<ufunc absolute > should clear the sign bit.
|
|
Trac 608).
|
|
Original ticket http projects.scipy.org numpy ticket 608 on 2007 11 07 by trac user lambertdw, assigned to unknown.
|
|
>>> a array [ 0., 0., 0.])
|
|
>>> absolute a) array [ 0., 0., 0.])
|
|
>>> array [0 if x == 0 else abs x) for x in a]) oops, this alters data type array [0, 0, 0]) This affects my docstring tests.
|
|
from scipy import def spherical to Cartesian r,phi,psi) >>> from scipy import >>> a = spherical to Cartesian 0,3,23) >>> list a) fails, but ok. [0.0, 0.0, 0.0] >>> list absolute a)) still fails at first glance surprising.
|
|
[0.0, 0.0, 0.0] >>> [abs x) for x in a] Maybe Guido should participate?
|
|
[0.0, 0.0, 0.0] >>> alltrue a == [0,0,0]) Success!
|
|
True t = r cos psi) return asarray t cos phi),t sin phi),r sin psi)))","enhancement"
|
|
"6013","1207","numpy linalg lapack litemodule.c lapack lite zgeqrf int vs. long Trac 609).
|
|
Original ticket http projects.scipy.org numpy ticket 609 on 2007 11 08 by trac user kjetijor, assigned to unknown.
|
|
When building numpy 1.0.4 and for that matter 1.0.3.1, 1.0.2 and svn trunk) on x86 64 some of my users had trouble with numpy.linalg.qr for matrices containing complex numbers.
|
|
After a little bit of testing and digging I ended up doing the following, in changeset 3850 and changeset 3851 parsing of arguments to zgeqrf was changed to take integer values instead of long values, however in the return statement for Py BuildValue it s still using long values which seems to break on my x86 64 boxes with various pythons and various distros.
|
|
Example would be CentOS 5, x86 64 and Python 2.4.3) numpy.linalg.qr numpy.array [[1.0j,2],[3,4]])) will raise LinAlgError with numpy.linalg.linalg.LinAlgError zgeqrf returns 18038862643 What I ended up doing was changing the return statement for numpy linalg lapack litemodule.c lapack lite zgeqrf return Py BuildValue {s i,s i,s i,s i,s i,s i} , .....) instead of return Py BuildValue {s l,s l,s l,s l,s l,s l} , .....) What also seemed to work were to leave the conversions as long s and cast the values to as long.
|
|
Given that I don t really know the inner workings of numpy all that well, this may of course be badness, although apparently it seemed to fix my problem.","bug"
|
|
"6013","1208","Let numpy.interp accept scalar values as first argument Trac 610).
|
|
Original ticket http projects.scipy.org numpy ticket 610 on 2007 11 09 by trac user Elby, assigned to unknown.
|
|
The current version of numpy.interp does not accept a single value as first argument, so it is not possible to make a single interpolation >>> import numpy as n >>> n. version 1.0.5.dev4420 >>> xp = n.arange 10) >>> yp = 2.5 xp 2 xp >>> x = 3.2 >>> n.interp x, xp, yp) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError object of too small depth for desired array It would be a lot more nice to let x be a float, an int or another numpy scalar object, as the attached patch does.","enhancement"
|
|
"6013","1209","[Vista NPY OWNDATA] data created in C code deleted on return to python Trac 611).
|
|
Original ticket http projects.scipy.org numpy ticket 611 on 2007 11 09 by trac user chafporte, assigned to cournape.
|
|
I create an array in C. I prepare it to be handled by numpy and python, I set the flag NPY OWNDATA.
|
|
then when it exits the C code and return to python, the array is correctly handled by numpy and python.
|
|
This works perfectly on every platform linux, macosx, XP, windows 2000) ... but not on VISTA.
|
|
where the array is deleted right away.","bug"
|
|
"6013","1210","NumPy 1.0.4 fails to compile on Cygwin Trac 612).
|
|
Original ticket http projects.scipy.org numpy ticket 612 on 2007 11 14 by mdehoon, assigned to unknown.
|
|
This is the output I get python setup.py build Running from numpy source directory.
|
|
non existing path in numpy distutils site.cfg F2PY Version 2 4422 blas opt info blas mkl info libraries mkl,vml,guide not found in usr local lib libraries mkl,vml,guide not found in usr lib NOT AVAILABLE atlas blas threads info Setting PTATLAS=ATLAS libraries ptf77blas,ptcblas,atlas not found in usr local lib libraries ptf77blas,ptcblas,atlas not found in usr lib NOT AVAILABLE atlas blas info libraries f77blas,cblas,atlas not found in usr local lib libraries f77blas,cblas,atlas not found in usr lib NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1340 UserWarn ing Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) blas info libraries blas not found in usr local lib libraries blas not found in usr lib NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1349 UserWarn ing Blas http www.netlib.org blas ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [blas]) or by setting the BLAS environment variable.
|
|
warnings.warn BlasNotFoundError.
|
|
doc ) blas src info NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1352 UserWarn ing Blas http www.netlib.org blas ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [blas src]) or by setting the BLAS SRC environment variable.
|
|
warnings.warn BlasSrcNotFoundError.
|
|
doc ) NOT AVAILABLE lapack opt info lapack mkl info mkl info libraries mkl,vml,guide not found in usr local lib libraries mkl,vml,guide not found in usr lib NOT AVAILABLE NOT AVAILABLE atlas threads info Setting PTATLAS=ATLAS libraries ptf77blas,ptcblas,atlas not found in usr local lib libraries lapack atlas not found in usr local lib libraries ptf77blas,ptcblas,atlas not found in usr lib libraries lapack atlas not found in usr lib numpy.distutils.system info.atlas threads info NOT AVAILABLE atlas info libraries f77blas,cblas,atlas not found in usr local lib libraries lapack atlas not found in usr local lib libraries f77blas,cblas,atlas not found in usr lib libraries lapack atlas not found in usr lib numpy.distutils.system info.atlas info NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1247 UserWarn ing Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) lapack info libraries lapack not found in usr local lib libraries lapack not found in usr lib NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1258 UserWarn ing Lapack http www.netlib.org lapack ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [lapack]) or by setting the LAPACK environment variable.
|
|
warnings.warn LapackNotFoundError.
|
|
doc ) lapack src info NOT AVAILABLE home mdehoon Software numpy 1.0.4 numpy distutils system info.py 1261 UserWarn ing Lapack http www.netlib.org lapack ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [lapack src]) or by setting the LAPACK SRC environment variable.
|
|
warnings.warn LapackSrcNotFoundError.
|
|
doc ) NOT AVAILABLE running build running config cc unifing config cc, config, build clib, build ext, build commands compiler opti ons running config fc unifing config fc, config, build clib, build ext, build commands fcompiler opt ions running build src building py modules sources building extension numpy.core.multiarray sources Generating build src.cygwin 1.5.24 i686 2.5 numpy core config.h customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 customize IntelVisualFCompiler Could not locate executable ifl customize AbsoftFCompiler Could not locate executable f90 don t know how to compile Fortran code on platform posix with compaqv compil er.
|
|
Supported compilers are none,absoft,compaq,gnu,vast,sun,nag,lahey,intelem,g nu95,intelv,g95,intele,pg,intel,mips,hpux,intelev,ibm) Traceback most recent call last) File setup.py , line 89, in <module> setup package ) File setup.py , line 82, in setup package configuration=configuration ) File home mdehoon Software numpy 1.0.4 numpy distutils core.py , line 176, i n setup return old setup new attr) File tmp python.6884 usr lib python2.5 distutils core.py , line 151, in setu p File tmp python.6884 usr lib python2.5 distutils dist.py , line 974, in run commands File tmp python.6884 usr lib python2.5 distutils dist.py , line 994, in run command File tmp python.6884 usr lib python2.5 distutils command build.py , line 112 , in run File usr lib python2.5 cmd.py , line 333, in run command del help[cmd] File tmp python.6884 usr lib python2.5 distutils dist.py , line 994, in run command File home mdehoon Software numpy 1.0.4 numpy distutils command build src.py , line 130, in run self.build sources ) File home mdehoon Software numpy 1.0.4 numpy distutils command build src.py , line 147, in build sources self.build extension sources ext) File home mdehoon Software numpy 1.0.4 numpy distutils command build src.py , line 250, in build extension sources sources = self.generate sources sources, ext) File home mdehoon Software numpy 1.0.4 numpy distutils command build src.py , line 307, in generate sources source = func extension, build dir) File numpy core setup.py , line 51, in generate config h library dirs = default lib dirs) File tmp python.6884 usr lib python2.5 distutils command config.py , line 27 8, in try run File home mdehoon Software numpy 1.0.4 numpy distutils command config.py , l ine 31, in check compiler c compiler=self.compiler) File home mdehoon Software numpy 1.0.4 numpy distutils fcompiler init .py , line 809, in new fcompiler c compiler=c compiler) File home mdehoon Software numpy 1.0.4 numpy distutils fcompiler init .py , line 791, in get default fcompiler c compiler=c compiler) File home mdehoon Software numpy 1.0.4 numpy distutils fcompiler init .py , line 740, in find existing fcompiler c.customize dist) AttributeError NoneType object has no attribute customize","bug"
|
|
"6013","1211","resize on an empty array fails in some cases, very weird Trac 613).
|
|
Original ticket http projects.scipy.org numpy ticket 613 on 2007 11 14 by tovrstra, assigned to unknown.
|
|
The problem is rather simple and probably not hurting many people.
|
|
This works >>> a = numpy.zeros 0) >>> a.resize 5) >>> a array [ 0., 0., 0., 0., 0.])
|
|
This doesn t >>> a = numpy.zeros 0) >>> a array [], dtype=float64) >>> a.resize 5) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
ValueError cannot resize an array that has been referenced or is referencing another array in this way.
|
|
Use the resize function I m not sure whether this is a real problem, but it is at least a bit confusing.","bug"
|
|
"6013","1212","SegFault double free with simple array mask operation Trac 614).
|
|
Original ticket http projects.scipy.org numpy ticket 614 on 2007 11 15 by trac user AchimGaedke, assigned to stefanv.
|
|
Please have a look at the program below start import numpy t array=numpy.ones 2048, dtype=numpy.float32) sinc array=numpy.array len t array),),dtype=numpy.float32) sinc array[ t array > 0.
|
|
)]=1.0 end If you execute this program, it crashes with Segmentation Fault or glibc detected python double free or corruption out) 0x081fe470 It depends on the circumstances, which error occurs, e.g.
|
|
you must quit your interpreter if you are in interactvie mode.
|
|
Obviously numpy.array ) should be numpy.zeros ) or numpy.empty ) ....
|
|
But this program should not crash with a core dump.
|
|
Used Linux Versions are Debian Testing with numpy 1.0.3, Debian Stable with numpy 1.0.1, Ubuntu Linux 6.10 with numpy 1.0 Also numpy 1.0.4 crashes.
|
|
The trac system does not know these versions?!)","bug"
|
|
"6013","1213","numpy.array [])[arange 10000)>=0] = 123 segfaults Trac 615).
|
|
Original ticket http projects.scipy.org numpy ticket 615 on 2007 11 15 by trac user rauli, assigned to unknown.
|
|
array [])[arange 10000)>=0] = 123 segfaults; apparently boolean array indices are used unchecked.
|
|
array [])[arange 10000)] instead raises IndexError as expected.
|
|
This happens on Ubuntu Feisty x86 32, numpy 1.0.1 as well as 1.0.4, and Ubuntu Gutsy x86 64 with numpy 1.0.3.
|
|
I haven t tested it with the trunk.","bug"
|
|
"6013","1214","arange incorrect output byteorder Trac 616).
|
|
Original ticket http projects.scipy.org numpy ticket 616 on 2007 11 17 by stefanv, assigned to stefanv.
|
|
Sebastian Haase found the following bug >> a = N.arange .5, dtype= >f ) >>> `a.dtype` dtype float32 ) >>> a = N.arange .5, dtype= <f ) >>> `a.dtype` dtype float32 )","bug"
|
|
"6013","1215","numpy.random.randint error with negative numbers Trac 617).
|
|
Original ticket http projects.scipy.org numpy ticket 617 on 2007 11 20 by trac user fawzi, assigned to stefanv.
|
|
import numpy.random print repr numpy.random.randint 3, 1)) will print a large positive number as 4294967293L) obviously negative 32 bit number interpreted as a positive 64bit number.
|
|
>>> numpy.version.version 1.0.3.1 Python 2.5 r25 51918, Sep 19 2006, 08 49 13) [GCC 4.0.1 Apple Computer, Inc. build 5341)] on darwin MacOSX 10.4.10 Intel Core 2 Duo Fawzi","bug"
|
|
"6013","1216","numpy freebsd fix Trac 618).
|
|
Original ticket http projects.scipy.org numpy ticket 618 on 2007 11 21 by trac user hawking, assigned to unknown.
|
|
Attached patch allows numpy to work on freebsd.
|
|
Note that it will still compile without the patch, but tests fail and it is not usable.
|
|
There s also a similar [http www.freebsd.org cgi cvsweb.cgi ports math py numpy files patch numpy core include numpy ufuncobject.h patch] in freebsd ports but it doesn t fix all problems.
|
|
Patch was written by Joe Peterson.","bug"
|
|
"6013","1217","remove split quoted in numpy.distutils.ccompiler Trac 619).
|
|
Original ticket http projects.scipy.org numpy ticket 619 on 2007 11 21 by jarrodmillman, assigned to jarrodmillman.
|
|
We have our own version of `distutils.util.split quoted )` in `numpy distutils ccompiler.py`, which just just makes a minor change to distutils default behavior.
|
|
Here is the relevant code from `numpy.distutils.ccompiler` def split quoted s) <snip> if has white re.search s[beg 1 end 1]) s = s[ beg] s[beg 1 end 1] s[end ] pos = m.end ) 2 else Keeping quotes when a quoted word does not contain white space.
|
|
XXX send a patch to distutils pos = m.end ) <snip> Here is the relevant code from `distutils.util` def split quoted s) <snip> s = s[ beg] s[beg 1 end 1] s[end ] pos = m.end ) 2 <snip> If this isn t necessary we should remove our version of `split quoted`.
|
|
If there is a reason for this check, we need to patch `split quoted` upstream and remove our version.","enhancement"
|
|
"6013","1218","Error in the typemaps in numpy.i Trac 620).
|
|
Original ticket http projects.scipy.org numpy ticket 620 on 2007 11 23 by trac user matthieu.brucher, assigned to stefanv.
|
|
Some typemaps are not C89 compatible variables are declared inside a block and not at its beginning).
|
|
For instance typemap in) DATA TYPE IN ARRAY1[ANY]) PyArrayObject array=NULL, int is new object=0) { array = obj to array contiguous allow conversion input, DATA TYPECODE, is new object); npy intp size[1] = { 1 dim0 }; if !array || !require dimensions array, 1) || !require size array, size, 1)) SWIG fail; 1 = 1 ltype) array data array); } size should have been declared first.","bug"
|
|
"6013","1219","wrapping f95 fails Trac 621).
|
|
Original ticket http projects.scipy.org numpy ticket 621 on 2007 11 23 by trac user jspaans, assigned to jarrodmillman.
|
|
When trying to wrap some fortran90 code on my platform OS X, python2.5 from macports, gfortran 4.3, using numpy latest downloaded today) I get an error which is fixed by applying the following obvious) patch opt local lib python2.5 site packages numpy f2py rules.py 2007 11 23 17 36 11.000000000 0100 opt local lib python2.5 site packages numpy f2py rules.py 2007 11 23 17 36 11.000000000 0100 1219,7 1219,7 f.write !
|
|
This file is autogenerated with f2py version s)\n f2py version)) f.write !
|
|
It contains Fortran 90 wrappers to fortran functions.\n ) lines = [] for l in \n\n .join funcwrappers2) \n .split \n ) for l in \n\n .join funcwrappers2) \n ).split \n ) if len l)>72 and l[0]== lines.append l[ 72] \n ) l = l[72 ]","bug"
|
|
"6013","1220","function for computing the condition number Trac 622).
|
|
Original ticket http projects.scipy.org numpy ticket 622 on 2007 11 30 by lebedov, assigned to unknown.
|
|
Trivial suggested addition to numpy.linalg from numpy.linalg import norm,inv def cond x,p=2) Compute the condition number of a matrix using the p norm.
|
|
return norm x,p) norm inv x),p)","enhancement"
|
|
"6013","1221","loadtxt fails with record arrays Trac 623).
|
|
Original ticket http projects.scipy.org numpy ticket 623 on 2007 12 01 by trac user mstarzyk, assigned to unknown.
|
|
Hi, there is a problem with loading records via loadtxt ).
|
|
See the example code below import numpy as N import StringIO c = StringIO.StringIO ) c.write 1 2\n3 4 ) N.loadtxt c, dtype=[ x , N.int32), y , N.int32)]) This raises the following exception Traceback most recent call last) File C \tmp\aa.py , line 6, in <module> N.loadtxt c, dtype=[ x , N.int32), y , N.int32)]) File c \python25\lib\site packages\numpy 1.0.4 py2.5 win32.egg\numpy\core\numeric.py , line 726, in loadtxt r,c = X.shape ValueError need more than 1 value to unpack","bug"
|
|
"6013","1222","Statistics related array method are not self consistent.
|
|
Trac 624).
|
|
Original ticket http projects.scipy.org numpy ticket 624 on 2007 12 03 by trac user usovalx, assigned to unknown.
|
|
Statistincs related array methods mean, std) are not self consistent.
|
|
The data type of the result does not reflects the way calculations are done.
|
|
Consider the example a = arange 1, 100, dtype=float32) r1 = a.mean ) r2 = a.astype float64).mean ) print type r1), r1) <type numpy.float64 >, 0.052296744452582464) print type r2), r2) <type numpy.float64 >, 0.052296743192004433) Both results have the same datatype, which is confusing.
|
|
As the exact method of calculation differs which might result in the dramatic differences in the results for some boundary cases).
|
|
Things are even more confusing when we deal with integer arrays a = repeat 2147483647, 100) print a.sum ) 100 print a.mean ) 2147483647.0 print a.astype float32).mean ) 2147483648.0 I believe the convention for the handling of intermediate variables should be clearly specified and obeyed.","bug"
|
|
"6013","1223","ImportError No module named test ufunc Trac 625).
|
|
Original ticket http projects.scipy.org numpy ticket 625 on 2007 12 03 by nilswagner01, assigned to stefanv.
|
|
python i usr local lib64 python2.5 site packages numpy core tests test multiarray.py Traceback most recent call last) File usr local lib64 python2.5 site packages numpy core tests test multiarray.py , line 538, in <module> from test ufunc import ImportError No module named test ufunc","bug"
|
|
"6013","1224","percentile ) and clamp ) Trac 626).
|
|
Original ticket http projects.scipy.org numpy ticket 626 on 2007 12 07 by trac user connellybarnes, assigned to unknown.
|
|
Two functions I am continually re implementing in my own number crunching projects are percentile ) and clamp ).
|
|
Here percentile array, p) returns the pth percentile of array, using linear interpolation p=0.5 is median, p=1.0 is max, p=0.25 is lower quartile, etc).
|
|
And clamp array, lo, hi) clamps to lower and upper bound scalars.
|
|
Would these functions be appropriate for numpy?
|
|
Will you accept patches?
|
|
Thanks, Connelly Barnes connellybarnes at symbol gmail dot com","enhancement"
|
|
"6013","1225","linalg.svd takes 100 CPU and never ends Trac 627).
|
|
Original ticket http projects.scipy.org numpy ticket 627 on 2007 12 09 by trac user Czarny, assigned to cournape.
|
|
When I call the svd function with an array it takes 100 of my CPU and never actually finishes the calculation.
|
|
Python needs to get killed manually Ctrl C doesn t work as usually).
|
|
I ve tested that on Linux Linux kacper 2.6.22.10 laptop 0.6 1 SMP PREEMPT Sun Oct 21 17 48 51 CEST 2007 i686 Intel R) Pentium R) M processor 1.70GHz PLD Linux python 2.5.1 4.i686 python numpy 1.0.4 1.i686 I ve tried with earlier 1.0.3.1 as well) Windows i686 python 2.5 python numpy 1.0.4 I ve used two scripts found on the net google with svd python ) and both behave in exactly the same way http www.thescripts.com forum post1806623 4.html http mail.python.org pipermail python list 2006 November 412156.html","bug"
|
|
"6013","1226","polyfit uses 100 CPU and does not stop Windows XP, Python 2.5.1 NumPy 1.0.4) Trac 628).
|
|
Original ticket http projects.scipy.org numpy ticket 628 on 2007 12 09 by trac user behrisch, assigned to cournape.
|
|
The attached code shows the very same behavior as 1225, although I can confirm that it works with numpy 1.0.3 on the same machine with te same Python distribution.
|
|
I used the binary packages from the sourceforge download site together with Python 2.5.1.
|
|
I did not install any LAPACK or BLAS packages.
|
|
Furthermore it works on openSUSE 10.3 with numpy 1.0.4, libblas3 3.1.1 29, liblapack3 3.1.1 29 which may be another hint, that it is no duplicate of 1225)","bug"
|
|
"6013","1227","repr ) fails on scalars float64) Trac 629).
|
|
Original ticket http projects.scipy.org numpy ticket 629 on 2007 12 15 by astraw, assigned to cournape.
|
|
The attached test fails, but I think it should pass.","bug"
|
|
"6013","1228","If float 123.45 ) works, so should numpy.float32 123.45 ) Trac 630).
|
|
Original ticket http projects.scipy.org numpy ticket 630 on 2007 12 16 by astraw, assigned to unknown.
|
|
We should make a constructor that takes strings for the numpy scalar types.
|
|
Note that float64 ) and int64 ) already do this.
|
|
This would let longdouble types be specified with all full precision.","bug"
|
|
"6013","1229","incorrect description of numpy.linalg.pinv in linalg info.py Trac 631).
|
|
Original ticket http projects.scipy.org numpy ticket 631 on 2007 12 17 by lebedov, assigned to unknown.
|
|
The function numpy.linalg.pinv uses svd.
|
|
However, the doc string in linalg info.py incorrectly describes the function as using lstsq.","bug"
|
|
"6013","1230","numpy.histogram fails with bin=<list> Trac 632).
|
|
Original ticket http projects.scipy.org numpy ticket 632 on 2007 12 17 by trac user orbeckst, assigned to unknown.
|
|
numpy.histogram ) in 1.0.4 with python 2.3) fails if the argument to bins is a list 1) any ) is not found[[BR]] 2) exception at line 155, `if any bins[1 ] bins[ 1] < 0)) ` Solution two changes in numpy lib function base.py 1) add from numpy.core import any at the top of the file[[BR]] 2) change line 155 from[[BR]] if any bins[1 ] bins[ 1] < 0)) to if any asarray bins[1 ]) asarray bins[ 1]) < 0))","bug"
|
|
"6013","1231","segfault in assignment to pyobject array Trac 633).
|
|
Original ticket http projects.scipy.org numpy ticket 633 on 2007 12 18 by trac user batripler, assigned to teoliphant.
|
|
Enter the following sequence >>> import numpy >>> m=numpy.zeros 1000000, dtype= object ) >>> m[ ]=1 >>> m[ ]=1 The 2nd assignment sometimes the 3rd) will cause a segfault.
|
|
This occurs on numpy v1.0.4 and py v2.5.1 installed from release binaries on Win32).","bug"
|
|
"6013","1232","masked array with dtype bool, sum to bool instead of int.
|
|
Trac 634).
|
|
Original ticket http projects.scipy.org numpy ticket 634 on 2007 12 18 by trac user zongzhi.liu ..., assigned to unknown.
|
|
masked array with dtype bool, sum to bool instead of int.
|
|
a = array [True, False, True]) a.sum ) 2 m = ma.array [True, False, True], mask=[False, False, False]) m.sum ) True numpy version 1.0.3.1 Linux Fedora 8 Intel Pentium M 1.4G laptop compaq presario X1000)","bug"
|
|
"6013","1233","numpy 1.0.4 crashes when doing linalg.inv on P3 Windows machine.
|
|
Invalid op code.
|
|
Trac 635).
|
|
Original ticket http projects.scipy.org numpy ticket 635 on 2007 12 18 by trac user brorson, assigned to cournape.
|
|
I have been testing numpy 1.0.4 on a variety of different machine configurations.
|
|
I have found a numpy crash in the following config Computer = Dell OptiPlex GX1p Processor = Intel Pentium 3, family 6, model 7, stepping 3, revision E. OS = Win XP 32 bit Python 2.5.1 win 32 bit version latest from python.org) numpy 1.0.4 The following code causes a crash, handled by windows import numpy A = numpy.random.rand 100, 100) B = numpy.linalg.inv A) > crash here The crash s error code indicates an illegal instruction windows code 0xc000001d).
|
|
Interestingly, Wiping the NumPy installation and installing numpy 1.0.3.1 works just fine.
|
|
It looks like you switched compile flags or even switched the compiler between 1.0.3.1 and 1.0.4, and the new compiler generates op codes which the P3 doesn t know about.","bug"
|
|
"6013","1234","Faster array version of ndindex Trac 636).
|
|
Original ticket http projects.scipy.org numpy ticket 636 on 2007 12 18 by jarrodmillman, assigned to jarrodmillman.
|
|
The following is from an email from Jonathan Taylor I was needing an array representation of ndindex since ndindex only gives an iterator but array list ndindex)) takes too long.
|
|
There is prob some obvious way to do this I am missing but if not feel free to include this code which is much faster.
|
|
In [252] time a=np.array list np.ndindex 10,10,10,10,10,10))) CPU times user 11.61 s, sys 0.09 s, total 11.70 s Wall time 11.82 In [253] time a=ndtuples 10,10,10,10,10,10) CPU times user 0.32 s, sys 0.21 s, total 0.53 s Wall time 0.60 def ndtuples dims) Fast implementation of array list ndindex dims))).
|
|
Need a list because we will go through it in reverse popping off the size of the last dimension.
|
|
dims = list dims) N will keep track of the current length of the indices.
|
|
N = dims.pop ) At the beginning the current list of indices just ranges over the last dimension.
|
|
cur = np.arange N) cur = cur[ ,np.newaxis] while dims != [] d = dims.pop ) This repeats the current set of indices d times.
|
|
e.g.
|
|
[0,1,2] > [0,1,2,0,1,2,...,0,1,2] cur = np.kron np.ones d,1)),cur) This ranges over the new dimension and stretches it by N. e.g.
|
|
[0,1,2] > [0,0,...,0,1,1,...,1,2,2,...,2] front = np.arange d).repeat N)[ ,np.newaxis] This puts these two together.
|
|
cur = np.column stack front,cur)) N = d return cur","enhancement"
|
|
"6013","1235","typo in scalarmathmodule.c.src for ctype negative Trac 637).
|
|
Original ticket http projects.scipy.org numpy ticket 637 on 2007 12 19 by trac user dpgrote, assigned to unknown.
|
|
I believe there is a typo in scalarmathmodule.c.src at line 382, which defines uns.
|
|
Should it be uns= 0,1) 5,0 3 with the 0 and 1 swapped, since the order of types in name gives the signed first, then unsigned.
|
|
Thanks!
|
|
Dave","bug"
|
|
"6013","1236","var should take absolute value for complex numbers.
|
|
Trac 638).
|
|
Original ticket http projects.scipy.org numpy ticket 638 on 2007 12 30 by trac user akumar, assigned to unknown.
|
|
Hi!
|
|
I was just wondering why numpy.var gives me complex variances for complex numbers.
|
|
e.g.
|
|
h1 = randn 1000,1) 1j randn 1000,1)) sqrt 2) var h1) gives ` 0.00757596036094 0.0234608549341j) ` While I actually expect the output to be similar to that which is from var square abs h1))) which gives ` 0.994899762171 ` I feel this is an error, though I may be wrong. )
|
|
Thanks.
|
|
Kumar","bug"
|
|
"6013","1237","test format.py 283 ImportError No module named nose.tools Trac 639).
|
|
Original ticket http projects.scipy.org numpy ticket 639 on 2008 01 02 by nilswagner01, assigned to teoliphant.
|
|
Warning FAILURE importing tests for <module numpy.lib.format from ... packages numpy lib format.pyc > usr local lib64 python2.5 site packages numpy lib tests test format.py 283 ImportError No module named nose.tools in <module>)","bug"
|
|
"6013","1238","Index assignment into pre allocated array does silent type cast, causing user confusion or worse) Trac 640).
|
|
Original ticket http projects.scipy.org numpy ticket 640 on 2008 01 04 by trac user brorson, assigned to unknown.
|
|
NumPy gurus I just discovered this today.
|
|
It looks like a bug to me.
|
|
Please flame me mercilessly if I am wrong! )
|
|
Sometimes you need to initialize an array using zeros ) before doing an assignment to it in a loop.
|
|
If you assign a complex value to the initialized array, the imaginary part of the array is dropped.
|
|
Does NumPy do a silent type cast which causes this behavior?
|
|
Is this typecast a feature?
|
|
Below I attach a session log showing the bug.
|
|
Note that I have boiled down my complex code to this simple case for ease of comprehension.
|
|
[1] I will also input this bug into the tracking system.
|
|
By the way, this is NumPy 1.0.4 In [39] numpy.
|
|
version Out[39] 1.0.4 Cheers, Stuart Brorson Interactive Supercomputing, inc. 135 Beaver Street | Waltham | MA | 02452 | USA http www.interactivesupercomputing.com <session log> In [29] A = numpy.random.rand 4) 1j numpy.random.rand 4) In [30] B = numpy.zeros 4)) In [31] In [31] for i in range 4) .... B[i] = A[i] ....
|
|
In [32] A Out[32] array [ 0.12150180 0.00577893j, 0.39792515 0.03607227j, 0.61933379 0.04506978j, 0.56751678 0.24576083j]) In [33] B Out[33] array [ 0.1215018 , 0.39792515, 0.61933379, 0.56751678]) < session log> [1] Yes, I know that I should use vectorized code as often as possible, and that this example is not vectorized.
|
|
This is a simple example illustrating the problem.
|
|
Moreover, many times the computation you wish to perform can t easily be vectorized, leaving the nasty old for loop as the only choice......","bug"
|
|
"6013","1239","numpy.sqrt numpy.array [ 1.0], dtype=numpy.complex256)) segfaults Trac 641).
|
|
Original ticket http projects.scipy.org numpy ticket 641 on 2008 01 07 by nilswagner01, assigned to unknown.
|
|
>>> numpy.sqrt numpy.array [ 1.0], dtype=numpy.complex256)) array [0.0 1.7100542e 4937j], dtype=complex256) >>> numpy.sqrt numpy.array [ 1.0], dtype=numpy.complex256)) array [0.0 9.1807977e 4929j], dtype=complex256) >>> numpy.sqrt numpy.array [ 1.0], dtype=numpy.complex256)) array [0.0 5.3847568e 4675j], dtype=complex256) >>> numpy.sqrt numpy.array [ 1.0], dtype=numpy.complex256)) Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 46912509653888 LWP 17547)] 0x00002aaaab5a4c86 in mpn mul 1 ) from lib64 tls libc.so.6 gdb) gdb) bt 0 0x00002aaaab5a4c86 in mpn mul 1 ) from lib64 tls libc.so.6 1 0x00002aaaab5ad355 in printf fp ) from lib64 tls libc.so.6","bug"
|
|
"6013","1240","maskedarray concatenate does not handle nomask right unlike numpy.core.ma) Trac 642).
|
|
Original ticket http projects.scipy.org numpy ticket 642 on 2008 01 09 by trac user bsulman, assigned to unknown.
|
|
When concatenating masked arrays in maskedarray, if one array has mask nomask, all arrays after are treated as nomask.
|
|
Example In [1] import maskedarray as MA In [2] x=MA.zeros 5);y=MA.ones 5);m=MA.zeros 5);m[3]=1;y=MA.array y,mask=m) In [3] MA.concatenate x,y)) Out[3] masked array data = [ 0.
|
|
0.
|
|
0.
|
|
0.
|
|
0.
|
|
1.
|
|
1.
|
|
1.
|
|
1.
|
|
1.
|
|
], mask = False, fill value=1e 20) In [6] MA.concatenate y,x)) Out[6] masked array data = [1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0], mask = [False False False True False False False False False False], fill value=1e 20) The reason for this in the code branches maskedarray numpy ma core.py line 2735 as of revision 4625) ...
|
|
Check whether one of the arrays has a non empty mask... for x in arrays if getmask x) is not nomask break return data OK, so we have to concatenate the masks ...
|
|
This returns the data as soon as it hits an array that has mask=nomask return data should be put in an else clause as appears in numpy.core.ma for x in arrays if getmask x) is not nomask break else return masked array d) corrected maskedarray code ...
|
|
Check whether one of the arrays has a non empty mask... for x in arrays if getmask x) is not nomask break else return data OK, so we have to concatenate the masks ...
|
|
I filed this under scipy originally.
|
|
Looks like maskedarray was moved to numpy.
|
|
Sorry if this showed up twice for people.","bug"
|
|
"6013","1241","Tracebacks in Turkish locales Trac 643).
|
|
Original ticket http projects.scipy.org numpy ticket 643 on 2008 01 10 by trac user marco, assigned to rkern.
|
|
The following result in a traceback.
|
|
I suspect it might be related to the fact that importing gtk changes the default encoding to utf 8.
|
|
I m using python 2.5.1 and numpy 1.0.3.1 LANG=tr TR.UTF 8 python c import gtk, numpy Traceback most recent call last) File <string> , line 1, in <module> File usr lib64 python2.5 site packages numpy init .py , line 39, in <module> import core File usr lib64 python2.5 site packages numpy core init .py , line 8, in <module> import numerictypes as nt File usr lib64 python2.5 site packages numpy core numerictypes.py , line 241, in <module> void = allTypes[ void ] KeyError void","bug"
|
|
"6013","1242","distutils cpuinfo fails to detect some nocona based hardware Trac 644).
|
|
Original ticket http projects.scipy.org numpy ticket 644 on 2008 01 11 by trac user jsbronder, assigned to cournape.
|
|
Reference https bugs.gentoo.org show bug.cgi?id=183236 Basically, cpuinfo is failed to verify that the following cpuinfo is, in fact, nocona.
|
|
This has been verified against numpy 1.0.4. processor 0 vendor id GenuineIntel cpu family 6 model 15 model name Genuine Intel R) CPU 2.40GHz stepping 4 cpu MHz 2400.130 cache size 4096 KB physical id 0 siblings 2 core id 0 cpu cores 2 fpu yes fpu exception yes cpuid level 10 wp yes flags fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant tsc pni monitor ds cpl vmx est tm2 ssse3 cx16 xtpr lahf lm bogomips 4802.95 clflush size 64 cache alignment 64 address sizes 36 bits physical, 48 bits virtual power management","bug"
|
|
"6013","1243","maskedarray extrema operations break on non array types Trac 645).
|
|
Original ticket http projects.scipy.org numpy ticket 645 on 2008 01 14 by trac user bsulman, assigned to unknown.
|
|
Extrema operations in maskedarray raise an exception when used on a non array object.
|
|
Example In [1] import maskedarray as ma In [2] import numpy.core.ma as MA In [3] l=[1,2,3] In [4] max l) Out[4] 3 In [5] MA.maximum l) Out[5] 3 In [6] ma.max l) <type exceptions.AttributeError > Traceback most recent call last) home bsulman <ipython console> in <module> ) home bsulman lib python maskedarray core.py in max obj, axis, out) 2386 raise TypeError Output arrays Unsupported for masked arrays ) 2387 if axis is None > 2388 return maximum obj) 2389 else 2390 return maximum.reduce obj, axis) home bsulman lib python maskedarray core.py in call self, a, b) 2307 Executes the call behavior.
|
|
2308 if b is None > 2309 return self.reduce a) 2310 return where self.compare a, b), a, b) 2311 ......... home bsulman lib python maskedarray core.py in reduce self, target, axis) 2317 else 2318 kargs = {} > 2319 target = target.ravel ) 2320 if not m is nomask) 2321 m = m.ravel ) <type exceptions.AttributeError > list object has no attribute ravel The maskedarray version of extrema operation.reduce line 2535 in branches maskedarray numpy ma core.py) calls target.ravel ) for axis=None.
|
|
I m using the older scipy.sandbox version of maskedarray but the line number I gave is for the maskedarray branch version and the problem still appears to be there.)
|
|
This is a problem because doing a ` from maskedarray import ` will break the builtin max and min functions for non array objects.
|
|
Probably the easiest fix is just to call array object) in reduce?
|
|
Thanks!","bug"
|
|
"6013","1244","incorrect calculation on int64 array Trac 646).
|
|
Original ticket http projects.scipy.org numpy ticket 646 on 2008 01 19 by trac user latrine00, assigned to unknown.
|
|
I get curious results on array calculation in numpy 1.0.4 Example code in python import numpy a = numpy.arange 0, 200, 2, dtype= int64 ) np = a 1) 1 print a print np Second array np contains even numbers should be odd only array [ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80, 82, 84, 86, 88, 90, 92, 94, 96, 98, 100, 102, 104, 106, 108, 110, 112, 114, 116, 118, 120, 122, 124, 126, 128, 130, 132, 134, 136, 138, 140, 142, 144, 146, 148, 150, 152, 154, 156, 158, 160, 162, 164, 166, 168, 170, 172, 174, 176, 178, 180, 182, 184, 186, 188, 190, 192, 194, 196, 198], dtype=int64) array [ 1, 3, 5, 7, 9, 11, 12, 15, 17, 19, 20, 22, 24, 27, 28, 30, 33, 35, 37, 39, 41, 43, 45, 47, 49, 51, 53, 54, 57, 58, 60, 62, 64, 66, 68, 71, 72, 74, 77, 79, 81, 82, 85, 86, 89, 91, 93, 95, 96, 99, 101, 103, 105, 107, 108, 111, 112, 115, 117, 119, 120, 123, 124, 127, 129, 131, 132, 135, 137, 139, 141, 142, 144, 147, 148, 151, 153, 154, 157, 159, 161, 163, 164, 166, 168, 171, 173, 175, 176, 179, 181, 182, 185, 187, 188, 190, 192, 194, 197, 199], dtype=int64)","bug"
|
|
"6013","1245","Bus Error with object arrays on big endian Solaris system Trac 647).
|
|
Original ticket http projects.scipy.org numpy ticket 647 on 2008 01 21 by chanley, assigned to teoliphant.
|
|
The following will cause a bus error on a big endian machine Solaris 10 Sun in this case) > Python 2.5.1 r251 54863, Jun 29 2007, 15 29 55) [C] on sunos5 > Type help , copyright , credits or license for more information.
|
|
>>>> import numpy >>>> o = numpy.ndarray shape=3,dtype=[ SEGMENT , |S4 ), SPEC FOUND , |i1 )]) >>>> o1 = o.getfield numpy.dtype |S4 ),0) >>>> print o1[0] > UXxc3x90 >>>> print o1[1] > 4 >>>> print o1[2] > NT >>>> print o1 > Bus error core dumped) There are no issues on Linux or Mac OS X Intel based systems.
|
|
This example was done on the latest svn version of numpy r1.0.5.dev47360).","bug"
|
|
"6013","1246","Bus Error with object arrays on big endian system Trac 648).
|
|
Original ticket http projects.scipy.org numpy ticket 648 on 2008 01 21 by chanley, assigned to unknown.
|
|
The following will cause a bus error on a big endian machine Solaris 10 Sun in this case) > Python 2.5.1 r251 54863, Jun 29 2007, 15 29 55) [C] on sunos5 > Type help , copyright , credits or license for more information.
|
|
>>>> import numpy >>>> o = numpy.ndarray shape=3,dtype=[ SEGMENT , |S4 ), SPEC FOUND , |i1 )]) >>>> o1 = o.getfield numpy.dtype |S4 ),0) >>>> print o1[0] > UXxc3x90 >>>> print o1[1] > 4 >>>> print o1[2] > NT >>>> print o1 > Bus error core dumped) There are no issues on Linux or Mac OS X Intel based systems.
|
|
This example was done on the latest svn version of numpy r1.0.5.dev47360).","bug"
|
|
"6013","1247","numpy.linalg.lstsq segfaults on i686 32 bit and ia64 linux Trac 649).
|
|
Original ticket http projects.scipy.org numpy ticket 649 on 2008 01 24 by trac user psederberg, assigned to unknown.
|
|
Howdy Everybody I am coding a simple ridge regression with numpy and seem to have uncovered a bug that exists in numpy.linalg.lstsq, but NOT in scipy.linalg,lstsq.
|
|
Running the following code gives rise to a segfault on 32 bit i686 and 64 bit ia64 linux, but not on a amd64 linux machine START CODE import numpy as N set dimensions nsamples = 53 nfeatures = 1000 create random data data = N.random.rand nsamples,nfeatures) lab = N.random.rand nsamples) run a ridge regression with constant term Lambda = .05 nfeatures N.eye nfeatures) a = N.concatenate \ N.concatenate data, N.ones nsamples, 1))), 1), N.concatenate Lambda, N.zeros nfeatures, 1))), 1))) b = N.concatenate lab, N.zeros nfeatures))) w = N.linalg.lstsq a,b)[0] here is the segfault END CODE All machines were tested with Debian testing and numpy version 1.0.4.
|
|
I ve attached the results of running valgrind on the above code.
|
|
It should also be added that replacing the numpy lstsq with the scipy.linalg.lstsq prevents the segfault and gives rise to the correct result.
|
|
Thanks, Per","bug"
|
|
"6013","1248","fromfile and fromstring differ in parsing ascii integer strings Trac 650).
|
|
Original ticket http projects.scipy.org numpy ticket 650 on 2008 01 29 by efiring, assigned to unknown.
|
|
When parsing ascii with an integer dtype specified, fromfile uses fscanf with a format string that assumes base 10, while fromstring uses PyOS strtol with its third parameter, base, set to 0.
|
|
This causes auto detection of the base, allowing hex and octal to be recognized automatically.
|
|
The disadvantages are that this can cause the same string to be parsed differently by fromfile and fromstring, and that in common cases, such as parsing a date or time string, the leading 0 triggers unwanted interpretation of a number as octal which can fail outright) when it is actually base 10.
|
|
In [2] numpy.fromstring 12 09 09 , dtype=int, sep= ) Out[2] array [12, 0]) Changing the base argument to PyOS strtol from 0 to 10 in arraytypes.inc.src would change the behavior of fromstring to match that of fromfile and eliminate the surprising behavior illustrated above.
|
|
This has been discussed in the thread [http thread.gmane.org gmane.comp.python.numeric.general 18630]","bug"
|
|
"6013","1249","typos in compiled base.c function arr bincount Trac 651).
|
|
Original ticket http projects.scipy.org numpy ticket 651 on 2008 01 30 by sebhaase, assigned to unknown.
|
|
Hi, in compiled base.c current SVN) I might have found 2 typos in function arr bincount 1) missing `f` irst argument of bincount must be non negative ); 2) the starting comment says histogram a few times), should it say arr bincount !
|
|
?","bug"
|
|
"6013","1250","linalg.eig takes 100 cpu and never ends Trac 652).
|
|
Original ticket http projects.scipy.org numpy ticket 652 on 2008 02 01 by trac user PeterC, assigned to cournape.
|
|
Using the official binaries for 1.0.4 py2.4 the following code never ends from numpy import arange, float64 from numpy.linalg import eig a = arange 13 13, dtype = float64) a.shape = 13,13) a = a 17 eig a) however this completes on the official 1.0.3.1 py2.4 This may be related to 1225 but I do not have python 2.5 so I cannot try the binaries posted.
|
|
My system is Windows XP, Intel Xeon 5140","bug"
|
|
"6013","1251","v1.0.4 incorrectly identifies Xeon as 64 bit Trac 653).
|
|
Original ticket http projects.scipy.org numpy ticket 653 on 2008 02 04 by trac user doneal24, assigned to dmcooke.
|
|
Platform Gentoo Linux on Intel Xeon, 32 bit biowolf cat proc cpuinfo processor 0 vendor id GenuineIntel cpu family 15 model 2 model name Intel R) Xeon TM) CPU 2.80GHz stepping 9 cpu MHz 2799.930 cache size 512 KB physical id 0 siblings 2 core id 0 cpu cores 1 fdiv bug no hlt bug no f00f bug no coma bug no fpu yes fpu exception yes cpuid level 2 wp yes flags fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr bogomips 5572.09 clflush size 64 biowolf python Python 2.4.4 1, Jan 15 2008, 12 27 23) [GCC 4.1.2 Gentoo 4.1.2 p1.0.2)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy.distutils.cpuinfo import cpu >>> cpu.is 64bit ) True >>> Commenting out lines 282 283 in usr lib python2.4 site packages numpy distutils cpuinfo.py if self.info[0].get clflush size , )== 64 return True will then correctly return the system type as 32bit but this is not the optimal solution.
|
|
Incorrect identification of the system as 64 bit causes scipy to fail in many modules.","bug"
|
|
"6013","1252","Numpy tests failure on solaris Trac 654).
|
|
Original ticket http projects.scipy.org numpy ticket 654 on 2008 02 05 by trac user yippi, assigned to cournape.
|
|
I find that numpy has test programs in this directory usr lib python2.4 vendor packages numpy core tests Two tests seem to have failures.
|
|
test scalarmath.py and test numeric.py.
|
|
The other tests seem to pass okay.
|
|
I am running on Solaris Nevada x86.
|
|
Are you seeing these errors on other platforms?
|
|
I see the following output with the two failed tests.
|
|
python test scalarmath.py Found 6 6 tests for main .F.... ====================================================================== FAIL check large types main .TestPower) Traceback most recent call last) File test scalarmath.py , line 52, in check large types assert b == 6765201, error with r got r t,b) AssertionError error with <type numpy.float96 > got 6765201.00000000000364 Ran 6 tests in 0.078s FAILED failures=1) python test numeric.py Found 62 62 tests for main .......................................................................Warning invalid value encountered in absolute Warning invalid value encountered in absolute Warning invalid value encountered in less equal .....
|
|
Ran 76 tests in 0.124s OK","bug"
|
|
"6013","1253","Library for padding arrays Trac 655).
|
|
Original ticket http projects.scipy.org numpy ticket 655 on 2008 02 11 by trac user timcera, assigned to stefanv.
|
|
The attached pad module contains a group of functions to pad values onto the edges of an n dimensional array.
|
|
I use it for FFT analysis in TAPPy http tappy.sf.net) to have some hope of reasonable values all the way to the edge of the array.
|
|
I thought it might be useful.","enhancement"
|
|
"6013","1254","numpy.histogram fails if bins is a sequence Trac 656).
|
|
Original ticket http projects.scipy.org numpy ticket 656 on 2008 02 11 by trac user jochen.eppler, assigned to jarrodmillman.
|
|
With numpy 1.04, the following code import numpy xs = numpy.arange 0.0, 20.0, 0.1) hist = numpy.histogram xs, bins=numpy.linspace 0.0, 20.0, 4.0)) produces an exception.
|
|
exceptions.NameError Traceback most recent call last) usr lib python2.4 site packages numpy lib function base.py in histogram a, bins, range, normed) 152 bins = linspace mn, mx, bins, endpoint=False) 153 else > 154 if any bins[1 ] bins[ 1] < 0)) 155 raise AttributeError, bins must increase monotonically.
|
|
156 NameError global name any is not defined With 1.0.3 the function works nicely.","bug"
|
|
"6013","1255","x.flat[ ] interprets non native byteorder arrays incorrectly Trac 657).
|
|
Original ticket http projects.scipy.org numpy ticket 657 on 2008 02 11 by rkern, assigned to unknown.
|
|
In [9] from numpy import In [10] import sys In [11] sys.byteorder Out[11] little In [12] arange 10).astype >i4 ) Out[12] array [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) In [13] arange 10).astype >i4 ).flat[ ] Out[13] array [ 0, 16777216, 33554432, 50331648, 67108864, 83886080, 100663296, 117440512, 134217728, 150994944]) In [14] arange 10).astype >f8 ) Out[14] array [ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9.])
|
|
In [15] arange 10).astype >f8 ).flat[ ] Out[15] array [ 0.00000000e 000, 3.03865194e 319, 3.16202013e 322, 1.04346664e 320, 2.05531309e 320, 2.56123631e 320, 3.06715953e 320, 3.57308275e 320, 4.07900597e 320, 4.33196758e 320])","bug"
|
|
"6013","1256","Segfault on reshape Trac 658).
|
|
Original ticket http projects.scipy.org numpy ticket 658 on 2008 02 12 by stefanv, assigned to unknown.
|
|
import numpy as np np.indices 0,3,4)).T.reshape 1,3) ==10171== 233675 errors in context 93 of 95 ==10171== Invalid read of size 4 ==10171== at 0x4550E79 PyArray Newshape multiarraymodule.c 516) ==10171== by 0x4551B17 array reshape arraymethods.c 87) ==10171== by 0x80C8EEB PyEval EvalFrameEx in usr bin python2.5) ==10171== by 0x80CA114 PyEval EvalCodeEx in usr bin python2.5) ==10171== by 0x80CA186 PyEval EvalCode in usr bin python2.5) ==10171== by 0x80E9697 PyRun FileExFlags in usr bin python2.5) ==10171== by 0x80E9926 PyRun SimpleFileExFlags in usr bin python2.5) ==10171== by 0x805938F Py Main in usr bin python2.5) ==10171== by 0x80588C1 main in usr bin python2.5) ==10171== Address 0x59A3F50 is 0 bytes after a block of size 8 alloc d ==10171== at 0x4022765 malloc vg replace malloc.c 149) ==10171== by 0x453885E PyArray IntpConverter multiarraymodule.c 4549) ==10171== by 0x4551AF9 array reshape arraymethods.c 79) ==10171== by 0x80C8EEB PyEval EvalFrameEx in usr bin python2.5) ==10171== by 0x80CA114 PyEval EvalCodeEx in usr bin python2.5) ==10171== by 0x80CA186 PyEval EvalCode in usr bin python2.5) ==10171== by 0x80E9697 PyRun FileExFlags in usr bin python2.5) ==10171== by 0x80E9926 PyRun SimpleFileExFlags in usr bin python2.5) ==10171== by 0x805938F Py Main in usr bin python2.5) ==10171== by 0x80588C1 main in usr bin python2.5) ==10171== ==10171== 639859 errors in context 94 of 95 ==10171== Conditional jump or move depends on uninitialised value s) ==10171== at 0x4550E8A PyArray Newshape multiarraymodule.c 515) ==10171== by 0x4551B17 array reshape arraymethods.c 87) ==10171== by 0x80C8EEB PyEval EvalFrameEx in usr bin python2.5) ==10171== by 0x80CA114 PyEval EvalCodeEx in usr bin python2.5) ==10171== by 0x80CA186 PyEval EvalCode in usr bin python2.5) ==10171== by 0x80E9697 PyRun FileExFlags in usr bin python2.5) ==10171== by 0x80E9926 PyRun SimpleFileExFlags in usr bin python2.5) ==10171== by 0x805938F Py Main in usr bin python2.5) ==10171== by 0x80588C1 main in usr bin python2.5) ==10171== 639859 errors in context 95 of 95 ==10171== ==10171== Conditional jump or move depends on uninitialised value s) ==10171== at 0x4550E82 PyArray Newshape multiarraymodule.c 514) ==10171== by 0x4551B17 array reshape arraymethods.c 87) ==10171== by 0x80C8EEB PyEval EvalFrameEx in usr bin python2.5) ==10171== by 0x80CA114 PyEval EvalCodeEx in usr bin python2.5) ==10171== by 0x80CA186 PyEval EvalCode in usr bin python2.5) ==10171== by 0x80E9697 PyRun FileExFlags in usr bin python2.5) ==10171== by 0x80E9926 PyRun SimpleFileExFlags in usr bin python2.5) ==10171== by 0x805938F Py Main in usr bin python2.5) ==10171== by 0x80588C1 main in usr bin python2.5)","bug"
|
|
"6013","1257","Some strings containing zeros are displayed incorrectly.
|
|
Trac 659).
|
|
Original ticket http projects.scipy.org numpy ticket 659 on 2008 02 13 by charris, assigned to unknown.
|
|
The array display routine has some problems when it is showing strings containing zeros.
|
|
In [8] array [ \x00\x50 , \x50\x50 ]) Out[8] array [ \x00P , PP ], dtype= |S2 ) In [9] array [ \x00\x00 , \x50\x50 ]) Out[9] array [ , PP ], dtype= |S2 )","bug"
|
|
"6013","1258","Crash in numpy.unique1D release only) Trac 660).
|
|
Original ticket http projects.scipy.org numpy ticket 660 on 2008 02 19 by trac user Kaloo, assigned to jarrodmillman.
|
|
I have one example which causes a crash with unique1d .
|
|
I re compiled numpy into debug mode, and it didn t crash.
|
|
However, the bug is perfectly reproducible with release version numpy 1.0.4 ) I attached a script which loads data in a numpy array, and launches unique1d.","bug"
|
|
"6013","1259","maximum handles nan improperly Trac 661).
|
|
Original ticket http projects.scipy.org numpy ticket 661 on 2008 02 21 by trac user tmb, assigned to unknown.
|
|
Reproduce by >>> maximum 3,array [1,nan,3.0])) Warning invalid value encountered in maximum array [ 3. , nan, 3. ])
|
|
>>> Printing a warning is about the worst thing the system can do the error message is not informative and it messes up interactive and CGI output.
|
|
There is also no obvious way of disabling it, and the message doesn t even make it clear that it comes from numpy.maximum, leading people on a wild goose chase through their sources.
|
|
The correct behavior would be either to raise an exception if the condition is detected, or to quietly carry out the operation.
|
|
I think the default should be to raise an exception, with an option to disable.","bug"
|
|
"6013","1260","eigh hangup Trac 662).
|
|
Original ticket http projects.scipy.org numpy ticket 662 on 2008 02 21 by trac user Numpty, assigned to unknown.
|
|
numpy version[[BR]] `1.0.4` sys.version[[BR]] `2.5 r25 51908, Sep 19 2006, 09 52 17) [MSC v.1310 32 bit Intel)]` !python import numpy cov = numpy.array [ [ 77.70273908, 3.51489954, 15.64602427], [ 3.51489954, 88.97013878, 1.07431931], [ 15.64602427, 1.07431931, 98.18223512], ]) or any array I ve tried This never returns not within minutes anyway) vals, vecs = numpy.linalg.eigh cov) eig ) works and using the gui examples tool from the LAPACK 3.1.1 Windows installation, it seems that the DSYEVD routine mentioned in the docstring succeeds with these inputs.","bug"
|
|
"6013","1261","Support for multi formatting elements in savetxt Trac 663).
|
|
Original ticket http projects.scipy.org numpy ticket 663 on 2008 02 21 by huard, assigned to unknown.
|
|
Here is a patch to add support for multiple formats in savetxt.
|
|
Index numpy lib io.py =================================================================== numpy lib io.py revision 4815) numpy lib io.py working copy) 326,6 326,10 the file is automatically saved in compressed gzip format.
|
|
The load ) command understands gzipped files transparently.
|
|
fmt can be a single format 10.5f), a sequence of formats, or a single multi format string, e.g.
|
|
Iteration d 10.5f , in which case delimiter is ignored.
|
|
Example usage save test.out , X) X is an array 354,8 358,23 if len X.shape)==1 origShape = X.shape X.shape = len X), 1 Fmt can be a string with multiple insertion points or a list of formats.
|
|
E.g.
|
|
10.5f\t 10d or 10.5f , 10d ) if type fmt) in list, tuple) if len fmt) != X.shape[1] raise AttributeError, fmt has wrong shape.
|
|
str fmt) format = delimiter.join fmt) elif type fmt) is str if fmt.count ) == 1 fmt = [fmt,] X.shape[1] format = delimiter.join fmt) elif fmt.count ) != X.shape[1] raise AttributeError, fmt has wrong number of formats.
|
|
fmt else format = fmt for row in X fh.write delimiter.join [fmt val for val in row]) \n ) fh.write format tuple row) \n ) if origShape is not None X.shape = origShape","enhancement"
|
|
"6013","1262","more accurate representation of polynomials Trac 664).
|
|
Original ticket http projects.scipy.org numpy ticket 664 on 2008 02 22 by pv, assigned to unknown.
|
|
numpy.poly1d represents polynomials by storing the polynomial coefficients.
|
|
I believe that for high order polynomials, this is not always the optimal way to represent them.
|
|
Consider one of the orthogonal polynomial functions in scipy.special >>> import scipy, scipy.special >>> for n in xrange 0,70,10) print n, scipy.special.chebyt n) scipy.cos z)) scipy.cos n z) 0 0.0 10 3.80806497446e 14 20 1.94978477808e 10 30 2.65364360286e 06 40 0.0237953834481 50 114.714051465 60 272370.465462 It can be seen that as the order of the polynomial increases, the accuracy of evaluation decreases very rapidly, and for n > 40 the result is basically numerical noise.
|
|
The scipy.special.chebyt function generates a poly1d object by passing the roots of the polynomial to poly1d.
|
|
In poly1d.
|
|
init , polynomial coefficients are calculated from the roots.
|
|
For high order Chebyshev polynomials, the coefficients are large and of varying sign which leads to loss of precision when evaluating the polynomial.
|
|
It might be useful if there was a poly1d compatible object that would retain a more accurate representation of the polynomial.
|
|
Using the roots, or a Horner scheme?)
|
|
This would come useful when more information than the coefficients is initially available.
|
|
See also http scipy.org scipy scipy ticket 581","enhancement"
|
|
"6013","1263","100 cpu error consolidation, tickets 627, 628, 652 Trac 665).
|
|
Original ticket http projects.scipy.org numpy ticket 665 on 2008 02 24 by charris, assigned to unknown.
|
|
Tickets 627, 628, 652 are probably related, if not identical.
|
|
If one gets fixed, test the others also so that we can close all of them.","bug"
|
|
"6013","1264","Sort memcpy reminder Trac 666).
|
|
Original ticket http projects.scipy.org numpy ticket 666 on 2008 02 24 by charris, assigned to unknown.
|
|
Reminder to self to use straight memcpy in string sorts in a later release.
|
|
I m waiting on this to make sure the compiler library releases with good memcpy implementations are widespread.","bug"
|
|
"6013","1265","dotblas.so only built if ATLAS is installed Trac 667).
|
|
Original ticket http projects.scipy.org numpy ticket 667 on 2008 02 24 by trac user otizonaizit, assigned to unknown.
|
|
Dear numpy devs, after changeset 3854 http projects.scipy.org scipy numpy changeset 3854 ) dotblas only gets built when ATLAS is installed LAPACK BLAS is not enough anymore).
|
|
This prevents the python numpy debian package http packages.debian.org sid python numpy ) from being built with dotblas support see debian bug for details http bugs.debian.org cgi bin bugreport.cgi?bug=464784 ).
|
|
What is the reason for such a dependency?
|
|
Shouldn t the API for BLAS and ATLAS routines be exactly the same so that they can be used interchangeably?
|
|
Simply adding a dependency on ATLAS for python numpy in debian at the moment is not feasible see the debian bug report for more details).
|
|
Thank you for your attention and keep up with the very good work!
|
|
tiziano ps I ve set the priority for this bug report to high and severity to critical , because numpy on debian is presently really slow, at the point that is almost unusable, forcing users to build it from source.","bug"
|
|
"6013","1266","Explain why numpy headers are installed using add data dir Trac 668).
|
|
Original ticket http projects.scipy.org numpy ticket 668 on 2008 02 24 by charris, assigned to jarrodmillman.
|
|
Someone with sufficient privilege should add this explanation to the FAQs.
|
|
Why numpy headers are installed using add data dir and not add headers?
|
|
To enable several versions of numpy to be installed at the same time, as well as to deal more easily with eggs.
|
|
More precisely add headers install headers system wide e.g.
|
|
in usr include on unix if usr is the prefix for installation), whereas add data dir install the headers in package specific location for example somewhere in usr lib python2.5 site packages numpy ).
|
|
Installing the headers system wide prevents multiple version of numpy to be installed at the same time, and that s why add headers use is discouraged for numpy scipy.","bug"
|
|
"6013","1267","Undesired upcasting of uint64 to float64 in NumPy scalars Trac 669).
|
|
Original ticket http projects.scipy.org numpy ticket 669 on 2008 02 25 by FrancescAlted, assigned to unknown.
|
|
In the middle of integer arithmetic, I find this quite misleading In [8] 5 numpy.int64 3) Out[8] 2 In [9] 5 numpy.uint64 3) Out[9] 2.0 But the next is definitely worse In [10] 5 numpy.int64 3) Out[10] 1 In [11] 5 numpy.uint64 3) Out[11] 1.66666666667 In my case, this lead to many programming errors.
|
|
I do think that the outcome of an operation between integers should be an integer, as most of programming languages do.","bug"
|
|
"6013","1268","cmp arg types bug Trac 670).
|
|
Original ticket http projects.scipy.org numpy ticket 670 on 2008 02 25 by nbecker, assigned to teoliphant.
|
|
I dont think the current cmp arg types works at all, it doesn t even increment the pointers to examine all the types.
|
|
I think this is correct static int cmp arg types int arg1, int arg2, int n) { for ;n > 0; n , arg1, arg2) { if PyArray EquivTypenums arg1, arg2) || PyArray CanCastSafely arg1, arg2)) continue; return 1; } return 0; }","bug"
|
|
"6013","1269","linalg ImportError on Solaris 10 systems Trac 671).
|
|
Original ticket http projects.scipy.org numpy ticket 671 on 2008 02 26 by chanley, assigned to unknown.
|
|
When building numpy from svn on a Sun Ultra 10 running Solaris 10 the build completes without error.
|
|
However, when attempting to then import numpy I receive the following ImportError exception Python 2.5.1 r251 54863, Feb 8 2008, 15 02 09) [C] on sunos5 Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File data basil5 site pckages lib python numpy init .py , line 51, in <module> File data basil5 site pckages lib python numpy linalg init .py , line 4, in <module> File data basil5 site pckages lib python numpy linalg linalg.py , line 28, in <module> ImportError ld.so.1 python fatal relocation error file data basil5 site packages lib python numpy linalg lapack lite.so symbol s cat referenced symbol not found >>> Since numpy doesn t have a required dependency on lapack I found this error puzzling.
|
|
The exception also didn t occur previously.","bug"
|
|
"6013","1270","Fatal Python error PyEval RestoreThread NULL tstate Trac 672).
|
|
Original ticket http projects.scipy.org numpy ticket 672 on 2008 02 27 by trac user ggenellina, assigned to cournape.
|
|
This looks like an error in NumPy, altough I ve found it using matplotlib.
|
|
With Windows XP SP2, Python 2.5.1 from www.python.org, NumPy 1.0.4, matplotlib 0.91.2, all installed using the distributed binaries.
|
|
All matplotlib examples fail with an application error box, and this message Fatal Python error PyEval RestoreThread NULL tstate This application has requested the Runtime to terminate it in an unusual way.
|
|
Please contact the application s support team for more information.
|
|
The error happens inside dotblas.pyd.
|
|
Downgrading to NumPy 1.0.3.1 fixed the problem.","bug"
|
|
"6013","1271","change the error handling code in lapack lite to return a python exception Trac 673).
|
|
Original ticket http projects.scipy.org numpy ticket 673 on 2008 02 28 by trac user Zbyszek Szmek, assigned to unknown.
|
|
Function xerbla is called when an error e.g.
|
|
improper argument) is detected in functions converted to C from Fortran.
|
|
Citing the manpage for xerbla from the lapack3 library XERBLA is an error handler for the LAPACK routines.
|
|
It is called by an LAPACK routine if an input parameter has an invalid value.
|
|
A message is printed and execution stops.
|
|
Installers may consider modifying the STOP statement in order to call system specific exception handling facilities.
|
|
I believe that it should indeed be customized for lapack lite in numpy, because a. it tries to print a message using Fortran I O, which doesn t always work b. calls s stop, which simply exits the program returning 0 The Python interface validates all arguments, so that the error handling code in functions converted from Fortran is not called, but in case it happens, an exception should be thrown.
|
|
An example import numpy Ainv = numpy.linalg.inv numpy.matrix numpy.zeros 0,0)) ) ) print Ainv On suse 10.3 64bits) with python numpy 1.0.4 0.pm.2 running such a program gives no output Fortran I O doesn t work properly?)
|
|
and returns 0.
|
|
On debian with python numpy 10.0.3 1 it says On entry to DGESV parameter number 4 had an illegal value and returns 0.
|
|
With the attached patch, python returns 1 after printing Traceback most recent call last) File test linalg inv.py , line 2, in <module> Ainv = numpy.linalg.inv numpy.matrix numpy.zeros 0,0)) ) ) File usr lib64 python2.5 site packages numpy linalg linalg.py , line 246, in inv return wrap solve a, identity a.shape[0], dtype=a.dtype))) File usr lib64 python2.5 site packages numpy linalg linalg.py , line 193, in solve return b.transpose ).astype result t) ValueError On entry to DGESV parameter number 4 had an illegal value Of course all this is caused by a small bug in the python wrapper for dsegv.","bug"
|
|
"6013","1272","Impossible to rename object fields Trac 674).
|
|
Original ticket http projects.scipy.org numpy ticket 674 on 2008 02 28 by trac user sameer, assigned to unknown.
|
|
It is not possible to change the name of a field if it is an object array.
|
|
Please see the code block below or reference the discussion at http projects.scipy.org pipermail numpy discussion 2008 February 031654.html In [1] from numpy import In [2] olddt = dtype [ foo , |O4 ), bar , float)]) In [3] a = zeros 10, olddt) In [4] a Out[4] array [ 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0), 0, 0.0)], dtype=[ foo , |O4 ), bar , <f8 )]) In [5] newdt = dtype [ notfoo , |O4 ), notbar , float)]) In [6] b = a.view newdt) TypeError Traceback most recent call last) home titan sameer projects <ipython console> TypeError Cannot change data type for object array.
|
|
Also, it would be useful to have an API to rename record array fields.
|
|
The current method of creating a new dtype and calling the view method of the record array seems like too much of a hack.","bug"
|
|
"6013","1273","0 d array does not allow masked bool) indexing Trac 675).
|
|
Original ticket http projects.scipy.org numpy ticket 675 on 2008 03 01 by stefanv, assigned to teoliphant.
|
|
import numpy as np x = np.array [3]) x[[True]] yields IndexError index 1 out of bounds 0<=index<1 The boolean index should not be converted to int.","bug"
|
|
"6013","1274","fortran order .flatten ) fails for multidimensional arrays Trac 676).
|
|
Original ticket http projects.scipy.org numpy ticket 676 on 2008 03 03 by pv, assigned to unknown.
|
|
Fortran order .flatten F ) does not actually output fortran order flattened data for multidimensional ndim > 2) arrays.
|
|
The fortran order flatten should return [1, 3, 2, 4] also in the second case >>> import numpy as N >>> N. version 1.0.4 >>> x = N.array [[1, 2], [3, 4]]) >>> x.flatten ) array [1, 2, 3, 4]) >>> x.flatten F ) array [1, 3, 2, 4]) >>> x = N.array [[[1, 2], [3, 4]]]) >>> x.flatten ) array [1, 2, 3, 4]) >>> x.flatten F ) array [1, 2, 3, 4]) The code responsible for flattening appears to be in arrayobject.c flat copyinto","bug"
|
|
"6013","1275","format string substitutable parameter mismatch in numpy core src arrayobject.c Trac 677).
|
|
Original ticket http projects.scipy.org numpy ticket 677 on 2008 03 04 by trac user paul.metcalfe, assigned to unknown.
|
|
Changeset 4839 introduced a call to `PyErr Format` at line 11006 of `numpy core src arrayobject.c`.
|
|
The format string has one substitutable parameter; two are given.
|
|
Attached very trivial patch fixes it.
|
|
HTH.","bug"
|
|
"6013","1276","fromfile ) reading past end of file causes unhandled exception Trac 678).
|
|
Original ticket http projects.scipy.org numpy ticket 678 on 2008 03 05 by trac user lcasburn, assigned to unknown.
|
|
Reading binary file past end of file causes crash; needs better error handling to catch unhandled exception.
|
|
Example interactive input that causes crash >>> from numpy import >>> x = arange 20).reshape 4,5) >>> x.tofile testBinary.mtx ) >>> >>> myfile = open testBinary.mtx , rb ) >>> print fromfile myfile, dtype=int, count=5) [0 1 2 3 4] >>> print fromfile myfile, dtype=int, count=5) [5 6 7 8 9] >>> print fromfile myfile, dtype=int, count=5) [10 11 12 13 14] >>> print fromfile myfile, dtype=int, count=5) [15 16 17 18 19] >>> print fromfile myfile, dtype=int, count=5)","bug"
|
|
"6013","1277","ERROR check MyPyLong AsUnsignedLongLong test regression.TestRegression) Trac 679).
|
|
Original ticket http projects.scipy.org numpy ticket 679 on 2008 03 06 by chanley, assigned to unknown.
|
|
As of r4851 I have been receiving this error on a RHE4 64 bit system.
|
|
This same error also occurs in r4852.
|
|
====================================================================== ERROR check MyPyLong AsUnsignedLongLong test regression.TestRegression) Traceback most recent call last) File data thor1 chanley dev site packages lib python numpy core tests test regression.py , line 806, in check MyPyLong AsUnsignedLongLong assert equal np.uint64 2), np.uint64 18446744073709551614)) OverflowError long too big to convert","bug"
|
|
"6013","1278","Bug in numpy core defmatrix.py Trac 680).
|
|
Original ticket http projects.scipy.org numpy ticket 680 on 2008 03 09 by nilswagner01, assigned to unknown.
|
|
>>> bmat [[identity 2),zeros 2,2))],[zeros 2,2)),zeros 2,2))]]) matrix [[ 1., 0., 0., 0.
|
|
], [ 0., 1., 0., 0.
|
|
], [ 0., 0., 0., 0.
|
|
], [ 0., 0., 0., 0.]])
|
|
works fine, but >>> bmat identity 2), zeros 2,2)); zeros 2,2)), zeros 2,2)) ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File usr lib python2.4 site packages numpy core defmatrix.py , line 503, in bmat return matrix from string obj, glob dict, loc dict)) File usr lib python2.4 site packages numpy core defmatrix.py , line 472, in from string raise KeyError, s not found col,) KeyError identity 2) not found","bug"
|
|
"6013","1279","numpy.random support for ACML Trac 681).
|
|
Original ticket http projects.scipy.org numpy ticket 681 on 2008 03 10 by trac user eilifmuller, assigned to unknown.
|
|
numpy.random performance is weak compared to matlab.
|
|
The AMD Core Math Library ACML) provides fast random number generator functions for use with AMD64 architectures, and it would be nice to somehow incorporate these into numpy.random.
|
|
Attached is a simple implementation of some of the basic functions of numpy.random using ACML.
|
|
The performance is superior to matlab rand,randn.","enhancement"
|
|
"6013","1280","fromfile tofile broken under Windows Trac 682).
|
|
Original ticket http projects.scipy.org numpy ticket 682 on 2008 03 11 by stefanv, assigned to unknown.
|
|
I O roundtrip fails under all Windows platforms see http buildbot.scipy.org) FAIL test file numpy.core.tests.test multiarray.TestFromToFile) x array [ 1.04871876e 316 0.65881535j, 0.00000000e 000 0.76613079j, 0.00000000e 000 0.80686635j, 0.00000000e 000 0.03426517j, 0.00000000e 000 0.31677123j, 0.00000000e 000 0.15563303j,... y array [ 0.58894384 0.65881535j, 0.68303040 0.76613079j, 0.41126205 0.80686635j, 0.99830670 0.03426517j, 0.19038197 0.31677123j, 0.90550957 0.15563303j,...
|
|
Priority set to highest in view of 1.0.5 release that has to be made.","bug"
|
|
"6013","1281","MyPyLong AsUnsignedLongLong fails on 64bit nix platforms Trac 683).
|
|
Original ticket http projects.scipy.org numpy ticket 683 on 2008 03 11 by jarrodmillman, assigned to charris.
|
|
The buildbot for Linux x86 64 Ubuntu and FreeBSD x86 64 has the following error ====================================================================== ERROR check MyPyLong AsUnsignedLongLong test regression.TestRegression) Traceback most recent call last) File home stefan Buildbot slave numpy b1 numpy install lib python2.5 site packages numpy core tests test regression.py , line 806, in check MyPyLong AsUnsignedLongLong assert equal np.uint64 2), np.uint64 18446744073709551614)) OverflowError long too big to convert I can verify this error on my 64bit Fedora 8 machine >>> numpy.uint64 2) 18446744073709551614 >>> numpy.uint64 18446744073709551614) Traceback most recent call last) File <stdin> , line 1, in <module> OverflowError long too big to convert","bug"
|
|
"6013","1282","Intp takes 2 arguments on all platforms except Windows x86 64 MSVC Trac 684).
|
|
Original ticket http projects.scipy.org numpy ticket 684 on 2008 03 11 by stefanv, assigned to unknown.
|
|
ERROR Ticket 99 Traceback most recent call last) File c \numpy buildbot\numpy\b11\install\Lib\site packages\numpy\core\tests\test regression.py , line 197, in check intp np.intp 0x f i width,16) TypeError function takes at most 1 argument 2 given) Marked as blocker before release of numpy 1.0.5.","bug"
|
|
"6013","1283","Intp takes 2 arguments on all platforms except Windows x86 64 MSVC Trac 685).
|
|
Original ticket http projects.scipy.org numpy ticket 685 on 2008 03 11 by stefanv, assigned to unknown.
|
|
ERROR Ticket 99 Traceback most recent call last) File c \numpy buildbot\numpy\b11\install\Lib\site packages\numpy\core\tests\test regression.py , line 197, in check intp np.intp 0x f i width,16) TypeError function takes at most 1 argument 2 given) Marked as blocker before release of numpy 1.0.5.","bug"
|
|
"6013","1284","Intp takes 2 arguments on all platforms except Windows x86 64 MSVC Trac 686).
|
|
Original ticket http projects.scipy.org numpy ticket 686 on 2008 03 11 by stefanv, assigned to unknown.
|
|
ERROR Ticket 99 Traceback most recent call last) File c \numpy buildbot\numpy\b11\install\Lib\site packages\numpy\core\tests\test regression.py , line 197, in check intp np.intp 0x f i width,16) TypeError function takes at most 1 argument 2 given) Marked as blocker before release of numpy 1.0.5.","bug"
|
|
"6013","1285","site.cfg ConfigParser in Python 2.6 does not allow section DEFAULT Trac 687).
|
|
Original ticket http projects.scipy.org numpy ticket 687 on 2008 03 11 by stefanv, assigned to unknown.
|
|
http www.gossamer threads.com lists python checkins 624896 At the moment, we use the DEFAULT section in site.cfg, which won t work under Python 2.6 any longer.","bug"
|
|
"6013","1286","Numpy and scipy should give informative error messages when run from source dir Trac 688).
|
|
Original ticket http projects.scipy.org numpy ticket 688 on 2008 03 11 by stefanv, assigned to stefanv.
|
|
Currently, running numpy or scipy from the source dir causes Traceback most recent call last) File <stdin> , line 1, in <module> File scipy init .py , line 54, in <module> from config import show as show config ImportError No module named config This should be replaced by a proper error message.","bug"
|
|
"6013","1287","SyntaxError while installing v1.0.5 on MacTel running OSX v10.4.11 Trac 689).
|
|
Original ticket http projects.scipy.org numpy ticket 689 on 2008 03 11 by trac user msousa, assigned to unknown.
|
|
byte compiling System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy ma extras.py to extras.pyc File System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy ma extras.py , line 759 tmp = numpy.fromiter g[1] for g in group), int ) SyntaxError invalid syntax","bug"
|
|
"6013","1288","No test for r4806 Trac 690).
|
|
Original ticket http projects.scipy.org numpy ticket 690 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
Changelog states For 4D samples, histogramdd returned badly shaped histograms once in a while.
|
|
This should be fixed now.
|
|
But no test has been submitted.
|
|
Marking as blocker since we cannot release 1.0.5 before testing added fixed functionality.","bug"
|
|
"6013","1289","No test for r4819 Trac 691).
|
|
Original ticket http projects.scipy.org numpy ticket 691 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
The bug came in through the Debian bug tracker, which means our test machinery didn t catch it.
|
|
Marking as blocker since we shouldn t release 1.0.5 before testing added fixed functionality.","bug"
|
|
"6013","1290","No test for r4822 Trac 692).
|
|
Original ticket http projects.scipy.org numpy ticket 692 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
Changes to C code without corresponding tests.
|
|
Marking as blocker since we should test all new features before releasing 1.0.5.","bug"
|
|
"6013","1291","No test for r4826 Trac 693).
|
|
Original ticket http projects.scipy.org numpy ticket 693 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
This changeset implements printing of complicated dtypes, but no test has been submitted.
|
|
Marking as a blocker, since 1.0.5 should not be released without testing all new functionality.","bug"
|
|
"6013","1292","No test for r4827 possibly) Trac 694).
|
|
Original ticket http projects.scipy.org numpy ticket 694 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
r4827 itself does not contain a test for the fix it implements, but it could be that a roundtrip test has been added later.
|
|
Make sure.","bug"
|
|
"6013","1293","No test for r4836 Trac 695).
|
|
Original ticket http projects.scipy.org numpy ticket 695 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
Locale problems are notoriously hard to test.
|
|
Investigate whether it is possible in this case.","bug"
|
|
"6013","1294","No test for r4853 Trac 696).
|
|
Original ticket http projects.scipy.org numpy ticket 696 on 2008 03 12 by stefanv, assigned to stefanv.
|
|
New ddof parameter should be tested.
|
|
Marking as blocker for 1.0.5 since all new functionality should be tested before release.","bug"
|
|
"6013","1295","ValueError shape mismatch objects cannot be broadcast to a single shape Trac 697).
|
|
Original ticket http projects.scipy.org numpy ticket 697 on 2008 03 13 by nilswagner01, assigned to charris.
|
|
usr bin python i julien.py Traceback most recent call last) File julien.py , line 36, in ?
|
|
ar,br) = numpy.polyfit T,XXX,1) File usr lib python2.4 site packages numpy lib polynomial.py , line 310, in polyfit c = vander [scale], order)[0] ValueError shape mismatch objects cannot be broadcast to a single shape","bug"
|
|
"6013","1296","segfault float96 None Trac 698).
|
|
Original ticket http projects.scipy.org numpy ticket 698 on 2008 03 14 by trac user gabriel, assigned to stefanv.
|
|
If I accidentally) try to add None to a float96, I get a segmentation fault instead of an error In [1] import numpy In [2] numpy.
|
|
version Out[2] 1.0.4 In [3] numpy.float96 3.0) None Segmentation fault core dumped) Ubuntu linux on a pentium 4.
|
|
Numpy compiled from tarball.
|
|
Best, Gabriel","bug"
|
|
"6013","1297","corrcoef incorrect behaviour Trac 699).
|
|
Original ticket http projects.scipy.org numpy ticket 699 on 2008 03 15 by trac user kazboris021, assigned to unknown.
|
|
In [1] import numpy as N In [2] N.corrcoef [0,0,0],[1,2,3]) Warning invalid value encountered in divide Out[2] array [[ NaN, NaN], [ NaN, 1.]])
|
|
In [3] numpy.corrcoef [0,0,0],[1,2,3])[0,0] is numpy.nan Warning invalid value encountered in divide Out[3] False Should be True I think","bug"
|
|
"6013","1298","Test for mean fails, incorrect processing of weights Trac 700).
|
|
Original ticket http projects.scipy.org numpy ticket 700 on 2008 03 15 by stefanv, assigned to stefanv.
|
|
David Huard added the following test which exposes some broken behaviour in mean ====================================================================== ERROR check shape numpy.lib.tests.test function base.TestAverage) Traceback most recent call last) File home stefan Buildbot slave numpy b1 numpy install lib python2.5 site packages numpy lib tests test function base.py , line 77, in check shape assert array equal average y, weights=w1, axis=1), desired) File .. numpy install lib python2.5 site packages numpy lib function base.py , line 383, in average ValueError axis not in array This ticket blocks the release of numpy 1.0.5.","bug"
|
|
"6013","1299","The returned keyword in average looks useless.
|
|
Trac 701).
|
|
Original ticket http projects.scipy.org numpy ticket 701 on 2008 03 16 by charris, assigned to unknown.
|
|
Remove the returned keyword from average.
|
|
This would clean up the code a bit and get rid of one more function with multiple return types.
|
|
I don t see any use for the sum of the weights except for debugging average itself.","enhancement"
|
|
"6013","1300","Memory leak in generic 1d array Trac 702).
|
|
Original ticket http projects.scipy.org numpy ticket 702 on 2008 03 17 by trac user robertwb, assigned to unknown.
|
|
The following file illustrates the issue.
|
|
It looks like the elements of the numpy array are not getting freed as they should.","bug"
|
|
"6013","1301","ma.masked where can make a defective array Trac 703).
|
|
Original ticket http projects.scipy.org numpy ticket 703 on 2008 03 17 by efiring, assigned to unknown.
|
|
As illustrated below, masked where can generate what appears at first to be a valid masked array, but which will later fail with an obscure looking error.
|
|
The invalid condition argument needs to be trapped in masked where, and an informative exception raised.
|
|
In [52] x = [1,2] In [53] y = ma.masked where False, x) In [54] y Out[54] masked array data = [1 2], mask = False, fill value=999999) In [55] y[1] IndexError Traceback most recent call last) home efiring <ipython console> in <module> ) usr local lib python2.5 site packages numpy ma core.pyc in getitem self, indx) 1307 if not getattr dout, ndim , False) 1308 Just a scalar............ > 1309 if m is not nomask and m[indx] 1310 return masked 1311 else IndexError 0 d arrays can t be indexed","bug"
|
|
"6013","1302","Building SciPy crashes if no FORTRAN compiler available Trac 704).
|
|
Original ticket http projects.scipy.org numpy ticket 704 on 2008 03 18 by jpaalasm, assigned to dmcooke.
|
|
Bue to a faulty checking of self.fcompiler being None self.compiler is checked instead), building scipy crashes when no fortran compiler if installed.
|
|
Output from running python2.5 setup.py build in SciPy s source directory don t know how to compile Fortran code on platform posix Traceback most recent call last) File setup.py , line 53, in <module> setup package ) File setup.py , line 45, in setup package configuration=configuration ) File usr lib python2.5 site packages numpy distutils core.py , line 184, in setup return old setup new attr) File usr lib python2.5 distutils core.py , line 151, in setup dist.run commands ) File usr lib python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr lib python2.5 site packages numpy distutils command build.py , line 40, in run old build.run self) File usr lib python2.5 distutils command build.py , line 112, in run self.run command cmd name) File usr lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr lib python2.5 site packages numpy distutils command build clib.py , line 83, in run self.fcompiler.customize self.distribution) AttributeError NoneType object has no attribute customize","bug"
|
|
"6013","1303","[ numpy Bugs 601052 ] SVD did not converge Trac 705).
|
|
Original ticket http projects.scipy.org numpy ticket 705 on 2008 03 19 by trac user menscher, assigned to unknown.
|
|
This was reported and supposedly fixed by dubois) in 2002, but the fix isn t in current code.
|
|
So reporting again....
|
|
The routine singular value decomposition ) sometimes dies with the error SVD did not converge .
|
|
But there is always by definition) an SVD inverse, so this is not an acceptable error message.
|
|
Tracing things back, it comes from an max iteration count that is too small in the file dlapack lite.c.
|
|
For the latest version, it s on line 22562.
|
|
When I change the value from 20 to 50, my matrix can be inverted.
|
|
There s no reason to have this number so small, since it s really just there to prevent infinite loops.
|
|
I ve changed it in my copy) to 100, and recommend you do the same.
|
|
I m attaching a pickle of a matrix I used for testing purposes.
|
|
Damian","bug"
|
|
"6013","1304","[ numpy Bugs 601052 ] SVD did not converge Trac 706).
|
|
Original ticket http projects.scipy.org numpy ticket 706 on 2008 03 19 by trac user menscher, assigned to stefanv.
|
|
This was reported and supposedly fixed by dubois) in 2002, but the fix isn t in current code.
|
|
So reporting again....
|
|
The routine singular value decomposition ) sometimes dies with the error SVD did not converge .
|
|
But there is always by definition) an SVD inverse, so this is not an acceptable error message.
|
|
Tracing things back, it comes from an max iteration count that is too small in the file dlapack lite.c.
|
|
For the latest version, it s on line 22562.
|
|
When I change the value from 20 to 50, my matrix can be inverted.
|
|
There s no reason to have this number so small, since it s really just there to prevent infinite loops.
|
|
I ve changed it in my copy) to 100, and recommend you do the same.
|
|
I m attaching a pickle of a matrix I used for testing purposes.
|
|
Damian","bug"
|
|
"6013","1305","numpy.array failes if the input is a list of matrixes Trac 707).
|
|
Original ticket http projects.scipy.org numpy ticket 707 on 2008 03 19 by trac user markmat, assigned to unknown.
|
|
numpy.array failes if the input is a list of matrixes with more then one column).
|
|
>>> import sys, numpy >>> sys.version 2.5 r25 51908, Sep 19 2006, 09 52 17) [MSC v.1310 32 bit Intel)] >>> numpy.version.version 1.0.4 >>> numpy.array [numpy.mat 1 0 )]) Traceback most recent call last) File <pyshell 3> , line 1, in <module> numpy.array [numpy.mat 1 0 )]) ValueError setting an array element with a sequence.
|
|
>>> numpy.array [numpy.mat 1; 0 )]) array [[[1], [0]]])","bug"
|
|
"6013","1306","Setuptools call to numpy trunk causes numpy.scons error Trac 708).
|
|
Original ticket http projects.scipy.org numpy ticket 708 on 2008 03 20 by matthew brett, assigned to cournape.
|
|
I am trying to install some packages using setuptools in fact from enthought, and I get a traceback into numpy distutils, ending with [mb312 localhost ] easy install f http code.enthought.com enstaller eggs source enthought.tvtk Searching for enthought.tvtk Reading http code.enthought.com enstaller eggs source Best match enthought.tvtk 2.0.2 Downloading http code.enthought.com enstaller eggs source enthought.tvtk 2.0.2.tar.gz Processing enthought.tvtk 2.0.2.tar.gz Running enthought.tvtk 2.0.2 setup.py q bdist egg dist dir home mb312 tmp easy install pPImLs enthought.tvtk 2.0.2 egg dist tmp nDpb2A Traceback most recent call last) File usr bin easy install , line 8, in <module> load entry point setuptools==0.6c6 , console scripts , easy install ) ) File usr lib python2.5 site packages easy install.py , line 1670, in main File usr lib python2.5 site packages easy install.py , line 1659, in with ei usage File usr lib python2.5 site packages easy install.py , line 1674, in <lambda> File usr lib64 python2.5 distutils core.py , line 151, in setup dist.run commands ) File usr lib64 python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File usr lib64 python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr lib python2.5 site packages easy install.py , line 211, in run File usr lib python2.5 site packages easy install.py , line 446, in easy install File usr lib python2.5 site packages easy install.py , line 471, in install item File usr lib python2.5 site packages easy install.py , line 655, in install eggs File usr lib python2.5 site packages easy install.py , line 930, in build and install File usr lib python2.5 site packages easy install.py , line 919, in run setup File home mandrake rpm tmp python setuptools root usr lib python2.5 site packages setuptools sandbox.py , line 27, in run setup File home mandrake rpm tmp python setuptools root usr lib python2.5 site packages setuptools sandbox.py , line 63, in run File home mandrake rpm tmp python setuptools root usr lib python2.5 site packages setuptools sandbox.py , line 29, in <lambda> File setup.py , line 70, in <module> File setup.py , line 16, in configuration File home mb312 lib64 python2.5 site packages numpy distutils misc util.py , line 837, in add subpackage caller level = 2) File home mb312 lib64 python2.5 site packages numpy distutils misc util.py , line 820, in get subpackage caller level = caller level 1) File home mb312 lib64 python2.5 site packages numpy distutils misc util.py , line 767, in get configuration from setup py config = setup module.configuration args) File enthought tvtk setup.py , line 18, in configuration File home mb312 lib64 python2.5 site packages numpy distutils misc util.py , line 850, in add subpackage dist = self.get distribution ) File home mb312 lib64 python2.5 site packages numpy distutils misc util.py , line 717, in get distribution return get distribution ) File home mb312 lib64 python2.5 site packages numpy distutils core.py , line 100, in get distribution raise NotImplementedError setuptools not supported yet for numpy.scons branch ) NotImplementedError setuptools not supported yet for numpy.scons branch","bug"
|
|
"6013","1307","isnan should not raise TypeErrors Trac 709).
|
|
Original ticket http projects.scipy.org numpy ticket 709 on 2008 03 20 by trac user cjw296, assigned to unknown.
|
|
I m faily sure that numpy.isnan datetime.datetime.now )) ...should just return False and not raise an exception.
|
|
rationale anything that is not nan should just return False, not raise TypeErrors.","bug"
|
|
"6013","1308","Reference count error, in array reshape?
|
|
Trac 710).
|
|
Original ticket http projects.scipy.org numpy ticket 710 on 2008 03 21 by matthew brett, assigned to unknown.
|
|
In some runs of numpy.test ) Reference count error detected an attempt was made to deallocate 17 O) Reference count error detected an attempt was made to deallocate 17 O) Reference count error detected an attempt was made to deallocate 17 O) etc) In [2] numpy.
|
|
version Out[2] 1.0.5.dev4899 Linux localhost 2.6.25 rc6 x61 1 SMP Tue Mar 18 18 22 11 CET 2008 x86 64 Intel R) Core TM)2 Duo CPU L7300 1.40GHz GNU Linux Valgrind session excerpt Ticket 658==26607== ==26607== Conditional jump or move depends on uninitialised value s) ==26607== at 0x659D56A PyArray UpdateFlags arrayobject.c 5043) ==26607== by 0x65CBC54 PyArray Newshape multiarraymodule.c 742) ==26607== by 0x65CC2A3 array reshape arraymethods.c 87) ==26607== by 0x51D626D PyEval EvalFrameEx in usr lib64 libpython2.5.so.1.0) ==26607== by 0x51D83F7 PyEval EvalCodeEx in usr lib64 libpython2.5.so.1.0) ==26607== by 0x51D67CC PyEval EvalFrameEx in usr lib64 libpython2.5.so.1.0) ==26607== by 0x51D83F7 PyEval EvalCodeEx in usr lib64 libpython2.5.so.1.0) ==26607== by 0x517F780 within usr lib64 libpython2.5.so.1.0) ==26607== by 0x51633C2 PyObject Call in usr lib64 libpython2.5.so.1.0) ==26607== by 0x51D504A PyEval EvalFrameEx in usr lib64 libpython2.5.so.1.0) ... ok","bug"
|
|
"6013","1309","object array refcount bug Trac 711).
|
|
Original ticket http projects.scipy.org numpy ticket 711 on 2008 03 21 by pv, assigned to unknown.
|
|
In r4906 there is another object array refcount bug >>> import numpy as np >>> a = object ) >>> m = np.zeros 5, dtype=np.object ) >>> m[ ] = a >>> del a >>> m[ ] = m >>> m[ ] = m Segmentation fault This occurs because in the object array copy pattern in r4906, all elements of m[ ] are DECREF d on slice assignment before copying, and the refcount of a so hits zero before the copy is done.","bug"
|
|
"6013","1310","scalar math test failure on solaris Trac 712).
|
|
Original ticket http projects.scipy.org numpy ticket 712 on 2008 03 23 by cournape, assigned to unknown.
|
|
See 1252.
|
|
I think the error is in the test, but someone who knows about this part of numpy should look at it more carefully.","bug"
|
|
"6013","1311","numpy.any segfaults for object arrays with >=10000 elements Trac 713).
|
|
Original ticket http projects.scipy.org numpy ticket 713 on 2008 03 25 by trac user mmanns, assigned to unknown.
|
|
The numpy.any function segfaults for object arrays with >=10000 elements.
|
|
2 dimensional arrays of the respective size also crash.
|
|
The type of object in the array does not seem to influence the problem.
|
|
> python > Python 2.5.1 r251 54863, Jan 26 2008, 01 34 00) > [GCC 4.1.2 Gentoo 4.1.2)] on linux2 > Type help , copyright , credits or license for more information.
|
|
> >>> import numpy > >>> numpy.version.version > 1.0.4 > >>> numpy.version.release > True > >>> small zero = [0] 9999 > >>> large zero = [0] 10000 > >>> small none = [None] 9999 > >>> large none = [None] 10000 > >>> any small zero) > False > >>> any large zero) > False > >>> any small none) > False > >>> any large none) > False > >>> any numpy.array small zero)) > False > >>> any numpy.array large zero)) > False > >>> any numpy.array small none)) > False > >>> any numpy.array large none)) > False > >>> numpy.any numpy.array small zero)) > False > >>> numpy.any numpy.array large zero)) > False > >>> numpy.any numpy.array small none)) > False > >>> numpy.any numpy.array large none)) > Segmentation fault","bug"
|
|
"6013","1312","segfault when using 0 d array as index Trac 714).
|
|
Original ticket http projects.scipy.org numpy ticket 714 on 2008 04 01 by trac user gabriel, assigned to unknown.
|
|
In [1] import numpy In [2] numpy.
|
|
version Out[2] 1.0.4 In [3] numpy.zeros 10)[numpy.array 0)] Segmentation fault core dumped) I work with Linux, pentium 4 machine.
|
|
But another pc in the lab with windows had the same problem.
|
|
Best, Gabriel","bug"
|
|
"6013","1313","The interactive mode of setup.py incorrect lower cases prefix settings Trac 715).
|
|
Original ticket http projects.scipy.org numpy ticket 715 on 2008 04 01 by rkern, assigned to pearu.
|
|
The menu system automatically lower cases inputs.
|
|
On case sensitive file systems, this means that the path set for the prefix is also lower cased.
|
|
I would prefer simply to remove the interactive setup.
|
|
It will always be a non standard method of installation, and I don t see evidence of people using it except for people who have problems with it.","bug"
|
|
"6013","1314","Inconsistent behavior, i686 vs x86 64 Trac 716).
|
|
Original ticket http projects.scipy.org numpy ticket 716 on 2008 04 02 by trac user jamie ..., assigned to unknown.
|
|
Gentoo, numpy 1.0.4 r2.
|
|
Possible related to http scipy.org scipy numpy ticket 669 Simple example, [either i686 or x86 64] >>> numpy.uint64 418465321) 8928000 46.871115703405017 >>> numpy.uint32 418465321) 8928000 46 Example that actually bit me, [i686] >>> A = numpy.array [1, 2, 3, 4], numpy.uint8) >>> B = numpy.array [2, 4, 6, 8], numpy.uint8) >>> C = B A >>> S = C.sum ) >>> type S) <type numpy.uint32 > >>> L = len C) >>> type L) <type int > >>> S L 2 [x86 64] >>> A = numpy.array [1, 2, 3, 4], numpy.uint8) >>> B = numpy.array [2, 4, 6, 8], numpy.uint8) >>> C = B A >>> S = C.sum ) >>> type S) <type numpy.uint64 > >>> type L) <type int > >>> S L 2.5 Comments 1) numpy.uint64 int > float numpy.uint32 int > int It seems like these should at least evaluate to the same type.
|
|
2) Should .sum ) return a numpy.uint64, even on 32 bit platforms, for the sake of consistency?","bug"
|
|
"6013","1315","numpy.loadtxt fails when missing values are present Trac 717).
|
|
Original ticket http projects.scipy.org numpy ticket 717 on 2008 04 04 by trac user lesserwhirls, assigned to unknown.
|
|
== Problem == numpy.loadtxt fails when missing values are present.
|
|
For example, assume your data file is well behaved val1,val2,val3,val4,val5\n loadtxt works great for this example no surprise).
|
|
Now, if your data file is not well behaved and contains missing values val1,val2,,val4,val5\n loadtxt fails.
|
|
== Solution == 1) Add keyword fill to def def loadtxt ...,fill= 999) 2) add the following after the line vals = line.split delimiter) line 713 in core numeric.py , numpy 1.0.4) vals = [ z, fill)[z is ] for z in vals] == Performace == Load an 18,000 line ascii dataset, 22 float variables on each line, skipping the first column its a time stamp).
|
|
Timings using timeit in ipython Reading an ascii file with no missing values using the current version of loadtxt [[BR]] 10 loops, best of 3 703 ms per loop Reading an ascii file with no missing values using the proposed changes to loadtxt [[BR]] 10 loops, best of 3 801 ms per loop The changes do create a slight performance hit for those who use loadtxt to read in nicely behaving ascii data.
|
|
If this is an issue, could a loadtxt2 function be added?","enhancement"
|
|
"6013","1316","Bug with numpy.float32.tolist Trac 718).
|
|
Original ticket http projects.scipy.org numpy ticket 718 on 2008 04 04 by trac user Elby, assigned to unknown.
|
|
numpy.float32 has a tolist method that standard python s float does not have, but this method does not return a list.
|
|
This can leads to bugs as reported in [http groups.google.com group Numpy discussion browse thread thread acc413d7089ddca5 e9ad794878b5572b?hl=fr lnk=gst q=tolist e9ad794878b5572b this thread], and IMHO this should be fixed or clearly explained.","bug"
|
|
"6013","1317","No test for fromregex in r4960 Trac 719).
|
|
Original ticket http projects.scipy.org numpy ticket 719 on 2008 04 05 by stefanv, assigned to teoliphant.
|
|
No test for fromregex in r4960.
|
|
Blocks release of 1.0.5.","bug"
|
|
"6013","1318","Changeset 4945 requires docstrings and tests Trac 720).
|
|
Original ticket http projects.scipy.org numpy ticket 720 on 2008 04 05 by stefanv, assigned to teoliphant.
|
|
r4945 changes the behaviour in calculating variance and standard deviation of complex numbers.
|
|
This requires explanation in the docstring, as well as tests.","bug"
|
|
"6013","1319","0 dimensional boolean arrays should work as masks for array scalars Trac 721).
|
|
Original ticket http projects.scipy.org numpy ticket 721 on 2008 04 05 by stefanv, assigned to teoliphant.
|
|
According to the changelog of 4822.","bug"
|
|
"6013","1320","Write tests to confirm that methods preserve arrays matrices Trac 722).
|
|
Original ticket http projects.scipy.org numpy ticket 722 on 2008 04 06 by stefanv, assigned to unknown.
|
|
Many functions accept either arrays or matrices and input, and should return corresponding output.
|
|
Write a set of tests to confirm that this is the current overall behaviour.","bug"
|
|
"6013","1321","gfortran installed in Program Files is not found Trac 723).
|
|
Original ticket http projects.scipy.org numpy ticket 723 on 2008 04 06 by pearu, assigned to pearu.
|
|
The issue was reported here http cens.ioc.ee pipermail f2py users 2008 April 001577.html","bug"
|
|
"6013","1322","numpy.ma.std and var should match numpy.std and var behaviour Trac 724).
|
|
Original ticket http projects.scipy.org numpy ticket 724 on 2008 04 07 by stefanv, assigned to unknown.
|
|
Quoting from the mailing list numpy.ma has the functions stdu and varu , for computing the unbiased standard deviation and variance where one divides by N 1 rather than N).
|
|
These are now available in numpy via std and var with ddof=1.
|
|
More seriously, they still provide the peculiar older definition of var, where varu [1,1.j, 1, 1.j])==0.
|
|
Update numpy.ma.var and std to accept ddof parameter, and to work correctly with complex numbers.
|
|
Remove stdu and varu from numpy.ma.","bug"
|
|
"6013","1323","Comparing string arrays to scalar numbers returns a single boolean value instead of an array of booleans Trac 725).
|
|
Original ticket http projects.scipy.org numpy ticket 725 on 2008 04 07 by rkern, assigned to teoliphant.
|
|
In [10] array [ foo , barbaz ]) == 0 Out[10] False In [11] array [1,2]) == 0 Out[11] array [False, False], dtype=bool) This has downstream effects; `isreal )` is implemented as `imag x) == 0`.","bug"
|
|
"6013","1324","r4980 doesn t test assert functions with nan arrays Trac 726).
|
|
Original ticket http projects.scipy.org numpy ticket 726 on 2008 04 08 by pv, assigned to unknown.
|
|
r4980 added a test to numpy testing tests test utils.py 47 def test nan array self) 48 Test two arrays with different shapes are found not equal.
|
|
49 a = N.array [1, 2]) 50 b = N.array [[1, 2], [1, 2]]) 51 52 self.
|
|
test not equal a, b) This is repetition of the previous test array diffshape and does not actually test the functions for nans.
|
|
Small thing, but let s file a ticket nonetheless.)","bug"
|
|
"6013","1325","Change mtrand to use Cython generated C code instead of Pyrex generated) Trac 727).
|
|
Original ticket http projects.scipy.org numpy ticket 727 on 2008 04 09 by fperez, assigned to rkern.
|
|
See http projects.scipy.org pipermail numpy discussion 2008 April 032567.html for background.
|
|
We re considering updating the pyrex based support to use Cython instead, and this patch updates the mtrand component.
|
|
There are no changes to the original pyrex sources, only the generated C code is different, as well as a trivial change to the code genearation script to call cython instead of pyrexc.","enhancement"
|
|
"6013","1326","numpy.r incorrectly casts with mixed types Trac 728).
|
|
Original ticket http projects.scipy.org numpy ticket 728 on 2008 04 09 by bsouthey, assigned to unknown.
|
|
Numpy.r appears to use the type of the array argument in determining the return types.
|
|
So when arguments have mixed types, there can be a loss of precision Correct when all arguments have the same type import numpy ra=numpy.r [ 10, numpy.array [2, 3, 4]), 10] type ra), type ra[0]), type ra[1]) Provides <type numpy.ndarray >, <type numpy.int64 >, <type numpy.int64 >) It is also correct if at least one array has the correct type ra=numpy.r [ 10.1, numpy.array [1]), numpy.array [2, 3, 4], dtype=float), 10] type ra), type ra[0]), type ra[1]), ra[0], ra[4] Provides <type numpy.ndarray >, <type numpy.float64 >, <type numpy.float64 >, 10.1, 4.0) But if mixed types are used import numpy ra=numpy.r [ 10.1, numpy.array [2, 3, 4]), 10.0] type ra), type ra[0]), type ra[1]), ra[0], ra[4] Provides <type numpy.ndarray >, <type numpy.int64 >, <type numpy.int64 >, 10, 10)[[BR]] So the float has been down cast to a integer so 10.1 becomes 10.
|
|
The workaround is to ensure that all numpy arrays have the same type as required import numpy ra=numpy.r [ 10.1, numpy.array [2, 3, 4], dtype=float), 10] type ra), type ra[0]), type ra[1]), ra[0], ra[4] Provides <type numpy.ndarray >, <type numpy.float64 >, <type numpy.float64 >, 10.1, 10.0)","bug"
|
|
"6013","1327","Complex types can t be set from strings.
|
|
Trac 729).
|
|
Original ticket http projects.scipy.org numpy ticket 729 on 2008 04 09 by charris, assigned to unknown.
|
|
In [1] csingle 1 ) TypeError Traceback most recent call last) home charris <ipython console> in <module> ) TypeError a float is required The same applies to the other complex types.","bug"
|
|
"6013","1328","long doubles display differently than the other float types Trac 730).
|
|
Original ticket http projects.scipy.org numpy ticket 730 on 2008 04 09 by charris, assigned to charris.
|
|
In [6] single 1.234 ) Out[6] 1.234 In [7] double 1.234 ) Out[7] 1.234 In [8] longdouble 1.234 ) Out[8] 1.233999999999999985789","bug"
|
|
"6013","1329","numpy.linalg.eigvals can t take nested lists while numpy.linalg.eig can Trac 731).
|
|
Original ticket http projects.scipy.org numpy ticket 731 on 2008 04 09 by trac user nmb, assigned to charris.
|
|
I was a bit surprised by the following this afternoon >>> import numpy >>> numpy.linalg.eig [[1,2],[3,4]]) array [ 0.37228132, 5.37228132]), array [[ 0.82456484, 0.41597356], [ 0.56576746, 0.90937671]])) >>> numpy.linalg.eigvals [[1,2],[3,4]]) Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.5 lib python2.5 site packages numpy linalg linalg.py , line 548, in eigvals assertRank2 a) File Library Frameworks Python.framework Versions 2.5 lib python2.5 site packages numpy linalg linalg.py , line 120, in assertRank2 if len a.shape) != 2 AttributeError list object has no attribute shape >>> The attached patch solves the problem for all of the functions in linalg and adds tests to that effect.","bug"
|
|
"6013","1330","error string on IndexError is wrong Trac 732).
|
|
Original ticket http projects.scipy.org numpy ticket 732 on 2008 04 09 by astraw, assigned to unknown.
|
|
The error string in the following is wrong.
|
|
>>> import numpy >>> x=numpy.arange 10) >>> x.shape=5,2 >>> x[5,1] Traceback most recent call last) File <stdin> , line 1, in <module> IndexError index 5) out of range 0<=index<=5) in dimension 0 The final <= should be < .
|
|
The attached patch fixes the issue.","bug"
|
|
"6013","1331","check object casting failing on Python 2.6 Trac 733).
|
|
Original ticket http projects.scipy.org numpy ticket 733 on 2008 04 10 by alberts, assigned to unknown.
|
|
As can be seen here http buildbot.scipy.org builders Linux x86 Fedora Py2.6 builds 332 steps shell 2 logs stdio there is a test failure on Python 2.6 ====================================================================== FAIL check object casting test regression.TestRegression) Traceback most recent call last) File home buildbot numpy b6 numpy install lib python2.6 site packages numpy core tests test regression.py , line 756, in check object casting self.failUnlessRaises TypeError,rs) AssertionError TypeError not raised It would be worthwhile to investigate this, as this test failure happened for the first time relatively recently, which might mean the test is new or there is a problem with Python 2.6.","bug"
|
|
"6013","1332","[PATCH] interactive docstring search lookfor) Trac 734).
|
|
Original ticket http projects.scipy.org numpy ticket 734 on 2008 04 10 by pv, assigned to unknown.
|
|
It d be nice to have a docstring search that could be used interactively >>> import numpy as np >>> np.lookfor eigenvalue hermitean ) Search results for eigenvalue hermitean numpy.linalg.eigvalsh Compute the eigenvalues of a Hermitean or real symmetric matrix.
|
|
numpy.linalg.eigh Compute eigenvalues for a Hermitian or real symmetric matrix.
|
|
numpy.linalg.eigvals Compute the eigenvalues of a general matrix.
|
|
>>> np.lookfor spline interpolat , module= scipy ) Search results for spline interpolat scipy.interpolate.spline Interpolate a curve xk,yk) at points xnew using a spline fit.
|
|
scipy.interpolate.RectBivariateSpline Bivariate spline approximation over a rectangular mesh.
|
|
scipy.interpolate.InterpolatedUnivariateSpline Interpolated univariate spline approximation.
|
|
Identical to scipy.ndimage.shift Shift an array.
|
|
scipy.ndimage.zoom Zoom an array.
|
|
scipy.interpolate.splrep Find the B spline representation of 1 D curve.
|
|
... long result, shown in a pager ...
|
|
This patch implements a simple minded docstring search lookfor generate cache walks the contents of the module to search, collecting docstrings lookfor splits the search string and searchs for entries that contain all parts.
|
|
No stemming etc.
|
|
is done.
|
|
lookfor sorts the results by a Harrison Stetson relevancy metric ie.
|
|
I just wrote up some heuristics to distinguish more relevant results from less relevant ).
|
|
It appears to work ok, but probably could be improved.
|
|
lookfor either prints or calls a pager via pydoc, for long listings) to show the results No automated tests at this time, sorry, I only did manual testing now.","enhancement"
|
|
"6013","1333","FAIL test record numpy.lib.tests.test io.Testloadtxt) On Solaris 10 Trac 735).
|
|
Original ticket http projects.scipy.org numpy ticket 735 on 2008 04 12 by chanley, assigned to unknown.
|
|
The following test case fails on Solaris 10 with 1.0.5.dev5024 ====================================================================== FAIL test record numpy.lib.tests.test io.Testloadtxt) Traceback most recent call last) File data basil5 site packages lib python numpy lib tests test io.py , line 42, in test record assert array equal x, a) File numpy testing utils.py , line 248, in assert array equal File numpy testing utils.py , line 240, in assert array compare AssertionError Arrays are not equal mismatch 100.0 ) x array [ 1, 2), 3, 4)], dtype=[ x , >i4 ), y , >i4 )]) y array [ 1, 2), 3, 4)], dtype=[ x , <i4 ), y , <i4 )]) Ran 895 tests in 9.468s FAILED failures=1) <unittest.
|
|
TextTestResult run=895 errors=0 failures=1> >>> numpy.
|
|
version 1.0.5.dev5024","bug"
|
|
"6013","1334","Inconsistent integer conversion from strings Trac 736).
|
|
Original ticket http projects.scipy.org numpy ticket 736 on 2008 04 12 by stefanv, assigned to teoliphant.
|
|
Pauli Virtanen noticed the following behaviour mentioned as part of 1317 discussion) In [8] np.array [ 123 ,), 456 ,)], dtype=[ num , <i8 )]) Out[8] array [ 123L,), 456L,)], dtype=[ num , <i8 )]) vs.
|
|
In [9] np.array [ 123 ,), 456 ,)], dtype=[ num , <i4 )]) TypeError expected a readable buffer object I believe this is related to the following inconsistent integer from string conversions In [27] np.int32 12 ) Out[27] 12 In [28] np.int64 12 ) Out[28] array [1, 2], dtype=int64)","bug"
|
|
"6013","1335","String to integer conversion inconsistent.
|
|
Trac 737).
|
|
Original ticket http projects.scipy.org numpy ticket 737 on 2008 04 13 by charris, assigned to unknown.
|
|
To wit In [1] int8 12 ) Out[1] array [1, 2], dtype=int8) In [2] int32 12 ) Out[2] 12 I suspect int32 behaves differently because it is a subtype of the python integer and so behaves as I) would expect.
|
|
If int32 is indeed a subtype of the python int I think that is also an error as Python ints are longs.","bug"
|
|
"6013","1336","Test trying to write file in wrong location Trac 738).
|
|
Original ticket http projects.scipy.org numpy ticket 738 on 2008 04 14 by cournape, assigned to unknown.
|
|
On vista, this causes an error ERROR test ValidHTTP numpy.lib.tests.test datasource.TestDataSourceOpen) Traceback most recent call last) File C \Program Files\Python25\lib\site packages\numpy\lib\tests\test datasource.py , line 79, in test ValidHTTP assert self.ds.open valid httpurl )) File C \Program Files\Python25\lib\site packages\numpy\lib\ datasource.py , line 366, in open found = self.
|
|
findfile path) File C \Program Files\Python25\lib\site packages\numpy\lib\ datasource.py , line 243, in findfile name = self.
|
|
cache name) File C \Program Files\Python25\lib\site packages\numpy\lib\ datasource.py , line 203, in cache file upath, w ).write openedurl.read )) IOError [Errno 13] Permission denied index.html Xp does not complain, but that just should not happen","bug"
|
|
"6013","1337","Please don t use yellow in the output of setup.py Trac 739).
|
|
Original ticket http projects.scipy.org numpy ticket 739 on 2008 04 16 by trac user WW, assigned to dmcooke.
|
|
I just ran setup.py on my Mac running OSX 2.4.
|
|
My poor eyes!
|
|
Please don t use yellow text in the output of setup.py.
|
|
On my Mac and Linux computers, the default background for a terminal is white.
|
|
Yellow text on a white background results in very low contrast and is hard to read.
|
|
Unless you can make setup.py smart enough to detect the background color of the terminal, please don t mess with colored text.","enhancement"
|
|
"6013","1338","numpy setup.py too restrictive, prevents use of fblas with cblas Trac 740).
|
|
Original ticket http projects.scipy.org numpy ticket 740 on 2008 04 17 by trac user gnurser, assigned to cournape.
|
|
For blas atlas etc in numpy scipy on an opteron I use the AMD libraries which only have fblas) from http developer.amd.com together with cblas from http www.netlib.org blas blast forum cblas.tgz.
|
|
This works very well, since the AMD libraries are well optimized.
|
|
Details of how I did it are in the attached file acml2.log However numpy from current SVN will not build linked to these libraries.
|
|
The problem is that SVN v4779) in line 295 296 of numpy core setup.py has if NO ATLAS INFO ,1) in blas info.get define macros ,[]) return None dotblas needs ATLAS, Fortran compiled blas will not be sufficient NO ATLAS INFO is set, and so dotblas is not created.
|
|
To get my AMD fblas cblas approach to work, I have to comment out these two lines.","bug"
|
|
"6013","1339","Extend numpy.dot ) to accept more than 2 arrays.
|
|
Trac 741).
|
|
Original ticket http projects.scipy.org numpy ticket 741 on 2008 04 17 by trac user joris, assigned to unknown.
|
|
if a,b,c,d are numpy arrays, than their matrix multiplication can be computed with dot a, dot b, dot c, d))) or by array mat a) mat b) mat c) mat d)) More convenient would be dot a,b,c,d).
|
|
An easy way to implement this would be def dot args) return reduce olddot, args) where olddot is the current implementation of dot that only takes 2 arguments.
|
|
The suggestion above would not break any code, as dot a,b) == olddot a,b).","enhancement"
|
|
"6013","1340","Don t require bz2 or gzip Trac 742).
|
|
Original ticket http projects.scipy.org numpy ticket 742 on 2008 04 17 by rkern, assigned to rkern.
|
|
The gzip and bz2 modules are optional components of the standard library and may not be built on some platforms.","bug"
|
|
"6013","1341","Array content zeroed by |= even though exception raised Trac 743).
|
|
Original ticket http projects.scipy.org numpy ticket 743 on 2008 04 18 by charris, assigned to unknown.
|
|
This bug is also dependent on the array size In [9] a = ones 10000,)) In [10] a |= a TypeError Traceback most recent call last) home charris <ipython console> in <module> ) TypeError unsupported operand type s) for | float and float In [11] a Out[11] array [ 0., 0., 0., ..., 0., 0., 0.])
|
|
In [14] a = ones 9999,)) In [15] a |= a TypeError Traceback most recent call last) home charris <ipython console> in <module> ) TypeError unsupported operand type s) for | float and float In [16] a Out[16] array [ 1., 1., 1., ..., 1., 1., 1.])","bug"
|
|
"6013","1342","Loss of dimension bug in numpy.ma.
|
|
getitem for arrays of shape n, 1) or 1, n) or n, 1, 1, ...) Trac 744).
|
|
Original ticket http projects.scipy.org numpy ticket 744 on 2008 04 19 by trac user dharland, assigned to pierregm.
|
|
There is a bug in MaskedArray.
|
|
getitem that causes a loss in dimension for masked arrays of shape n, 1) or 1, n) or more generally, any shape of the form 1,) a n,) 1,) b Heres an example A 2d column vector x = ma.arange 5).reshape 1, 1) ) mask a point x[3] = ma.masked Different behavior for unmasked and masked items print x[0], x[0].shape print x[3], x[3].shape this produces as output [0] 1,) Correct ) This should be [ ] As a result code cannot reliably iterate over rows of a 2d array in case its in fact a n,1) array) The error appears in the following code snippet in numpy core ma.py mi = m[i] if mi.size == 1 < Should this be if mi.size == 1 and mi.ndim == 0 if mi return masked < dimension loss here if mi.ndim > 0 else return dout","bug"
|
|
"6013","1343","assert almost equal fails for object arrays.
|
|
Trac 745).
|
|
Original ticket http projects.scipy.org numpy ticket 745 on 2008 04 19 by charris, assigned to cournape.
|
|
This traceback refers to a test function that isn t yet committed.
|
|
ERROR Test generic loops.
|
|
Traceback most recent call last) File usr lib python2.5 site packages numpy core tests test ufunc.py , line 120, in check generic loops assert almost equal np.abs x), 1, err msg=msg) File usr lib python2.5 site packages numpy testing utils.py , line 156, in assert almost equal return assert array almost equal actual, desired, decimal, err msg) File usr lib python2.5 site packages numpy testing utils.py , line 255, in assert array almost equal header= Arrays are not almost equal ) File usr lib python2.5 site packages numpy testing utils.py , line 225, in assert array compare val = comparison x,y) File usr lib python2.5 site packages numpy testing utils.py , line 253, in compare return around abs x y),decimal) <= 10.0 decimal) File usr lib python2.5 site packages numpy core fromnumeric.py , line 1516, in around return round decimals, out) AttributeError float object has no attribute rint","bug"
|
|
"6013","1344","Segfault calling logical not on own object Trac 746).
|
|
Original ticket http projects.scipy.org numpy ticket 746 on 2008 04 19 by charris, assigned to unknown.
|
|
The following causes a segfault class foo def logical not self) return np.bool 1) def logical and self, obj) return np.bool 1) check unary PyUFunc O O method msg = PyUFunc O O method x = np.zeros 10, dtype=np.object)[0 2] x[...] = foo ) assert np.all np.logical not x) == True), msg","bug"
|
|
"6013","1345","Reference counting bug in call to ufunc object loops.
|
|
Trac 747).
|
|
Original ticket http projects.scipy.org numpy ticket 747 on 2008 04 20 by charris, assigned to unknown.
|
|
The following code tests out fine if I comment out the Py XDECREF.
|
|
static void PyUFunc O O method char args, intp dimensions, intp steps, void func) { intp n = dimensions[0]; intp is1 = steps[0]; intp is2 = steps[1]; char ip1 = args[0]; char op = args[1]; char meth = char )func; intp i; for i = 0; i < n; i , ip1 = is1, op = is2) { PyObject in1 = PyObject )ip1; PyObject out = PyObject )op; PyObject ret = PyObject CallMethod in1, meth, NULL); if ret == NULL) { return; } Py XDECREF out); out = ret; } }","bug"
|
|
"6013","1346","Ifft pads incorrectly Trac 748).
|
|
Original ticket http projects.scipy.org numpy ticket 748 on 2008 04 20 by stefanv, assigned to unknown.
|
|
Ifft should pad with zeros in the middle of the input, not at the end.
|
|
import numpy as np from matplotlib import pylab as pp from math import def pad x,n,centre=False) if not centre return np.hstack x,np.zeros n))) else return np.hstack x[ len x) 2],np.zeros n),x[len x) 2 ])) p = pp.hamming 10) Fp = np.fft.fft p,n=2 10) p1 = np.fft.ifft Fp) p2=np.fft.ifft Fp,n=2 12) p3=np.fft.ifft pad Fp, 2 12, centre=True)) pp.subplot 311) pp.title Input ) pp.plot p) pp.subplot 312) pp.title ifft padding ) pp.plot p2[ 45]) pp.subplot 313) pp.title Correct ifft padding ) pp.plot p3[ 45]) pp.show )","bug"
|
|
"6013","1347","incorrect behaviour when using mask to index masked array Trac 749).
|
|
Original ticket http projects.scipy.org numpy ticket 749 on 2008 04 20 by trac user mattknox ca, assigned to unknown.
|
|
Below is some code that illustrates the problem ========================================= import numpy as np from numpy import ma a = ma.array [55]) a[0] = ma.masked b = ma.array [66]) print repr a[a.mask]) print repr b[a.mask]) idx = a.mask .copy ) print repr idx) a[ idx] = b[ idx] a[0] is equal to 55 still, but should be 66 print repr a) print repr b) ========================================= If you change the code such that idx = a.mask.copy ), then it works as expected ie.
|
|
a[0] == 66) Even wierder is that if you explicitly pass mask=[True] when constructing a , and remove the line a[0] = ma.masked , then it works without needing to copy the mask prior to indexing.","bug"
|
|
"6013","1348","Swig interface Bug in FORTRAN ordering support Trac 750).
|
|
Original ticket http projects.scipy.org numpy ticket 750 on 2008 04 22 by trac user kishorg, assigned to unknown.
|
|
I am applying following typemap for passing address of a Fortran style Numpy matrix to C function using Swig.
|
|
apply double IN FARRAY2, int DIM1,int DIM2) { double ptr in, index t n rows in, index t n cols in)}; I always get C style ordered array in my C code with a new copy of Numpy matrix.
|
|
The problem is typemap definition for IN FARRAY2 in numpy.i at line number 713 calls function array = obj to array contiguous allow conversion input, DATA TYPECODE, is new object); which copies it into new C Style array.
|
|
If I replace this call with function array = obj to array no conversion input, DATA TYPECODE); then it works correctly.
|
|
This is a bug in Numpy.i and need to fix it.","bug"
|
|
"6013","1349","distutils does not seem to handle the ALL section in site.cfg Trac 751).
|
|
Original ticket http projects.scipy.org numpy ticket 751 on 2008 04 22 by cournape, assigned to cournape.
|
|
I can t make numpy.distutils use my custom paths set in site.cfg if set in the ALL section.
|
|
If I use the [DEFAULT] section, it does work.
|
|
It seems something was not set up correctly when the default section name was changed.","bug"
|
|
"6013","1350","Order of compilation of fortran 90 modules Trac 752).
|
|
Original ticket http projects.scipy.org numpy ticket 752 on 2008 04 22 by huard, assigned to pearu.
|
|
When compiling a fortran 90 extension using disutils, the build process sometimes fails because the dependence between modules seems to be ignored.
|
|
That is, if one module uses another module, there seems to be no guarantee the latter will be compiled first.","bug"
|
|
"6013","1351","numpy.ma.morestats requires scipy Trac 753).
|
|
Original ticket http projects.scipy.org numpy ticket 753 on 2008 04 23 by rkern, assigned to pierregm.
|
|
This code needs to be removed before 1.1.0 is released.
|
|
Although the module itself is optional, the test suite picks it up and causes errors for people without scipy.","bug"
|
|
"6013","1352","setup.py not recognizing gcc on install Trac 754).
|
|
Original ticket http projects.scipy.org numpy ticket 754 on 2008 04 24 by trac user Baphijmm, assigned to unknown.
|
|
I am running a SPARCstation 20 under Solaris 9; upon attempting sudo python setup.py install , I continuously get the following error cc DNDEBUG x03 xtarget=ultra xarch=v8 IInclude IPackages FFT Include IPackages RNG Include I opt csw include python2.5 c Src numpymodule.c o build temp.solaris 2.9 sun4m 2.5 Src numpymodule.o usr ucb cc language optional software package not installed error command cc failed with exit status 1 I know this means that cc isn t installed on the machine; in fact, it is not.
|
|
However, gcc is, and according to what I m seeing in the setup.py code, it should be able to recognize this fact I m somewhat new to Python though, so I m not at all certain).
|
|
What am I doing wrong, and what can I do to fix this?
|
|
I m sorry if this is the wrong place to ask this; I have no other ideas where to go, and the internet turns up nothing.","bug"
|
|
"6013","1353","Support for elemental procedures Trac 755).
|
|
Original ticket http projects.scipy.org numpy ticket 755 on 2008 04 24 by huard, assigned to unknown.
|
|
Elemental functions could be converted into ufuncs using numpy.frompyfunc.
|
|
The resulting function could then accept ndarray arguments, the result being conformable with the array argument.
|
|
newf = frompyfunc f, nin, 1)","enhancement"
|
|
"6013","1354","String conversion should be supported for arithmetic operators.
|
|
Trac 756).
|
|
Original ticket http projects.scipy.org numpy ticket 756 on 2008 04 25 by charris, assigned to unknown.
|
|
Here is the problem In [1] a = ones 3) In [2] a 1 TypeError Traceback most recent call last) home charris <ipython console> in <module> ) TypeError unsupported operand type s) for numpy.ndarray and str I think the general rule for combining strings with any numeric type should be casting the string.
|
|
The casting now works in numby, we just need to finish making it work with everything else.","enhancement"
|
|
"6013","1355","numpy.sort sometimes doesn t handle strings right Trac 757).
|
|
Original ticket http projects.scipy.org numpy ticket 757 on 2008 04 25 by trac user jloper, assigned to unknown.
|
|
numpy.sort numpy.array [ \x01\x00\x02 , \x01\x00\x04 , \x01\x00\x02 ])) array [ \x01\x00\x02 , \x01\x00\x04 , \x01\x00\x02 ], dtype= |S3 ) \x01\x00\x02 < \x01\x00\x04 True \x01\x00\x02 > \x01\x00\x04 False ... important because people are using a.view S8 ) to run unique ) across rows...","bug"
|
|
"6013","1356","g3 f2py clean up Trac 758).
|
|
Original ticket http projects.scipy.org numpy ticket 758 on 2008 04 26 by pearu, assigned to pearu.
|
|
The development of g3 f2py tool is being moved to http launchpad.net f2py see http www.f2py.org for details).
|
|
It means that 1) numpy f2py lib directory can be removed after the move is complete.
|
|
2) f2py script must be updated to use f2py package instead of numpy.f2py.lib when using the switch g3 numpy.
|
|
3) We also need a way to enable the g3 f2py tool for numpy.distutils.
|
|
I can think of the following options 3.1) introduce environment variable F2PY than can hold strings g3 numpy , 2d numpy , 2d numeric , 2d numarray that determine which f2py backends will be used by numpy.distutils.
|
|
3.2) introduce a function `use f2py <backend name>)` to numpy.distutils that can be used from setup.py files as well as from python session to set the f2py backend.
|
|
The second option may be most appropiate.
|
|
Both options have a problem of being global options.
|
|
Theoretically, there may exist situations where different f2py backends must be used for different Fortran sources.
|
|
The best solution would be to avoid such situations, meaning that g3 f2py must be backward compatible to numpy f2py.","bug"
|
|
"6013","1357","spurious underflow warning in numpy.lib.machar Trac 759).
|
|
Original ticket http projects.scipy.org numpy ticket 759 on 2008 04 26 by cournape, assigned to unknown.
|
|
I sometimes use seterr all = warn ) to detect potential problems in my algorithms, and I always get warning in numpy.lib.machar when importing e.g.
|
|
scipy.io.
|
|
For example, this minimal script shows this behaviour !python import numpy as np from scipy.testing import a = np.seterr all = warn ) from scipy.io import savemat, loadmat I attached a patch which should solve the problem.
|
|
As I am not familiar with this part of numpy codebase, I did not apply it right away.","bug"
|
|
"6013","1358","scalar indexing of matrices > deprecation warning Trac 760).
|
|
Original ticket http projects.scipy.org numpy ticket 760 on 2008 04 26 by trac user aisaac, assigned to unknown.
|
|
Given current discussions, it seems that scalar indexing of matrices will change behavior.
|
|
Therefore there should be a warning.
|
|
The attached patch to defmatrix.py produces a deprecation warning.","bug"
|
|
"6013","1359","operator = fails silently when adding arrays with diffenent base types Trac 761).
|
|
Original ticket http projects.scipy.org numpy ticket 761 on 2008 04 27 by trac user drevicko, assigned to unknown.
|
|
Example >>> a=array [1,2,3]) >>> b=array [4,5,6]) >>> b =a >>> b b and a are both integer arrays and = succeeds as expected array [5, 7, 9]) >>> b =0.1 a >>> b b and integer array, 0.1 a a float array = failed silently array [5, 7, 9]) >>> b=b 0.1 a >>> b works when written longhand.
|
|
array [ 5.1, 7.2, 9.3]) >>> I have numpy version 1.04 this is not in the trac version list...) Python 2.5.2 on win32","bug"
|
|
"6013","1360","Figure out why fixing 743 fixed 733 too.
|
|
Trac 762).
|
|
Original ticket http projects.scipy.org numpy ticket 762 on 2008 04 27 by charris, assigned to cournape.
|
|
The results for python2.5 and python2.6 should have been the same raising an error and incorrectly filling in the lhs of x |= x for large arrays.
|
|
For small arrays a shorter path might be taken and it isn t clear what fixed that for python2.6.
|
|
Possibilities 1) Putting the threading macros containing if s in do {} while 0) fixed something in the way numpy compiled on the 2.6 machine.
|
|
2) Timing change?
|
|
Hope not...","bug"
|
|
"6013","1361","Need tests for linalg.tensorsolve and linalg.tensorinv Trac 763).
|
|
Original ticket http projects.scipy.org numpy ticket 763 on 2008 04 29 by charris, assigned to cournape.
|
|
Add test for linalg.norm.
|
|
There are currently three functions untested in the linalg tests norm, tensorsolve, and tensorinv.","bug"
|
|
"6013","1362","Inconsistent behavior of float32, float64 Trac 764).
|
|
Original ticket http projects.scipy.org numpy ticket 764 on 2008 04 29 by charris, assigned to unknown.
|
|
In [1] float32 array [[1]])) Out[1] array [[ 1.
|
|
]], dtype=float32) In [2] float64 array [[1]])) Out[2] 1.0 In [3] int64 array [[1]])) Out[3] array [[1]], dtype=int64) In [4] int32 array [[1]])) Out[4] 1 But also In [5] float64 [[1]]) Out[5] array [[ 1.]])
|
|
In [6] int32 [[1]]) Out[6] array [[1]])","bug"
|
|
"6013","1363","Implement logsumexp Trac 765).
|
|
Original ticket http projects.scipy.org numpy ticket 765 on 2008 04 30 by cournape, assigned to cournape.","bug"
|
|
"6013","1364","fastCopyAndTranspose segfaults Trac 766).
|
|
Original ticket http projects.scipy.org numpy ticket 766 on 2008 04 30 by trac user dave.hirschfeld, assigned to unknown.
|
|
The below code causes fastCopyAndTranspose to segfault.
|
|
> It looks like the segfault occurs in the printing of the result.
|
|
If > you set y = fastCopyAndTranspose x) you don t get a segfault until you > try to print it actually repr it).
|
|
> > Travis from datetime import datetime, timedelta from numpy import asarray, fastCopyAndTranspose timestamps = [datetime 2007,1,1)] while timestamps[ 1] < datetime 2007,12,31) timestamps.append timestamps[ 1] timedelta 1)) x = [asarray timestamps),asarray range len timestamps)))] fastCopyAndTranspose x)","bug"
|
|
"6013","1365","numpy.i typo in error messages Trac 767).
|
|
Original ticket http projects.scipy.org numpy ticket 767 on 2008 04 30 by tovrstra, assigned to unknown.
|
|
In the file numpy.i, there is a small typo two times the same) Array must be have s dimensions.
|
|
...
|
|
Array must be have shape of s. ...
|
|
The be should be removed.","enhancement"
|
|
"6013","1366","numpy.i suggestion for improvement require shape require size Trac 768).
|
|
Original ticket http projects.scipy.org numpy ticket 768 on 2008 04 30 by tovrstra, assigned to unknown.
|
|
I would suggest that `require size` is renamed to `require shape`.
|
|
The current situation is confusing because one easily thinks that `require size` is similar to `array size`.
|
|
It would be nice to have a `require size` that works like this Require the given PyArrayObject to have a specified size in dimension i.
|
|
If the array has the specified size in dimension i, return 1.
|
|
Otherwise, set the python error string and return 0. int require size PyArrayObject ary, int i, npy intp size) { int success = 1; if size != 1 size != array size ary,i)) { success = 0; } if !success) { PyErr Format PyExc TypeError, Array must have size of d in dimension d. Given array has size d in dimension d. , size, i, array size ary,i), i ); } return success; }","enhancement"
|
|
"6013","1367","add output argument to numpy.outer Trac 769).
|
|
Original ticket http projects.scipy.org numpy ticket 769 on 2008 04 30 by trac user aisaac, assigned to unknown.
|
|
An output argument for ``numpy.outer`` would be a nice enhancement.
|
|
Motivation I m using ``outer`` in a loop body.)","enhancement"
|
|
"6013","1368","numpy.core.tests.test multiarray.TestView failures on big endian machines Trac 770).
|
|
Original ticket http projects.scipy.org numpy ticket 770 on 2008 05 05 by chanley, assigned to unknown.
|
|
We are receiving the following numpy self test failures on our big endian Solaris 10 system.
|
|
The numpy version is 1.1.0.dev5125.
|
|
====================================================================== FAIL test basic numpy.core.tests.test multiarray.TestView) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test multiarray.py , line 843, in test basic assert array equal y, [67305985, 134678021]) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 248, in assert array equal verbose=verbose, header= Arrays are not equal ) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 240, in assert array compare assert cond, msg AssertionError Arrays are not equal mismatch 100.0 ) x array [16909060, 84281096]) y array [ 67305985, 134678021]) ====================================================================== FAIL test keywords numpy.core.tests.test multiarray.TestView) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test multiarray.py , line 852, in test keywords assert array equal y,[[513]]) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 248, in assert array equal verbose=verbose, header= Arrays are not equal ) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 240, in assert array compare assert cond, msg AssertionError Arrays are not equal mismatch 100.0 ) x array [[258]], dtype=int16) y array [[513]]) Ran 986 tests in 21.707s FAILED failures=2) errors failures <numpy.core.tests.test multiarray.TestView testMethod=test basic>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test multiarray.py , line 843, in test basic\n assert array equal y, [67305985, 134678021])\n File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 248, in assert array equal\n verbose=verbose, header=\ Arrays are not equal\ )\n File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 240, in assert array compare\n assert cond, msg\nAssertionError \nArrays are not equal\n\n mismatch 100.0 )\n x array [16909060, 84281096])\n y array [ 67305985, 134678021])\n ) <numpy.core.tests.test multiarray.TestView testMethod=test keywords>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test multiarray.py , line 852, in test keywords\n assert array equal y,[[513]])\n File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 248, in assert array equal\n verbose=verbose, header=\ Arrays are not equal\ )\n File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 240, in assert array compare\n assert cond, msg\nAssertionError \nArrays are not equal\n\n mismatch 100.0 )\n x array [[258]], dtype=int16)\n y array [[513]])\n ) if 1 != 0 ) then echo <font color=orange>selftest< font> exit 0","bug"
|
|
"6013","1369","Valgrind warning when running test for ticket 658 Trac 771).
|
|
Original ticket http projects.scipy.org numpy ticket 771 on 2008 05 07 by alberts, assigned to unknown.
|
|
Valgrind warning when running test for ticket 1256 ==14758== Conditional jump or move depends on uninitialised value s) ==14758== at 0x433E2D2 PyArray UpdateFlags arrayobject.c 5172) ==14758== by 0x43679C8 PyArray Newshape multiarraymodule.c 742) ==14758== by 0x436856E array reshape arraymethods.c 87) ==14758== by 0x80CF775 PyEval EvalFrameEx ceval.c 3645) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x80CF002 PyEval EvalFrameEx ceval.c 3741) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x81235D5 function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CB503 PyEval EvalFrameEx ceval.c 3944) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x812350B function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x8062C76 instancemethod call classobject.c 2558) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CBE34 PyEval EvalFrameEx ceval.c 3856) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x812350B function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x8062C76 instancemethod call classobject.c 2558) Found by the Python 2.6 buildbot.","bug"
|
|
"6013","1370","Valgrind warning when running test for compress2d Trac 772).
|
|
Original ticket http projects.scipy.org numpy ticket 772 on 2008 05 07 by alberts, assigned to unknown.
|
|
Valgrind warning when running test for compress2d ==14758== Conditional jump or move depends on uninitialised value s) ==14758== at 0x4330DB6 PyArray MapIterReset arrayobject.c 10190) ==14758== by 0x437D85E array subscript arrayobject.c 2549) ==14758== by 0x437E491 array subscript nice arrayobject.c 3173) ==14758== by 0x80CC7BD PyEval EvalFrameEx ceval.c 1207) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x80CF002 PyEval EvalFrameEx ceval.c 3741) ==14758== by 0x80D04A2 PyEval EvalFrameEx ceval.c 3731) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x81235D5 function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CB503 PyEval EvalFrameEx ceval.c 3944) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x812350B function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x8062C76 instancemethod call classobject.c 2558) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CBE34 PyEval EvalFrameEx ceval.c 3856) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x812350B function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) Found by Python 2.6 buildbot.","bug"
|
|
"6013","1371","Valgrind causes some tests to fail Trac 773).
|
|
Original ticket http projects.scipy.org numpy ticket 773 on 2008 05 07 by alberts, assigned to unknown.
|
|
Valgrind causes the following tests to fail ====================================================================== FAIL test divide test errstate.TestErrstate) Traceback most recent call last) File <string> , line 38, in test divide AssertionError ====================================================================== FAIL test invalid test errstate.TestErrstate) Traceback most recent call last) File <string> , line 24, in test invalid AssertionError ====================================================================== FAIL test divideerr numpy.core.tests.test numeric.TestSeterr) Traceback most recent call last) File home buildbot numpy b6 numpy install lib python2.6 site packages numpy core tests test numeric.py , line 196, in test divideerr self.fail ) AssertionError It would be useful if all tests could pass when run under Valgrind.","bug"
|
|
"6013","1372","Memory leak when running unit tests Trac 774).
|
|
Original ticket http projects.scipy.org numpy ticket 774 on 2008 05 07 by alberts, assigned to unknown.
|
|
Valgrind reports the following memory leak when running the unit tests ==14758== 88 80 direct, 8 indirect) bytes in 2 blocks are definitely lost in loss record 30 of 75 ==14758== at 0x4005400 malloc vg replace malloc.c 149) ==14758== by 0x4334B92 array alloc arrayobject.c 6929) ==14758== by 0x433F63A PyArray NewFromDescr arrayobject.c 5575) ==14758== by 0x4358EEC PyArray FromScalar scalartypes.inc.src 249) ==14758== by 0x437E991 gen arrtype subscript scalartypes.inc.src 2405) ==14758== by 0x80CC7BD PyEval EvalFrameEx ceval.c 1207) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x81235D5 function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CB503 PyEval EvalFrameEx ceval.c 3944) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x80CF002 PyEval EvalFrameEx ceval.c 3741) ==14758== by 0x80D04A2 PyEval EvalFrameEx ceval.c 3731) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x81235D5 function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) ==14758== by 0x80CB503 PyEval EvalFrameEx ceval.c 3944) ==14758== by 0x80D0CA5 PyEval EvalCodeEx ceval.c 2908) ==14758== by 0x812350B function call funcobject.c 524) ==14758== by 0x805BE24 PyObject Call abstract.c 2490) Found using the Python 2.6 buildbot.","bug"
|
|
"6013","1373","array resize shifts values in array Trac 775).
|
|
Original ticket http projects.scipy.org numpy ticket 775 on 2008 05 07 by trac user nick, assigned to unknown.
|
|
When resize is called on a 2d array, if the value of the column resize is greater than the current column size, a wrapping of the values occurs.
|
|
I noticed this while trying to resize a diagonal matrix.
|
|
Example Python 2.5.1 r251 54863, Apr 18 2007, 08 51 08) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> from numpy.lib import twodim base >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4]] >>> a diag.resize 5,4)) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4] [0 0 0 0]] >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> a diag.resize 4,5)) >>> print a diag [[1 0 0 0 0] [2 0 0 0 0] [3 0 0 0 0] [4 0 0 0 0]] >>> The documentation led me to believe that a column of 0 s would be added much like how a row of 0 s was added when the number of rows was increased).
|
|
Nick Loadholtes nick ironboundsoftware.com","bug"
|
|
"6013","1374","array resize shifts values in array Trac 776).
|
|
Original ticket http projects.scipy.org numpy ticket 776 on 2008 05 07 by trac user nick, assigned to unknown.
|
|
When resize is called on a 2d array, if the value of the column resize is greater than the current column size, a wrapping of the values occurs.
|
|
I noticed this while trying to resize a diagonal matrix.
|
|
Example Python 2.5.1 r251 54863, Apr 18 2007, 08 51 08) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> from numpy.lib import twodim base >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4]] >>> a diag.resize 5,4)) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4] [0 0 0 0]] >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> a diag.resize 4,5)) >>> print a diag [[1 0 0 0 0] [2 0 0 0 0] [3 0 0 0 0] [4 0 0 0 0]] >>> The documentation led me to believe that a column of 0 s would be added much like how a row of 0 s was added when the number of rows was increased).
|
|
Nick Loadholtes nick ironboundsoftware.com","bug"
|
|
"6013","1375","array resize shifts values in array Trac 777).
|
|
Original ticket http projects.scipy.org numpy ticket 777 on 2008 05 07 by trac user nick, assigned to unknown.
|
|
When resize is called on a 2d array, if the value of the column resize is greater than the current column size, a wrapping of the values occurs.
|
|
I noticed this while trying to resize a diagonal matrix.
|
|
Example Python 2.5.1 r251 54863, Apr 18 2007, 08 51 08) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> from numpy.lib import twodim base >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4]] >>> a diag.resize 5,4)) >>> print a diag [[1 0 0 0] [0 2 0 0] [0 0 3 0] [0 0 0 4] [0 0 0 0]] >>> a = [1,2,3,4] >>> a diag = twodim base.diag a) >>> a diag.resize 4,5)) >>> print a diag [[1 0 0 0 0] [2 0 0 0 0] [3 0 0 0 0] [4 0 0 0 0]] >>> The documentation led me to believe that a column of 0 s would be added much like how a row of 0 s was added when the number of rows was increased).
|
|
Nick Loadholtes nick ironboundsoftware.com","bug"
|
|
"6013","1376","Markup errors in trunk numpy ma API CHANGES.txt Trac 778).
|
|
Original ticket http projects.scipy.org numpy ticket 778 on 2008 05 08 by vnoel, assigned to jarrodmillman.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1377","Markup errors in trunk numpy ma API CHANGES.txt Trac 779).
|
|
Original ticket http projects.scipy.org numpy ticket 779 on 2008 05 08 by vnoel, assigned to unknown.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1378","Markup errors in trunk numpy ma API CHANGES.txt Trac 780).
|
|
Original ticket http projects.scipy.org numpy ticket 780 on 2008 05 08 by vnoel, assigned to unknown.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1379","Markup errors in trunk numpy ma API CHANGES.txt Trac 781).
|
|
Original ticket http projects.scipy.org numpy ticket 781 on 2008 05 08 by vnoel, assigned to unknown.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1380","Markup errors in trunk numpy ma API CHANGES.txt Trac 782).
|
|
Original ticket http projects.scipy.org numpy ticket 782 on 2008 05 08 by vnoel, assigned to unknown.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1381","Markup errors in trunk numpy ma API CHANGES.txt Trac 783).
|
|
Original ticket http projects.scipy.org numpy ticket 783 on 2008 05 08 by vnoel, assigned to unknown.
|
|
There are two errors in the markup in trunk numpy ma API CHANGES.txt, see http scipy.org scipy numpy wiki MaskedArrayApiChanges I ve noticed there is no numpy.ma Trac component for bugs in masked arrays.
|
|
Should one be created ?","bug"
|
|
"6013","1382","Make test infrastructure Valgrind friendly Trac 784).
|
|
Original ticket http projects.scipy.org numpy ticket 784 on 2008 05 08 by alberts, assigned to alanmcintyre.
|
|
From time to time Valgrind finds a memory leak when running the NumPy tests.
|
|
To find these leaks, it would be useful if it were possible to run exactly one test per Python invocation.
|
|
A buildbot might do something like the following in a shell script for t in `python numpytest.py list` valgrind python numpytest.py t end So there should be a little script that produces a list of all tests that can be used to run exactly one test at a time.","bug"
|
|
"6013","1383","Error in linalg.norm ) Trac 785).
|
|
Original ticket http projects.scipy.org numpy ticket 785 on 2008 05 09 by trac user nick, assigned to unknown.
|
|
While working on a unit test for linalg.norm ) see ticket 1361), I discovered that if a vector is passed into the norm ) method with fro as the argument for the ord, an error occurs.
|
|
Example >>> from numpy import linalg >>> a = [1,2,3,4] >>> linalg.norm a, fro ) Traceback most recent call last) File <stdin> , line 1, in <module> File tmp lib python2.5 site packages numpy linalg linalg.py , line 1262, in norm return abs x) ord).sum )) 1.0 ord) TypeError unsupported operand type s) for or pow ) numpy.ndarray and str >>> Performing the same test but omiting the parameter fro allows the method to execute normally.
|
|
A test case that exposes this is available in ticket 1361.","bug"
|
|
"6013","1384","TypeError in linalg.norm ) Trac 786).
|
|
Original ticket http projects.scipy.org numpy ticket 786 on 2008 05 09 by trac user nick, assigned to unknown.
|
|
While working on a unit test for linalg.norm ) see ticket 1361), I discovered that if a vector is passed into the norm ) method with fro as the argument for the ord, an error occurs.
|
|
Example >>> from numpy import linalg >>> a = [1,2,3,4] >>> linalg.norm a, fro ) Traceback most recent call last) File <stdin> , line 1, in <module> File tmp lib python2.5 site packages numpy linalg linalg.py , line 1262, in norm return abs x) ord).sum )) 1.0 ord) TypeError unsupported operand type s) for or pow ) numpy.ndarray and str >>> Performing the same test but omiting the parameter fro allows the method to execute normally.
|
|
A test case that exposes this is available in ticket 1361.","bug"
|
|
"6013","1385","Matrix power augmented assignment doesn t work correctly Trac 787).
|
|
Original ticket http projects.scipy.org numpy ticket 787 on 2008 05 10 by trac user jprey ..., assigned to cournape.
|
|
The power augmented assignment for matrix objects doesn t work correctly.
|
|
It is inherited from ndarray and compute element by element power instead of matrix power.
|
|
>>> m = numpy.matrix 1 2 ; 3 4 ) >>> m = 2 >>> print m [[ 1 4] [ 9 16]] == Proposed fix == in numpy core defmatrix.py, add a new method to class matrix def ipow self, other) self[ ] = self other return self","bug"
|
|
"6013","1386","astype sometimes fails to return a copy Trac 788).
|
|
Original ticket http projects.scipy.org numpy ticket 788 on 2008 05 10 by efiring, assigned to unknown.
|
|
The astype ndarray method is supposed to return a copy, but sometimes it returns the original array.
|
|
Here is an example session illustrating this,using a pickled array.
|
|
Note that the array interface [ data ] points to the same location for the original array and for the result of the astype d ) method.
|
|
The badarray pickle file is attached.
|
|
In [1] import cPickle In [3] xx = cPickle.load open badarray )) In [4] xx Out[4] array [ 23.73140116, 23.72649352, 23.72382685, 23.72405394, 23.72166806, 23.6917912 , 23.62796481, 23.59340417, 23.66240509, 23.78591019, 23.78932755, 23.7951713 , 23.800875 , 23.80596065, 23.7897963 , 23.79318796, 23.79156296, 23.78386296, 23.76369236, 23.76529907, 23.7601169 , 23.76050602, 23.75864167, 23.75500278, 23.75078102, 23.74768727, 23.74270162, 23.73668056, 23.68076759, 23.59250231, 23.54354653, 23.59569167, 23.65320718, 23.74927338, 23.80626389, 23.83042199, 23.82865718, 23.82814468, 23.82959745, 23.83229907, 23.8211838 , 23.87093542, 23.88018032, 23.89005787, 23.89711852, 23.89482917, 23.89307963, 23.89227315, 23.89159653, 23.89037407, 23.88799167, 23.88153796, 23.80637824, 23.74419097, 23.73798194, 23.75471204, 23.87062662, 23.91047731, 23.94460741, 23.9374419 , 23.89750278, 23.81816829, 23.81924861, 23.83898287, 23.84776829, 23.85297662, 23.85658449, 23.86032917, 23.86472083, 23.87044954, 23.87483171, 23.88724005, 23.96076778]) In [5] xx.
|
|
array interface Out[5] { data 137505876, False), descr [ , <f8 )], shape 73,), strides None, typestr <f8 , version 3} In [6] xxx = xx.astype d ) In [7] xxx.
|
|
array interface Out[7] { data 137505876, False), descr [ , <f8 )], shape 73,), strides None, typestr <f8 , version 3} In [8] import numpy as np In [9] np.version Out[9] <module numpy.version from usr local lib python2.5 site packages numpy version.pyc > In [10] np.
|
|
version Out[10] 1.2.0.dev5150","bug"
|
|
"6013","1387","Segfault if ndarray.compress can t safely cast into the out array Trac 789).
|
|
Original ticket http projects.scipy.org numpy ticket 789 on 2008 05 12 by alanmcintyre, assigned to cournape.
|
|
If you give ndarray.compress an out argument of a type with lower precision than self, it causes a segfault.
|
|
Here s a simple example > from numpy import > a = array [[1,2,3],[4,5,6]]) > b = array [[0.,0.],[0.,0.
|
|
]],dtype=single) > a.compress [False,False,True],axis=1,out=c) Segmentation fault This appears to happen because the call to PyArray FromArray from PyArray TakeFrom at multiarraymodule.c line 3896) fails but there is no NULL check for the return variable.
|
|
I m attaching a patch that adds a null check, and just ends up raising a ValueError for the above example.","bug"
|
|
"6013","1388","import error with latest numpy Trac 790).
|
|
Original ticket http projects.scipy.org numpy ticket 790 on 2008 05 12 by chanley, assigned to unknown.
|
|
I am unable to import numpy after building from svn.
|
|
I receive the following error In [1] import numpy <type exceptions.ImportError > Traceback most recent call last) home chanley <ipython console> in <module> ) user chanley dev site packages lib python numpy init .py in <module> ) 50 import lib 51 from lib import > 52 import linalg 53 import fft 54 import random user chanley dev site packages lib python numpy linalg init .py in <module> ) 2 from info import doc 3 > 4 from linalg import 5 6 def test level=1, verbosity=1) user chanley dev site packages lib python numpy linalg linalg.py in <module> ) 27 isfinite, size 28 from numpy.lib import triu > 29 from numpy.linalg import lapack lite 30 from numpy.core.defmatrix import matrix power, matrix 31 <type exceptions.ImportError > user chanley dev site packages lib python nump y linalg lapack lite.so undefined symbol e wsfe This is a RedHat Enterprise system running on a 64 bit processor.
|
|
The build is from r5157.
|
|
This has never been an issue before.","bug"
|
|
"6013","1389","Possible bug std var not properly calling array finalize Trac 791).
|
|
Original ticket http projects.scipy.org numpy ticket 791 on 2008 05 13 by trac user psederberg, assigned to unknown.
|
|
Hi Folks My colleagues and I have been making a subclass of ndarray that has the ability to save custom attributes similar to the example given at http www.scipy.org Subclasses), however I think I ve found a bug where array finalize does not get called properly when you use the std or var methods all other methods I have tried work fine.)
|
|
Here s a sample class that is even simpler than the example import numpy as np that s what was adopted right?
|
|
class TestArray np.ndarray) def new cls, data, info) ensure ndarray result = np.array data) transform the data to the new class result = result.view cls) set the custom attribute result.info = info return new custom array return result def array finalize self, obj) provide info for what s happening print finalize \t s\n\t\t s self.
|
|
class , obj.
|
|
class ) set the custom attribute self.info = getattr obj, info , ) provide more info if hasattr obj, info ) print \t\t s s self.info, obj.info) else print \t\t s None self.info) So, if you run it like this, you ll see the problem the debug messages are a little verbose) dat = TestArray [[1,2,3,4],[5,6,7,8]], jubba ) dat.info gives you jubba dat.mean 1).info gives you jubba dat.std 1).info gives you After calling array finalize with the proper self and obj, the last time it is called in std is with a ndarray that is not an instance of TestArray.
|
|
I m not sure why this is, but I don t think it s correct and if it is, I think it prevents all subclassing with custom attributes like the example given on the scipy page) from working.
|
|
Thanks for looking at this issue because it s preventing us from continuing work.
|
|
If you do find and fix it, could you also suggest any workarounds for the current release because we would ideally like to use the numpy releases in Debian testing.
|
|
Oh, I forgot to tell you my version info numpy.version.version = 1.0.4 python = Python 2.5.2 This is the current Debian testing.
|
|
Thanks, Per","bug"
|
|
"6013","1390","set D FORTIFY SOURCE=1 instead of 2 Trac 792).
|
|
Original ticket http projects.scipy.org numpy ticket 792 on 2008 05 14 by huard, assigned to pearu.
|
|
On Fedora, the flag D FORTIFY SOURCE is set to 2 by default, with for effect the removal of all docstrings from f2py generated fortran extensions.
|
|
The following patch sets the flag to 1.
|
|
Note to devs I have no knowledge of distutils so this patch may be seriously flawed or have unintended consequences.
|
|
On the upside, it works for me.
|
|
Index numpy distutils ccompiler.py =================================================================== numpy distutils ccompiler.py revision 5077) numpy distutils ccompiler.py working copy) 219,6 219,12 if hasattr self, compiler ) log.warn s self.compiler,)) log.warn Missing compiler cxx fix for self.
|
|
class .
|
|
name ) try i = self.compiler so.index Wp, D FORTIFY SOURCE=2 ) self.compiler so[i] = Wp, D FORTIFY SOURCE=1 except AttributeError, ValueError) pass return replace method CCompiler, customize , CCompiler customize)","bug"
|
|
"6013","1391","recarray.list ) raises RunTimeError Trac 793).
|
|
Original ticket http projects.scipy.org numpy ticket 793 on 2008 05 16 by huard, assigned to unknown.
|
|
[10] from numpy import [11] num = 2 [12] a = recarray num, formats= i4,f8,f8 ,names= id,x,y ) [13] a [13] recarray [ 15987136, 1.1449730052959239e 12, 6.7903865339544522e 313), 0, 4.7430302000759668e 322, 2.4190752022043574e 312)], dtype=[ id , <i4 ), x , <f8 ), y , <f8 )]) [14] a.tolist ) RuntimeError Traceback most recent call last) aos home huardda <ipython console> in <module> ) RuntimeError array item not returning smaller dimensional array","bug"
|
|
"6013","1392","Patch to fix behavior of numpy.sign numpy.nan) Trac 794).
|
|
Original ticket http projects.scipy.org numpy ticket 794 on 2008 05 17 by trac user brorson, assigned to charris.
|
|
As was discussed on numpy discuss earlier today, I am submitting a patch to fix the behavior of numpy.sign numpy.nan).
|
|
With this patch, the following behavior obtains >>> R = numpy.array [ numpy.inf, 2, 0, 2, numpy.inf, numpy.nan]) >>> numpy.sign R) array [ 1., 1., 0., 1., 1., NaN]) >>> >>> C = numpy.array [ numpy.inf, 2 1j, 0, 2 1j, numpy.inf, numpy.nan]) >>> numpy.sign C) array [ 1.
|
|
0.j, 1.
|
|
0.j, 0.
|
|
0.j, 1.
|
|
0.j, 1.
|
|
0.j, NaN 0.j]) Cheers, Stuart Brorson[[BR]] Interactive Supercomputing, inc.[[BR]] 135 Beaver Street | Waltham | MA | 02452 | USA[[BR]] http www.interactivesupercomputing.com [[BR]]","bug"
|
|
"6013","1393","How to document classes, methods, attributes?
|
|
Trac 795).
|
|
Original ticket http projects.scipy.org numpy ticket 795 on 2008 05 17 by pearu, assigned to stefanv.
|
|
The example.py in numpy doc explains how to document Python functions and extension functions but not how to document classes and their methods as well as attributes.
|
|
For example, it is not clear to me where one should describe constructor parameters in the class documentation string or in the ` init ` method documentation string?
|
|
What if class defines also ` new ` method?
|
|
An example how to document the following class definition would be nice to have in example.py class Class BaseClass) ... attribute = ... def init self, x, y=v, args, kwargs) ... def foo self, x, y=v, args, kwargs) ... def add self, other) ...","enhancement"
|
|
"6013","1394","Fake ticket to test Trac s ability to send notification emails Trac 796).
|
|
Original ticket http projects.scipy.org numpy ticket 796 on 2008 05 20 by trac user pwang, assigned to unknown.
|
|
This is a test ticket.","bug"
|
|
"6013","1395","Follow up on changes to histogram Trac 797).
|
|
Original ticket http projects.scipy.org numpy ticket 797 on 2008 05 21 by huard, assigned to cournape.
|
|
In 1.1, a new version of histogram was introduced with a `new` keyword to toggle a new behavior for histogram function.
|
|
This new behavior is compatible with histogram2d and histogramdd and should reduce confusion with respect to the handling of outliers.
|
|
In !NumPy 1.1 , `new` is set to False by default, thus preserving the old behavior by default but allowing users to switch to the new version.
|
|
In version 1.2, `new` will be set to True by default.
|
|
For version 1.3, `new=False` will raise a deprecation warning.
|
|
For version 1.4, the new keyword will disappear.
|
|
Overall, this will leave users around 1 year to convert their code assuming a 4 months cycle between releases).","bug"
|
|
"6013","1396","`piecewise` exposes raw memory Trac 798).
|
|
Original ticket http projects.scipy.org numpy ticket 798 on 2008 05 21 by stefanv, assigned to unknown.
|
|
`piecewise` uses `empty` to allocate output memory.
|
|
If the conditions do not sufficiently cover the output, then raw memory is returned, e.g., import numpy as np np.piecewise [0,1,2],[True,False,False],[1]) A patch which addresses the issue is available here for review http codereview.appspot.com 1105 Documentation is being updated on the wiki.","bug"
|
|
"6013","1397","Segfault when subclassing numpy.float64 with multiple inheritance Trac 799).
|
|
Original ticket http projects.scipy.org numpy ticket 799 on 2008 05 23 by trac user ReinderKamphorst, assigned to cournape.
|
|
With numpy 1.0.4 python 2.4.4 on win XP), the following code causes python to crash import numpy class A object) pass class B A, numpy.float64) pass x = B 1.0) Apparently numpy.float64 doesn t like to be subclassed the way class B does.
|
|
However, no crash occurs with the following snippet import numpy class A object) pass class B numpy.float64, A) pass x = B 1.0) The same is true if you substitute numpy.float64 with numpy.int64.
|
|
I haven t tried this yet with other numpy.
|
|
types.
|
|
Subclassing numpy.float64 and the like like this might seem a bit strange; I use it in a tool that traces variables and their values throughout their existence in an arbitrary python script.
|
|
The trick above works with regular int and float.
|
|
Cheers, Reinder","bug"
|
|
"6013","1398","numpy.float64 has wrong value when subclassed with multiple inheritance Trac 800).
|
|
Original ticket http projects.scipy.org numpy ticket 800 on 2008 05 23 by trac user ReinderKamphorst, assigned to unknown.
|
|
When subclassed with multiple inheritance, an instance of the subclass doesn t get the wanted value >>> import numpy >>> class A object) pass >>> class B numpy.float64, A) def init self, val) numpy.float64.
|
|
init self, val) >>> x = B 1.0) >>> x 0.0 The behavior is correct with single inheritance >>> import numpy >>> class A object) pass >>> class B numpy.float64) def init self, val) numpy.float64.
|
|
init self, val) >>> x = B 1.0) >>> x 1.0","bug"
|
|
"6013","1399","load library cannot load libraries with extensions in the name Trac 801).
|
|
Original ticket http projects.scipy.org numpy ticket 801 on 2008 05 25 by cournape, assigned to unknown.
|
|
!python import numpy as np np.ctypeslib.load library libc.so , lib) Raises an error <type exceptions.UnboundLocalError > Traceback most recent call last) usr media misc local stow <ipython console> in <module> ) home david local stow numpy.work lib python2.5 site packages numpy ctypeslib.py in load library libname, loader path) 46 libdir = loader path 47 > 48 for ln in libname ext 49 try 50 libpath = os.path.join libdir, ln) <type exceptions.UnboundLocalError > local variable libname ext referenced before assignment","bug"
|
|
"6013","1400","`zeros like` does not support record arrays Trac 802).
|
|
Original ticket http projects.scipy.org numpy ticket 802 on 2008 05 28 by stefanv, assigned to stefanv.
|
|
>>> x array [ 1, 2), 3, 4)], dtype=[ x , <i4 ), y , <i4 )]) >>> np.zeros like x) Traceback most recent call last) File <ipython console> , line 1, in <module> File Users stefan lib python2.5 site packages numpy core numeric.py , line 56, in zeros like res.fill 0) TypeError expected a readable buffer object This should work, and be equivalent to >>> np.zeros x.shape, x.dtype) array [ 0, 0), 0, 0)], dtype=[ x , <i4 ), y , <i4 )])","bug"
|
|
"6013","1401","Assignment problem on matrix advanced selection Trac 803).
|
|
Original ticket http projects.scipy.org numpy ticket 803 on 2008 05 31 by trac user jprey ..., assigned to unknown.
|
|
I found a problem when assigning a matrix to a matrix advanced selection in numpy 1.0.4. m = numpy.matrix [[1, 2], [3, 4]]) m[[0, 0], [0, 1]] = numpy.matrix [ 1, 2]) ValueError array is not broadcastable to correct shape This is specific to assigning a matrix to a matrix advanced selection.
|
|
The following examples all work fine m = numpy.matrix [[1, 2], [3, 4]]) m[0] = numpy.matrix [ 1, 2]) m = numpy.matrix [[1, 2], [3, 4]]) m[[0, 0], [0, 1]] = numpy.array [ 1, 2]) Most surprisingly, the following example also works m = numpy.matrix [[1, 2], [3, 4]]) m[numpy.matrix [0, 0]), numpy.matrix [0, 1])] = numpy.matrix [ 1, 2]) I discovered this poblem while trying m = numpy.matrix [[1, 2], [3, 4]]) m[m < 3] = m[m < 3] ValueError array is not broadcastable to correct shape I believe that if the integer advanced selection is fixed, the boolean advanced selection will also be corrected.","bug"
|
|
"6013","1402","numpy 1.1.0 does not build as a python egg easy install.exe) Trac 804).
|
|
Original ticket http projects.scipy.org numpy ticket 804 on 2008 06 02 by trac user djkawa, assigned to cournape.
|
|
Hi, i have been trying to build the numpy 1.1.0 release as an egg with the easy install.exe) but it fails both from the tar.gz source archive a strange error) and the binary install because it is not anymore a simple distutils .exe).
|
|
the one for python 2.4) any help on that, or having the 1.1.0 for python 2.4 available as an egg even if i don t build it myself, would be much appreciated.
|
|
thanks gilles","bug"
|
|
"6013","1403","PyArray FromDims should use intp instead of int type for the dims on 64 bit machines Trac 805).
|
|
Original ticket http projects.scipy.org numpy ticket 805 on 2008 06 02 by trac user gnatty7, assigned to charris.
|
|
In order to get the num util library working on our 64 bit amd64) linux machines, I had to make some changes to arrayobject.c, and fftpack litemodule.c Specifically, changed the d argument in PyArray FromDimsandDataAndDescr and PyArray FromDims to use the intp type rather than int.
|
|
The allows proper casting from the 8 byte pointer to the 8 byte long type that is used when Python is compiled on the 64 bit machine Py intptr t is set to long) which intp and npy intp are thus set to.
|
|
In addition, fftpack litemodule.c was patched to use npt intp types for the dim variable in the fftpack rffti and fftp pack cffti functions.
|
|
I have compiled and ran the unit tests on our machine against Python 2.5.1, and with numpy compiled against Intel MKL 10.0.1.014, and all tests pass with these edits.
|
|
I hope that I m not breaking some other functionality or what not with this change.
|
|
Any help or advice would be greatly appreciated.
|
|
Thank You for developing such an amazing library and suite of tools!
|
|
Nate","bug"
|
|
"6013","1404","Conversion of masked arrays from one dtype to another fails because of fill value Trac 806).
|
|
Original ticket http projects.scipy.org numpy ticket 806 on 2008 06 02 by huard, assigned to pierregm.
|
|
Converting an array of char to an integer arrays fails.
|
|
The problem seems to be due to the fill value attribute which is not correctly modified when going from one type to another.
|
|
> [29] a = ma.array [ 3 , 4 , 5 ]) > [30] ma.array a, dtype=int) ValueError Traceback most recent call last) storage huardda buoy buoy metocean.py in <module> ) > 1 2 3 4 5 usr lib64 python2.5 site packages numpy ma core.pyc in array data, dtype, copy, order, mask, fill value, keep mask, hard mask, shrink, subok, ndmin) 2680 return MaskedArray data, mask=mask, dtype=dtype, copy=copy, subok=subok, 2681 keep mask=keep mask, hard mask=hard mask, > 2682 fill value=fill value, ndmin=ndmin, shrink=shrink) 2683 array.
|
|
doc = masked array.
|
|
doc 2684 usr lib64 python2.5 site packages numpy ma core.pyc in new cls, data, mask, dtype, copy, subok, ndmin, fill value, keep mask, hard mask, flag, shrink, options) 1218 if fill value is None 1219 fill value = getattr data, fill value , None) > 1220 data.
|
|
fill value = check fill value fill value, data.dtype) 1221 Process extra options .. 1222 data.
|
|
hardmask = hard mask usr lib64 python2.5 site packages numpy ma core.pyc in check fill value fill value, dtype) 199 for f,d) in zip fval, descr)] 200 else > 201 fill value = narray fval, copy=False, dtype=dtype).item ) 202 return fill value 203 ValueError invalid literal for int ) with base 10 N A","bug"
|
|
"6013","1405","massked array value = scalar value > mask and data must be same size Trac 807).
|
|
Original ticket http projects.scipy.org numpy ticket 807 on 2008 06 03 by trac user Bill5107, assigned to pierregm.
|
|
I have a case where I cannot do array value = scalar value but I can to array value = array value scalar value It seems the near negative limit values are somehow involved in the failure.
|
|
Here is a short example.
|
|
>>> numpy.
|
|
version 1.1.0 >>> import numpy >>> b = b 2 >>> a=numpy.array [1.23, 3.40282346639e 35,3.45], numpy.float32) >>> b=numpy.ma.masked values a, 3.40282346639e 038) >>> b =2 Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python25\Lib\site packages\numpy\ma\core.py , line 1743, in idiv numpy.putmask other data, dom mask, 1) ValueError putmask mask and data must be the same size >>> b = b 2 But if I slightly change the value in a , things seem ok >>> b = b 2 >>> a=numpy.array [1.23, 3.40282346639e 34,3.45], numpy.float32) >>> b=numpy.ma.masked values a, 3.40282346639e 038) >>> b = b 2 Does this look like a bug or am I doing something wrong here?
|
|
Thanks!","bug"
|
|
"6013","1406","inserting multiple values using numpy.insert ) Trac 808).
|
|
Original ticket http projects.scipy.org numpy ticket 808 on 2008 06 03 by trac user markbakker, assigned to unknown.
|
|
The insert function works incorrectly when trying to insert multiple values after one index.
|
|
It works if there are multiple positions but only one value a = arange 5) insert a,[3,3],4) Returns array [0, 1, 2, 4, 4, 3, 4]) But not when you give one position with multiple values insert a,3,[7,7]) Returns array [0, 1, 2, 7, 3, 4]) while it should have returned array [0, 1, 2, 7, 7, 3, 4])","bug"
|
|
"6013","1407","dtype ) should accept strings like >int32 Trac 809).
|
|
Original ticket http projects.scipy.org numpy ticket 809 on 2008 06 03 by rkern, assigned to rkern.
|
|
The long form dtype strings like int32 or double are useful for their readability when constructing record dtypes.
|
|
However, since they do not accept endianness specifiers, their utility is restricted to native endian types and cannot be used for specifying cross platform, endian specific file formats.","enhancement"
|
|
"6013","1408","Verbose argument not effective for ndarrays in assert equal Trac 810).
|
|
Original ticket http projects.scipy.org numpy ticket 810 on 2008 06 05 by huard, assigned to unknown.
|
|
The assert equal function provides a verbose keyword but it has no effect for ndarrays >>> a = np.array [1,2]) >>> np.testing.assert equal a, a 1, verbose=False) <type exceptions.AssertionError > Traceback most recent call last) [...] <type exceptions.AssertionError > Arrays are not equal mismatch 100.0 ) x array [1, 2]) y array [2, 3]) This seems to be simply due to the fact that verbose is not given as an argument to assert array equal.
|
|
If nobody objects, I ll fix that and report the revision number in a comment below.","bug"
|
|
"6013","1409","install error Error suffix or operands invalid for `fnstsw Trac 811).
|
|
Original ticket http projects.scipy.org numpy ticket 811 on 2008 06 05 by trac user nbraspen, assigned to cournape.
|
|
My installation of numpy on a fresh cygwin failed on this Error suffix or operands invalid for `fnstsw .
|
|
I think the error occured during the gcc step for umathmodule.
|
|
Browsing the internet on this error also appeared for other packages than numpy), I found the following reason at http www.nabble.com The Linux binutils 2.18.50.0.4 is released td15360254.html Starting from the 2.18.50.0.4 release, the x86 assembler no longer accepts fnstsw eax fnstsw stores 16bit into ax and the upper 16bit of eax is unchanged.
|
|
Please use fnstsw ax Since installation of numpy on cygwin worked on another pc, I compared the installed binutils of cygwin and the fresh cygwin had version 20080523 while the other pc had version 20060817.
|
|
By installing the older binutils package installation of numpy went fine.
|
|
So I guess that numpy should use the correct new fnstsw suffix operand when a newer version of binutils is installed on the system.","bug"
|
|
"6013","1410","numpy.sum, numpy.std, numpy.var fail on masked arrays Trac 812).
|
|
Original ticket http projects.scipy.org numpy ticket 812 on 2008 06 06 by trac user bsulman, assigned to pierregm.
|
|
The numpy sum, std, and var functions attempt to call the corresponding method of a masked array with arguments axis, dtype, out), but the corresponding numpy.ma methods only accept the two keywords axis and dtype.
|
|
This would be fixed by adding the out keyword to the numpy.ma methods x.mean already accepts the out keyword and does not have this problem).
|
|
This is current as of version 1.1.0 Example In [1] import numpy In [2] x=numpy.ma.arange 10) In [3] x.sum ) Out[3] 45 In [4] numpy.sum x) <type exceptions.TypeError > Traceback most recent call last) data projects cheas lcreek data lic <ipython console> in <module> ) home bsulman lib python numpy core fromnumeric.py in sum a, axis, dtype, out) 992 except AttributeError 993 return wrapit a, sum , axis, dtype, out) > 994 return sum axis, dtype, out) 995 996 <type exceptions.TypeError > sum ) takes at most 3 arguments 4 given) In [7] numpy.
|
|
version Out[7] 1.1.0","bug"
|
|
"6013","1411","zerofill has a path with no return value Trac 813).
|
|
Original ticket http projects.scipy.org numpy ticket 813 on 2008 06 07 by alanmcintyre, assigned to unknown.
|
|
I happened to see a compiler warning about zerofill in core src arrayobject.c) not having a return value, and lo and behold the success path doesn t have a return statement.
|
|
I m attaching a patch that adds one.","bug"
|
|
"6013","1412","self test failuers on Sun Solaris 10 System, version 1.2.0.dev5255 Trac 814).
|
|
Original ticket http projects.scipy.org numpy ticket 814 on 2008 06 07 by chanley, assigned to unknown.
|
|
umpy is installed in usr ra pyssg 2.5.1 numpy Numpy version 1.2.0.dev5255 Python version 2.5.1 r251 54863, Jun 4 2008, 15 48 19) [C] Found 3 3 tests for numpy.core.tests.test ufunc Found 145 145 tests for numpy.core.tests.test regression Found 63 63 tests for numpy.core.tests.test unicode Found 36 36 tests for numpy.core.tests.test numerictypes Found 16 16 tests for numpy.core.tests.test umath Found 289 289 tests for numpy.core.tests.test multiarray Found 12 12 tests for numpy.core.tests.test records Found 70 70 tests for numpy.core.tests.test numeric Found 18 18 tests for numpy.core.tests.test defmatrix Found 3 3 tests for numpy.core.tests.test errstate Found 7 7 tests for numpy.core.tests.test scalarmath Found 3 3 tests for numpy.core.tests.test memmap Found 4 4 tests for numpy.distutils.tests.test fcompiler gnu Found 5 5 tests for numpy.distutils.tests.test misc util Found 2 2 tests for numpy.fft.tests.test fftpack Found 3 3 tests for numpy.fft.tests.test helper Found 15 15 tests for numpy.lib.tests.test twodim base Found 1 1 tests for numpy.lib.tests.test regression Found 4 4 tests for numpy.lib.tests.test polynomial Found 43 43 tests for numpy.lib.tests.test type check Found 6 6 tests for numpy.lib.tests.test index tricks Found 1 1 tests for numpy.lib.tests.test financial Found 1 1 tests for numpy.lib.tests.test machar Found 53 53 tests for numpy.lib.tests.test function base Found 24 24 tests for numpy.lib.tests.test datasource Found 15 15 tests for numpy.lib.tests.test io Found 10 10 tests for numpy.lib.tests.test arraysetops Found 1 1 tests for numpy.lib.tests.test ufunclike Found 6 6 tests for numpy.lib.tests.test getlimits Found 49 49 tests for numpy.lib.tests.test shape base Found 3 3 tests for numpy.linalg.tests.test regression Found 89 89 tests for numpy.linalg.tests.test linalg Found 36 36 tests for numpy.ma.tests.test old ma Found 99 99 tests for numpy.ma.tests.test core Found 15 15 tests for numpy.ma.tests.test extras Found 4 4 tests for numpy.ma.tests.test subclassing Found 18 18 tests for numpy.ma.tests.test mrecords Found 7 7 tests for numpy.tests.test random Found 16 16 tests for numpy.testing.tests.test utils Found 6 6 tests for numpy.tests.test ctypeslib .E..............................................................................................E.EE......................................................................................................................................................................................................................................................................................................................................................................E.EE.....................................................................................E..................................................................................E...........................................................................................................................................................................................................................................................................................................................EE...............................................................................................................................................................................................................................................................................................E.....EE...................................................... ====================================================================== ERROR Test generic loops.
|
|
Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops x = np.zeros 10, dtype=np.object)[0 2] SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 86 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 174, in check object array fill x = np.zeros 1, O ) SystemError error return without exception set ====================================================================== ERROR Ticket 711 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 633 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 86 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 174, in check object array fill x = np.zeros 1, O ) SystemError error return without exception set ====================================================================== ERROR Ticket 711 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 633 Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Test generic loops.
|
|
Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops x = np.zeros 10, dtype=np.object)[0 2] SystemError NULL result without error in PyObject Call ====================================================================== ERROR check recarray from obj numpy.core.tests.test records.TestFromrecords) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy core tests test records.py , line 44, in check recarray from obj a = zeros count, dtype= O ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR check 3d numpy.lib.tests.test shape base.TestApplyAlongAxis) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy lib tests test shape base.py , line 18, in check 3d assert array equal apply along axis sum,0,a), [[27,30,33],[36,39,42],[45,48,51]]) File usr stsci pyssgdev 2.5.1 numpy lib shape base.py , line 24, in apply along axis i = zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR check simple numpy.lib.tests.test shape base.TestApplyAlongAxis) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy lib tests test shape base.py , line 11, in check simple assert array equal apply along axis len,0,a),len a) ones shape a)[1])) File usr stsci pyssgdev 2.5.1 numpy lib shape base.py , line 24, in apply along axis i = zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR check 3d numpy.ma.tests.test extras.TestApplyAlongAxis) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 325, in check 3d xa = apply along axis myfunc,2,a) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR Tests median w 2D Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 343, in test 2d assert equal median z[ ,0]), 0) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median result = apply along axis median1D, axis, asorted) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR Tests median w 3D Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 350, in test 3d assert equal median x,0), [[12,9],[6,15],[12,9],[18,15]]) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median result = apply along axis median1D, axis, asorted) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set Ran 1297 tests in 25.295s FAILED errors=14) ctypes is not available on this python skipping the test import error was ctypes is not available.)
|
|
No distutils available, skipping test.
|
|
errors <numpy.core.tests.test ufunc.TestUfunc testMethod=check generic loops>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops\n x = np.zeros 10, dtype=np.object)[0 2]\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array fill>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 174, in check object array fill\n x = np.zeros 1, \ O\ )\nSystemError error return without exception set\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcount self assign>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcounting>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array fill>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 174, in check object array fill\n x = np.zeros 1, \ O\ )\nSystemError error return without exception set\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcount self assign>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcounting>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test ufunc.TestUfunc testMethod=check generic loops>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops\n x = np.zeros 10, dtype=np.object)[0 2]\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test records.TestFromrecords testMethod=check recarray from obj>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy core tests test records.py , line 44, in check recarray from obj\n a = zeros count, dtype=\ O\ )\nSystemError NULL result without error in PyObject Call\n ) <numpy.lib.tests.test shape base.TestApplyAlongAxis testMethod=check 3d>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy lib tests test shape base.py , line 18, in check 3d\n assert array equal apply along axis sum,0,a), [[27,30,33],[36,39,42],[45,48,51]])\n File usr stsci pyssgdev 2.5.1 numpy lib shape base.py , line 24, in apply along axis\n i = zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.lib.tests.test shape base.TestApplyAlongAxis testMethod=check simple>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy lib tests test shape base.py , line 11, in check simple\n assert array equal apply along axis len,0,a),len a) ones shape a)[1]))\n File usr stsci pyssgdev 2.5.1 numpy lib shape base.py , line 24, in apply along axis\n i = zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.ma.tests.test extras.TestApplyAlongAxis testMethod=check 3d>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 325, in check 3d\n xa = apply along axis myfunc,2,a)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.ma.tests.test extras.TestMedian testMethod=test 2d>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 343, in test 2d\n assert equal median z[ ,0]), 0)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median\n result = apply along axis median1D, axis, asorted)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.ma.tests.test extras.TestMedian testMethod=test 3d>, Traceback most recent call last) \n File usr ra pyssg 2.5.1 numpy ma tests test extras.py , line 350, in test 3d\n assert equal median x,0), [[12,9],[6,15],[12,9],[18,15]])\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median\n result = apply along axis median1D, axis, asorted)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) failures if 1 != 0 ) then echo <font color=orange>selftest< font> exit 0","bug"
|
|
"6013","1413","self test failuers on RHE 4 32 bit, version 1.2.0.dev5255 Trac 815).
|
|
Original ticket http projects.scipy.org numpy ticket 815 on 2008 06 07 by chanley, assigned to unknown.
|
|
Numpy is installed in usr stsci pyssgdev 2.5.1 numpy Numpy version 1.2.0.dev5255 Python version 2.5.1 r251 54863, Jan 16 2008, 14 28 22) [GCC 3.4.6 20060404 Red Hat 3.4.6 9)] Found 3 3 tests for numpy.core.tests.test ufunc Found 145 145 tests for numpy.core.tests.test regression Found 63 63 tests for numpy.core.tests.test unicode Found 36 36 tests for numpy.core.tests.test numerictypes Found 16 16 tests for numpy.core.tests.test umath Found 289 289 tests for numpy.core.tests.test multiarray Found 12 12 tests for numpy.core.tests.test records Found 70 70 tests for numpy.core.tests.test numeric Found 18 18 tests for numpy.core.tests.test defmatrix Found 3 3 tests for numpy.core.tests.test errstate Found 7 7 tests for numpy.core.tests.test scalarmath Found 3 3 tests for numpy.core.tests.test memmap Found 4 4 tests for numpy.distutils.tests.test fcompiler gnu Found 5 5 tests for numpy.distutils.tests.test misc util Found 2 2 tests for numpy.fft.tests.test fftpack Found 3 3 tests for numpy.fft.tests.test helper Found 15 15 tests for numpy.lib.tests.test twodim base Found 1 1 tests for numpy.lib.tests.test regression Found 4 4 tests for numpy.lib.tests.test polynomial Found 43 43 tests for numpy.lib.tests.test type check Found 6 6 tests for numpy.lib.tests.test index tricks Found 1 1 tests for numpy.lib.tests.test financial Found 1 1 tests for numpy.lib.tests.test machar Found 53 53 tests for numpy.lib.tests.test function base Found 24 24 tests for numpy.lib.tests.test datasource Found 15 15 tests for numpy.lib.tests.test io Found 10 10 tests for numpy.lib.tests.test arraysetops Found 1 1 tests for numpy.lib.tests.test ufunclike Found 6 6 tests for numpy.lib.tests.test getlimits Found 49 49 tests for numpy.lib.tests.test shape base Found 3 3 tests for numpy.linalg.tests.test regression Found 89 89 tests for numpy.linalg.tests.test linalg Found 36 36 tests for numpy.ma.tests.test old ma Found 99 99 tests for numpy.ma.tests.test core Found 15 15 tests for numpy.ma.tests.test extras Found 4 4 tests for numpy.ma.tests.test subclassing Found 18 18 tests for numpy.ma.tests.test mrecords Found 7 7 tests for numpy.tests.test random Found 16 16 tests for numpy.testing.tests.test utils Found 6 6 tests for numpy.tests.test ctypeslib .E..............................................................................................E.EE......................................................................................................................................................................................................................................................................................................................................................................E.EE.....................................................................................E..................................................................................E............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E.....EE...................................................... ====================================================================== ERROR Test generic loops.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops x = np.zeros 10, dtype=np.object)[0 2] SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 86 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 174, in check object array fill x = np.zeros 1, O ) SystemError error return without exception set ====================================================================== ERROR Ticket 711 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 633 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 86 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 174, in check object array fill x = np.zeros 1, O ) SystemError error return without exception set ====================================================================== ERROR Ticket 711 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Ticket 633 Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting arr = np.zeros 5, dtype=np.object ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR Test generic loops.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops x = np.zeros 10, dtype=np.object)[0 2] SystemError NULL result without error in PyObject Call ====================================================================== ERROR check recarray from obj numpy.core.tests.test records.TestFromrecords) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy core tests test records.py , line 44, in check recarray from obj a = zeros count, dtype= O ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR check 3d numpy.ma.tests.test extras.TestApplyAlongAxis) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 325, in check 3d xa = apply along axis myfunc,2,a) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR Tests median w 2D Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 343, in test 2d assert equal median z[ ,0]), 0) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median result = apply along axis median1D, axis, asorted) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set ====================================================================== ERROR Tests median w 3D Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 350, in test 3d assert equal median x,0), [[12,9],[6,15],[12,9],[18,15]]) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median result = apply along axis median1D, axis, asorted) File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis i = np.zeros nd, O ) SystemError error return without exception set Ran 1297 tests in 5.435s FAILED errors=12) errors <numpy.core.tests.test ufunc.TestUfunc testMethod=check generic loops>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops\n x = np.zeros 10, dtype=np.object)[0 2]\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array fill>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 174, in check object array fill\n x = np.zeros 1, \ O\ )\nSystemError error return without exception set\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcount self assign>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcounting>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array fill>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 174, in check object array fill\n x = np.zeros 1, \ O\ )\nSystemError error return without exception set\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcount self assign>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 961, in check object array refcount self assign\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test regression.TestRegression testMethod=check object array refcounting>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test regression.py , line 863, in check object array refcounting\n arr = np.zeros 5, dtype=np.object )\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test ufunc.TestUfunc testMethod=check generic loops>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test ufunc.py , line 131, in check generic loops\n x = np.zeros 10, dtype=np.object)[0 2]\nSystemError NULL result without error in PyObject Call\n ) <numpy.core.tests.test records.TestFromrecords testMethod=check recarray from obj>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy core tests test records.py , line 44, in check recarray from obj\n a = zeros count, dtype=\ O\ )\nSystemError NULL result without error in PyObject Call\n ) <numpy.ma.tests.test extras.TestApplyAlongAxis testMethod=check 3d>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 325, in check 3d\n xa = apply along axis myfunc,2,a)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.ma.tests.test extras.TestMedian testMethod=test 2d>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 343, in test 2d\n assert equal median z[ ,0]), 0)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median\n result = apply along axis median1D, axis, asorted)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) <numpy.ma.tests.test extras.TestMedian testMethod=test 3d>, Traceback most recent call last) \n File usr stsci pyssgdev 2.5.1 numpy ma tests test extras.py , line 350, in test 3d\n assert equal median x,0), [[12,9],[6,15],[12,9],[18,15]])\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 430, in median\n result = apply along axis median1D, axis, asorted)\n File usr stsci pyssgdev 2.5.1 numpy ma extras.py , line 176, in apply along axis\n i = np.zeros nd,\ O\ )\nSystemError error return without exception set\n ) failures if 1 != 0 ) then echo <font color=orange>selftest< font> exit 0","bug"
|
|
"6013","1414","Constructing numpy.object causes a crash Trac 816).
|
|
Original ticket http projects.scipy.org numpy ticket 816 on 2008 06 09 by pv, assigned to unknown.
|
|
The following code causes a segmentation fault import numpy as np print np.
|
|
version np.object 3) The reason apparently is see the Valgrind trace below) that the object constructor expects a memory address as its parameter.
|
|
This functionality probably shouldn t be user visible.
|
|
Valgrind trace valgrind suppressions=.valgrind valgrind python.supp python c import numpy; print numpy.
|
|
version ; numpy.object 3) ... 1.1.0 ==23495== ==23495== Invalid read of size 4 ==23495== at 0x4A6830F object arrtype new scalartypes.inc.src 1877) ==23495== by 0x809EDCC type call typeobject.c 422) ==23495== by 0x805CB36 PyObject Call abstract.c 1861) ==23495== by 0x80C7986 PyEval EvalFrameEx ceval.c 3784) ==23495== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==23495== by 0x80CB226 PyEval EvalCode ceval.c 494) ==23495== by 0x80EB474 PyRun StringFlags pythonrun.c 1273) ==23495== by 0x80EB534 PyRun SimpleStringFlags pythonrun.c 900) ==23495== by 0x8058C80 Py Main main.c 512) ==23495== by 0x80587F1 main python.c 23) ==23495== Address 0x3 is not stack d, malloc d or recently) free d ==23495== ==23495== Process terminating with default action of signal 11 SIGSEGV) ==23495== Access not within mapped region at address 0x3 ==23495== at 0x4A6830F object arrtype new scalartypes.inc.src 1877) ==23495== by 0x809EDCC type call typeobject.c 422) ==23495== by 0x805CB36 PyObject Call abstract.c 1861) ==23495== by 0x80C7986 PyEval EvalFrameEx ceval.c 3784) ==23495== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==23495== by 0x80CB226 PyEval EvalCode ceval.c 494) ==23495== by 0x80EB474 PyRun StringFlags pythonrun.c 1273) ==23495== by 0x80EB534 PyRun SimpleStringFlags pythonrun.c 900) ==23495== by 0x8058C80 Py Main main.c 512) ==23495== by 0x80587F1 main python.c 23) ==23495== ==23495== ERROR SUMMARY 34 errors from 9 contexts suppressed 1194 from 7) ==23495== malloc free in use at exit 4,800,420 bytes in 4,902 blocks.
|
|
==23495== malloc free 104,174 allocs, 99,272 frees, 48,061,410 bytes allocated.
|
|
==23495== For counts of detected errors, rerun with v ==23495== searching for pointers to 4,902 not freed blocks.
|
|
==23495== checked 6,128,912 bytes.
|
|
==23495== ==23495== LEAK SUMMARY ==23495== definitely lost 0 bytes in 0 blocks.
|
|
==23495== possibly lost 110,244 bytes in 210 blocks.
|
|
==23495== still reachable 4,690,176 bytes in 4,692 blocks.
|
|
==23495== suppressed 0 bytes in 0 blocks.
|
|
==23495== Rerun with leak check=full to see details of leaked memory.
|
|
Segmentation fault","bug"
|
|
"6013","1415","0 d arrays ignore printoptions Trac 817).
|
|
Original ticket http projects.scipy.org numpy ticket 817 on 2008 06 10 by stefanv, assigned to unknown.
|
|
import numpy as np np.set printoptions precision=1) print np.array [f]) print np.array f) yields [ 0.]
|
|
0.039485656484 }}}","bug"
|
|
"6013","1416","floats cast to integer overwrites two entries in a numpy integer matrix Trac 818).
|
|
Original ticket http projects.scipy.org numpy ticket 818 on 2008 06 12 by trac user deparvius, assigned to unknown.
|
|
In version 1.2.0.dev5243, A = matrix [2,3],[4,5]]) A[0][0] = NaN results in the matrix 0 0 4 5 A[0][0] = 4.8 results in 4 4 4 5 Result is A[0][0] = x int x) int x) 4 5 A[1][1] = NaN results in an array index out of bounds error.
|
|
No such problems if A is defined as an array rather than a matrix.","bug"
|
|
"6013","1417",".flat reorders data Trac 819).
|
|
Original ticket http projects.scipy.org numpy ticket 819 on 2008 06 13 by trac user ahaldane, assigned to unknown.
|
|
Hi, I get an unexpected effect from doing .flat on arrays created by c .
|
|
In the code below, the values in the variable b seem to get re ordered during the call to array.
|
|
Using .flatten ) instead works as expected.
|
|
from numpy import c , array from numpy.random import rand b = c [ rand 2,2), rand 2)] print b array b.flat) print b This prints [[ 0.8606752 0.11848512 0.95805826] [ 0.84665539 0.7989498 0.20762479]] [[ 0.8606752 0.95805826 0.7989498 ] [ 0.11848512 0.84665539 0.20762479]]","bug"
|
|
"6013","1418","numpy.ma.array ignores mask when asembling masked array inputs Trac 820).
|
|
Original ticket http projects.scipy.org numpy ticket 820 on 2008 06 16 by astraw, assigned to pierregm.
|
|
The following test fails with r5283.
|
|
I assume perhaps wrongly) that this is a bug.
|
|
It appears numpy.ma.array has an argument keep mask, which defaults to True.
|
|
That implied to me that the expected behavior would have passed this test.
|
|
{{{import numpy x = numpy.arange 10) y = numpy.arange 10) 20 z = numpy.arange 10) 40 x = numpy.ma.masked where x>5,x) X1=numpy.ma.array [x,y,z]) X2=numpy.ma.vstack [x[numpy.newaxis, ], y[numpy.newaxis, ], z[numpy.newaxis, ]]) assert numpy.ma.allclose X1.mask,X2.mask) }}}","bug"
|
|
"6013","1419","Numpy 1.1.0 broken for Python 2.3, because of unsupported dictionary update Trac 821).
|
|
Original ticket http projects.scipy.org numpy ticket 821 on 2008 06 17 by trac user dhendriks, assigned to pearu.
|
|
After I just upgraded my Numpy from 1.0.3.1 to 1.1.0 and executed using Python 2.3.4) python c import numpy; numpy.test ) I got Traceback most recent call last) File <string> , line 1, in ?
|
|
File lib python2.3 site packages numpy init .py , line 107, in ?
|
|
import ma File lib python2.3 site packages numpy ma init .py , line 14, in ?
|
|
import core File lib python2.3 site packages numpy ma core.py , line 114, in ?
|
|
max filler.update [ k, numpy.inf) for k in [numpy.float32, numpy.float64]]) AttributeError keys It seems Python 2.3 doesn t allow updating dictionaries using a list of key value pair tuples; it only allows a dictionary as parameter to update ).
|
|
See also [[BR]] http www.python.org doc 2.3.4 lib typesmapping.html[[BR]] http www.python.org doc 2.4 lib typesmapping.html","bug"
|
|
"6013","1420","Segfault deallocating long chains of views Trac 822).
|
|
Original ticket http projects.scipy.org numpy ticket 822 on 2008 06 17 by rkern, assigned to unknown.
|
|
In [1] from numpy import In [2] A = zeros 1, dtype=uint8) In [3] B = A[ ] In [4] for i in xrange 1000000) ... B = B[ ] ... ...
|
|
In [5] del B zsh segmentation fault ipython This has caused a problem for someone using PyMC.
|
|
The segfault occurs on line 2079 of arrayobject.c .
|
|
On 32 bit Intel OS X, I only get the segfault if the number of iterations is >= 261935.","bug"
|
|
"6013","1421","Size 1 arrays should be indexable by numpy.bool objects Trac 823).
|
|
Original ticket http projects.scipy.org numpy ticket 823 on 2008 06 17 by teoliphant, assigned to unknown.
|
|
The numpy.bool scalar should be allowed for masked indexing of size 1 arrays.
|
|
Currently, if x is a 0 d array, then x>3 is a numpy.bool objects and for a 0 d array y[x>3] will fail, while it should succeed as if numpy.bool were a 0 d array of boolean datatype.
|
|
This is related to Ticket 1319.
|
|
There are some questions about whether or not 0 d arrays should be be maskable .
|
|
I think they should be maskable and return a size 0 array if the mask does not succeed.","bug"
|
|
"6013","1422","Infinite loop in floating point addition Trac 824).
|
|
Original ticket http projects.scipy.org numpy ticket 824 on 2008 06 18 by stefanv, assigned to unknown.
|
|
In r5264, the following code enters an infinite loop import numpy as np z = np.float128 np.inf) np.nan to num z) 200000000000000000000000000000000000000 0 0x000d141e in construct loop self=0x4188f0, args=0x1326a30, kwds=0x0, mps=0xbf800a50) at ufuncobject.c 1787 1 0x000d30ef in PyUFunc GenericFunction self=0x4188f0, args=0x1326a30, kwds=0x0, mps=0xbf800a50) at ufuncobject.c 1953 2 0x000d47a9 in ufunc generic call self=0x4188f0, args=0x1326a30, kwds=0x4188f0) at ufuncobject.c 3316 3 0x001a3289 in call function tail ) 4 0x001a726a in PyObject CallFunction SizeT ) 5 0x00520d47 in array add m1=0x144a720, m2=0x79100) at arrayobject.c 3543 6 0x006dcc73 in longdouble add a=0x144a720, b=0x79100) at scalarmathmodule.c.src 575 7 0x001a1f92 in binary op1 ) 8 0x001a2b96 in PyNumber Add ) 9 0x000cd66e in PyUFunc OO O args=0x1f58648, dimensions=0x4188f0, steps=0x1326a30, func=0x1a2b70) at ufuncobject.c 535 10 0x000d3790 in PyUFunc GenericFunction self=0x4188f0, args=0x1326a30, kwds=0x0, mps=0xbf8014a0) at ufuncobject.c 1971 11 0x000d47a9 in ufunc generic call self=0x4188f0, args=0x13269e0, kwds=0x4188f0) at ufuncobject.c 3316 12 0x001a3289 in call function tail ) 13 0x001a726a in PyObject CallFunction SizeT ) 14 0x00520d47 in array add m1=0x144a640, m2=0x79100) at arrayobject.c 3543 15 0x006dcc73 in longdouble add a=0x144a640, b=0x79100) at scalarmathmodule.c.src 575 16 0x001a1f92 in binary op1 ) 17 0x001a2b96 in PyNumber Add ) ...","bug"
|
|
"6013","1423","SIGBUS in UNICODE getatter on Sparc Linux Trac 825).
|
|
Original ticket http projects.scipy.org numpy ticket 825 on 2008 06 18 by drnlm, assigned to charris.
|
|
The following throws a SIGBUS on Sparc Linux import numpy as np desc str = [ x , u1 ), s , S2 ), ] desc uni2 = [ x , u4 ), s , U2 ), ] desc uni = [ x , u1 ), s , U2 ), ] buffer = [ 5, cc ), 6, dd ) ] ta = np.array buffer, dtype=np.dtype desc str).newbyteorder > )) print ta ta = np.array buffer, dtype=np.dtype desc uni2).newbyteorder > )) print ta ta = np.array buffer, dtype=np.dtype desc uni).newbyteorder > )) tb = ta print assigmenet succeeds print ta python .
|
|
scipy sparc sugbus.py [ 5, cc ) 6, dd )] [ 5L, u cc ) 6L, u dd )] assigmenet succeeds Bus error From gdb Program received signal SIGBUS, Bus error.
|
|
[Switching to Thread 0xf7faba00 LWP 7000)] UNICODE getitem ip=0x2426b9 , ap=0x30be38) at numpy core src arraytypes.inc.src 264 264 while mysize > 0 dptr == 0) mysize ; This occurs with the latest svn snapshot it s causing the bus error on the Sparc Linux buildbots), but also occurs with the debian 1.0.1 pacakges, so this likes like an existing bug uncovered by the new test suite.","bug"
|
|
"6013","1424","Error in basic arithmetic involving masked objects and scalar types Trac 826).
|
|
Original ticket http projects.scipy.org numpy ticket 826 on 2008 06 20 by trac user dharland, assigned to unknown.
|
|
[Running numpy 1.1 against python 2.4.5 on OSX 10.4.11] Basic arithmetic involving some numpy types and ma.masked doesn t propagate the masked value.
|
|
For example >>> a = ma.arange 10, dtype=float) >>> a[0] = ma.masked >>> print a[2] a[0] 2.0 Should be Whether ma.masked propagates depends on the array type and order of arithmetic >>> print numpy.float64 10) ma.masked 10.0 WRONG, doesn t enter into ma core.py array wrap >>> print ma.masked numpy.float64 10) RIGHT, MaskedArray.
|
|
add handles the operation >>> print numpy.float32 10) ma.masked RIGHT, enters into ma core.py array wrap >>> print numpy.float128 10) ma.masked 10.0 WRONG, doesn t enter into ma core.py array wrap >>> print numpy.int32 10) ma.masked RIGHT, enters into ma core.py array wrap The error is ultimately caused by array wrap not being called for some numpy types which handles the propagation of the mask) Although it works correctly for an int with an ma.masked a similar error can be produced if we effectively transform ma.masked.dtype to numpy.int >>> print numpy.int 1) ma.MaskedArray 0, dtype=numpy.int , mask=True) 1 Sould be Its speculating but I believe the error is in scalarmathmodule.c name convert to ctype when called on the ma.masked PyObject) which I believe will call PyArray ScalarFromObject which does not work correctly for an ma.masked it will detect its zero dimensional and just return its data element","bug"
|
|
"6013","1425","numpy.test fails nose traceback on windows Trac 827).
|
|
Original ticket http projects.scipy.org numpy ticket 827 on 2008 06 20 by cournape, assigned to unknown.
|
|
I got the following error while testing numpy on windows !python File <string> , line 1, in <module> File C \Python25\Lib\site packages\numpy\testing\nosetester.py , line 150, in test nose.run argv=argv) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\core.py , line 368, in run return TestProgram arg, kw).success File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\core.py , line 219, in init argv=argv, testRunner=testRunner, testLoader=testLoader) File C \Python25\lib\unittest.py , line 768, in init self.runTests ) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\core.py , line 298, in runTests result = self.testRunner.run self.test) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\core.py , line 62, in run test result) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 162, in run test orig) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 162, in run test orig) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 162, in run test orig) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 162, in run test orig) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 162, in run test orig) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 132, in call return self.run arg, kw) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 155, in run for test in self.
|
|
tests File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 277, in get wrapped tests for test in self.
|
|
get tests ) File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\suite.py , line 76, in get tests for test in self.test generator File c \python25\lib\site packages\nose 0.10.3 py2.5.egg\nose\loader.py , line 221, in generate for test in g ) File C \Python25\lib\site packages\numpy\lib\tests\test format.py , line 430, in test memmap roundtrip fp = open nfn, wb ) OError [Errno 2] No such file or directory c \\docume 1\\admini 1\\locals 1\\temp\\tmp9fvihc\\normal.npy The file does exists, so I expect some problems with the path.","bug"
|
|
"6013","1426","npy file format test failures on open solaris x86, 32 bits) Trac 828).
|
|
Original ticket http projects.scipy.org numpy ticket 828 on 2008 06 21 by cournape, assigned to unknown.
|
|
Big list of test failures on open solaris.
|
|
I attached the compressed log of numpy.test verbose = 10)","bug"
|
|
"6013","1427","SEGFAULT when attempting to use arrays as indices into multi dimensional arrays Trac 829).
|
|
Original ticket http projects.scipy.org numpy ticket 829 on 2008 06 22 by cournape, assigned to unknown.
|
|
Forwarded from scipy ticket 670 I ve tried several methods of using arrays as indices into a multi dimensional array.
|
|
One of them, taken from http mathesaurus.sourceforge.net matlab numpy.html simply causes a ValueError?, the other two cause a segfault.
|
|
Sample 1 !python from pylab import import numpy print numpy.version, numpy.
|
|
file a = array [[ 11, 12, 13, 14 ], [ 21, 22, 23, 24 ], [ 31, 32, 33, 34 ]]) try a.take [0,2]).take [0,3], axis=1) except ValueError, e print Got a Value Error s e Sample 2 !python from pylab import import numpy print numpy.version, numpy.
|
|
file nrows = 648 ncols = 690 data = rand nrows,ncols) i,j = data < 0.14).nonzero ) print I get a segfault here data[ix i,j)] = data[ix i,j)] 0 Sample 3 !python from pylab import import numpy print numpy.version, numpy.
|
|
file nrows = 648 ncols = 690 data = rand nrows,ncols) i,j = data < 0.14).nonzero ) print I get a segfault here data[i, ][ ,j] = data[i, ][ ,j] 0","bug"
|
|
"6013","1428","Divide by zero memory leak Trac 830).
|
|
Original ticket http projects.scipy.org numpy ticket 830 on 2008 06 24 by cournape, assigned to cournape.
|
|
See ticket 493 from scipy trac Dividing by zero e.g.
|
|
1 scipy.float64 0.))
|
|
yields scipy.inf as it is supposed to, but leaks memory in the process.
|
|
I ll attach a sample program where memory usage grows without bound.
|
|
!python import scipy x = scipy.float64 0.)
|
|
while 1 1 x","bug"
|
|
"6013","1429","numpy.float64 poly1d = array Trac 831).
|
|
Original ticket http projects.scipy.org numpy ticket 831 on 2008 06 24 by cournape, assigned to unknown.
|
|
Initially reported on scipy trac 495) I ran into a problem with multiplying a float by a poly1d and getting an array instead of a poly1d returned.
|
|
It only actually happens if multiplying by a numpy.float64 and not just a regular float or int temp = poly1d [1.0,2.0]) temp 10 < This is fine, returns poly1d 10 temp < returns poly1d 10.0 temp < returns poly1d numpy.float64 10.0) temp < returns array of coeffs","bug"
|
|
"6013","1430","numpy.ndarray relational operators do not seem to) respect array priority Trac 832).
|
|
Original ticket http projects.scipy.org numpy ticket 832 on 2008 06 24 by cournape, assigned to unknown.
|
|
redirected from ticket 521 of scipy) Suppose I have an object class of Foo, say a that can be compared to an ndarray object, b .
|
|
I can make a b or b a both call the plus or rplus methods in class Foo.
|
|
However, there is no way to make this happen for a == b and b == a or any of the relational operators) I have not been able to figure out where to put this logic.","bug"
|
|
"6013","1431","No array function equivalent to matrix [PATCH] Trac 833).
|
|
Original ticket http projects.scipy.org numpy ticket 833 on 2008 06 24 by cournape, assigned to unknown.
|
|
See ticket 625 on scipy.","bug"
|
|
"6013","1432","numpy.sort sometimes doesn t handle strings right Trac 834).
|
|
Original ticket http projects.scipy.org numpy ticket 834 on 2008 06 24 by cournape, assigned to unknown.
|
|
See ticket 656 on scipy trac.","bug"
|
|
"6013","1433","bus error from numpy self tests on Solaris Trac 835).
|
|
Original ticket http projects.scipy.org numpy ticket 835 on 2008 06 25 by chanley, assigned to unknown.
|
|
We have been seeing a bus error on Solaris in the numpy self tests for a couple of days now.
|
|
The error is occurring with the HEAD of the svn trunk.
|
|
The message is below Writing usr stsci pyssgdev 2.5.1 numpy 1.2.0.dev5313 py2.5.egg info if 0 != 0 ) then echo BUILD SUCCEEDED STARTING TEST BUILD SUCCEEDED STARTING TEST python eng ssb auto prog lib numpy test.py usr ra pyssg 2.5.1 numpy lib function base.py 166 FutureWarning The semantics of histogram will be modified in release 1.2 to improve outlier handling.
|
|
The new behavior can be obtained using new=True.
|
|
Note that the new version accepts returns the bin edges instead of the left bin edges.
|
|
Please read the docstring for more information.
|
|
Please read the docstring for more information.
|
|
, FutureWarning) usr ra pyssg 2.5.1 numpy lib function base.py 193 FutureWarning The semantic for bins will change in version 1.2.
|
|
The bins will become the bin edges, instead of the left bin edges.
|
|
, FutureWarning) Bus Error core dumped) if 138 != 0 ) then echo <font color=orange>selftest< font> exit 0 A full build log can be found at http stsdas.stsci.edu lore log build grail numpy 06 24.23.14.log","bug"
|
|
"6013","1434","Powers of very large complex scalars Trac 836).
|
|
Original ticket http projects.scipy.org numpy ticket 836 on 2008 06 25 by stefanv, assigned to unknown.
|
|
Current behaviour In [17] z = complex 0, inf) In [18] z128 = complex128 z) In [19] z128 2 Out[19] nan nanj) In [20] z64 = complex64 z) In [21] z64 2 Out[21] nan nanj) Desired behaviour In [19] z128 2 Out[19] inf nanj)","enhancement"
|
|
"6013","1435","Infinite loop in fromfile fromstring Trac 837).
|
|
Original ticket http projects.scipy.org numpy ticket 837 on 2008 06 28 by pv, assigned to unknown.
|
|
In numpy 1.1.0 This causes an infinite loop >>> import numpy as np >>> f = open foo.txt , w ); f.write 1.234 ); f.close ) >>> np.fromfile foo.txt , sep= ) array [ 1.234]) >>> f = open foo.txt , w ); f.write 1,234 ); f.close ) >>> np.fromfile foo.txt , sep= ) [infinite loop ensues] Another way, which highlights a locale dependency >>> import numpy as np, locale >>> locale.setlocale locale.LC ALL, fi FI.UTF 8 ) >>> f = open foo.txt , w ); f.write 1,234 ); f.close ) >>> np.fromfile foo.txt , sep= ) array [ 1.234]) >>> f = open foo.txt , w ); f.write 1.234 ); f.close ) >>> np.fromfile foo.txt , sep= ) [infinite loop ensues] fromstring also hangs >>> import numpy as np >>> np.fromstring 1.234 , sep= ) array [ 1.234]) >>> np.fromstring 1,234 , sep= ) [hangs] but it s not locale dependent >>> import numpy as np, locale >>> locale.setlocale locale.LC ALL, fi FI.UTF 8 ) >>> np.fromstring 1.234 , sep= ) array [ 1.234]) >>> np.fromstring 1,234 , sep= ) [hangs] gdb shows gdb) bt 0 0xb7e4bd73 in IO vfscanf ) from lib tls i686 cmov libc.so.6 1 0xb7e4f353 in vfscanf ) from lib tls i686 cmov libc.so.6 2 0xb7e54ce2 in fscanf ) from lib tls i686 cmov libc.so.6 3 0xb7cf9cae in DOUBLE scan fp=0x8320960, ip=0xb6102cd0, ignore=0x0, ignore2=0xb7d35aa0) at numpy core src arraytypes.inc.src 877 4 0xb7ce509a in fromfile next element fp=0xbfc9fe94, dptr=0xb6102cd0, dtype=0xb7d35aa0, stream data=0x0) at numpy core src multiarraymodule.c 5994 5 0xb7cfe62b in array from text dtype=0xb7d35aa0, num= 1, sep=<value optimized out>, nread=0xbfc9fec8, stream=0x8320960, next=0xb7ce5070 <fromfile next element>, skip sep=0xb7d02f90 <fromfile skip separator>, stream data=0x0) at numpy core src multiarraymodule.c 6137 6 0xb7d02b46 in PyArray FromFile fp=0x8320960, dtype=0xb7d35aa0, num= 1, sep=0xbfc9fdb8 xefxbfxbd,\020xefxbfxbdT2xd3xb7\211\234xcfxb7T2xd3xb7xefxbfxbdxefxbfxbdxc9xbf\232Pxcexb7`\t2\bxefxbfxbd,\020xefxbfxbd ) at numpy core src multiarraymodule.c 6384 7 0xb7d02ec6 in array fromfile ignored=0x0, args=0xb7de56cc, keywds=0xb64051c4) at numpy core src multiarraymodule.c 6448 8 0x0805cb37 in PyObject Call func=0xbfc9fdb8, arg=0xb7de56cc, kw=0xb64051c4) at ..
|
|
Objects abstract.c 1861 9 0x080c7987 in PyEval EvalFrameEx f=0x81a823c, throwflag=0) at .. Python ceval.c 3784 Fixing this In any case there shouldn t be an infinite loop I think that fromfile and fromstring should behave equivalently wrt.
|
|
locales I think it might be preferable if fromfile and fromstring didn t depend on the locale The fix should probably be backported to 1.1.1","bug"
|
|
"6013","1436","Improving numpy import time Trac 838).
|
|
Original ticket http projects.scipy.org numpy ticket 838 on 2008 07 03 by cournape, assigned to cournape.
|
|
Ok, here is a serie of patches to speed up numpy import times.
|
|
First one, postponing urllib2 import.
|
|
All test pass with this one.","enhancement"
|
|
"6013","1437","DataSource tests require network access Trac 839).
|
|
Original ticket http projects.scipy.org numpy ticket 839 on 2008 07 03 by rkern, assigned to rkern.
|
|
The unit tests for the URL handling capability require network access.
|
|
We should stub in a replacement for `urllib2.urlopen )` to avoid accessing the network.","bug"
|
|
"6013","1438","SystemError in ndarray.resize, order= argument doesn t work Trac 840).
|
|
Original ticket http projects.scipy.org numpy ticket 840 on 2008 07 04 by pv, assigned to cournape.
|
|
The third order=) parameter in ndarray.resize apparently doesn t work >>> import numpy as np >>> np.
|
|
version 1.1.0 >>> a = np.array [[0,1],[2,3]]) >>> a.resize 2,2), order=False) Traceback most recent call last) File <stdin> , line 1, in <module> SystemError NULL result without error in PyObject Call Also, what is the third argument supposed to do?
|
|
Its use is not explained in the docstring.
|
|
Reported originally here http sd 2116.dedibox.fr pydocweb doc numpy.ndarray.resize )","bug"
|
|
"6013","1439","Inconsistency in complex string representation introduced in NumPy 1.1 Trac 841).
|
|
Original ticket http projects.scipy.org numpy ticket 841 on 2008 07 04 by FrancescAlted, assigned to charris.
|
|
With !NumPy 1.0 series, the representation of complex types where identical for !NumPy and Python >>> numpy.
|
|
version 1.0.3 >>> repr numpy.complex 0)) The Python type 0j >>> repr numpy.complex128 0)) The NumPy type 0j Starting with !NumPy 1.1, it diverges from the Python standard >>> numpy.
|
|
version 1.2.0.dev5313 >>> repr numpy.complex 0)) 0j >>> repr numpy.complex128 0)) 0.0 0.0j) I think that, in order to ensure good interactivity of the !NumPy complex type with that of Python, the previous way to represent complex number should be reverted.","bug"
|
|
"6013","1440","Weird behavior of conjugate Trac 842).
|
|
Original ticket http projects.scipy.org numpy ticket 842 on 2008 07 04 by huard, assigned to unknown.
|
|
The string representation of complex arrays returned by conjugate is buggy.
|
|
np.conjugate np.array 8 7j)) <[83] 8.0 7.0j) > [84] np.array 8 7j) <[84] array 8 7j)) Note also that integers are converted to floats.","bug"
|
|
"6013","1441","Reference count error in PyArray CanCoerceScalar Trac 843).
|
|
Original ticket http projects.scipy.org numpy ticket 843 on 2008 07 04 by pv, assigned to charris.
|
|
Forwarded here from numpy mailing list Michael Abbott wrote The bug itself is very easy to see in Python debug mode adding arrays of differing shapes causes the reference count to increase on each operation.
|
|
For example array 1) 1 Leaks one ref count per call 1 array 1) Leaks 12 ref counts per call !)
|
|
It turns out that PyArray CanCoerceScalar is being called rather a lot of times some of the lower level numpy code is really not nice, which is a pity as there s some really clean code there too) ... and it has a ref count leak.
|
|
The patch below fixes this problem.
|
|
Attached patch was made against version 5331 from subversion.
|
|
The leak was introduced in revision 2575 in June 06, over two years ago I m a little surprised this hasn t been discovered already.","bug"
|
|
"6013","1442","failure with numpy.inner Trac 844).
|
|
Original ticket http projects.scipy.org numpy ticket 844 on 2008 07 05 by trac user gregorth, assigned to cournape.
|
|
After upgrading to NumPy 1.1.0 I installed numpy 1.1.0 win32 superpack pyhon2.5) I observed a fatal failure with the following code which uses numpy.inner import numpy F = numpy.zeros shape = 1,79), dtype = numpy.float64) this suceeds FtF = numpy.inner F,F.copy )) this fails FtF = numpy.inner F,F) The failure Exception code 0xc0000005) happens in dotblas.pyd.
|
|
I use Windows XP on a Intel Core2Duo system.","bug"
|
|
"6013","1443","deepcopy not identical to source Trac 845).
|
|
Original ticket http projects.scipy.org numpy ticket 845 on 2008 07 07 by trac user Goswin, assigned to unknown.
|
|
Instances created by calling deepcopy on ojects including numpy arrays do not always behave identically as their source object.
|
|
An example follows !
|
|
usr bin python import numpy as n Arrays from copy import deepcopy class Data object) def init self,rows,cols) self.background = n.ones rows 2,cols),) self.matrix = self.background[ rows, ] self.addon = self.background[rows rows 2, ] def show self) print \n\nState of the data print \n , self.background prototype = Data rows=4,cols=8) data = prototype WORKS OK DESIRED BEHAVIOR) data = deepcopy prototype) DIFFERENT OUTPUT, uncomment in order to compare data.show ) data.matrix[1,1] = 2 data.addon = 2 data.matrix = 5 data.show )","bug"
|
|
"6013","1444","Topical documentation as `numpy.doc` Trac 846).
|
|
Original ticket http projects.scipy.org numpy ticket 846 on 2008 07 08 by stefanv, assigned to stefanv.
|
|
Provide a new sub module, numpy.doc, which contains docstrings for topics such as indexing and broadcasting.
|
|
Please review the proposed patch at http codereview.appspot.com 2485.","bug"
|
|
"6013","1445","Warnings on printing pointers during compilation Trac 847).
|
|
Original ticket http projects.scipy.org numpy ticket 847 on 2008 07 08 by stefanv, assigned to cournape.
|
|
Since d is used to print pointers, during compilation we get many warnings like this numpy core src arrayobject.c 3165 warning format xe2x80x98 dxe2x80x99 expects type xe2x80x98intxe2x80x99, but argument 3 has type xe2x80x98long intxe2x80x99 Ideally, one would use z to print pointers instead, but according to http mail.python.org pipermail python dev 2006 January 059516.html some Microsoft compilers don t support that.","enhancement"
|
|
"6013","1446","Another ref count leak using PyArray DescrFromType Trac 848).
|
|
Original ticket http projects.scipy.org numpy ticket 848 on 2008 07 08 by trac user Araneidae, assigned to teoliphant.
|
|
The attached patch fixes one of the two reference count leaks that occur every time a new datatype is created eg, `float32 )` leaks two references this patch fixes one).","bug"
|
|
"6013","1447","Possible reference to deallocated object Trac 849).
|
|
Original ticket http projects.scipy.org numpy ticket 849 on 2008 07 09 by trac user Araneidae, assigned to charris.
|
|
In `dotblas matrixproduct` there is a code sequence which can potentially set a reference count on an object to zero before incrementing it again this could occur if `PyArray DescrFromType` succeeds but doesn t retain its `newtype` argument.
|
|
A little marginal, but easy to fix and other code elsewhere does this right).
|
|
Patch attached.","bug"
|
|
"6013","1448","Memory leak in ` strings richcompare` Trac 850).
|
|
Original ticket http projects.scipy.org numpy ticket 850 on 2008 07 09 by trac user Araneidae, assigned to charris.
|
|
When comparing `str` and `unicode` arrays the routine ` strings richcompare` will leak a reference count if the type conversion fails.
|
|
This is fixed by the attached patch.","bug"
|
|
"6013","1449","binary repr bug for some input types Trac 851).
|
|
Original ticket http projects.scipy.org numpy ticket 851 on 2008 07 13 by rgommers, assigned to cournape.
|
|
binary repr ) behaves differently for different types of ints uints In [210] binary repr 255) Out[210] 11111111 In [211] binary repr uint32 255)) Out[211] 11111111 In [212] binary repr uint16 255)) Out[212] 1 This is due to the use of hex ) checked that it is still present in trunk), which expects int or long int In [9] hex np.uint16 255)) Out[9] 0x1","bug"
|
|
"6013","1450","Avoid temporary files when writing NPZ files.
|
|
Trac 852).
|
|
Original ticket http projects.scipy.org numpy ticket 852 on 2008 07 15 by rkern, assigned to rkern.
|
|
Currently, when writing out NPZ zip files, we make temporary files in order to use the ZipFile.write ) API.
|
|
These file names can conflict with other processes writing similar files.
|
|
ZipFile.writestr ) could be used to avoid writing temporary files at all, at the cost of memory.
|
|
An alternative is to simply make a new temporary directory with tempfile.mkdtemp ) and always put files in there, instead.
|
|
Chuck, is this a fix you want to allow for 1.1.1?
|
|
Or should we just do it in 1.2?
|
|
Thoughts on which fix to do?","bug"
|
|
"6013","1451","numpy.sum is broken for anything with wrong a.sum ) signature Trac 853).
|
|
Original ticket http projects.scipy.org numpy ticket 853 on 2008 07 16 by pv, assigned to unknown.
|
|
Originally reported in Scipy Trac http scipy.org scipy scipy ticket 700) by alsuren.
|
|
The report is with numpy 1.0.4, but I believe this particular piece of code hasn t changed since.)
|
|
This is for python numpy 1.0.4 6ubuntu3, but neither are on the version list below.
|
|
== Test Case == from scipy import import numarray a = numarray.numarraycore.ones 5,5)) fail = sum a) usr lib python2.5 site packages numpy core fromnumeric.py in sum a, axis, dtype, out) 862 try 863 sum = a.sum 864 except AttributeError 865 return wrapit a, sum , axis, dtype, out) > 866 return sum axis, dtype, out) 867 868 == Solution == I personally don t like all this horrible flow control crap I wouldn t even attempt to look for a sum ) function, because the chances of it having the right signature is reasonably slim.
|
|
I would simply call wrapit ) straight off.
|
|
Something tells me that this may be expensive though.
|
|
The solution looks like this note that I have also added some assert statements.
|
|
If these are also bugs, tell me.)
|
|
def sum a, axis=None, dtype=None, out=None) ... if isinstance a, gentype) assert axis==None, axis is ignored for generator types.
|
|
res = sum a) if out is not None out[...] = res return out return res assert out == None, out is ignored for non generator types.
|
|
try return a.sum axis, dtype, out) except AttributeError Doesn t look like an ndarray except TypeError Doesn t quack like an ndarray return wrapit a, sum , axis, dtype, out) == Notes == Also, looking at the code, I m not quite sure what the point of out is My first guess was that it would avoid stupid memory allocation, but surely the temporary array res will have been allocated on the heap anyway?).
|
|
Work around Just convert to ndarray using asarray ) or array ) when is one or the other likely to be faster in the long run?)","bug"
|
|
"6013","1452","Non standard branch cuts Trac 854).
|
|
Original ticket http projects.scipy.org numpy ticket 854 on 2008 07 17 by pv, assigned to unknown.
|
|
At least numpy.arccosh has a very nonstandard choice of branch cuts.
|
|
I think there are also no tests present that verify that the numpy functions with branch cuts behave according to some standard eg.
|
|
it would be nice to follow the ISO C99 standard).
|
|
Tests should be written, and all functions with branch cuts should be audited.
|
|
Consider arccosh >>> import numpy as np >>> numpy.
|
|
version 1.2.0.dev5436.e45a7627a39d >>> np.arccosh 1e 9 0.1j) 0.099834078899207618 1.5707963277899337j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963257998594j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207618 1.5707963277899337j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963257998594j) It should be >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963277899337j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963257998594j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963277899337j) >>> np.arccosh 1e 9 0.1j) 0.099834078899207576 1.5707963257998594j) I see revisions r4881 and r4882 were made to fix this, but a braino slipped in as the new code is equivalent to the old one ) Here s a quick fix for arccosh Index umathmodule.c.src =================================================================== umathmodule.c.src revision 5439) umathmodule.c.src working copy) 825,15 825,16 c typ t; nc sum c x, nc 1 c , t); nc sqrt c t, t); nc diff c x, nc 1 c , r); nc sqrt c r, r); nc prod c t, r, r); nc sqrt c r, r); nc sum c x, r, r); nc log c r, r); return; return nc log nc sum x, nc sqrt nc prod nc sum x,nc 1), nc diff x,nc 1))))); nc prod nc sqrt nc sum x,nc 1)), nc sqrt nc diff x,nc 1))))); } This is a very naive fix; we should be able to avoid calling sqrt twice by choosing the branch of the sqrt manually according to real and imaginary parts of x.
|
|
Also, I didn t have time to write any tests yet, but I ll try to find time for writing and committing them.
|
|
Please, Don t close this ticket before there are test cases for all umath functions that have branch cuts.
|
|
Don t commit any fixes to behavior if there are no tests are in SVN.
|
|
References ISO C standard draft) is here http www.open std.org jtc1 sc22 wg14 www standards Abramowitz Stegun is here http www.math.sfu.ca cbm aands","bug"
|
|
"6013","1453","chararray mul error Trac 855).
|
|
Original ticket http projects.scipy.org numpy ticket 855 on 2008 07 18 by alanmcintyre, assigned to unknown.
|
|
Multiplication of a chararray by integers doesn t behave properly if any strings in the result should be longer than the size of a platform int or so it seems).
|
|
For example, on a 32 bit machine >>> A=np.array [ 1 , ab , xyz ]).view np.chararray) >>> A 1 chararray [ 1 , ab , xyz ], dtype= |S4 ) >>> A 2 chararray [ 11 , abab , xyzx ], dtype= |S4 ) >>> A 3 chararray [ 111 , abab , xyzx ], dtype= |S4 ) The dtype is always |S4 , and the longest strings are truncated to length 4.
|
|
This seems to be caused by the way ndarray.
|
|
new uses the itemsize argument when a numpy integer type is given for this argument, the itemsize of the argument rather than the integer value it contains) is used as the itemsize of the new array.
|
|
Attached is a patch that fixes this for chararray by forcing the itemsize argument given to ndarray.
|
|
new to be a Python int.","bug"
|
|
"6013","1454","chararray mod error Trac 856).
|
|
Original ticket http projects.scipy.org numpy ticket 856 on 2008 07 19 by alanmcintyre, assigned to unknown.
|
|
The operator doesn t seem to work as I d expect for chararray >>> A=np.array [[ .3f , d ],[ s , r ]]).view np.chararray) >>> A np.array [[1,2],[3,4]]) Traceback most recent call last) File <stdin> , line 1, in <module> File opt local lib python2.5 site packages numpy core defchararray.py , line 126, in mod newarr[ ] = res ValueError shape mismatch objects cannot be broadcast to a single shape Attached is a patch that fixes this behavior and some associated tests.","bug"
|
|
"6013","1455","dividing masked arrays with newaxis fails Trac 857).
|
|
Original ticket http projects.scipy.org numpy ticket 857 on 2008 07 22 by trac user fitz, assigned to pierregm.
|
|
Dividing masked arrays fails when the denominator is broadcast with newaxis r5497).
|
|
Example script attached.
|
|
test ma.py in <module> ) 6 y = o.arange 3,dtype=n.float) 7 print x y[n.newaxis, ] > 8 print x y[n.newaxis, ] 9 10 lib python numpy ma core.pyc in div self, other) 1882 def div self, other) 1883 Divide other into self, and return a new masked array.
|
|
> 1884 return divide self, other) 1885 1886 def truediv self, other) lib python numpy ma core.pyc in call self, a, b) 640 mb = mask or mb, t) 641 The following line controls the domain filling > 642 d2 = np.where t,self.filly,d2) 643 m = mask or ma, mb) 644 if not m.ndim) and m ValueError array dimensions must agree","bug"
|
|
"6013","1456","Removing some warnings from numpy.i GCC 4.2.3) Trac 858).
|
|
Original ticket http projects.scipy.org numpy ticket 858 on 2008 07 23 by trac user matthieu.brucher, assigned to cournape.
|
|
Some const correctness issues need to be fixed.
|
|
This patch suppress warnings that arise with GCC 4.2.3.
|
|
The warning are sound, as reference to contants are returned as char instead of const char .","enhancement"
|
|
"6013","1457","Indexing uniinitialized arrays in PyArrayIter Reset Trac 859).
|
|
Original ticket http projects.scipy.org numpy ticket 859 on 2008 07 24 by cournape, assigned to cournape.
|
|
Detected under valgrind, and looking at the code, it looks like there is a problem, but I don t know enough about that code to solve it quickly Tests compress2d==5148== ==5148== Conditional jump or move depends on uninitialised value s) ==5148== at 0x4636A27 PyArray MapIterReset arrayobject.c 10242) ==5148== by 0x466B511 array subscript arrayobject.c 2549) ==5148== by 0x466C07B array subscript nice arrayobject.c 3173) ==5148== by 0x80C7235 PyEval EvalFrameEx ceval.c 1193) ==5148== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==5148== by 0x80C92DD PyEval EvalFrameEx ceval.c 3669) ==5148== by 0x80C95C4 PyEval EvalFrameEx ceval.c 3659) ==5148== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==5148== by 0x81133BA function call funcobject.c 517) ==5148== by 0x805CB36 PyObject Call abstract.c 1861) ==5148== by 0x80C7CE3 PyEval EvalFrameEx ceval.c 3853) ==5148== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) The problem is inside the macro PyArrayIter GOTO some items of the argument destination are accessed without having been set previously when they are swapped with the copyswap function).","bug"
|
|
"6013","1458","Numpy as setuptools dependency Trac 860).
|
|
Original ticket http projects.scipy.org numpy ticket 860 on 2008 07 24 by trac user phsilva, assigned to dmcooke.
|
|
Numpy 1.1.0 works well with easy install if called from command line, like easy install numpy==1.1.0 BUT, if I declare numpy==1.1.0 as a dependency in my setup.py setup ) function, then I got this error actually I get the same errors with Numpy 1.0.4) Traceback most recent call last) File setup.py , line 116, in <module> platforms = chimera platform ) File usr lib python2.5 distutils core.py , line 151, in setup dist.run commands ) File usr lib python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr lib python2.5 site packages setuptools 0.6c8 py2.5.egg setuptools command develop.py , line 27, in run File usr lib python2.5 site packages setuptools 0.6c8 py2.5.egg setuptools command develop.py , line 102, in install for development File usr lib python2.5 site packages easy install.py , line 519, in process distribution File usr lib python2.5 site packages pkg resources.py , line 522, in resolve dist = best[req.key] = env.best match req, self, installer) File usr lib python2.5 site packages pkg resources.py , line 758, in best match return self.obtain req, installer) try and download install File usr lib python2.5 site packages pkg resources.py , line 770, in obtain return installer requirement) File usr lib python2.5 site packages easy install.py , line 446, in easy install File usr lib python2.5 site packages easy install.py , line 476, in install item File usr lib python2.5 site packages easy install.py , line 655, in install eggs File usr lib python2.5 site packages easy install.py , line 930, in build and install File usr lib python2.5 site packages easy install.py , line 919, in run setup File usr lib python2.5 site packages setuptools 0.6c8 py2.5.egg setuptools sandbox.py , line 27, in run setup File usr lib python2.5 site packages setuptools 0.6c8 py2.5.egg setuptools sandbox.py , line 63, in run File usr lib python2.5 site packages setuptools 0.6c8 py2.5.egg setuptools sandbox.py , line 29, in <lambda> File setup.py , line 96, in <module> numpy == 1.1.0 , File setup.py , line 89, in setup package Elixir >= 0.5.2 , File tmp easy install kOtWpP numpy 1.1.0 numpy distutils core.py , line 150, in setup File setup.py , line 56, in configuration setup File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 839, in add subpackage File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 822, in get subpackage File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 769, in get configuration from setup py File home henrique work chimera chimera setup.py , line 9, in configuration as published by the Free Software Foundation; either version 2 File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 839, in add subpackage File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 822, in get subpackage File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 769, in get configuration from setup py File home henrique work chimera chimera setup.py , line 223, in configuration File tmp easy install kOtWpP numpy 1.1.0 numpy distutils misc util.py , line 1039, in add include dirs AttributeError NoneType object has no attribute extend This seems similar to issue 679.
|
|
Numpy distutils was expecting something which setuptools distutils only in dependency mode) overrides and throw away.
|
|
Follows a patch hack against 1.1.0 that worked to me for 1.0.4 it s just the same hack on different chunck).
|
|
What the plan for distutils based installers like setuptools) when numscons became the default?
|
|
I m not going to integrate my package in distribution yet, so I m using easy install which seems to be the easiest solution available.
|
|
Any other idea about how to distribute Numpy?
|
|
I was thinking about try to create an egg, but then I ll need to have one for each different setup distro arch) that I would like to use, right?
|
|
Thanks a lot, and congrats for the great work with Numpy!","bug"
|
|
"6013","1459","numpy.matrix.var does not use ddof Trac 861).
|
|
Original ticket http projects.scipy.org numpy ticket 861 on 2008 07 24 by jpaalasm, assigned to unknown.
|
|
numpy.matrix.var does not pass the ddof parameter to N.ndarray.var.
|
|
See the patch for a fix.
|
|
Index numpy core defmatrix.py =================================================================== numpy core defmatrix.py revision 5516) numpy core defmatrix.py working copy) 445,7 445,7 is taken before squaring, so that the result is always real and nonnegative.
|
|
return N.ndarray.var self, axis, dtype, out).
|
|
align axis) return N.ndarray.var self, axis, dtype, out, ddof).
|
|
align axis) def prod self, axis=None, dtype=None, out=None) return N.ndarray.prod self, axis, dtype, out).
|
|
align axis)","bug"
|
|
"6013","1460","bad detection of AMD Phenom procesor ...in 64bit mode Trac 862).
|
|
Original ticket http projects.scipy.org numpy ticket 862 on 2008 07 24 by trac user mimi.vx, assigned to dmcooke.
|
|
Phenom processor is in 64bit mode detected as amd k6 2, detection routine in distutils cpuinfo ... Phenom processor will be detected as family 16 model 2 on line 130 return self.
|
|
is AMD ) and self.info[0][ model ] == 2 and self.info[0][ family ] != 16 and on line 146 return self.is AMD ) and self.info[0][ family ] == 15 or self.info[0][ family ] == 16 PS sorry for my english","bug"
|
|
"6013","1461","Regression in 658 ?
|
|
valgrind warning) Trac 863).
|
|
Original ticket http projects.scipy.org numpy ticket 863 on 2008 07 26 by cournape, assigned to unknown.
|
|
1.1.1rc2 Ticket 658==7735== ==7735== Conditional jump or move depends on uninitialised value s) ==7735== at 0x4639E7A PyArray UpdateFlags arrayobject.c 5172) ==7735== by 0x4658FB3 PyArray Newshape multiarraymodule.c 758) ==7735== by 0x465A097 array reshape arraymethods.c 87) ==7735== by 0x80C9992 PyEval EvalFrameEx ceval.c 3573) ==7735== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==7735== by 0x80C92DD PyEval EvalFrameEx ceval.c 3669) ==7735== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==7735== by 0x81133BA function call funcobject.c 517) ==7735== by 0x805CB36 PyObject Call abstract.c 1861) ==7735== by 0x80C7CE3 PyEval EvalFrameEx ceval.c 3853) ==7735== by 0x80CB0D6 PyEval EvalCodeEx ceval.c 2836) ==7735== by 0x811342F function call funcobject.c 517)","bug"
|
|
"6013","1462","MAError is formatted badly when displayed patch included) Trac 864).
|
|
Original ticket http projects.scipy.org numpy ticket 864 on 2008 07 27 by astraw, assigned to pierregm.
|
|
`python c import numpy as np; np.ma.array [1,2,3],mask=[1,2]) ` gives the following error python c import numpy as np; np.ma.array [1,2,3],mask=[1,2]) Traceback most recent call last) File <string> , line 1, in <module> File home astraw PY flydra lib python2.5 site packages numpy ma core.py , line 3021, in array fill value=fill value, ndmin=ndmin, shrink=shrink) File home astraw PY flydra lib python2.5 site packages numpy ma core.py , line 1231, in new raise MAError, msg nd, nm) numpy.ma.core.MAError M , a , s , k , , a , n , d , , d , a , t , a , , n , o , t , , c , o , m , p , a , t , i , b , l , e , , , d , a , t , a , , s , i , z , e , , i , s , , 3 , , , , m , a , s , k , , s , i , z , e , , i , s , , 2 , . )
|
|
With the following patch, I get a much nicer formatted error Index numpy ma core.py =================================================================== numpy ma core.py revision 5543) numpy ma core.py working copy) 86,15 86,7 Exceptions class MAError Exception) Class for MA related errors.
|
|
def init self, args=None) Creates an exception.
|
|
Exception.
|
|
init self, args) self.args = args def str self) Calculates the string representation.
|
|
return str self.args) repr = str pass Filling options Is there any reason MAError needs to implement it s own machinery?
|
|
I found r3514, which suggests somebody might be using the internals of MAError.)
|
|
Tested on the 1.1.x svn branch r5507.
|
|
With the patch applied, all tests pass.","bug"
|
|
"6013","1463","Recent C99 changes complex functions) break tests on windows Trac 865).
|
|
Original ticket http projects.scipy.org numpy ticket 865 on 2008 07 29 by cournape, assigned to unknown.
|
|
I have 12 failed tests related to recent test umath on win32 with mingw).
|
|
====================================================================== FAIL test umath.TestC99.test cacos <ufunc arccos >, 1.0, 1.
|
|
IND), 1.
|
|
IND, 1.
|
|
IND), invalid optional ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
QNAN, 1.
|
|
IND) , 1.
|
|
IND, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test cacos <ufunc arccos >, 1.
|
|
IND, 1.0), 1.
|
|
IND, 1.
|
|
IND), invalid optional ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
QNAN, 1.
|
|
IND) , 1.
|
|
IND, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test cacosh <ufunc arccosh >, 1.0, 1.
|
|
IND), 1.
|
|
IND, 1.
|
|
IND), invalid optional ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
QNAN, 1.
|
|
QNAN) , 1.
|
|
IND, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test cacosh <ufunc arccosh >, 1.
|
|
IND, 1.0), 1.
|
|
IND, 1.
|
|
IND), invalid optional ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
QNAN, 1.
|
|
QNAN) , 1.
|
|
IND, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 0.0, 0.0), 1.
|
|
INF, 3.1415926535897931), divide ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 399, in check assert raises FloatingPointError, func, point) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 347, in assert raises return nose.tools.assert raises args, kwargs) AssertionError FloatingPointError not raised ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 1.
|
|
INF, 1.
|
|
INF), 1.
|
|
INF, 2.3561944901923448), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
INF, 1.
|
|
IND) , 1.
|
|
INF, 2.3561944901923448) ) ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 1.
|
|
INF, 1.
|
|
INF), 1.
|
|
INF, 0.78539816339744828), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
INF, 1.
|
|
IND) , 1.
|
|
INF, 0.78539816339744828) ) ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 1.
|
|
INF, 1.
|
|
IND), 1.
|
|
INF, 1.
|
|
IND), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
IND, 1.
|
|
IND) , 1.
|
|
INF, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 1.
|
|
INF, 1.
|
|
IND), 1.
|
|
INF, 1.
|
|
IND), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
IND, 1.
|
|
IND) , 1.
|
|
INF, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test clog <ufunc log >, 1.
|
|
IND, 1.
|
|
INF), 1.
|
|
INF, 1.
|
|
IND), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
IND, 1.
|
|
IND) , 1.
|
|
INF, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test csqrt <ufunc sqrt >, 1.
|
|
INF, 1.
|
|
IND), 1.
|
|
INF, 1.
|
|
IND), ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
IND, 1.
|
|
IND) , 1.
|
|
INF, 1.
|
|
IND) ) ====================================================================== FAIL test umath.TestC99.test csqrt <ufunc sqrt >, 1.
|
|
IND, 1.0), 1.
|
|
IND, 1.
|
|
IND), invalid optional ) Traceback most recent call last) File c \python25\lib\site packages\nose 0.10.1 py2.5.egg\nose\case.py , line 203, in runTest self.test self.arg) File C \Python25\lib\site packages\numpy\core\tests\test umath.py , line 393, in check assert got == expected, got, expected) AssertionError 1.
|
|
QNAN, 1.
|
|
QNAN) , 1.
|
|
IND, 1.
|
|
IND) )","bug"
|
|
"6013","1464","ldexp broken Trac 866).
|
|
Original ticket http projects.scipy.org numpy ticket 866 on 2008 07 29 by huard, assigned to cournape.
|
|
> [3] np.ldexp .4, 5) TypeError Traceback most recent call last) storage huardda 10kmf90 python <ipython console> in <module> ) TypeError function not supported for these types, and can t coerce safely to supported types On Fedora 9, gcc 4.3.0, NumPy 1.2.0.dev5566","bug"
|
|
"6013","1465","Array generation fail Trac 867).
|
|
Original ticket http projects.scipy.org numpy ticket 867 on 2008 07 31 by trac user davyzhu, assigned to unknown.
|
|
>>> a = array [0x0000000E,0xFFFFFEB3],int32) Traceback most recent call last) File <interactive input> , line 1, in <module> ValueError setting an array element with a sequence.","bug"
|
|
"6013","1466","ipmt is not implemented in lib.financial) Trac 868).
|
|
Original ticket http projects.scipy.org numpy ticket 868 on 2008 07 31 by pv, assigned to unknown.
|
|
The ipmt function is not implemented in numpy r5584, there s a stub instead.
|
|
def ipmt rate, per, nper, pv, fv=0.0, when= end ) total = pmt rate, nper, pv, fv, when) Now, compute the nth step in the amortization raise NotImplementedError As it appears in the main namespace, it probably should be implemented.
|
|
Also update http sd 2116.dedibox.fr pydocweb doc numpy.lib.financial.ipmt when implemented.","bug"
|
|
"6013","1467","defer import tempfile Trac 869).
|
|
Original ticket http projects.scipy.org numpy ticket 869 on 2008 07 31 by trac user dalke, assigned to unknown.
|
|
Currently as of CVS for July 31), the import for tempfile takes 7 of the startup costs.
|
|
This module is only used by one function.
|
|
I deferred the import cost with this patch.","enhancement"
|
|
"6013","1468","np.asarray does not create new array view when different but compatible type code is passed in the dtype argument Trac 870).
|
|
Original ticket http projects.scipy.org numpy ticket 870 on 2008 08 01 by trac user damian.eads, assigned to stefanv.
|
|
Examining `ndarray` type codes is common in C extensions before passing array contents to functions expecting `int ` or `long `.
|
|
The type codes are ideal ensure the contents can be safely casted to `int ` or `long ` on the host machine.
|
|
In some cases, the `int` and `long` types are of the same size but their type codes are different.
|
|
`np.asarray ..., dtype= i )` should be used to create a new array view with the type code the C type code expected by the C extension; however, it does not appear to work when using it for this purpose.
|
|
Problem I construct X as a C long array and then I cast it to a C int array Y using `np.asarray`) but the type code does not change.
|
|
However, when constructing the array from scratch as a C int using `np.zeros`), I get the right type code ie 5).
|
|
Expectation of Behavior I assumed that when X gets casted to a C int, no copying should occur but a new array view should be constructed with the C int type code.
|
|
Workaround Casting from a C long type code 7) to a double to a C int returns an array with the right type code but an ugly double copy occurs.
|
|
Recreation Construct X as a C long.
|
|
In [16] X=numpy.zeros 10,10),dtype= l ) Now cast X to a C int.
|
|
In [17] Y=numpy.asarray X, dtype= i ) Check X and Y s data type; they are the same.
|
|
In [18] X.dtype Out[18] dtype int32 ) In [19] Y.dtype Out[19] dtype int32 ) Their type codes are the same.
|
|
In [20] X.dtype.num Out[20] 7 In [21] Y.dtype.num Out[21] 7 Constructing with dtype= i , gives the right type code.
|
|
In [22] Z=numpy.zeros 10,10),dtype= i ) In [23] Z.dtype Out[23] dtype int32 ) In [24] Z.dtype.num Out[24] 5 Stefan van der Walt observes that the `dtype` objects are compared for equality but not the type codes.
|
|
This explains why the `asarray` function does not create a new view.
|
|
In [4] np.dtype i ) == np.dtype l ) Out[4] True In [5] np.dtype i ).num == np.dtype l ).num Out[5] False Fix Check the type codes for equality, not the `dtype` objects, to decide whether to create a new array view in `np.asarray`.","bug"
|
|
"6013","1469","f2py and or distutils broken unpack tuple of wrong size) Trac 871).
|
|
Original ticket http projects.scipy.org numpy ticket 871 on 2008 08 04 by trac user dhendriks, assigned to unknown.
|
|
Using Python 2.3 on r5602 of the 1.1.x branch, after solving the problems of ticket 1419, I now try to use f2py.
|
|
However, I get the following running build running scons customize UnixCCompiler Found executable usr bin gcc customize GnuFCompiler Found executable usr bin g77 gnu no Fortran 90 compiler found Traceback most recent call last) File XXX bin f2py , line 24, in ?
|
|
main ) File XXX lib python2.3 site packages numpy f2py f2py2e.py , line 560, in main run compile ) File XXX lib python2.3 site packages numpy f2py f2py2e.py , line 547, in run compile setup ext modules = [ext]) File XXX lib python2.3 site packages numpy distutils core.py , line 184, in setup return old setup new attr) File usr lib python2.3 distutils core.py , line 149, in setup dist.run commands ) File usr lib python2.3 distutils dist.py , line 907, in run commands self.run command cmd) File usr lib python2.3 distutils dist.py , line 927, in run command cmd obj.run ) File XXX lib python2.3 site packages numpy distutils command build.py , line 38, in run self.run command scons ) File usr lib python2.3 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.3 distutils dist.py , line 926, in run command cmd obj.ensure finalized ) File usr lib python2.3 distutils cmd.py , line 117, in ensure finalized self.finalize options ) File XXX lib python2.3 site packages numpy distutils command scons.py , line 310, in finalize options force = self.force) File XXX lib python2.3 site packages numpy distutils fcompiler init .py , line 809, in new fcompiler c compiler=c compiler) File XXX lib python2.3 site packages numpy distutils fcompiler init .py , line 791, in get default fcompiler c compiler=c compiler) File XXX lib python2.3 site packages numpy distutils fcompiler init .py , line 740, in find existing fcompiler c.customize dist) File XXX lib python2.3 site packages numpy distutils fcompiler init .py , line 491, in customize get flags debug , dflags) File XXX lib python2.3 site packages numpy distutils fcompiler init .py , line 478, in get flags flags.extend getattr self.flag vars, tag)) File XXX lib python2.3 site packages numpy distutils environment.py , line 37, in getattr return self.
|
|
get var name, conf desc) File XXX lib python2.3 site packages numpy distutils environment.py , line 50, in get var hook, envvar, confvar, convert = conf desc ValueError unpack tuple of wrong size","bug"
|
|
"6013","1470","numpy.lib.setmember1d contains unneeded code and does work with duplicated elements in arguments Trac 872).
|
|
Original ticket http projects.scipy.org numpy ticket 872 on 2008 08 04 by trac user novak, assigned to unknown.
|
|
I believe that setmember1d contains unnecessary code.
|
|
Also, the docstring of setmember1d asserts that the arguments must not contain duplicated values, but I believe that this is incorrect.
|
|
The function works as expected even with duplication.
|
|
Below is my post to the numpy discussion mailing list to this effect.
|
|
I have two arrays of integers, and would like to know where they have elements in common, not just which elements are in common.
|
|
This is because the entries in the integer array are aligned with other arrays.
|
|
This seems very close to what member1d advertises as its function.
|
|
However, member1d says that it expects arrays with only unique elements.
|
|
First of all, my desired operation is well posed I d like f ar1, ar2) to return something in the shape of ar1 with True if the value at that position appears anywhere in ar2 regardless of duplication) and False otherwise.
|
|
So I looked at the code and have two questions 1) What is this code trying to achieve?
|
|
aux = perm[ii 1] perm[ii 1] = perm[ii] perm[ii] = aux Here perm is the stable argsort of the two concatenated arguments perm = concatenate ar1, ar2)).argsort kind= mergesort ).
|
|
arr is the array of combined inputs in sorted order arr = concatenate ar1, ar2))[perm] and ii is a list of indices into arr where the value of arr is equal to the next value in the array arr[ii] == arr[ii 1]) and arr[ii] came from the second input ar2).
|
|
Now, this last bit looking for elements of arr that are equal and both came from the second array) is clearly trying to deal with duplication, which is why I m interested...
|
|
So, the code snippet is trying to swap perm[ii 1] with perm[ii], but I don t see why.
|
|
Furthermore, there are funny results if a value is duplicated three times, not just twice perm is no longer a permutation vector.
|
|
Eg, member1d [1], [2,2,2]) results perm=[0,1,2,3] and ii=[1,2] before the above snippet, and the above snippet makes perm into [0,2,3,2] I ve commented those three lines, and I ve never seen any changes to the output of member1d.
|
|
The new value of perm is used to compute the expression perm.argsort kind= mergesort )[ len ar1 )], but the changes to that expression as a result of the above three lines are always at the high end of the array, which is sliced off by the last [ len ar1)].
|
|
Finally, my second question is 2) Does anyone have a test case where member1d fails as a result of duplicates in the input?
|
|
So far I haven t found any, with the above lines commented or not.
|
|
Upon reflection and review of the changelog, another theory occurs to me member1d did not originally use a stable sort.
|
|
What I ve written above for interpretation of the value ii indicates duplication within ar2) is true for a stable sort, but for an unstable sort the same condition has the interpretation that ii holds the values where the sorting algorithm swapped the order of equal values unstably.
|
|
Then the code snippet in question 1) looks like an attempt to swap those values in the permutation array to make the sort stable again.
|
|
The attempt would fail if there was duplication in either array.
|
|
So, I would propose deleting those three lines since they seem to be a non functional relic) and declaring in the docstring that member1d doesn t require unique elements.
|
|
Also, if this is correct, then the function simplifies considerably since several values don t need to be computed anymore def setmember1d ar1, ar2 ) ar = nm.concatenate ar1, ar2 ) ) perm = ar.argsort kind= mergesort ) aux = ar[perm] flag = nm.concatenate aux[1 ] == aux[ 1], [False] ) ) indx = perm.argsort kind= mergesort )[ len ar1 )] return flag[indx]","enhancement"
|
|
"6013","1471","how to test if a file is a filename or a file object?
|
|
Trac 873).
|
|
Original ticket http projects.scipy.org numpy ticket 873 on 2008 08 05 by trac user dalke, assigned to unknown.
|
|
This is a minor point I came across when looking through the code.
|
|
There are a few places where numpy takes either a filename or a file handle.
|
|
It s inconsistent about how to test for that case.
|
|
Some of the methods are test for string via isinstance obj, basestring) test for string by checking if obj is value test for presence of the method name read or readline Which is to be preferred?
|
|
Here are some direct examples numpy ma core.py if not hasattr F, readline ) F = open F, w ) numpy lib io.py if isinstance file, basestring) fid = file file, rb ) else fid = file numpy lib io.py def string like obj) try obj except TypeError, ValueError) return 0 return 1 def loadtxt ...) ... if string like fname) if fname.endswith .gz ) import gzip fh = gzip.open fname) else fh = file fname) elif hasattr fname, seek ) fh = fname else raise ValueError fname must be a string or file handle )","enhancement"
|
|
"6013","1472","quicker import numpy time Trac 874).
|
|
Original ticket http projects.scipy.org numpy ticket 874 on 2008 08 05 by trac user dalke, assigned to cournape.
|
|
Here s a patch containing many small changes which together on my machine reduce the import numpy overhead costs from 0.15 seconds to 0.09 seconds.
|
|
All tests continue to pass with my changes.
|
|
Here are the details of what I did 1) remove import ctypeslib in numpy init .py Looking in the SVN log I see r3027 | oliphant | 2006 08 15 11 53 49 0200 Tue, 15 Aug 2006) | 1 line import ctypeslib on numpy load and change name from ctypes load library to load library However, I don t see why this module is required.
|
|
People using ctypes almost certainly know that they should be importing the modules directly.
|
|
This saves 0.02 seconds.
|
|
2) removed import gzip bzip from top level of datasource.py This saves 0.06 seconds.
|
|
3) defer import of urlparse in datasource.py Saves 0.03 seconds 4) Removed the import doc in numpy init .py because it seems to only import placeholder documentation.
|
|
Why is it there?
|
|
Saved 0.035 seconds.
|
|
5) replaced test = numpy.testing.Tester ).test and bench bound method definitions in various init .py files with equivalent functions.
|
|
This defers importing numpy.testing until needed and saves about 0.01 seconds.
|
|
6) remove unused glob in import tools 7) remove import string in numerictype.py; replace with hard coded transtables.
|
|
Saves a small 0.007 seconds 8) In the memmap module, defer importing Python s mmap module until needed.
|
|
This saves the even smaller 0.002 seconds.
|
|
9) In the format module, defer importing struct until needed.
|
|
This saves 0.005 seconds.
|
|
I know especially the last few changes are very small, but they add up.
|
|
There s at least one other, more contentious, change I can make, which I ll do in another ticket.","enhancement"
|
|
"6013","1473","numpy.lib.io.save fails with NameError Trac 875).
|
|
Original ticket http projects.scipy.org numpy ticket 875 on 2008 08 05 by trac user hvontres, assigned to unknown.
|
|
While Testing the 1.1.1 relase on Python 2.3.5 for Windows I encountered the following problem from numpy import test=ones 10,10)) save test ,test) Traceback most recent call last) File <pyshell 21> , line 1, in toplevel save test ,test) File C \Python23\Lib\site packages\numpy\lib\io.py , line 159, in save format.write array fid, arr) File C \Python23\Lib\site packages\numpy\lib\format.py , line 285, in write array write array header 1 0 fp, header data from array 1 0 array)) File C \Python23\Lib\site packages\numpy\lib\format.py , line 172, in write array header 1 0 for key, value in sorted d.items )) NameError global name sorted is not defined Any Ideas?","bug"
|
|
"6013","1474","Convert isposinf and isneginf from ufunclike to proper ufunc objects Trac 876).
|
|
Original ticket http projects.scipy.org numpy ticket 876 on 2008 08 06 by bsouthey, assigned to unknown.
|
|
Both isposinf ) and isneginf ) return only return a boolean array.
|
|
In contrast, both isinf ) and isnan ) return a numpy boolean for a scalar input and a numpy boolean array for array like input.
|
|
import numpy as np np.
|
|
version returns 1.1.0 type np.isinf np.inf)) returns <type numpy.bool > type np.isposinf np.inf)) returns <type numpy.ndarray > type np.isneginf np.inf)) returns <type numpy.ndarray >","bug"
|
|
"6013","1475","Segfault when using recarray O entries Trac 877).
|
|
Original ticket http projects.scipy.org numpy ticket 877 on 2008 08 06 by mdboom, assigned to unknown.
|
|
This bug report began on the mailing list here http projects.scipy.org pipermail numpy discussion 2008 July 035998.html) I ve run into a segfault that occurs in the array destructor with arrays containing object references type O ) that have both names and titles.
|
|
When a field contains both and name and a title, the fields dictionary contains two entries for that field.
|
|
This means that the array item destructor which iterates through the fields dictionary) will decref the pointed to object twice.
|
|
If the first decref causes the object to be deleted, the second decref has the potential to segfault.
|
|
It seems the simplest patch is to set the object pointer to NULL after decref ing, so the second decref will do nothing.
|
|
However, perhaps there is a way to avoid decref ing twice in the first place.
|
|
I ve attached a script that exercises the segfault, a gdb backtrace, and a patch.
|
|
You may need to adjust the number of rows until it is high enough to create a segfault on your system.
|
|
This is on RHEL4 Python 2.5.2 Numpy SVN r5497","bug"
|
|
"6013","1476","np.fix does not work properly for negative integers Trac 878).
|
|
Original ticket http projects.scipy.org numpy ticket 878 on 2008 08 06 by trac user badland, assigned to cournape.
|
|
np.fix is supposed to round floats towards the nearest integer closer to zero.
|
|
In particular, it should leave integers unchanged.
|
|
However, at least in version 1.1.0, np.fix 2) returns 1.
|
|
In the python function np.fix, a negative argument x is handled as np.floor x) 1 This is correct if x is not an integer.
|
|
An easy fix pun intended) would be to use np.floor x) in the negative case.
|
|
BTW The web interface did not allow me to put version = 1.1.0 Bjxc3xb8rn xc3x85dlandsvik","bug"
|
|
"6013","1477","Finite integer scalar arguments to nan to num gives incorrect result Trac 879).
|
|
Original ticket http projects.scipy.org numpy ticket 879 on 2008 08 06 by bsouthey, assigned to unknown.
|
|
If an finite integer number is passed to nan to num then an array is returned.
|
|
But if a float or nonfinite number nan or inf) is passed, then a NumPy float64 is returned.
|
|
import numpy as np type np.nan to num np.inf)) returns <type numpy.float64 > type np.nan to num 1.0)) returns <type numpy.float64 > type np.nan to num 1)) returns <type numpy.ndarray >","bug"
|
|
"6013","1478","An error is given if a integer list is passed to nan to num Trac 880).
|
|
Original ticket http projects.scipy.org numpy ticket 880 on 2008 08 06 by bsouthey, assigned to unknown.
|
|
When an integer list is passed to nan to num then an error occurs.
|
|
This does not occur if the list contains floats or a numpy integer array is used.
|
|
np.nan to num [1.0,3]) returns array [ 1., 3.])
|
|
n=np.array [1,3]) np.nan to num n) returns array [1, 3]) np.nan to num [1,3]) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib64 python2.5 site packages numpy lib type check.py , line 135, in nan to num maxf, minf = getmaxmin y.dtype.type) File usr lib64 python2.5 site packages numpy lib type check.py , line 103, in getmaxmin f = getlimits.finfo t) File usr lib64 python2.5 site packages numpy lib getlimits.py , line 46, in new raise ValueError, data type r not inexact dtype) ValueError data type <type numpy.int64 > not inexact","bug"
|
|
"6013","1479","Segmentation fault with numpy.object0 Trac 881).
|
|
Original ticket http projects.scipy.org numpy ticket 881 on 2008 08 07 by trac user mgogoulos ..., assigned to cournape.
|
|
numpy.object0 and numpy.object result in segmentation fault, see the following gdb python ... gdb) run c import numpy; obj = numpy.object0 ) Starting program usr bin python c import numpy; obj = numpy.object0 ) ... [New Thread 0xb7d498c0 LWP 17024)] Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 0xb7d498c0 LWP 17024)] object arrtype new type=0xb7c7d1c0, args=0xb7d0902c, kwds=0x0) at numpy core src scalartypes.inc.src 1977 1977 Py INCREF PyObject )dest)); Tested on versions 1.0.4 and 1.1.1 with scipy 0.6.0, on Ubuntu hardy python 2.5.2) Best regards, Markos Gogoulos","bug"
|
|
"6013","1480","Segmentation fault in array.choose with empty list as argument Trac 882).
|
|
Original ticket http projects.scipy.org numpy ticket 882 on 2008 08 09 by trac user nvj, assigned to cournape.
|
|
array.choose fails when given an empty list as argument.
|
|
This code reliably provokes a segmentation fault [[BR]] [[BR]] from numpy import numarray a = numarray.array 1) a.choose []) [[BR]][[BR]] gdb on core dump gives [[BR]] Core was generated by `python test.py .
|
|
[[BR]] Program terminated with signal 11, Segmentation fault.
|
|
[[BR]] [New process 10356][[BR]] 0 PyArray ConvertToCommonType op=0xb6d95e0c, retn=0xbfa2bb88)[[BR]] at numpy core src multiarraymodule.c 2231 [[BR]][[BR]] I m running debian sid packages [[BR]] python dbg 2.5.2 2)[[BR]] python numpy dbg 1 1.1.0 3)[[BR]] Running in vmware under XP) [[BR]][[BR]] Hope this helps, I haven t checked the latest version, too much trouble and I m pretty new to this.","bug"
|
|
"6013","1481","fromstring fromfile behave badly with malformed data Trac 883).
|
|
Original ticket http projects.scipy.org numpy ticket 883 on 2008 08 09 by pv, assigned to unknown.
|
|
See the discussion at http permalink.gmane.org gmane.comp.python.numeric.general 24115 In summary 1.
|
|
Empty string is understood as zero except at the end of a string) >>> np.fromstring xxxxx , sep= x ) array [ 0., 0., 0., 0., 0.])
|
|
Missing values are treated as zeros.
|
|
This is probably accidental, it s more likely that the memory allocated just happens to be filled with zeros... 2.
|
|
If count is given and not enough entries were read, the rest is padded with uninitialized data >>> np.fromstring 23 , dtype=np.float, sep= x , count=3) array [ 2.30000000e 001, 8.19397657e 269, 1.94625366e 312]) 3.
|
|
Fromfile behaves differently from fromstring Assume foo.dat contains string 23xxx ) >>> np.fromfile foo.dat , dtype=np.float, sep= x ) array [ 2.30000000e 001, 8.36783574e 269, 2.12885402e 313]) >>> np.fromfile foo.dat , dtype=np.float, count=5, sep= x ) 5 items requested but only 3 read array [ 2.30000000e 001, 6.04468244e 269, 1.54891786e 039]) Again, the last value is ignored, but the former ones are not treated as zero; instead, the array is filled with uninitialized data for the missing entries.
|
|
Note also that neither warning or error is raised; instead, a message is printed to stdout.
|
|
Question what is the correct behavior A) Raise ValueError on malformed data.
|
|
B) Raise ValueError on malformed data if count is given, otherwise truncate to valid part.
|
|
C) Truncate to valid part of data.","bug"
|
|
"6013","1482","fromfile and fromstring locale dependency Trac 884).
|
|
Original ticket http projects.scipy.org numpy ticket 884 on 2008 08 09 by pv, assigned to cournape.
|
|
Behavior of fromfile depends on the C locale used, fromstring does not depend on the locale assume foo.dat contains the string 1,234 .
|
|
Note that in Finnish locale, the decimal separator is comma, not period.)
|
|
>>> import numpy as np >>> np.fromstring 1,234 , sep= x ) array [ 1.])
|
|
>>> np.fromfile foo.dat , sep= x ) array [ 1.])
|
|
>>> import locale >>> locale.setlocale locale.LC NUMERIC, fi FI ) >>> g 1.2345 python doesn t care about the C locale 1.2345 >>> np.fromstring 1,234 , sep= x ) array [ 1.])
|
|
>>> np.fromfile foo.dat , sep= x ) array [ 1.234]) In summary fromfile respects the C locale because it uses fscanf fromstring doesn t respect the locale as it uses PyOS strtol fromstring still breaks with malformed data; some part of its code might rely on the C locale The main question is whether we want to rely on C locale or not in I O. Python itself appears to try to avoid relying on C locale.
|
|
Personally, I think also we should not rely on C locale.
|
|
The simplest way to fix this probably is to wrap the fromfile fromstring in pairs of setlocale calls that set and revert the locale.
|
|
Similar audit might also be needed for all parts of the numpy code that use the C stdio library services.","bug"
|
|
"6013","1483","sort does not accept axis=None Trac 885).
|
|
Original ticket http projects.scipy.org numpy ticket 885 on 2008 08 10 by trac user ctw, assigned to unknown.
|
|
This bug refers to ticket [http www.scipy.org scipy numpy ticket 391 989] and the following mailing list message [http projects.scipy.org pipermail numpy discussion 2008 February 031384.html] Currently the sort docstring states that axis=None is valid, but it raises a TypeError exception with the message an integer is required .
|
|
Either axis=None functionality should be implemented or the docstring should be changed to reflect the fact that None is not a valid value for axis.","bug"
|
|
"6013","1484","inf nan handling in umath is undefined Trac 886).
|
|
Original ticket http projects.scipy.org numpy ticket 886 on 2008 08 13 by pv, assigned to unknown.
|
|
The complex valued ufuncs in numpy.core.umath do not handle inf, nan, and floating point signals in a well defined way.
|
|
The results are platform dependent.
|
|
See eg.
|
|
1463 and discussion following [1]).
|
|
It might be useful to follow the C99 standard [2] here, possibly looking up how things were recently improved in Python s cmath module.
|
|
Some test cases for C99 behavior were backed out in r5644; when this issue is addressed, it would make sense to resurrect them.
|
|
[1] http article.gmane.org gmane.comp.python.numeric.general 23497 [2] http www.open std.org jtc1 sc22 wg14 www standards","bug"
|
|
"6013","1485","Generalized Universal Functions Trac 887).
|
|
Original ticket http projects.scipy.org numpy ticket 887 on 2008 08 13 by trac user fuwenjie, assigned to unknown.
|
|
= Generalized Universal Functions = There is a general need for looping over not only functions on scalars but also over functions on vectors or arrays), as explained on http scipy.org scipy numpy wiki GeneralLoopingFunctions.
|
|
We propose to realize this concept by generalizing the universal functions ufuncs), and provide a C implementation that adds 500 lines to the numpy code base.
|
|
In current specialized) ufuncs, the elementary function is limited to element by element operations, whereas the generalized version supports sub array by sub array operations.
|
|
The Perl vector library PDL provides a similar functionality and its terms are re used in the following.
|
|
Each generalized ufunc has information associated with it that states what the core dimensionality of the inputs is, as well as the corresponding dimensionality of the outputs the element wise ufuncs have zero core dimensions).
|
|
The list of the core dimensions for all arguments is called the signature of a ufunc.
|
|
For example, the ufunc numpy.add has signature ` ), ) > ) ` defining two scalar inputs and one scalar output.
|
|
Another example is see the GeneralLoopingFunctions page) the function `inner1d a,b)` with a signature of ` i), i) > ) `.
|
|
This applies the inner product along the last axis of each input, but keeps the remaining indices intact.
|
|
For example, where `a` is of shape ` 3,5,N)` and `b` is of shape ` 5,N)`, this will return an output of shape ` 3,5)`.
|
|
The underlying elementary function is called 3 5 times.
|
|
In the signature, we specify one core dimension ` i) ` for each input and zero core dimensions ` ) ` for the output, since it takes two 1 d arrays and returns a scalar.
|
|
By using the same name ` i `, we specify that the two corresponding dimensions should be of the same size or one of them is of size 1 and will be broadcasted).
|
|
The dimensions beyond the core dimensions are called loop dimensions.
|
|
In the above example, this corresponds to ` 3,5)`.
|
|
The usual numpy broadcasting rules apply, where the signature determines how the dimensions of each input output object are split into core and loop dimensions 1.
|
|
While an input array has a smaller dimensionality than the corresponding number of core dimensions, 1 s are pre pended to its shape.
|
|
1.
|
|
The core dimensions are removed from all inputs and the remaining dimensions are broadcasted; defining the loop dimensions.
|
|
1.
|
|
The output is given by the loop dimensions plus the output core dimensions.
|
|
= Definitions = Elementary Function Each ufunc consists of an elementary function that performs the most basic operation on the smallest portion of array arguments e.g.
|
|
adding two numbers is the most basic operation in adding two arrays).
|
|
The ufunc applies the elementary function multiple times on different parts of the arrays.
|
|
The input output of elementary functions can be vectors; e.g., the elementary function of inner1d takes two vectors as input.
|
|
Signature A signature is a string describing the input output dimensions of the elementary function of a ufunc.
|
|
See section below for more details.
|
|
Core Dimension The dimensionality of each input output of an elementary function is defined by its core dimensions zero core dimensions correspond to a scalar input output).
|
|
The core dimensions are mapped to the last dimensions of the input output arrays.
|
|
Dimension Name A dimension name represents a core dimension in the signature.
|
|
Different dimensions may share a name, indicating that they are of the same size or are broadcastable).
|
|
Dimension Index A dimension index is an integer representing a dimension name.
|
|
It enumerates the dimension names according to the order of the first occurrence of each name in the signature.
|
|
= Details of Signature = The signature defines core dimensionality of input and output variables, and thereby also defines the contraction of the dimensions.
|
|
The signature is represented by a string of the following format Core dimensions of each input or output array are represented by a list of dimension names in parentheses, ` i 1,...,i N) `; a scalar input output is denoted by ` ) `.
|
|
Instead of ` i 1 `, ` i 2 `, etc, one can use any valid Python variable name.
|
|
Dimension lists for different arguments are separated by ` , `.
|
|
Input output arguments are separated by ` > `.
|
|
If one uses the same dimension name in multiple locations, this enforces the same size or broadcastable size) of the corresponding dimensions.
|
|
The formal syntax of signatures is as follows.
|
|
<Signature> = <Input arguments> > <Output arguments> <Input arguments> = <Argument list> <Output arguments> = <Argument list> <Argument list> = nil | <Argument> | <Argument> , <Argument list> <Argument> = <Core dimension list> ) <Core dimension list> = nil | <Dimension name> | <Dimension name> , <Core dimension list> <Dimension name> = valid Python variable name Notes 1.
|
|
All quotes are for clarity.
|
|
1.
|
|
Core dimensions that share the same name must be broadcastable, as the two ` i `s in our example above.
|
|
Each dimension name typically corresponding to one level of looping in the elementary function s implementation.
|
|
1.
|
|
White spaces are ignored.
|
|
Here are some examples of signatures.
|
|
|| add || ` ), ) > ) ` || || || inner1d || ` i), i) > ) ` || || || sum1d || ` i) > ) ` || || || dot2d || ` m,n), n,p) > m,p) ` || matrix multiplication) || || outer inner || ` i,t), j,t) > i,j) ` || inner over the last dimension, outer over the second to last, and loop broadcast over the rest.)
|
|
|| = C API for implementing Elementary Functions = The current interface remains unchanged, and `PyUFunc FromFuncAndData` can still be used to implement specialized) ufuncs, consisting of scalar elementary functions.
|
|
One can use `PyUFunc FromFuncAndDataAndSignature` to declare a more general ufunc.
|
|
The argument list is the same as `PyUFunc FromFuncAndData`, with an additional argument specifying the signature as C string.
|
|
Furthermore, the callback function is of the same type as before, `void foo) char args, intp dimensions, intp steps, void func)`.
|
|
When invoked, `args` is a list of length `nargs` containing the data of all input output arguments.
|
|
For a scalar elementary function, `steps` is also of length `nargs`, denoting the strides used for the arguments.
|
|
`dimensions` is a pointer to a single integer defining the size of the axis to be looped over.
|
|
For a non trivial signature, `dimensions` will also contain the sizes of the core dimensions as well, starting at the second entry.
|
|
Only one size is provided for each unique dimension name and the sizes are given according to the first occurrence of a dimension name in the signature.
|
|
The first `nargs` elements of `steps` remain the same as for scalar ufuncs.
|
|
The following elements contain the strides of all core dimensions for all arguments in order.
|
|
For example, consider a ufunc with signature ` i,j), i) > ) `.
|
|
In this case, `args` will contain three pointers to the data of the input output arrays `a`, `b`, `c`.
|
|
Furthermore, `dimensions` will be `[N, I, J]` to define the size of `N` of the loop and the sizes `I` and `J` for the core dimensions `i` and `j`.
|
|
Finally, `steps` will be `[a N, b N, c N, a i, a j, b i]`, containing all necessary strides.
|
|
= License = The attached code and the above documentation was written by Wenjie Fu <fuw deshaw.com> and Hans Andreas Engel <engelh deshaw.com>, and is provided under the numpy license Copyright c) 2005, NumPy Developers All rights reserved.
|
|
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
|
|
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and or other materials provided with the distribution.
|
|
Neither the name of the NumPy Developers nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission.
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
|
|
IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.","enhancement"
|
|
"6013","1486","Incorrect value returned by numpy.array.max Trac 888).
|
|
Original ticket http projects.scipy.org numpy ticket 888 on 2008 08 13 by trac user lefranck, assigned to unknown.
|
|
Hello, I ve been trying to solve a weird problem with some of my data, and as far as I can tell, the problem seems to come from the numpy.array.max function.
|
|
With my particular data block 640.000 float32 values between 0.0 and 1.0, pickle file attached), data.max ) returns 0.545826792717, while the actual max value is 1.0.
|
|
The built in function max data) returns 1.0, and if I loop through the data myself I find 1.0.
|
|
If I call data.sort ) before hand, data.max ) then correctly returns 1.0.
|
|
I m using python 2.5.2 and numpy 1.0.4 on a standard Ubuntu 8.04 system with a core 2 duo CPU.
|
|
The problem persists with numpy 1.2.0.dev5495.
|
|
For instance In [11] H.max ) Out[11] 0.545826792717 In [12] H.argmax ) Out[12] 24400 In [13] H[24400] Out[13] 1.0 Also In [41] H.max ) Out[41] 0.545826792717 In [42] H.sort ) In [43] H.max ) Out[43] 1.0 Cheers, Franck","bug"
|
|
"6013","1487","Incorrect value returned by numpy.array.max Trac 889).
|
|
Original ticket http projects.scipy.org numpy ticket 889 on 2008 08 13 by trac user lefranck, assigned to unknown.
|
|
Hello, I ve been trying to solve a weird problem with some of my data, and as far as I can tell, the problem seems to come from the numpy.array.max function.
|
|
With my particular data block 640.000 float32 values between 0.0 and 1.0, pickle file attached), data.max ) returns 0.545826792717, while the actual max value is 1.0.
|
|
The built in function max data) returns 1.0, and if I loop through the data myself I find 1.0.
|
|
If I call data.sort ) before hand, data.max ) then correctly returns 1.0.
|
|
I m using python 2.5.2 and numpy 1.0.4 on a standard Ubuntu 8.04 system with a core 2 duo CPU.
|
|
The problem persists with numpy 1.2.0.dev5495.
|
|
A compressed version of the pickle file can be found here [http le franck.com datablock.tar.bz2] For instance In [11] H.max ) Out[11] 0.545826792717 In [12] H.argmax ) Out[12] 24400 In [13] H[24400] Out[13] 1.0 Also In [41] H.max ) Out[41] 0.545826792717 In [42] H.sort ) In [43] H.max ) Out[43] 1.0 Cheers, Franck","bug"
|
|
"6013","1488","test failures with mingw when SIZEOF LONG DOUBLE == 8 Trac 890).
|
|
Original ticket http projects.scipy.org numpy ticket 890 on 2008 08 16 by cournape, assigned to cournape.
|
|
Since mingw long double support is broken, we should force the configuration SIZEOF LONG DOUBLE to 8.
|
|
Doing so cause the followng test failures ====================================================================== ERROR Test generic loops.
|
|
Traceback most recent call last) File C \Python25\lib\site packages\numpy\core\tests\test ufunc.py , line 79, in test generic loops assert almost equal fone x), fone val, err msg=msg) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 205, in assert almost equal return assert array almost equal actual, desired, decimal, err msg) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 304, in assert array almost equal header= Arrays are not almost equal ) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 272, in assert array compare val = comparison x[ xnanid], y[ ynanid]) IndexError 0 d arrays can t be indexed ====================================================================== FAIL test large types test scalarmath.TestPower) Traceback most recent call last) File C \Python25\lib\site packages\numpy\core\tests\test scalarmath.py , line 54, in test large types assert almost equal b, 6765201, err msg=msg) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 207, in assert almost equal assert round abs desired actual),decimal) == 0, msg AssertionError Items are not equal error with <type numpy.float64 > got 1.0 ACTUAL 1.0 DESIRED 6765201","bug"
|
|
"6013","1489","configuration problems on mingw Trac 891).
|
|
Original ticket http projects.scipy.org numpy ticket 891 on 2008 08 16 by cournape, assigned to cournape.
|
|
There are problems when configure numpy with VS 2003 vs mingw mingw !c define SIZEOF SHORT 2 define SIZEOF INT 4 define SIZEOF LONG 4 define SIZEOF FLOAT 4 define SIZEOF DOUBLE 8 define SIZEOF LONG DOUBLE 8 define SIZEOF PY INTPTR T 4 define SIZEOF LONG LONG 8 define SIZEOF PY LONG LONG 8 define CHAR BIT 8 define MATHLIB define HAVE LONGDOUBLE FUNCS define HAVE FLOAT FUNCS define HAVE LOG1P define HAVE EXPM1 define HAVE INVERSE HYPERBOLIC define HAVE INVERSE HYPERBOLIC FLOAT define HAVE INVERSE HYPERBOLIC LONGDOUBLE define HAVE RINT define NPY PRIVATE NO SIGNAL vs 2003 !c define SIZEOF SHORT 2 define SIZEOF INT 4 define SIZEOF LONG 4 define SIZEOF FLOAT 4 define SIZEOF DOUBLE 8 define SIZEOF LONG DOUBLE 8 define SIZEOF PY INTPTR T 4 define SIZEOF LONG LONG 8 define SIZEOF PY LONG LONG 8 define CHAR BIT 8 define MATHLIB define NPY PRIVATE NO SIGNAL","bug"
|
|
"6013","1490","One line summary in docstring isn t one line.
|
|
Trac 892).
|
|
Original ticket http projects.scipy.org numpy ticket 892 on 2008 08 22 by trac user timcera, assigned to unknown.
|
|
The example docstring for the foo function has two lines for the one line summary.
|
|
http projects.scipy.org scipy numpy browser trunk numpy doc example.py L37","bug"
|
|
"6013","1491","Numpy incompatible with python OO Trac 893).
|
|
Original ticket http projects.scipy.org numpy ticket 893 on 2008 08 26 by trac user cscott, assigned to pierregm.
|
|
In line 3256 of numpy ma core.py, the following code appears inner.
|
|
doc = doc note Masked values are replaced by 0. )
|
|
This line fails when python is invoked with the OO option, which removes doc strings to save memory, since inner.
|
|
doc will be None and a TypeError will be raised trying to add a string to None.
|
|
Adding a guard like if inner.
|
|
doc would fix this bug.","bug"
|
|
"6013","1492","OO optimization fails with nosetester Trac 894).
|
|
Original ticket http projects.scipy.org numpy ticket 894 on 2008 08 27 by pierregm, assigned to unknown.
|
|
Looks like `nosetester.docmethod` fails when the docstring are explicitly suppressed with the OO flag.
|
|
python OO c import numpy as np Traceback most recent call last) File <string> , line 1, in ?
|
|
File usr lib64 python2.4 site packages numpy init .py , line 125, in ?
|
|
import add newdocs File usr lib64 python2.4 site packages numpy add newdocs.py , line 9, in ?
|
|
from lib import add newdoc File usr lib64 python2.4 site packages numpy lib init .py , line 4, in ?
|
|
from type check import File usr lib64 python2.4 site packages numpy lib type check.py , line 8, in ?
|
|
import numpy.core.numeric as nx File usr lib64 python2.4 site packages numpy core init .py , line 34, in ?
|
|
from numpy.testing import Tester File usr lib64 python2.4 site packages numpy testing init .py , line 12, in ?
|
|
from utils import File usr lib64 python2.4 site packages numpy testing utils.py , line 9, in ?
|
|
from nosetester import import nose File usr lib64 python2.4 site packages numpy testing nosetester.py , line 88, in ?
|
|
class NoseTester object) File usr lib64 python2.4 site packages numpy testing nosetester.py , line 292, in NoseTester docmethod test argv, testtype) ) File usr lib64 python2.4 site packages numpy testing nosetester.py , line 85, in docmethod meth.
|
|
doc = meth.
|
|
doc { test header test header} TypeError unsupported operand type s) for NoneType and dict Note I m just kicking the baby from ticket 1491...","bug"
|
|
"6013","1493","Import error for matplotlib Trac 895).
|
|
Original ticket http projects.scipy.org numpy ticket 895 on 2008 08 27 by charris, assigned to unknown.
|
|
We should have no changes to the API and ABI for 1.2, yet matplotlib fails to import on account of a version error.
|
|
RuntimeError module compiled against version 100000a of C API but this version of numpy is 1000009","bug"
|
|
"6013","1494","Geometric, negative binomial and poisson fail for extreme arguments Trac 896).
|
|
Original ticket http projects.scipy.org numpy ticket 896 on 2008 08 29 by trac user anand, assigned to rkern.
|
|
I m guessing these are all related and have something to do with overflow In [1] import numpy In [2] p=1e 13 In [3] numpy.random.geometric p) Out[3] 2147483648 In [4] numpy.random.poisson 1 p) p) Out[4] 2147483648 In [5] numpy.random.negative binomial 1,p) Out[5] 2147483648","bug"
|
|
"6013","1495","Numpy mrecords unpickle causes unhandled win32 exception Trac 897).
|
|
Original ticket http projects.scipy.org numpy ticket 897 on 2008 08 29 by trac user chrisshucksmith, assigned to pierregm.
|
|
Putting a datetime object into a masked record array causing one or more entries in the dtype to be |O4 ) seems to trigger a bug in unpickling, resulting in either a TypeError object pickle not returning list or sudden death python.exe crash.
|
|
Crash import os from pickle import Pickler, Unpickler, HIGHEST PROTOCOL from StringIO import StringIO from numpy import array, ones, zeros from numpy.ma import mrecords from datetime import datetime cache dir = os.curdir rows = [] rowmasks = [] for i in xrange 5) x = datetime 2008,6,5) rows.append [x,x,x,x,x]) mask = zeros [5], bool ) mask[2] = True rowmasks.append mask) print data \n , rows print Mask \n , rowmasks recarr = mrecords.fromrecords rows, names=[ Ones , Twos , Threes , Fours , Fives ], mask=rowmasks) print Records; print recarr.dtype print recarr[0] print recarr.Twos print recarr.Threes print Picking sio = StringIO ) p = Pickler sio, HIGHEST PROTOCOL) p.dump recarr) print Unpickling sio.seek 0) u = Unpickler sio) recarr2 = u.load ) print recarr2[0] Exception case Traceback most recent call last) File C \workspace\eclipse\hacking\src\pycrash.py , line 33, in <module> recarr2 = u.load ) File c \python25\lib\pickle.py , line 858, in load dispatch[key] self) File c \python25\lib\pickle.py , line 1217, in load build setstate state) File C \Python25\Lib\site packages\numpy\ma\mrecords.py , line 520, in setstate ndarray.
|
|
setstate self, shp, typ, isf, raw)) TypeError object pickle not returning list Code import os from pickle import Pickler, Unpickler, HIGHEST PROTOCOL from StringIO import StringIO from numpy import array, ones, zeros from numpy.ma import mrecords from datetime import datetime cache dir = os.curdir rows = [] rowmasks = [] for i in xrange 6) rows.append [ datetime 2008,6,5) ] ) rowmasks.append zeros [1], bool )) print data \n , rows print Mask \n , rowmasks recarr = mrecords.fromrecords rows, names=[ Dates ], mask=rowmasks) print Records; print recarr.dtype print recarr[0] print recarr.Dates print Picking sio = StringIO ) p = Pickler sio, HIGHEST PROTOCOL) p.dump recarr) print Unpickling sio.seek 0) u = Unpickler sio) recarr2 = u.load ) print recarr2[0] print recarr2[1] I m aware neither of these cases represent good code, the datetime should concisely be stored in a numerical representation.
|
|
Tested with python 2.5 and various Numpy versions including numpy 1.1.1 py2.5.egg","bug"
|
|
"6013","1496","NumPy trunk doesn t compile on Windows Trac 898).
|
|
Original ticket http projects.scipy.org numpy ticket 898 on 2008 08 29 by alberts, assigned to cournape.
|
|
!NumPy trunk r5726 fails to compile on Windows Vista Business 64 bit with Python 2.6b2 and Visual Studio 2008. umathmodule.c numpy\core\src\umathmodule.c.src 64) error C2059 syntax error type numpy\core\src\umathmodule.c.src 70) error C2059 syntax error type numpy\core\src\ufuncobject.c 1701) warning C4244 = conversion from npy intp to int , possible loss of data numpy\core\src\ufuncobject.c 2422) warning C4244 = conversion from npy intp to int , possible loss of data Build log attached.","bug"
|
|
"6013","1497","NumPy test suite fails to run with Python 2.6b3 and nose 0.10.3 Trac 899).
|
|
Original ticket http projects.scipy.org numpy ticket 899 on 2008 08 30 by alberts, assigned to unknown.
|
|
The NumPy test suite fails to run with Python 2.6b3 and nose 0.10.3 python2.6 c import numpy; numpy.test ) Running unit tests for numpy NumPy version 1.2.0.dev5726 NumPy is installed in opt python26 lib python2.6 site packages numpy Python version 2.6b3 r26b3 65922, Aug 29 2008, 23 35 24) [GCC 4.3.0 20080428 Red Hat 4.3.0 8)] nose version 0.10.3 Traceback most recent call last) File <string> , line 1, in <module> File opt python26 lib python2.6 site packages numpy testing nosetester.py , line 278, in test t = NumpyTestProgram argv=argv, exit=False, plugins=plugins) File opt python26 lib python2.6 site packages nose core.py , line 219, in init argv=argv, testRunner=testRunner, testLoader=testLoader) File opt python26 lib python2.6 unittest.py , line 816, in init self.parseArgs argv) File opt python26 lib python2.6 site packages nose core.py , line 270, in parseArgs self.createTests ) File opt python26 lib python2.6 site packages nose core.py , line 284, in createTests self.test = self.testLoader.loadTestsFromNames self.testNames) File opt python26 lib python2.6 site packages nose loader.py , line 422, in loadTestsFromNames return unittest.TestLoader.loadTestsFromNames self, names, module) File opt python26 lib python2.6 unittest.py , line 613, in loadTestsFromNames suites = [self.loadTestsFromName name, module) for name in names] File opt python26 lib python2.6 site packages nose loader.py , line 377, in loadTestsFromName module, discovered=discovered) File opt python26 lib python2.6 site packages nose loader.py , line 301, in loadTestsFromModule tests.extend self.loadTestsFromDir path)) File opt python26 lib python2.6 site packages nose loader.py , line 165, in loadTestsFromDir entry path, discovered=True) File opt python26 lib python2.6 site packages nose loader.py , line 377, in loadTestsFromName module, discovered=discovered) File opt python26 lib python2.6 site packages nose loader.py , line 306, in loadTestsFromModule return self.suiteClass ContextList tests, context=module)) File opt python26 lib python2.6 site packages nose suite.py , line 332, in call return self.makeSuite tests, context) File opt python26 lib python2.6 site packages nose suite.py , line 377, in makeSuite self.context.setdefault suite, []).append context) TypeError unhashable type ContextSuite","bug"
|
|
"6013","1498","Compiler warnings Trac 900).
|
|
Original ticket http projects.scipy.org numpy ticket 900 on 2008 08 30 by alberts, assigned to unknown.
|
|
Build !NumPy trunk r5726 on 64 Bit Linux with GCC 4.3 gcc o build scons numpy core src multiarraymodule.os c O2 fno strict aliasing DNDEBUG fomit frame pointer g Wall Wstrict prototypes fPIC I opt python26 include python2.6 Ibuild scons numpy core include Ibuild scons numpy core I opt python26 include python2.6 I opt intel mkl 10.0.4.023 include I opt intel mkl 10.0.4.023 include build scons numpy core src multiarraymodule.c In file included from build scons numpy core src arrayobject.c 536, from build scons numpy core src multiarraymodule.c 97 build scons numpy core src scalartypes.inc In function object arrtype new build scons numpy core src scalartypes.inc 1903 warning unused variable src build scons numpy core src scalartypes.inc 1903 warning unused variable dest build scons numpy core src scalartypes.inc 1902 warning unused variable itemsize","bug"
|
|
"6013","1499","segment axis Trac 901).
|
|
Original ticket http projects.scipy.org numpy ticket 901 on 2008 08 31 by cournape, assigned to charris.
|
|
A patch to have segment axis contributed by Anne Archibald I believe) in numpy with tests.
|
|
It uses as strided","enhancement"
|
|
"6013","1500","inconsistent float to string conversion with german locale Trac 902).
|
|
Original ticket http projects.scipy.org numpy ticket 902 on 2008 08 31 by trac user ChristianK, assigned to cournape.
|
|
The string representation of numpy.floatxx object is inconsistent on systems with german locale import locale locale.str 1.2) 1,2 a = N.linspace 0,1,5) a array [ 0. , 0.25, 0.5 , 0.75, 1. ])
|
|
a.astype float) array [ 0. , 0.25, 0.5 , 0.75, 1. ])
|
|
a[0].astype float) 0.0 a[1].astype float) 0,25 Note that 0 got converted to 0.0 whereas 0.25 yielded 0,25 using a comma as floating point in accordance with german locale.
|
|
I vote against accounting for locale settings at all when printing numpy floats.
|
|
Christian","bug"
|
|
"6013","1501","Compilation fails on Cygwin Trac 903).
|
|
Original ticket http projects.scipy.org numpy ticket 903 on 2008 09 01 by mdehoon, assigned to unknown.
|
|
compiling C sources C compiler gcc fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes compile options Ibuild src.cygwin 1.5.25 i686 2.5 numpy core src Inumpy core include Ibuild src.cygwin 1.5.25 i686 2.5 numpy core include numpy Inumpy cor e src Inumpy core include I usr include python2.5 I usr include python2.5 c gcc build src.cygwin 1.5.25 i686 2.5 numpy core src umathmodule.c In file included from numpy core src umathmodule.c.src 2182 numpy core src ufuncobject.c In function ` extract pyvals numpy core src ufuncobject.c 1166 warning int format, long int arg arg 4) numpy core src ufuncobject.c 1166 warning int format, long int arg arg 5) cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s Assembler messages cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s 72840 Error suffix or op erands invalid for `fnstsw cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s 73095 Error suffix or op erands invalid for `fnstsw In file included from numpy core src umathmodule.c.src 2182 numpy core src ufuncobject.c In function ` extract pyvals numpy core src ufuncobject.c 1166 warning int format, long int arg arg 4) numpy core src ufuncobject.c 1166 warning int format, long int arg arg 5) cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s Assembler messages cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s 72840 Error suffix or op erands invalid for `fnstsw cygdrive c DOCUME 1 mdehoon LOCALS 1 Temp ccpLcg0M.s 73095 Error suffix or op erands invalid for `fnstsw error Command gcc fno strict aliasing DNDEBUG g O3 Wall Wstrict prototyp es Ibuild src.cygwin 1.5.25 i686 2.5 numpy core src Inumpy core include Ibuil d src.cygwin 1.5.25 i686 2.5 numpy core include numpy Inumpy core src Inumpy c ore include I usr include python2.5 I usr include python2.5 c build src.cygwi n 1.5.25 i686 2.5 numpy core src umathmodule.c o build temp.cygwin 1.5.25 i686 2.5 build src.cygwin 1.5.25 i686 2.5 numpy core src umathmodule.o failed with e xit status 1","bug"
|
|
"6013","1502","Ctypes Test failure on cygwin Trac 904).
|
|
Original ticket http projects.scipy.org numpy ticket 904 on 2008 09 01 by cournape, assigned to rgommers.
|
|
====================================================================== ERROR test basic test ctypeslib.TestLoadLibrary) Traceback most recent call last) File usr lib python2.5 site packages numpy tests test ctypeslib.py , line 9, in test basic np.core.multiarray.
|
|
file ) File usr lib python2.5 site packages numpy ctypeslib.py , line 111, in load library raise e OSError No such file or directory Ran 1573 tests in 19.562s With revision r5738","bug"
|
|
"6013","1503","numpy.loadtxt usecols argument no longer accepts numpy arrays Trac 905).
|
|
Original ticket http projects.scipy.org numpy ticket 905 on 2008 09 04 by dopplershift, assigned to unknown.
|
|
t appears that the usecols argument to loadtxt no longer accepts numpy arrays >>>from StringIO import StringIO >>>text = StringIO 1 2 3\n4 5 6\n ) >>>data = np.loadtxt text, usecols=np.arange 1,3)) ValueError Traceback most recent call last) usr lib64 python2.5 site packages numpy lib io.py in loadtxt fname, dtype, comments, delimiter, converters, skiprows, usecols, unpack) 323 first line = fh.readline ) 324 first vals = split line first line) > 325 N = len usecols or first vals) 326 327 dtype types = flatten dtype dtype) ValueError The truth value of an array with more than one element is ambiguous.
|
|
Use a.any ) or a.all ) >>>data = np.loadtxt text, usecols=np.arange 1,3).tolist )) >>>data array [[ 2., 3.
|
|
], [ 5., 6.]])
|
|
Before the fix and refactoring of loadtxt in 1.1.1, converting to a list was not necessary.
|
|
I think the assumptions just kind of cropped in exploiting certain list functionality like list.find for ease of code).
|
|
I ve attached a simple patch against HEAD that does the obvious fix and converts usecols to a list if it s not None.
|
|
This should allow almost any iterable including tuples, which are also currently broken but used in the docstring) to be passed in to usecols.","bug"
|
|
"6013","1504","Deepcopy does not copy mask attribute of a masked array object Trac 906).
|
|
Original ticket http projects.scipy.org numpy ticket 906 on 2008 09 05 by trac user vladan, assigned to pierregm.
|
|
Changing mask attribute of a deep copy of a masked array also changes the original masked array.
|
|
Example In [1] import numpy In [2] from copy import deepcopy In [3] mar = numpy.ma.array [0,1,2], mask=[False,True,False]) In [4] mar copy = deepcopy mar) In [5] mar copy.
|
|
mask[1] = False In [6] mar Out[6] masked array data = [0 1 2], mask = [False False False], fill value=999999) Interestingly, changing mask of the original masked array does not change a deep copy of that array.","bug"
|
|
"6013","1505","Simplifying isnan and co replacement Trac 907).
|
|
Original ticket http projects.scipy.org numpy ticket 907 on 2008 09 07 by cournape, assigned to unknown.
|
|
Here is a patch to simplify isnan, isinf, isfinite and signbit for platforms which do not have it.
|
|
The patch does two things it detects whether the above are available as macro as mandated by C99) if not, define macro replacement, avoiding bit twiddling as much as possible only signbit is kept).","enhancement"
|
|
"6013","1506","Infinite loop in loadtxt of empty file Trac 908).
|
|
Original ticket http projects.scipy.org numpy ticket 908 on 2008 09 09 by pv, assigned to unknown.
|
|
loadtxt enters an infinite loop when one attempts to load an empty file > foo.dat python c import numpy; print numpy.
|
|
version ; numpy.loadtxt foo.dat ) 1.1.1 [... does not terminate ...]","bug"
|
|
"6013","1507","fromstring ) fromfile ) Enhancements Trac 909).
|
|
Original ticket http projects.scipy.org numpy ticket 909 on 2008 09 10 by trac user ChrisBarker, assigned to unknown.
|
|
== Proposed Enhancements and bug fixes for `fromfile )` and `fromstring )` text handling == === Motivation === The goal of the `fromfile )` text file handling capability is to enable users to write code that can read a lot of numbers from a text file into an array.
|
|
Python provides a lot of nifty text processing capabilities, and there are a number of higher level facilities for reading blocks of data including `numpy.loadtxt`).
|
|
These are very capable, but there really is a significant performance hit, at least when loading 10s of thousands of numbers into a file.
|
|
We don t want to write all of `loadtxt `) and friends in C. Rather, the goal is to allow the simple cases to be done very efficiently, and hopefully fancier text reading packages can build on it to add more features.
|
|
Unfortunately, the current numpy version 1.2) version has a few bugs and limitations that keep of from being nearly as useful as it could be.
|
|
=== Possible features === Create `fromtextfile )` and `fromtextstring` functions, distinct from `fromfile )` and `fromstring )`.
|
|
It really is a different functionality.
|
|
`fromfile `) could still call `fromtextfile )` for backward compatibility.
|
|
Allow more than one separator?
|
|
for example, a comma or whitespace?
|
|
In the general case, the user could perhaps specify any number of separators, though I doubt that would be useful in practice.
|
|
At the very least, however, `fromtextfile )` should support reading files that look like 43.5, 345.6, 123.456, 234.33 34.5, 22.57, 2345, 2345, 252 ... That is, comma separated, but being able to read multiple lines in one shot.
|
|
The easiest way to support that would probably be to always allow whitespace as a separator, and add the one passed in.
|
|
I can t think of a reason not to do this, but maybe I m not very imaginative.
|
|
Allow the user to specify a shape for the output array.
|
|
There may be little point, as all this does is save a calls to reshape ), but it may be another way to support the above.
|
|
i.e.
|
|
you could read that data with `a = np.fromtextfile infile, dtype=np.float, sep= , , shape= 1, 4))` Then it would know to skip the newlines every 4 elements.
|
|
Allow the user to specify a comment string.
|
|
The reader would then skip everything in the file between the comment string and a newline.
|
|
Maybe Universal newline any of \r, \n or \r\n.
|
|
Or simply expect that the user has opened the file with mode U if they want that.
|
|
This could also be extended to support C style comments with an opening and closing character sequence, but that s a lot less common.
|
|
Allow the user to specify a Locale.
|
|
It may be best to be able to specify a locale, rather than relying on the system on whether .
|
|
or , is the decimal separator, for instance.
|
|
ticket 1482) parsing of Inf and the like that doesn t depend on system ticket 1108).
|
|
This would be nice, but maybe too difficult would we need to write our own `scanf`?
|
|
=== Bugs to be fixed === `fromfile )` and `fromstring` handling malformed data poorly ticket 1481 Any others?","enhancement"
|
|
"6013","1508","Some test issues with Python 2.6 Trac 910).
|
|
Original ticket http projects.scipy.org numpy ticket 910 on 2008 09 10 by alberts, assigned to unknown.
|
|
I ve gotten nose 0.10.3 up and running with !Python 2.6 by setting ` hash ` in !ContextSuite to `object.
|
|
hash `.
|
|
Now, when I run the test suite with the latest Python and !NumPy trunks, I see the following issues opt python26 lib python2.6 site packages numpy lib tests test io.py 68 SyntaxWarning assertion is always true, perhaps remove parentheses?
|
|
opt python26 lib python2.6 site packages numpy ma tests test core.py 1350 SyntaxWarning assertion is always true, perhaps remove parentheses?
|
|
opt python26 lib python2.6 site packages numpy ma tests test core.py 1357 SyntaxWarning assertion is always true, perhaps remove parentheses?
|
|
opt python26 lib python2.6 site packages numpy ma tests test core.py 2024 SyntaxWarning assertion is always true, perhaps remove parentheses?
|
|
====================================================================== ERROR Tests the min max functions with explicit outputs Traceback most recent call last) File opt python26 lib python2.6 site packages numpy ma tests test core.py , line 688, in test minmax funcs with output result = npfunc xm,axis=0,out=nout) File opt python26 lib python2.6 site packages numpy core fromnumeric.py , line 1525, in amin return amin axis, out) File opt python26 lib python2.6 site packages numpy ma core.py , line 2882, in min np.putmask out, newmask, np.nan) ValueError cannot convert float NaN to integer The following failure happens on Windows with Python 2.6b2 too ====================================================================== FAIL test umath.TestComplexFunctions.test against cmath Traceback most recent call last) File opt python26 lib python2.6 site packages nose case.py , line 182, in runTest self.test self.arg) File opt python26 lib python2.6 site packages numpy core tests test umath.py , line 266, in test against cmath assert abs a b) < atol, s s s; cmath s fname,p,a,b) AssertionError arcsin 2 1.57079632679 1.31695789692j); cmath 1.57079632679 1.31695789692j)","bug"
|
|
"6013","1509","python2.4 indexing list with Numpy some integer scalars does not work Trac 911).
|
|
Original ticket http projects.scipy.org numpy ticket 911 on 2008 09 11 by pv, assigned to unknown.
|
|
On a 64 bit platform with Python 2.4 indexing a list with a Numpy integer scalar of non platform native size fails Python 2.4.4 2, Apr 16 2008, 17 58 59) [GCC 4.1.2 20061115 prerelease) Debian 4.1.1 21)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.2.0rc2 >>> x=[1,2,3] >>> x[np.int32 0)] Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError list indices must be integers >>> x[np.int64 0)] 1 >>> np.int is np.int64 True It works with Numpy on Python 2.5 Python 2.5 release25 maint, Jul 23 2008, 18 15 29) [GCC 4.1.2 20061115 prerelease) Debian 4.1.1 21)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.1.1 >>> x=[1,2,3] >>> x[np.int32 0)] 1 >>> x[np.int64 0)] 1 >>> np.int is np.int64 True","bug"
|
|
"6013","1510","numpy.test does not find any test on cygwin Trac 912).
|
|
Original ticket http projects.scipy.org numpy ticket 912 on 2008 09 13 by cournape, assigned to cournape.
|
|
I think it is because for some reason, all test scripts are executable when checking out numpy with subversion on cygwin, and nose do not execute those.
|
|
If I remove the executable bit of all test scripts, then it works.","bug"
|
|
"6013","1511","max is bogus if nan is in the array Trac 913).
|
|
Original ticket http projects.scipy.org numpy ticket 913 on 2008 09 19 by cournape, assigned to unknown.
|
|
!python import numpy as np np.max np.array [ 1, np.nan, 2])) Returns 2, not 1.","bug"
|
|
"6013","1512","min is bogus if nan is in the array Trac 914).
|
|
Original ticket http projects.scipy.org numpy ticket 914 on 2008 09 19 by cournape, assigned to unknown.
|
|
!python import numpy as np np.min np.array [ 1, np.nan, 2])) Returns 2, not the correct) 1.","bug"
|
|
"6013","1513","Simpler math configuration Trac 915).
|
|
Original ticket http projects.scipy.org numpy ticket 915 on 2008 09 20 by cournape, assigned to cournape.
|
|
Here is a patch to simplify math configuration instead of relying on heuristics for presence of some function, each function is checked separately To avoid long configuration times on standard compliant platforms, a check funcs once function to distutils config command has been added.
|
|
Build times with gcc on linux and mac os x are the same as before.
|
|
I have tested the patch on numpy it passes all tests on linux 32 bits), mac os X, mingw and Visual Studio 2003.","bug"
|
|
"6013","1514","Inconsistency in type equality Trac 916).
|
|
Original ticket http projects.scipy.org numpy ticket 916 on 2008 09 22 by FrancescAlted, assigned to unknown.
|
|
It seems to me that there is a inconsistency between the comparisons of dtype and dtype.type attributes of a !NumPy array.
|
|
The next shows the problem ran on a 32 bit machine) In [20] a=numpy.array [2], i ) In [21] b=numpy.array [2], l ) In [22] a.dtype == b.dtype Out[22] True In [23] a.dtype.type == b.dtype.type Out[23] False In [24] a.dtype.type Out[24] <type numpy.int32 > In [25] b.dtype.type Out[25] <type numpy.int32 > I d like to use arr.dtype.type instead of arr.dtype alone because there are situations where I want to check only for the raw type, without byteordering, shape or other considerations.
|
|
Thanks, Francesc","bug"
|
|
"6013","1515","allow dtype= c in chararray.
|
|
Trac 917).
|
|
Original ticket http projects.scipy.org numpy ticket 917 on 2008 09 24 by trac user brentp, assigned to unknown.
|
|
currently this fails np.array abc1 , dtype= c ).view np.chararray) because chararray checks for dtype.char in SUb the 1 character change in the patch test) allows chararray to accept arrays with dtype= c .","bug"
|
|
"6013","1516","SETITEM and GETITEM macros end with semicolons Trac 918).
|
|
Original ticket http projects.scipy.org numpy ticket 918 on 2008 09 25 by trac user jason, assigned to unknown.
|
|
The macros PyArray SETITEM and PyArray GETITEM) in ndarrayobject.h end in semicolons, which causes problems as detailed below.
|
|
Can we remove the semicolon, but wrap the entire macro body in parentheses?
|
|
Are there other macros that end in semicolons that shouldn t?
|
|
In the email below, the replies are from Chuck Harris) On Wed, Sep 24, 2008 at 9 25 PM, <jason sage creativetrax.com> wrote I m working on getting the Sage matrices for real complex doubles to use numpy as a backend.
|
|
In this, I m using the PyArray SETITEM macro from within Cython.
|
|
However, Cython wraps the macro in a function call to convert the output to a Python value pyx 1 = PyInt FromLong PyArray SETITEM pyx v self > matrix numpy, PyArray GETPTR2 pyx v self > matrix numpy, pyx v i, pyx v j), pyx v value)); However, after preprocessing, because of the semicolon at the end of the SETITEM macro, we get Hmm, not good.
|
|
IMHO, macros should be ended with semicolons when used , not in themselves.
|
|
PyInt FromLong PyArrayObject ) pyx v self > matrix numpy)) >descr >f >setitem PyObject ) pyx v value), char ) void ) PyArrayObject ) pyx v self > matrix numpy)) >data) pyx v i) PyArrayObject ) pyx v self > matrix numpy)) >strides)[0] pyx v j) PyArrayObject ) pyx v self > matrix numpy)) >strides)[1]))), PyArrayObject ) pyx v self > matrix numpy));); Note that at the end, we have a );); .
|
|
The file refuses to compile.
|
|
Presumably, since SETITEM returns a value, wrapping the return value in a function call seems to be a reasonable thing to do.
|
|
Would there be a problem in eliminating the semicolon and instead wrapping the entire function body in parenthesis?
|
|
We should do this if possible.
|
|
Open a ticket.
|
|
I noticed that GETITEM also ended in a semicolon, though I didn t have the same problem as above since Cython didn t automatically wrap it in a function call I m not even sure if it returns something).
|
|
On a side note, is the above the best way i.e., fastest way given an arbitrary numpy array) to set get an element?
|
|
Don t know.
|
|
Chuck","bug"
|
|
"6013","1517","Incorrect printing some complex complex arrays containing nans.
|
|
Trac 919).
|
|
Original ticket http projects.scipy.org numpy ticket 919 on 2008 09 30 by charris, assigned to cournape.
|
|
Example In [29] c = np.complex NAN NAN 1j) In [30] c Out[30] nan nanj) In [31] array [c,c]) Out[31] array [ NaN NaNj, NaN NaNj])","bug"
|
|
"6013","1518","np.nan has negative sign, should be positive.
|
|
Trac 920).
|
|
Original ticket http projects.scipy.org numpy ticket 920 on 2008 10 02 by charris, assigned to unknown.
|
|
On my reverse endian machine In [3] array [np.nan]).tostring ) Out[3] \x00\x00\x00\x00\x00\x00\xf8\xff","bug"
|
|
"6013","1519","numpy.random.hypergeometric error for some cases Trac 921).
|
|
Original ticket http projects.scipy.org numpy ticket 921 on 2008 10 02 by josef pkt, assigned to unknown.
|
|
In my fuzz testing of scipy stats, I get sometimes a test failure.
|
|
I think there is something wrong, random numbers are outside of support of the distribution, with numpy.random.hypergeometric for some cases >>> np.version.version 1.2.0rc2 signature hypergeometric ngood, nbad, nsample, size=None) when sample size number of draws, nsample) is small then random numbers look ok >>> np.random.hypergeometric 3,18,8,size=10) array [2, 1, 0, 2, 1, 1, 2, 1, 0, 3]) >>> np.random.hypergeometric 3,18,9,size=10) array [2, 2, 0, 0, 1, 0, 1, 1, 3, 1]) >>> np.random.hypergeometric 3,18,10,size=10) array [0, 2, 2, 0, 2, 2, 2, 0, 0, 2]) for sample size larger than 11 the random number are larger than possible there are only 3 good balls in urn) >>> np.random.hypergeometric 3,18,11,size=10) array [18, 16, 16, 17, 18, 17, 18, 17, 17, 16]) >>> np.random.hypergeometric 3,18,12,size=10) array [17, 16, 16, 18, 17, 17, 18, 17, 16, 17]) >>> np.random.hypergeometric 3,18,13,size=10) array [18, 17, 17, 16, 17, 17, 16, 16, 18, 16]) >>> np.random.hypergeometric 3,18,14,size=10) array [16, 17, 17, 17, 17, 18, 17, 16, 18, 17]) >>> np.random.hypergeometric 3,18,15,size=10) array [18, 18, 17, 16, 17, 16, 17, 17, 18, 18]) reversing the number of good and bad balls > ok for small sample size >>> np.random.hypergeometric 18,3,5,size=10) ok array [5, 4, 5, 5, 4, 4, 4, 4, 5, 5]) >>> np.random.hypergeometric 18,3,10,size=10) ok array [9, 8, 9, 7, 8, 8, 7, 8, 8, 9]) negative numbers for sample size >= 11 >>> np.random.hypergeometric 18,3,11,size=10) array [ 5, 5, 5, 7, 4, 5, 5, 4, 6, 6]) >>> np.random.hypergeometric 18,3,13,size=10) array [ 3, 5, 5, 4, 4, 4, 5, 5, 3, 4]) >>> np.random.hypergeometric 18,3,14,size=10) array [ 4, 3, 4, 3, 2, 4, 3, 4, 2, 4]) >>> np.random.hypergeometric 18,3,15,size=10) array [ 2, 2, 1, 1, 3, 2, 2, 2, 1, 2]) >>> np.random.hypergeometric 18,3,16,size=10) array [ 1, 1, 0, 1, 0, 2, 1, 2, 1, 2])","bug"
|
|
"6013","1520","numpy.random.hypergeometric error for some cases Trac 922).
|
|
Original ticket http projects.scipy.org numpy ticket 922 on 2008 10 02 by josef pkt, assigned to unknown.
|
|
In my fuzz testing of scipy stats, I get sometimes a test failure.
|
|
I think there is something wrong, random numbers are outside of support of the distribution, with numpy.random.hypergeometric for some cases >>> np.version.version 1.2.0rc2 signature hypergeometric ngood, nbad, nsample, size=None) when sample size number of draws, nsample) is small then random numbers look ok >>> np.random.hypergeometric 3,18,8,size=10) array [2, 1, 0, 2, 1, 1, 2, 1, 0, 3]) >>> np.random.hypergeometric 3,18,9,size=10) array [2, 2, 0, 0, 1, 0, 1, 1, 3, 1]) >>> np.random.hypergeometric 3,18,10,size=10) array [0, 2, 2, 0, 2, 2, 2, 0, 0, 2]) for sample size larger than 11 the random number are larger than possible there are only 3 good balls in urn) >>> np.random.hypergeometric 3,18,11,size=10) array [18, 16, 16, 17, 18, 17, 18, 17, 17, 16]) >>> np.random.hypergeometric 3,18,12,size=10) array [17, 16, 16, 18, 17, 17, 18, 17, 16, 17]) >>> np.random.hypergeometric 3,18,13,size=10) array [18, 17, 17, 16, 17, 17, 16, 16, 18, 16]) >>> np.random.hypergeometric 3,18,14,size=10) array [16, 17, 17, 17, 17, 18, 17, 16, 18, 17]) >>> np.random.hypergeometric 3,18,15,size=10) array [18, 18, 17, 16, 17, 16, 17, 17, 18, 18]) reversing the number of good and bad balls > ok for small sample size >>> np.random.hypergeometric 18,3,5,size=10) ok array [5, 4, 5, 5, 4, 4, 4, 4, 5, 5]) >>> np.random.hypergeometric 18,3,10,size=10) ok array [9, 8, 9, 7, 8, 8, 7, 8, 8, 9]) negative numbers for sample size >= 11 >>> np.random.hypergeometric 18,3,11,size=10) array [ 5, 5, 5, 7, 4, 5, 5, 4, 6, 6]) >>> np.random.hypergeometric 18,3,13,size=10) array [ 3, 5, 5, 4, 4, 4, 5, 5, 3, 4]) >>> np.random.hypergeometric 18,3,14,size=10) array [ 4, 3, 4, 3, 2, 4, 3, 4, 2, 4]) >>> np.random.hypergeometric 18,3,15,size=10) array [ 2, 2, 1, 1, 3, 2, 2, 2, 1, 2]) >>> np.random.hypergeometric 18,3,16,size=10) array [ 1, 1, 0, 1, 0, 2, 1, 2, 1, 2])","bug"
|
|
"6013","1521","numpy.random.logseries incorrect convergence for k=1, k=2 Trac 923).
|
|
Original ticket http projects.scipy.org numpy ticket 923 on 2008 10 02 by josef pkt, assigned to unknown.
|
|
random numbers generated by numpy.random.logseries do not converge to the theoretical distribution Note I checked with sample size 1 million, but numpy.random.logseries converges already for smaller sample sizes to the wrong values for k=1 and k=2.
|
|
For probability paramater pr = 0.8, the random number generator converges to a frequency for k=1 at 39.8 , while the theoretical probability mass is 49.71 , k=2 is oversampled, other k s look ok check frequency of k=1 and k=2 at N = 1000000 0.398406 0.296465 pmf at k = 1 and k=2 with formula [ 0.4971 0.1988] For probability paramater pr = 0.3, the results are not as bad, but are still off frequency for k=1 at 82.6 , while the theoretical probability mass is 84.11 check frequency of k=1 and k=2 at N = 1000000 0.826006 0.141244 pmf at k = 1 and k=2 with formula [ 0.8411 0.1262] below is a quick script for checking this Josef import numpy as np from scipy import stats pr = 0.8 np.set printoptions precision=2, suppress=True) calculation for N=1million takes some time for N in [1000, 10000, 10000, 1000000] rvsn=np.random.logseries pr,size=N) fr=stats.itemfreq rvsn) pmfs=stats.logser.pmf fr[ ,0],pr) 100 print log series sample frequency and pmf in ) with N = , N print np.column stack fr[ ,0],fr[ ,1] 100.0 N,pmfs)) np.set printoptions precision=4, suppress=True) print check frequency of k=1 and k=2 at N = , N print np.sum rvsn==1) float N), print np.sum rvsn==2) float N) k = np.array [1,2]) print pmf at k = 1 and k=2 with formula print pr k 1.0 k np.log 1 pr)","bug"
|
|
"6013","1522","problem with summing large array of float32 Trac 924).
|
|
Original ticket http projects.scipy.org numpy ticket 924 on 2008 10 03 by trac user emil, assigned to unknown.
|
|
In medical imaging large arrays of single floats are often used.
|
|
In summing over such an array I discovered the following problem a=ones [840,2200,60],float32) sum a.flatten )) yields 16777216 4096 2) ???
|
|
However, sum sum sum a,axis=0),axis=0),axis=0) yields 110880000 the correct result.
|
|
The problem doesn t occur with float64 dot a.flatten ) , a.flatten ) ) also gives 16777216 I don t think the problem lies only with numpy.
|
|
I tried to circumvent the problem by writing my own summation in fortran gfortran) and using f2py.
|
|
The following fortran snippet gave me also incorrect results Real 4 sum sum = 0. do i = 1,840 do j = 1,2200 do k = 1,60 sum = sum 1.0 end do end do end do sum ends up being 16777216 again!
|
|
the incorrect result!
|
|
on the other hand the following works Real 4 sum1, sum sum = 0. do i = 1,840 sum1 = 0. do j = 1,2200 do k = 1,60 sum1 = sum1 1.0 end do end do sum = sum sum1 end do The same problem seems to come up in matlab.
|
|
My computer has amd64 s opteron) and is running suse linux","bug"
|
|
"6013","1523","choose does not broadcast index array appropriately Trac 925).
|
|
Original ticket http projects.scipy.org numpy ticket 925 on 2008 10 03 by teoliphant, assigned to unknown.
|
|
The advertised behavior of choose choices, sequence of arrays) is that the choices array and all the arrays in the sequence must be broadcastable to the same shape which is the shape of the output.
|
|
However, NumPy does not look at the shape of the choices array, only the sequence of arrays.
|
|
This requires an adaptation of the code in PyArray Choose function in C. Probably the easiest thing is to construct a MultiIterator object to handle the broadcasting internal to that function, construct the output, and then fill appropriately.","bug"
|
|
"6013","1524","NumPy array constructors are not honoring the shape in dtype Trac 926).
|
|
Original ticket http projects.scipy.org numpy ticket 926 on 2008 10 06 by FrancescAlted, assigned to unknown.
|
|
The array constructors are shifting the shape in specified dtype into the ndarray object itself >>> a=numpy.empty [], dtype= float64 , 2, 2))) >>> a.shape 2, 2) >>> a.dtype.shape ) while I would expect >>> a=numpy.empty [], dtype= float64 , 2, 2))) >>> a.shape ) >>> a.dtype.shape 2,2) I need this in order to implement the capability to specify multidimensional atoms in !PyTables.
|
|
You can see a little discussion about this issue in http www.pytables.org trac ticket 133 Thanks","bug"
|
|
"6013","1525","automatic type conversin fails for numpy array Trac 927).
|
|
Original ticket http projects.scipy.org numpy ticket 927 on 2008 10 08 by trac user Qwertz, assigned to unknown.
|
|
usually, automatic type conversion in python works like this a = 2 a = 3.0 a = 0.66666, type a) is float in numpy an analogous conversion fails a = numpy.arange 4) a is array [0, 1, 2, 3]) a = 2.1 a is array [0, 0, 0, 1]) < bug a stays an int array, that s inconsistant with the type conversion shown above shall be like the following instead a = a 2.1 works array [ 0. , 0.47619048, 0.95238095, 1.42857143])","bug"
|
|
"6013","1526","histogramdd ) overflow error on high dimension arrays Trac 928).
|
|
Original ticket http projects.scipy.org numpy ticket 928 on 2008 10 13 by trac user jvansanten, assigned to unknown.
|
|
When histogramdd ) is called on a large array say, 5 dimensions with 100 bins each), it tries to create a flat array of the bins.
|
|
The size of the flat array is given by nbin.prod ).
|
|
prod ) uses the dtype of nbin by default int32), which on arrays like the example above overflows into a negative integer, throwing an error.
|
|
This is of course not a problem until one starts working with gigantic histograms, but it can be tricky nonetheless.
|
|
Forcing nbin.prod ) to use a long would clean this up nic","bug"
|
|
"6013","1527","empty like and zeros like behave differently from ones like Trac 929).
|
|
Original ticket http projects.scipy.org numpy ticket 929 on 2008 10 14 by trac user jbosch, assigned to unknown.
|
|
The `empty like` and `zeros like` creation functions, which are defined in python in numeric.py, differ in their handling of array subtype input from `ones like`, which is defined in C as a ufunc.
|
|
In particular, if a subtype of ndarray is passed to `empty like` or `zeros like`, neither ` array wrap ` or ` array finalize ` is called, which makes it impossible for subtypes to define any extra attributes they might need.
|
|
I believe this behavior could be fixed entirely by switching from `if isinstance a,ndarray) ` to `if type a)==ndarray ` at the top of those two functions, though there are probably more efficient ways to accomplish the same thing.","bug"
|
|
"6013","1528","empty like and zeros like behave differently from ones like Trac 930).
|
|
Original ticket http projects.scipy.org numpy ticket 930 on 2008 10 14 by trac user jbosch, assigned to unknown.
|
|
The `empty like` and `zeros like` creation functions, which are defined in python in numeric.py, differ in their handling of array subtype input from `ones like`, which is defined in C as a ufunc.
|
|
In particular, if a subtype of ndarray is passed to `empty like` or `zeros like`, neither ` array wrap ` or ` array finalize ` is called, which makes it impossible for subtypes to define any extra attributes they might need.
|
|
I believe this behavior could be fixed entirely by switching from `if isinstance a,ndarray) ` to `if type a)==ndarray ` at the top of those two functions, though there are probably more efficient ways to accomplish the same thing.","bug"
|
|
"6013","1529","Records containing zero length items pickle just fine, but cannot be unpickled Trac 931).
|
|
Original ticket http projects.scipy.org numpy ticket 931 on 2008 10 14 by njsmith, assigned to unknown.
|
|
In numpy 1.0.4 I don t have any newer version convenient to test), it is possible to create a structured array with a zero repeat count entry.
|
|
Such an array works fine, and can be pickled as normal.
|
|
However, attempting to unpickle such an array fails.
|
|
This is admittedly a somewhat esoteric bit of functionality, but the bug shows up in a nasty way I had some data that took all night to generate and that I saved to disk in a pickle, and it was only by chance that I still had the original array in memory when I attempted to re load the data and discovered the problem.
|
|
And it contained a zero length sub record because the code in question handles a variable number of distractor variables, which happened to be 0 in this run.)
|
|
Single line reproducer python c import numpy, cPickle; cPickle.loads cPickle.dumps numpy.array [ 1, [])], dtype=[ a , i ), b , i , 0)]))) Transcript python Python 2.5.2 r252 60911, Jul 31 2008, 17 31 22) [GCC 4.2.3 Ubuntu 4.2.3 2ubuntu7)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.0.4 >>> a = np.array [ 1, [])], dtype=[ a , np.int32), b , np.int32, 0)]) >>> a array [ 1, [])], dtype=[ a , <i4 ), b , <i4 , 0)]) >>> import cPickle >>> cPickle.dumps a) long string elided) >>> cPickle.loads cPickle.dumps a)) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError data type not understood , <type numpy.dtype >, V0 , 0, 1)) Thanks.","bug"
|
|
"6013","1530","Add new convolve method for faster computation of even functions Trac 932).
|
|
Original ticket http projects.scipy.org numpy ticket 932 on 2008 10 14 by trac user xamusk, assigned to unknown.
|
|
Convolve and other functions which depend on convolve like operations, like acorr, xcorr those from matplotlib), etc, could use a new method to operate only on one side of the full mode.
|
|
So, besides mode being full , same and valid , there could be a half mode name only as a proposition) that could compute only one side of the function lag>=0) so that it would return only len x) elements instead of 2 len x) 1.
|
|
This is particularly useful with the acorr function, for which a mode like that could compute the result much faster and represent only the wanted elements, much like Matlab s autocorr ).
|
|
Actual behavior in matlab autocorr x)[[BR]] What is currently needed to do in numpy pylab [[BR]] a=acorr x x.mean ),normed=True)[1][len x) 1 ][[BR]] close )[[BR]] plot a,marker= .
|
|
)[[BR]] And that doesn t count doing fancy markings in the plot like the matlab ones.","enhancement"
|
|
"6013","1531","Where is LU decomposition?
|
|
Trac 933).
|
|
Original ticket http projects.scipy.org numpy ticket 933 on 2008 10 15 by trac user moo, assigned to unknown.
|
|
http www.scipy.org NumPy for Matlab Users Suggests that your package includes LU decomposition.
|
|
It s a good page, aside from being hilariously out of date, as http scipy.org doc numpy api docs numpy.linalg.html seems to suggest it doesn t even exist anymore, which is a lovely bunch of coconuts.
|
|
Has it migrated to somewhere super obvious and I m just an idiot for relying on Google and common sense, or has that functionality been removed as a feature ?","bug"
|
|
"6013","1532","Documentation error in site.cfg.example Trac 934).
|
|
Original ticket http projects.scipy.org numpy ticket 934 on 2008 10 15 by trac user net20, assigned to dmcooke.
|
|
site.cfg.example claims that directories in library dirs should be separated by os.pathsep which is on FreeBSD 7.0, Python 2.5).
|
|
It appears that they should actually be separated by commas, and that colon separated directory lists are silently ignored.","bug"
|
|
"6013","1533","Failure to build install numpy Trac 935).
|
|
Original ticket http projects.scipy.org numpy ticket 935 on 2008 10 15 by trac user asbjorn, assigned to unknown.
|
|
Hi.
|
|
When running the following command python setup.py install prefix= private afel .local , the compiling process begins.
|
|
After 4 seconds, I get this error building py modules sources building extension numpy.core.multiarray sources adding build src.linux x86 64 2.5 numpy core include numpy config.h to sources.
|
|
adding build src.linux x86 64 2.5 numpy core include numpy numpyconfig.h to sources.
|
|
executing numpy core code generators generate array api.py adding build src.linux x86 64 2.5 numpy core include numpy multiarray api.h to sources.
|
|
conv template > build src.linux x86 64 2.5 numpy core src scalartypes.inc Traceback most recent call last) File setup.py , line 96, in <module> setup package ) File setup.py , line 89, in setup package configuration=configuration ) File work afel python numpy 1.1.1 numpy distutils core.py , line 184, in setup return old setup new attr) File private afel .local lib python2.5 distutils core.py , line 151, in setup dist.run commands ) File private afel .local lib python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File private afel .local lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File work afel python numpy 1.1.1 numpy distutils command build.py , line 40, in run old build.run self) File private afel .local lib python2.5 distutils command build.py , line 112, in run self.run command cmd name) File private afel .local lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File private afel .local lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File work afel python numpy 1.1.1 numpy distutils command build src.py , line 130, in run self.build sources ) File work afel python numpy 1.1.1 numpy distutils command build src.py , line 147, in build sources self.build extension sources ext) File work afel python numpy 1.1.1 numpy distutils command build src.py , line 252, in build extension sources sources = self.template sources sources, ext) File work afel python numpy 1.1.1 numpy distutils command build src.py , line 359, in template sources outstr = process c file source) File work afel python numpy 1.1.1 numpy distutils conv template.py , line 252, in process file code = process str .join lines)) File work afel python numpy 1.1.1 numpy distutils conv template.py , line 222, in process str code.extend parse string astr, global names, 0, 1)) File work afel python numpy 1.1.1 numpy distutils conv template.py , line 210, in parse string newcode = parse string text, newenv, newlevel, newline) File work afel python numpy 1.1.1 numpy distutils conv template.py , line 216, in parse string code.append replace re.sub replace, astr)) File work afel python numpy 1.1.1 numpy distutils conv template.py , line 187, in replace raise KeyError, msg KeyError line 558\n form I ve tried with both the numpy 1.1.0 and numpy 1.2.0, both gives the same output.
|
|
Does anyone have any idea regarding what causes this error?","bug"
|
|
"6013","1534","Hashing protocol violation.
|
|
Trac 936).
|
|
Original ticket http projects.scipy.org numpy ticket 936 on 2008 10 15 by charris, assigned to unknown.
|
|
Copied from Robert Kern s post on numpy discussion Actually, the problem is that we provide a hash function explicitly.
|
|
In multiarraymodule.c PyArrayDescr Type.tp hash = hashfunc) Py HashPointer; That is a violation of the hashing protocol objects which compare equal and are hashable need to hash equal), and should be fixed.","bug"
|
|
"6013","1535","Bad complex return values documentation for linalg.lstsq Trac 937).
|
|
Original ticket http projects.scipy.org numpy ticket 937 on 2008 10 17 by nilswagner01, assigned to unknown.
|
|
I am confused by the output of lstsq in case of complex matrices.
|
|
residues Euclidian norm) should be real.
|
|
{{{>>> numpy.
|
|
version 1.3.0.dev5945 }}} If I run the attached file I get residues should be real but is complex [ 0.04002398 0.01310083j] residues is empty [] 0","bug"
|
|
"6013","1536","choose ) much slower than in numarray Trac 938).
|
|
Original ticket http projects.scipy.org numpy ticket 938 on 2008 10 18 by trac user smoerz, assigned to unknown.
|
|
choose is much slower in numpy than in numarray, and even more if an output array is specified, as these tests show import timeit setups = { numarray import numarray as N n1, n2 = 4, 1000000 a1 = N.arange n1 n2, type= Float64 , shape= n1,n2)) a2 = N.arange n1 n2, type= Float64 , shape= n1,n2)) a3 = N.arange n1 n2, type= Float64 , shape= n1,n2)) b1 = N.arange n2, type= Float64 ) b2 = N.arange n2, type= Float64 ) b3 = N.arange n2, type= Float64 ) c = N.remainder N.arange n2, type= Int32 ),2) , numpy import numpy as N n1, n2 = 4, 1000000 a1 = N.arange n1 n2, dtype= Float64 ).reshape n1,n2)) a2 = N.arange n1 n2, dtype= Float64 ).reshape n1,n2)) a3 = N.arange n1 n2, dtype= Float64 ).reshape n1,n2)) b1 = N.arange n2, dtype= Float64 ) b2 = N.arange n2, dtype= Float64 ) b3 = N.arange n2, dtype= Float64 ) c = N.remainder N.arange n2, dtype= Int32 ),2) } stmta = N.choose c, a1, a2)) stmtao = N.choose c, a1, a2), a3) stmtb = N.choose c, b1, b2)) stmtbo = N.choose c, b1, b2), b3) timeit.Timer setup=setups[ numarray ], stmt=stmta).repeat 3,100) [3.3187780380249023, 3.2966721057891846, 3.3234250545501709] timeit.Timer setup=setups[ numpy ], stmt=stmta).repeat 3,100) [14.842453002929688, 14.833296060562134, 14.836632966995239] timeit.Timer setup=setups[ numarray ], stmt=stmtao).repeat 3,100) [3.1973719596862793, 3.2031948566436768, 3.2093629837036133] timeit.Timer setup=setups[ numpy ], stmt=stmtao).repeat 3,100) [17.546916007995605, 17.548220157623291, 17.536314010620117] timeit.Timer setup=setups[ numarray ], stmt=stmtb).repeat 3,100) [0.6694338321685791, 0.66939401626586914, 0.67307686805725098] timeit.Timer setup=setups[ numpy ], stmt=stmtb).repeat 3,100) [3.7615809440612793, 3.7627589702606201, 3.7547731399536133] timeit.Timer setup=setups[ numarray ], stmt=stmtbo).repeat 3,100) [0.67037606239318848, 0.67186903953552246, 0.66994881629943848] timeit.Timer setup=setups[ numpy ], stmt=stmtbo).repeat 3,100) [4.4750981330871582, 4.4650890827178955, 4.4679431915283203]","bug"
|
|
"6013","1537","Reference counting bug in PyArray TakeFrom in exceptional case Trac 939).
|
|
Original ticket http projects.scipy.org numpy ticket 939 on 2008 10 20 by mdboom, assigned to unknown.
|
|
When calling .take ) with an out kwarg and PyArray TakeFrom throws an exception because the indices array is not convertible to PyArray INTP, the out object is incorrectly decref d, causing a segfault in the Python garbage collector upon exit.
|
|
I have attached a script to reproduce this error, and a proposed patch.
|
|
The patch Py XINCREFs out before going to the cleanup fail section of PyArray TakeFrom, since fail assumes an extra reference count.
|
|
An alternative is to add a second goto statement below the PyArray XDECREF ERR ret) and go there instead.
|
|
Initially reported in matplotlib tracker by Dav Clark here https sourceforge.net tracker index.php?func=detail aid=2173204 group id=80706 atid=560720","bug"
|
|
"6013","1538","Add bzip2 support to numpy.loadtxt Trac 940).
|
|
Original ticket http projects.scipy.org numpy ticket 940 on 2008 10 22 by dopplershift, assigned to cournape.
|
|
numpy.loadtxt currently has support for gzipped files, but not for bz2 d files.
|
|
The attached 3 line patch adds in support for reading bzipped files.","enhancement"
|
|
"6013","1539","Add bzip2 support to numpy.loadtxt Trac 941).
|
|
Original ticket http projects.scipy.org numpy ticket 941 on 2008 10 22 by dopplershift, assigned to unknown.
|
|
numpy.loadtxt currently has support for gzipped files, but not for bz2 d files.
|
|
The attached 3 line patch adds in support for reading bzipped files.","enhancement"
|
|
"6013","1540","installation problem > cannot find lgcc s Trac 942).
|
|
Original ticket http projects.scipy.org numpy ticket 942 on 2008 10 23 by trac user christoph, assigned to unknown.
|
|
I m a newbie...
|
|
I have a problem installing the numpy packages tried 1.2.0 and down to 1.0.4).
|
|
I couldnt fix my problem although I searched information in the web...
|
|
There is a problem after python setup.py install , usr bin ld skipping incompatible usr lib liblapack.so when searching for lla pack usr bin ld skipping incompatible usr lib liblapack.a when searching for llap ack usr bin ld skipping incompatible usr lib libblas.so when searching for lblas usr bin ld skipping incompatible usr lib libblas.a when searching for lblas usr bin ld skipping incompatible usr lib libg2c.so when searching for lg2c usr bin ld skipping incompatible usr lib libg2c.a when searching for lg2c usr bin ld skipping incompatible usr lib libfrtbegin.a when searching for lf rtbegin usr bin ld skipping incompatible usr lib libg2c.so when searching for lg2c usr bin ld skipping incompatible usr lib libg2c.a when searching for lg2c usr bin ld cannot find lgcc s collect2 ld returned 1 exit status usr bin ld skipping incompatible usr lib liblapack.so when searching for lla pack usr bin ld skipping incompatible usr lib liblapack.a when searching for llap ack usr bin ld skipping incompatible usr lib libblas.so when searching for lblas usr bin ld skipping incompatible usr lib libblas.a when searching for lblas usr bin ld skipping incompatible usr lib libg2c.so when searching for lg2c usr bin ld skipping incompatible usr lib libg2c.a when searching for lg2c usr bin ld skipping incompatible usr lib libfrtbegin.a when searching for lf rtbegin usr bin ld skipping incompatible usr lib libg2c.so when searching for lg2c usr bin ld skipping incompatible usr lib libg2c.a when searching for lg2c usr bin ld cannot find lgcc s collect2 ld returned 1 exit status error Command usr local bin g77 g Wall g Wall shared build temp.linux x8 6 64 2.5 numpy linalg lapack litemodule.o build temp.linux x86 64 2.5 numpy lina lg python xerbla.o L usr lib llapack lblas lg2c o build lib.linux x86 64 2.
|
|
5 numpy linalg lapack lite.so failed with exit status 1 I saw the same problem here posted in another Ticket sorry, cant find it again!
|
|
), where the person posted unsetenv cflags was the solution.
|
|
I tried this, but I could not solve my problem.
|
|
Sorry, I m quite a newbie, so I would be very glad if you could help me...
|
|
Thank you very very much, christoph","bug"
|
|
"6013","1541","Add blkdiag command from MATLAB Trac 943).
|
|
Original ticket http projects.scipy.org numpy ticket 943 on 2008 10 27 by trac user chrismurf, assigned to unknown.
|
|
A command similar to MATLAB s blkdiag would be handy in Numpy.
|
|
I m not sure what the correct component is for this I can imagine it going in the normal namespace like diag , but tagged the numpy.linalg component in case it s deemed unworthy of the normal namespace.
|
|
I ve attached an example implementation, which almost certainly breaks every Numpy coding convention but is functional).
|
|
If there s interest but the code quality is not acceptable, I m happy to clean it up for inclusion.","enhancement"
|
|
"6013","1542","polyint ) does not work with complex arrays Trac 944).
|
|
Original ticket http projects.scipy.org numpy ticket 944 on 2008 10 27 by rkern, assigned to rkern.
|
|
The return array is always created with dtype=float.","bug"
|
|
"6013","1543","Convert dotblas to prefixed numpy headers Trac 945).
|
|
Original ticket http projects.scipy.org numpy ticket 945 on 2008 10 28 by cournape, assigned to cournape.","bug"
|
|
"6013","1544","Segmentation fault with numpy.clip char array Trac 946).
|
|
Original ticket http projects.scipy.org numpy ticket 946 on 2008 11 03 by pv, assigned to unknown.
|
|
This crashes eg.
|
|
on numpy trunk r5968) import numpy obj = numpy.complex ) obj.clip rrr ,1) See the report in Scipy Trac from http scipy.org scipy scipy ticket 716","bug"
|
|
"6013","1545","Segmentation fault with tile char array Trac 947).
|
|
Original ticket http projects.scipy.org numpy ticket 947 on 2008 11 03 by pv, assigned to unknown.
|
|
This crashes numpy r5968) import numpy numpy.tile 1, A 33) See the original bug report in Scipy Trac http scipy.org scipy scipy ticket 717","bug"
|
|
"6013","1546","Incorrect shape after basic arithmetic involving masked objects Trac 948).
|
|
Original ticket http projects.scipy.org numpy ticket 948 on 2008 11 04 by trac user dharland, assigned to unknown.
|
|
Adding ma.masked to an array of size 1 may change its shape.
|
|
a = ma.arange 10) print a[1 2] 1).shape > 1,) print a[1 2] ma.masked).shape > ) INCORRECT, should be 1,) Possible patch attached.","bug"
|
|
"6013","1547","gzip support for ndarray.tofile ) Trac 949).
|
|
Original ticket http projects.scipy.org numpy ticket 949 on 2008 11 05 by trac user hoffman, assigned to unknown.
|
|
It would be very useful if ndarray.tofile ) supported gzip.GzipFile objects as well as files.
|
|
Previous discussion on this issue http groups.google.com group comp.lang.python browse thread thread 55f817654332b57b","enhancement"
|
|
"6013","1548","multiply empty arrays gives nonzero result Trac 950).
|
|
Original ticket http projects.scipy.org numpy ticket 950 on 2008 11 05 by trac user pierre, assigned to unknown.
|
|
Hi, I found this bug on mac os 10.5, with power pc and also on an intel computer but not on a 64bit linux).
|
|
I used numpy 1.1.1, 1.2, and the latest svn.
|
|
In [6] np.dot np.array []).reshape 1,0), np.array []).reshape 0,1)) Out[6] array [[ 4.00002862]]) In [7] np.dot np.array []).reshape 2,0), np.array []).reshape 0,3)) Out[7] array [[ 0., 0., 0.
|
|
], [ 0., 0., 0.]])
|
|
Thanks!
|
|
Pierre","bug"
|
|
"6013","1549","test memmap.py leaves tmpfiles on RHE4 systems Trac 951).
|
|
Original ticket http projects.scipy.org numpy ticket 951 on 2008 11 06 by chanley, assigned to unknown.
|
|
This was discovered by one of my co workers.
|
|
I have been able to reproduce this bug on a RHE4 system.
|
|
I was not able to reproduce it on my Mac OS X 10.5 machine.
|
|
I believe the test just needs to do a little cleanup when finishing.
|
|
python import numpy numpy.test ) leaves 6 files behind in tmp.
|
|
rw r r 1 sienkiew science 448 Nov 6 09 40 tmphdFtvk rw r r 1 sienkiew science 11 Nov 6 09 40 tmp 9fU 9 rw 1 sienkiew science 48 Nov 6 09 40 mmapnt7S 3 rw 1 sienkiew science 48 Nov 6 09 40 mmapRJlNsL rw 1 sienkiew science 0 Nov 6 09 40 mmapHGlGTq rw 1 sienkiew science 48 Nov 6 09 40 mmapFllGqi I believe the offending source code is in numpy 1.1.0 numpy core tests test memmap.py This behavior exists in the current svn trunk.","bug"
|
|
"6013","1550","Pickling an element of a recarray is silently broken Trac 952).
|
|
Original ticket http projects.scipy.org numpy ticket 952 on 2008 11 07 by trac user philbinj, assigned to unknown.
|
|
As per the summary.
|
|
Unpickling a pickled recarray element returns garbage data.
|
|
With numpy.
|
|
version == 1.3.0.dev5883 In [1] import numpy as np In [2] dtype = [ r , f4 ), g , f4 ), b , f4 )] In [3] arr = np.ones 10,), dtype=dtype) In [4] arr Out[4] array [ 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0), 1.0, 1.0, 1.0)], dtype=[ r , <f4 ), g , <f4 ), b , <f4 )]) In [5] arr[0] Out[5] 1.0, 1.0, 1.0) In [6] import cPickle; cPickle.loads cPickle.dumps arr[0])) Out[6] 0.0, 0.0, 1.8643547392640242e 38)","bug"
|
|
"6013","1551","Typos in docstrings for matlib.ones and matlib.zeros cause doctest failures Trac 953).
|
|
Original ticket http projects.scipy.org numpy ticket 953 on 2008 11 12 by dopplershift, assigned to unknown.
|
|
Here s a quick diff to fix some typos in the docstrings for matlib.zeros and matlib.ones.
|
|
They re causing 2 of many) failures in the doctests for me.","bug"
|
|
"6013","1552","Add memapping to numpy.load Trac 954).
|
|
Original ticket http projects.scipy.org numpy ticket 954 on 2008 11 19 by GaelVaroquaux, assigned to unknown.
|
|
Attached is a patch to add memmapping to numpy.load.
|
|
This is a very rough implementation, but it is better than nothing, IMHO.","enhancement"
|
|
"6013","1553","[debian 505999] memory leak in exponentiation Trac 955).
|
|
Original ticket http projects.scipy.org numpy ticket 955 on 2008 11 20 by stefanv, assigned to unknown.
|
|
From http bugs.debian.org cgi bin bugreport.cgi?bug=505999 Version 1 1.1.0 3 Raising zero to a negative power sometimes leaks memory, depending on the types involved.
|
|
While this is easy to work around check for zero and deal with it separately), it shouldn t happen.
|
|
In particular, with the following combinations of types where the base is 0 and the exponent is 3 or 3.5), one object is leaked float numpy.float64 int numpy.float64 int numpy.int32 int numpy.int64 numpy.float32 numpy.float32 numpy.float64 float numpy.float64 int numpy.float64 numpy.float32 numpy.float64 numpy.float64 numpy.float64 numpy.int32 numpy.float64 numpy.int64 numpy.int32 int numpy.int32 numpy.int32 numpy.int64 int numpy.int64 numpy.int32 numpy.int64 numpy.int64 In addition, float numpy.int32 raises an exception; however, strictly speaking, that would be a separate issue.)
|
|
Code to reproduce this is enclosed below.
|
|
!
|
|
usr bin python import gc, numpy def mem ) gc.collect ) return len gc.get objects )) for zt in int, float, numpy.int32, numpy.int64, numpy.float32, numpy.float64 z = zt 0) for pt in int, float, numpy.int32, numpy.int64, numpy.float32, numpy.float64 if zt in int, float) and pt in int, float) continue p = pt 3.5) try before = mem ) z p if mem )>before print leak , type z), , type p) else print OK , type z), , type p) except print error , type z), , type p)","bug"
|
|
"6013","1554","ValueError cannot convert float NaN to integer Trac 956).
|
|
Original ticket http projects.scipy.org numpy ticket 956 on 2008 11 20 by cournape, assigned to pierregm.
|
|
Python 2.6 specific error ERROR Tests the min max functions with explicit outputs Traceback most recent call last) File export bbc8 python2.6 lib python2.6 site packages numpy ma tests test core.py , line 700, in test minmax funcs with output result = npfunc xm,axis=0,out=nout) File export bbc8 python2.6 stow numpy.svn lib python2.6 site packages numpy core fromnumeric.py , line 1569, in amin return amin axis, out) File export bbc8 python2.6 stow numpy.svn lib python2.6 site packages numpy ma core.py , line 3188, in min np.putmask out, newmask, np.nan) ValueError cannot convert float NaN to integer","bug"
|
|
"6013","1555","decorators don t work with test generators Trac 957).
|
|
Original ticket http projects.scipy.org numpy ticket 957 on 2008 11 24 by jarrodmillman, assigned to alanmcintyre.
|
|
Nose has some problems with test generators in classes inherited from `TestCase`.
|
|
For example, dec.knownfailureif True, This test is known to fail ) def test discrete rvs cdf fail ) distknownfail = [ [ logser , 0.6,)]] for distname, arg in distknownfail if debug print distname yield check discrete chisquare, distname, arg results in ...........E ====================================================================== ERROR test discrete chisquare.test discrete rvs cdf fail Traceback most recent call last) File C \Programs\Python25\lib\site packages\nose 0.10.4 py2.5.egg\nose\case.p y , line 182, in runTest self.test self.arg) File C \Programs\Python25\lib\site packages\numpy\testing\decorators.py , line 119, in skipper raise KnownFailureTest, msg KnownFailureTest This test is known to fail Ran 12 tests in 0.781s FAILED errors=1)","bug"
|
|
"6013","1556","numpy failures on mac os X python 2.6) Trac 958).
|
|
Original ticket http projects.scipy.org numpy ticket 958 on 2008 11 24 by cournape, assigned to cournape.
|
|
====================================================================== ERROR Tests the min max functions with explicit outputs Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy ma tests test core.py , line 704, in test minmax funcs with output result = npfunc xm,axis=0,out=nout) File Users david pylocal lib python2.6 site packages numpy core fromnumeric.py , line 1569, in amin return amin axis, out) File Users david pylocal lib python2.6 site packages numpy ma core.py , line 3192, in min np.putmask out, newmask, np.nan) ValueError cannot convert float NaN to integer ====================================================================== FAIL test basic test multiarray.TestClip) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 677, in test basic self.
|
|
clip type float ,1024, 12.8,100.2, inplace=inplace) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 671, in clip type assert equal x.dtype.byteorder,byteorder) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL > DESIRED = ====================================================================== FAIL test binary test multiarray.TestFromstring) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 120, in test binary assert array equal a, array [1,2,3,4])) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 303, in assert array equal verbose=verbose, header= Arrays are not equal ) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [ 4.60060299e 41, 8.96831017e 44, 2.30485571e 41, 4.60074312e 41], dtype=float32) y array [1, 2, 3, 4]) ====================================================================== FAIL test version0 float32 test multiarray.TestPickling) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 543, in test version0 float32 assert equal a, p) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 180, in assert equal return assert array equal actual, desired, err msg, verbose) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 303, in assert array equal verbose=verbose, header= Arrays are not equal ) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [ 1., 2., 3., 4.
|
|
], dtype=float32) y array [ 4.60060299e 41, 8.96831017e 44, 2.30485571e 41, 4.60074312e 41], dtype=float32) ====================================================================== FAIL test version1 float32 test multiarray.TestPickling) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 562, in test version1 float32 assert equal a, p) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 180, in assert equal return assert array equal actual, desired, err msg, verbose) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 303, in assert array equal verbose=verbose, header= Arrays are not equal ) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [ 1., 2., 3., 4.
|
|
], dtype=float32) y array [ 4.60060299e 41, 8.96831017e 44, 2.30485571e 41, 4.60074312e 41], dtype=float32) ====================================================================== FAIL test basic test multiarray.TestView) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 894, in test basic assert array equal y, [67305985, 134678021]) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 303, in assert array equal verbose=verbose, header= Arrays are not equal ) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [16909060, 84281096]) y array [ 67305985, 134678021]) ====================================================================== FAIL test keywords test multiarray.TestView) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test multiarray.py , line 904, in test keywords assert array equal y,[[513]]) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 303, in assert array equal verbose=verbose, header= Arrays are not equal ) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [[258]], dtype=int16) y array [[513]]) ====================================================================== FAIL Test non native double input with scalar min max.
|
|
Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 437, in test simple nonnative a = self.
|
|
generate non native data self.nr, self.nc) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 396, in generate non native data assert not data.dtype.isnative AssertionError ====================================================================== FAIL Test NON native with native array min max.
|
|
Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 632, in test type cast 07 assert not a s.dtype.isnative AssertionError ====================================================================== FAIL Test NON native with native scalar min max.
|
|
Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 643, in test type cast 08 assert not a s.dtype.isnative AssertionError ====================================================================== FAIL Test native with NON native array min max.
|
|
Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 654, in test type cast 09 assert not m s.dtype.isnative AssertionError ====================================================================== FAIL Test non native with native scalar, min max, out non native Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 671, in test type cast 11 a = self.
|
|
generate non native data self.nr, self.nc) File Users david pylocal lib python2.6 site packages numpy core tests test numeric.py , line 396, in generate non native data assert not data.dtype.isnative AssertionError ====================================================================== FAIL Check creation from list of list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 167, in test list of list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 158, in test list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 149, in test tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 167, in test list of list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 158, in test list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 149, in test tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 167, in test list of list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 158, in test list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 149, in test tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 167, in test list of list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from list of tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 158, in test list of tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation from tuples Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 149, in test tuple self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of 0 dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 106, in test zeros0D self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of multi dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 126, in test zerosMD self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of single dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 116, in test zerosSD self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of 0 dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 106, in test zeros0D self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of multi dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 126, in test zerosMD self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL Check creation of single dimensional objects Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test numerictypes.py , line 116, in test zerosSD self.assert normalize descr self.
|
|
descr) == h.dtype.descr) AssertionError ====================================================================== FAIL test method array test records.TestFromrecords) Traceback most recent call last) File Users david pylocal lib python2.6 site packages numpy core tests test records.py , line 13, in test method array assert equal r[1].item ), 25444, efg , 1633837924)) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 176, in assert equal assert equal actual[k], desired[k], item= r\n s k,err msg), verbose) File Users david pylocal lib python2.6 site packages numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal item=0 ACTUAL 25699 DESIRED 25444 ====================================================================== FAIL test umath.TestComplexFunctions.test against cmath Traceback most recent call last) File Users david pylocal lib python2.6 site packages nose case.py , line 182, in runTest self.test self.arg) File Users david pylocal lib python2.6 site packages numpy core tests test umath.py , line 423, in test against cmath assert abs a b) < atol, s s s; cmath s fname,p,a,b) AssertionError arcsin 2 1.57079632679 1.31695789692j); cmath 1.57079632679 1.31695789692j) Ran 1769 tests in 10.913s","bug"
|
|
"6013","1557","atanh 1) gives nan instead of inf on Windows Trac 959).
|
|
Original ticket http projects.scipy.org numpy ticket 959 on 2008 11 25 by FrancescAlted, assigned to unknown.
|
|
There is an inconsistency in the positive limit value 1.)
|
|
of arctanh between 1.2.x and 1.3.x in trunk.
|
|
When using !NumPy 1.2.1 one has Python 2.4.4 71, Oct 18 2006, 08 34 43) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.2.1 >>> numpy.arctanh 1.)
|
|
1.
|
|
INF >>> numpy.isinf <ufunc isinf > >>> numpy.isinf numpy.arctanh 1.))
|
|
True >>> numpy.arctanh 1.)
|
|
1.
|
|
INF >>> numpy.isinf numpy.arctanh 1.))
|
|
True while when using version 1.3.0.dev6085 Python 2.6 r26 66721, Oct 2 2008, 11 35 03) [MSC v.1500 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.3.0.dev6085 >>> numpy.arctanh 1.)
|
|
nan >>> numpy.isinf numpy.arctanh 1.))
|
|
False >>> numpy.arctanh 1.)
|
|
inf >>> numpy.isinf numpy.arctanh 1.))
|
|
True As you see, the trunk version returns ``nan`` for arctanh 1.
|
|
), while 1.2.1 returns ``inf`` the correct value).
|
|
For arctanh 1.)
|
|
both versions correctly returns `` inf``.
|
|
I used the official binaries for 1.2.1, while I ve used the MSVC 2008 32 bit) for compiling trunk the resuilting binaries works badly in both Windows XP 32 bit and Windows Vista 64 bit).
|
|
My experiments on Linux shows that they both return `` inf`` and `` inf``, so it seems that this is a Windows specific issue.","bug"
|
|
"6013","1558","Intel64 compiler=intelem fails to set correct compiler options Trac 960).
|
|
Original ticket http projects.scipy.org numpy ticket 960 on 2008 11 26 by trac user em101, assigned to dmcooke.
|
|
On an Intel64 Linux x86 64 platform, using compiler=intelem is not fully supported fails to set correct compiler options), however fcompiler=intelem is supported.
|
|
Setting CC and CXX and using compiler=unix is a work around.","bug"
|
|
"6013","1559","Masked array getdata ) and getmask ) resolve to get data ) and get mask ) in docs Trac 961).
|
|
Original ticket http projects.scipy.org numpy ticket 961 on 2008 11 28 by scottza, assigned to pierregm.
|
|
The getdata ) and getmask ) functions in the numpy.ma module resolve to get data ) and get mask ).
|
|
>>> import numpy.ma as ma >>> ma.getdata <function get data at 0xb7514e2c> >>> ma.getmask <function get mask at 0xb665f3e4> getmask ) and getdata ) are exposed in the top level namespace numpy.ma, while get data ) and get mask ) are only available under numpy.ma.core.
|
|
>>> ma.get mask Traceback most recent call last) File <stdin> , line 1, in <module> AttributeError module object has no attribute get mask >>> ma.core.get mask <function get mask at 0xb665f3e4> >>> ma.get data Traceback most recent call last) File <stdin> , line 1, in <module> AttributeError module object has no attribute get data >>> ma.core.get data <function get data at 0xb7514e2c> This is an issue in the documentation effort since the documentation for getmask getdata will link to the ma.core functions get mask get data.
|
|
There is no easy way to resolve this in the documentation.
|
|
Therefore, since this can be confusing to users and is easily fixed in a backward compatible way that doesn t break Numpy, I suggest that the attached patch against r6116 be applied.","bug"
|
|
"6013","1560","f2py external callbacks don t work Trac 962).
|
|
Original ticket http projects.scipy.org numpy ticket 962 on 2008 11 28 by trac user wieferink, assigned to pearu.
|
|
I have attached an extension module using a callback which relies on the existence of an external name the callback is not passed as an argument).
|
|
The f2py module in the tarball builds fine, but according to Pearu Peterson should not add cb func to the interface of cb func because intent callback) means something different in this context.
|
|
Pearu Peterson wrote Adding intent callback) cb func to argument list is a f2py bug.
|
|
Actually, the message analyzeline missing user module could be nothing) is indicating this bug I didn t implement the could be nothing part;) and missing module note is misleading.).
|
|
Could someone report this bug to numpy tracker?
|
|
The fix itself should be simple locate the line that outputs the message in f2py crackfortran.py file and intent the next line by one level if user not in groupcache[groupcounter 2][ name ] outmess analyzeline missing user module could be nothing)\n ) groupcache[groupcounter][ args ].append k)","bug"
|
|
"6013","1561","Object array comparisons eat exceptions Trac 963).
|
|
Original ticket http projects.scipy.org numpy ticket 963 on 2008 11 29 by inducer, assigned to cournape.
|
|
The attached Python file reports False instead of propagating the RuntimeError that s thrown in A s eq .
|
|
Errors should never pass silently.
|
|
Unless explicitly silenced.","bug"
|
|
"6013","1562","f2py error with python 2.6 Trac 964).
|
|
Original ticket http projects.scipy.org numpy ticket 964 on 2008 12 02 by trac user edlenz, assigned to unknown.
|
|
I am using the latest svn varsion of numpy and the following error ocoured when using f2py?
|
|
python Python 2.6 r26 66714, Oct 27 2008, 10 50 31) [GCC 4.3.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import hello >>> hello.foo 4) Traceback most recent call last) File <stdin> , line 1, in <module> RuntimeError more argument specifiers than keyword list entries remaining format | hello.foo ) >>> I also tried with other functions and the error is still the same.
|
|
Thanks for your help","bug"
|
|
"6013","1563","random integers error regarding integer value range Trac 965).
|
|
Original ticket http projects.scipy.org numpy ticket 965 on 2008 12 05 by cournape, assigned to cournape.
|
|
!python import numpy as np np.random.random integers 2 30, 2 30, 10) Raises ValueError low >= high It may not be obvious what that means for the user.","bug"
|
|
"6013","1564","Enhance meshgrid to generate 3D grids add option for sparse grids Trac 966).
|
|
Original ticket http projects.scipy.org numpy ticket 966 on 2008 12 09 by pbrod, assigned to charris.
|
|
I would suggest to replace numpy s meshgrid with the meshgrid found in SciTools at http code.google.com p scitools for the following reasons The `meshgrid` function in NumPy is limited to two dimensions only, while the SciTools version can also work with 3D grids.
|
|
In addition, the NumPy version of `meshgrid` has no option for generating sparse grids to conserve memory, like it is in SciTools by specifying the `sparse` argument >>> xv, yv = meshgrid linspace 2,2,5), linspace 1,1,3), sparse=True) >>> xv array [[ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1.
|
|
], [ 0.
|
|
], [ 1.]])
|
|
Actually, this is the default behavior for the `meshgrid` function in SciTools.
|
|
In NumPy, however, we will in this case get a full 2D grid >>> xv, yv = numpy.meshgrid linspace 2,2,5), linspace 1,1,3)) >>> xv array [[ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1., 1., 1., 1., 1.
|
|
], [ 0., 0., 0., 0., 0.
|
|
], [ 1., 1., 1., 1., 1.]])
|
|
This is the same result we get by setting `sparse=False` in `meshgrid` in SciTools.
|
|
The NumPy functions `mgrid` and `ogrid` does provide support for, respectively, full and sparse n dimensional meshgrids, however, these functions uses slices to generate the meshgrids rather than one dimensional coordinate arrays such as in Matlab.
|
|
With slices, the user does not have the option to generate meshgrid with, e.g., irregular spacings, like >>> x = array [ 1, 0.5,1,4,5], float) >>> y = array [0, 2, 5], float) >>> xv, yv = meshgrid x, y, sparse=False) >>> xv array [[ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5. ]])
|
|
>>> yv array [[ 0., 0., 0., 0., 0.
|
|
], [ 2., 2., 2., 2., 2.
|
|
], [ 5., 5., 5., 5., 5.]])
|
|
In addition to the reasons mentioned above, the meshgrid function in NumPy supports only Cartesian indexing, i.e., x and y, not matrix indexing, i.e., rows and columns `mgrid` and `ogrid` supports only matrix indexing).
|
|
The `meshgrid` function in SciTools supports both indexing conventions through the `indexing` keyword argument.
|
|
Giving the string ` ij ` returns a meshgrid with matrix indexing, while ` xy ` returns a meshgrid with Cartesian indexing.
|
|
The difference is illustrated by the following code snippet nx = 10 ny = 15 x = linspace 2,2,nx) y = linspace 2,2,ny) xv, yv = meshgrid x, y, sparse=False, indexing= ij ) for i in range nx) for j in range ny) treat xv[i,j], yv[i,j] xv, yv = meshgrid x, y, sparse=False, indexing= xy ) for i in range nx) for j in range ny) treat xv[j,i], yv[j,i] It is not entirely true that matrix indexing is not supported by the `meshgrid` function in NumPy because we can just switch the order of the first two input and output arguments yv, xv = numpy.meshgrid y, x) is the same as xv, yv = meshgrid x, y, sparse=False, indexing= ij ) However, we think it is clearer to have the logical x, y sequence on the left hand side and instead adjust a keyword argument.","enhancement"
|
|
"6013","1565","Enhance meshgrid to generate 3D grids add option for sparse grids Trac 967).
|
|
Original ticket http projects.scipy.org numpy ticket 967 on 2008 12 09 by pbrod, assigned to unknown.
|
|
I would suggest to replace numpy s meshgrid with the meshgrid found in SciTools at http code.google.com p scitools for the following reasons The `meshgrid` function in NumPy is limited to two dimensions only, while the SciTools version can also work with 3D grids.
|
|
In addition, the NumPy version of `meshgrid` has no option for generating sparse grids to conserve memory, like it is in SciTools by specifying the `sparse` argument >>> xv, yv = meshgrid linspace 2,2,5), linspace 1,1,3), sparse=True) >>> xv array [[ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1.
|
|
], [ 0.
|
|
], [ 1.]])
|
|
Actually, this is the default behavior for the `meshgrid` function in SciTools.
|
|
In NumPy, however, we will in this case get a full 2D grid >>> xv, yv = numpy.meshgrid linspace 2,2,5), linspace 1,1,3)) >>> xv array [[ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1., 1., 1., 1., 1.
|
|
], [ 0., 0., 0., 0., 0.
|
|
], [ 1., 1., 1., 1., 1.]])
|
|
This is the same result we get by setting `sparse=False` in `meshgrid` in SciTools.
|
|
The NumPy functions `mgrid` and `ogrid` does provide support for, respectively, full and sparse n dimensional meshgrids, however, these functions uses slices to generate the meshgrids rather than one dimensional coordinate arrays such as in Matlab.
|
|
With slices, the user does not have the option to generate meshgrid with, e.g., irregular spacings, like >>> x = array [ 1, 0.5,1,4,5], float) >>> y = array [0, 2, 5], float) >>> xv, yv = meshgrid x, y, sparse=False) >>> xv array [[ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5. ]])
|
|
>>> yv array [[ 0., 0., 0., 0., 0.
|
|
], [ 2., 2., 2., 2., 2.
|
|
], [ 5., 5., 5., 5., 5.]])
|
|
In addition to the reasons mentioned above, the meshgrid function in NumPy supports only Cartesian indexing, i.e., x and y, not matrix indexing, i.e., rows and columns `mgrid` and `ogrid` supports only matrix indexing).
|
|
The `meshgrid` function in SciTools supports both indexing conventions through the `indexing` keyword argument.
|
|
Giving the string ` ij ` returns a meshgrid with matrix indexing, while ` xy ` returns a meshgrid with Cartesian indexing.
|
|
The difference is illustrated by the following code snippet nx = 10 ny = 15 x = linspace 2,2,nx) y = linspace 2,2,ny) xv, yv = meshgrid x, y, sparse=False, indexing= ij ) for i in range nx) for j in range ny) treat xv[i,j], yv[i,j] xv, yv = meshgrid x, y, sparse=False, indexing= xy ) for i in range nx) for j in range ny) treat xv[j,i], yv[j,i] It is not entirely true that matrix indexing is not supported by the `meshgrid` function in NumPy because we can just switch the order of the first two input and output arguments yv, xv = numpy.meshgrid y, x) is the same as xv, yv = meshgrid x, y, sparse=False, indexing= ij ) However, we think it is clearer to have the logical x, y sequence on the left hand side and instead adjust a keyword argument.","enhancement"
|
|
"6013","1566","Enhance meshgrid to generate 3D grids add option for sparse grids Trac 968).
|
|
Original ticket http projects.scipy.org numpy ticket 968 on 2008 12 09 by pbrod, assigned to unknown.
|
|
I would suggest to replace numpy s meshgrid with the meshgrid found in SciTools at http code.google.com p scitools for the following reasons The `meshgrid` function in NumPy is limited to two dimensions only, while the SciTools version can also work with 3D grids.
|
|
In addition, the NumPy version of `meshgrid` has no option for generating sparse grids to conserve memory, like it is in SciTools by specifying the `sparse` argument >>> xv, yv = meshgrid linspace 2,2,5), linspace 1,1,3), sparse=True) >>> xv array [[ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1.
|
|
], [ 0.
|
|
], [ 1.]])
|
|
Actually, this is the default behavior for the `meshgrid` function in SciTools.
|
|
In NumPy, however, we will in this case get a full 2D grid >>> xv, yv = numpy.meshgrid linspace 2,2,5), linspace 1,1,3)) >>> xv array [[ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.
|
|
], [ 2., 1., 0., 1., 2.]])
|
|
>>> yv array [[ 1., 1., 1., 1., 1.
|
|
], [ 0., 0., 0., 0., 0.
|
|
], [ 1., 1., 1., 1., 1.]])
|
|
This is the same result we get by setting `sparse=False` in `meshgrid` in SciTools.
|
|
The NumPy functions `mgrid` and `ogrid` does provide support for, respectively, full and sparse n dimensional meshgrids, however, these functions uses slices to generate the meshgrids rather than one dimensional coordinate arrays such as in Matlab.
|
|
With slices, the user does not have the option to generate meshgrid with, e.g., irregular spacings, like >>> x = array [ 1, 0.5,1,4,5], float) >>> y = array [0, 2, 5], float) >>> xv, yv = meshgrid x, y, sparse=False) >>> xv array [[ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5.
|
|
], [ 1. , 0.5, 1. , 4. , 5. ]])
|
|
>>> yv array [[ 0., 0., 0., 0., 0.
|
|
], [ 2., 2., 2., 2., 2.
|
|
], [ 5., 5., 5., 5., 5.]])
|
|
In addition to the reasons mentioned above, the meshgrid function in NumPy supports only Cartesian indexing, i.e., x and y, not matrix indexing, i.e., rows and columns `mgrid` and `ogrid` supports only matrix indexing).
|
|
The `meshgrid` function in SciTools supports both indexing conventions through the `indexing` keyword argument.
|
|
Giving the string ` ij ` returns a meshgrid with matrix indexing, while ` xy ` returns a meshgrid with Cartesian indexing.
|
|
The difference is illustrated by the following code snippet nx = 10 ny = 15 x = linspace 2,2,nx) y = linspace 2,2,ny) xv, yv = meshgrid x, y, sparse=False, indexing= ij ) for i in range nx) for j in range ny) treat xv[i,j], yv[i,j] xv, yv = meshgrid x, y, sparse=False, indexing= xy ) for i in range nx) for j in range ny) treat xv[j,i], yv[j,i] It is not entirely true that matrix indexing is not supported by the `meshgrid` function in NumPy because we can just switch the order of the first two input and output arguments yv, xv = numpy.meshgrid y, x) is the same as xv, yv = meshgrid x, y, sparse=False, indexing= ij ) However, we think it is clearer to have the logical x, y sequence on the left hand side and instead adjust a keyword argument.","enhancement"
|
|
"6013","1567","gfortran does not work on windows when built with MS compiler Trac 969).
|
|
Original ticket http projects.scipy.org numpy ticket 969 on 2008 12 16 by cournape, assigned to cournape.
|
|
Looks like some distutils voodoo is needed c \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\link.exe DLL nologo INCREMENTAL NO LIBPATH c \mingw\lib\gcc\mingw32\4.3.0 LIBPATH C \Python25\libs LIBPATH C \Python25\PCBuild LIBPATH build\temp.win32 2.5\Release LIBPATH C \Python25\libs LIBPATH C \Python25\PCBuild gfortran.lib gcc.lib EXPORT initmultiarray build\temp.win32 2.5\Release\numpy\core\src\multiarraymodule.obj OUT build\lib.win32 2.5\numpy\core\multiarray.pyd IMPLIB build\temp.win32 2.5\Release\numpy\core\src\multiarray.lib LINK fatal error LNK1181 cannot open input file gfortran.lib This is a problem for win64, as g77 is not ported there, only gfortran is.","bug"
|
|
"6013","1568","Problems building NumPy with Python 2.6 MinGW 3.4.5 on Windows XP SP3 Trac 970).
|
|
Original ticket http projects.scipy.org numpy ticket 970 on 2008 12 18 by trac user fpetermaas, assigned to cournape.
|
|
I cannot build NumPy 1.2.1 with the aforementionend environment.
|
|
It works with Python 2.5 but with Python 2.6 I get >>> python setup.py build c mingw32 [lots of lines omitted] running build src building py modules sources building extension numpy.core.multiarray sources Generating build\src.win32 2.6\numpy\core\include numpy\config.h error None","bug"
|
|
"6013","1569","numpy.memmap offset parameter docs are almost entirely wrong.
|
|
Trac 971).
|
|
Original ticket http projects.scipy.org numpy ticket 971 on 2008 12 19 by trac user 0ion9, assigned to pv.
|
|
In the file, array data starts at this offset.
|
|
offset should be a multiple of the byte size of dtype.
|
|
Requires shape=None.
|
|
The default is 0 In actual fact, shape != None is usable with offset, and crucial when the file only contains some records of the type you want to memmap, with other data included in the file.
|
|
Supposing you have a file of 27 bytes, with a 7 byte header and 10 byte records.
|
|
Let s create one echo n HEADER!
|
|
onerecord tworecord > tmp npmmap If you want to read the records, you can do >>> mydtype = S10 >>> mmappedarray = np.memmap tmp npmmap , mode = r , offset = 7, dtype = mydtype) >>> mmappedarray memmap [ onerecord , tworecord ], dtype= |S10 ) Now supposing you had some other data at the end of the file echo n HEADER!
|
|
onerecord tworecordJUNK > tmp npmmap You can handle that by explicitly stating the shape >>> mmappedarray = np.memmap tmp npmmap , mode = r , offset = 7, dtype = mydtype, shape = 2) >>> mmappedarray memmap [ onerecord , tworecord ], dtype= |S10 ) Therefore, here is a statement to replace offset should be a multiple of the byte size of dtype.
|
|
Requires shape=None.
|
|
if number of bytes remaining after offset is evenly divisible into the byte size of dtype, shape can be omitted.
|
|
Otherwise, when you specify offset you must also specify shape.
|
|
When creating or overwriting files, offset can be any positive number, subject to file size limitations see Notes).
|
|
and for shape, When reading, the remaining bytes in the file must be <= the bytesize of an array of the specified dtype and shape, if shape is given.
|
|
When writing a new file, shape must always be given.
|
|
When overwriting mode == w ) or appending mode == r ) to an existing file, specifying a shape that extends beyond the end of the current file contents causes the file to be immediately extended to fit.
|
|
for w mode docs, overwriting an existing file may cause file size to grow or shrink as needed.
|
|
specified because, in combination with offset > 0, the total file size after memmapping may differ from the memmapped array size) I set the version for this bug to devel , but it is actually 1.2, which is not available in the bug tracker)","bug"
|
|
"6013","1570","Do not put the generated import library in python tree Trac 972).
|
|
Original ticket http projects.scipy.org numpy ticket 972 on 2008 12 19 by cournape, assigned to cournape.
|
|
While working on mingw w64 support, I noticed that libpython .a is put into C \Python25\libs, which is not so good for at least two reasons it is not regenerated when available, even if it is broken.
|
|
we should not put anything in the python tree anyway.","bug"
|
|
"6013","1571","Error compiling trunk using MSVS 9 Trac 973).
|
|
Original ticket http projects.scipy.org numpy ticket 973 on 2008 12 19 by FrancescAlted, assigned to unknown.
|
|
FYI.
|
|
After trying to compile the NumPy trunk with MSVS 9 32 bits), I m getting the next error C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\link.exe DLL nologo INCREMENTAL NO LIBPATH C \Python26\libs LIBPATH C \Python26\PCbuild EXPORT i nitmtrand build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.obj build\temp .win32 2.6\Release\numpy\random\mtrand\randomkit.obj build\temp.win32 2.6\Releas e\numpy\random\mtrand\initarray.obj build\temp.win32 2.6\Release\numpy\random\mt rand\distributions.obj OUT build\lib.win32 2.6\numpy\random\mtrand.pyd IMPLIB build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.lib MANIFESTFILE build\ temp.win32 2.6\Release\numpy\random\mtrand\mtrand.pyd.manifest Creating library build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.lib and object build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.exp randomkit.obj error LNK2019 unresolved external symbol imp CryptReleaseCon text 8 referenced in function rk devfill randomkit.obj error LNK2019 unresolved external symbol imp CryptGenRandom 12 referenced in function rk devfill randomkit.obj error LNK2019 unresolved external symbol imp CryptAcquireCon textA 20 referenced in function rk devfill build\lib.win32 2.6\numpy\random\mtrand.pyd fatal error LNK1120 3 unresolved externals error Command C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\link.e xe DLL nologo INCREMENTAL NO LIBPATH C \Python26\libs LIBPATH C \Python26\P Cbuild EXPORT initmtrand build\temp.win32 2.6\Release\numpy\random\mtrand\mtran d.obj build\temp.win32 2.6\Release\numpy\random\mtrand\randomkit.obj build\temp.
|
|
win32 2.6\Release\numpy\random\mtrand\initarray.obj build\temp.win32 2.6\Release \numpy\random\mtrand\distributions.obj OUT build\lib.win32 2.6\numpy\random\mtr and.pyd IMPLIB build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.lib MAN IFESTFILE build\temp.win32 2.6\Release\numpy\random\mtrand\mtrand.pyd.manifest failed with exit status 1120 I regularly use MSVS 9 to compile trunk, but I ve never seen this error before.","bug"
|
|
"6013","1572","FAIL test umath.TestComplexFunctions.test against cmath Trac 974).
|
|
Original ticket http projects.scipy.org numpy ticket 974 on 2008 12 20 by nilswagner01, assigned to unknown.
|
|
python2,6 >>> numpy.
|
|
version 1.3.0.dev6173 ====================================================================== FAIL test umath.TestComplexFunctions.test against cmath Traceback most recent call last) File home nwagner local lib64 python2.6 site packages nose 0.10.4 py2.6.egg nose case.py , line 182, in runTest self.test self.arg) File home nwagner local lib64 python2.6 site packages numpy core tests test umath.py , line 423, in test against cmath assert abs a b) < atol, s s s; cmath s fname,p,a,b) AssertionError arcsin 2 1.57079632679 1.31695789692j); cmath 1.57079632679 1.31695789692j) >>> numpy.show config ) atlas threads info libraries = [ lapack , ptf77blas , ptcblas , atlas ] library dirs = [ home nwagner src ATLAS3.8.2 mybuild lib ] language = f77 include dirs = [ home nwagner src ATLAS3.8.2 include ] blas opt info libraries = [ ptf77blas , ptcblas , atlas ] library dirs = [ home nwagner src ATLAS3.8.2 mybuild lib ] define macros = [ ATLAS INFO , \\ 3.8.2\\ )] language = c include dirs = [ home nwagner src ATLAS3.8.2 include ] atlas blas threads info libraries = [ ptf77blas , ptcblas , atlas ] library dirs = [ home nwagner src ATLAS3.8.2 mybuild lib ] language = c include dirs = [ home nwagner src ATLAS3.8.2 include ] lapack opt info libraries = [ lapack , ptf77blas , ptcblas , atlas ] library dirs = [ home nwagner src ATLAS3.8.2 mybuild lib ] define macros = [ ATLAS INFO , \\ 3.8.2\\ )] language = f77 include dirs = [ home nwagner src ATLAS3.8.2 include ] lapack mkl info NOT AVAILABLE blas mkl info NOT AVAILABLE mkl info NOT AVAILABLE","bug"
|
|
"6013","1573","Formatting unit tests failure on windows 2.6 Trac 975).
|
|
Original ticket http projects.scipy.org numpy ticket 975 on 2008 12 23 by cournape, assigned to unknown.
|
|
Harmless, but still annoying ====================================================================== FAIL Check formatting.
|
|
Traceback most recent call last) File C \Python26\Lib\site packages\numpy\core\tests\test print.py , line 28, in test complex types assert equal str t x)), str complex x))) File C \Python26\Lib\site packages\numpy\testing\utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL 1e 020 0j) DESIRED 1e 20 0j) ====================================================================== FAIL Check formatting.
|
|
Traceback most recent call last) File C \Python26\Lib\site packages\numpy\core\tests\test print.py , line 16, in test float types assert equal str t x)), str float x))) File C \Python26\Lib\site packages\numpy\testing\utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL 1e 020 DESIRED 1e 20","bug"
|
|
"6013","1574","incorrect repr of float32 and float64 Trac 976).
|
|
Original ticket http projects.scipy.org numpy ticket 976 on 2008 12 23 by trac user wxgeo, assigned to unknown.
|
|
Hi, On my Linux installation, numpy uses comma instead of dot in representation of float32, float64 and float96 types at least) >>> import numpy >>> numpy.version.version 1.2.1 >>> numpy.float64 4.3) 4,2999999999999998 >>> eval repr numpy.float64 4.3))) 4, 2999999999999998L) I use Ubuntu 8.10 Unbuntu Ibex) kernel 2.6.27 9 generic.
|
|
Default numpy version is 1.1.1, so I compiled 1.2.1, but the same bug occurs.
|
|
Not this bug does not appear on Windows XP.
|
|
Not also this bug might be french language relative although I can t actually figure how), since comma is used instead of dot for floating number representation in french speaking countries.
|
|
echo LANG fr FR.UTF 8","bug"
|
|
"6013","1575","Numpy test failure with Python 2.5.3 win32) Trac 977).
|
|
Original ticket http projects.scipy.org numpy ticket 977 on 2008 12 26 by trac user raybaut, assigned to unknown.
|
|
Running unit tests for numpy NumPy version 1.2.1 NumPy is installed in C \Program Files\pythonxy\python\lib\site packages\numpy Python version 2.5.3 r253 67855, Dec 19 2008, 16 58 30) [MSC v.1310 32 bit Intel)] nose version 0.10.4 ====================================================================== FAIL test umath.TestComplexFunctions.test against cmath Traceback most recent call last) File nose\case.py , line 182, in runTest self.test self.arg) File C \Program Files\pythonxy\python\lib\site packages\numpy\core\tests\test umath.py , line 268 , in test against cmath assert abs a b) < atol, s s s; cmath s fname,p,a,b) AssertionError arcsinh 2j 1.31695789692 1.57079632679j); cmath 1.31695789692 1.57079632679j) Ran 1592 tests in 11.484s FAILED KNOWNFAIL=3, failures=1) Out[1] <nose.result.TextTestResult run=1592 errors=0 failures=1>","bug"
|
|
"6013","1576","unit test failures for numpy 1.3.0.dev6294 Trac 978).
|
|
Original ticket http projects.scipy.org numpy ticket 978 on 2009 01 05 by chanley, assigned to unknown.
|
|
The following unit test failures are received with numpy 1.3.0.dev6294 with Python 2.5.1.
|
|
The test system is an Intel Mac running OS X 10.5.6.
|
|
However, the errors can be reproduced on RHE 4 and 4 systems as well.
|
|
[redcedar dev devCode] chanley python Python 2.5.1 r251 54863, Mar 13 2008, 13 22 51) [GCC 4.0.1 Apple Computer, Inc. build 5367)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy;numpy.test ) Running unit tests for lib.python.numpy NumPy version 1.3.0.dev6294 NumPy is installed in Users chanley dev site packages lib python numpy Python version 2.5.1 r251 54863, Mar 13 2008, 13 22 51) [GCC 4.0.1 Apple Computer, Inc. build 5367)] nose version 0.10.3 .............................................................................................................................................................................................................................................................F..F..F..FFF..................................................................................................................................................................................................K.............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................S.......................................................................................................................................................................................................................................................................................................................................................................ctypes is not available on this python skipping the test import error was ctypes is not available.)
|
|
.No distutils available, skipping test.
|
|
..... ====================================================================== FAIL Check formatting.
|
|
Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy core tests test print.py , line 74, in check complex type err msg= Failed str formatting for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed str formatting for type <type numpy.complex64 > ACTUAL 1e 10 0j) DESIRED 1e 10 ====================================================================== FAIL Check formatting when using print Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy core tests test print.py , line 108, in check float type print test redirected print float x), tp) File Users chanley dev site packages lib python numpy core tests test print.py , line 104, in test redirected print err msg= print failed for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal print failed for type<type numpy.float32 > ACTUAL 10000000000.0\n DESIRED 1e 10\n ====================================================================== FAIL Check formatting when using print Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy core tests test print.py , line 115, in check complex type print test redirected print complex x), tp) File Users chanley dev site packages lib python numpy core tests test print.py , line 104, in test redirected print err msg= print failed for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal print failed for type<type numpy.complex64 > ACTUAL 10000000000 0j)\n DESIRED 1e 10 0j)\n ====================================================================== FAIL test print.test locale single Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy testing decorators.py , line 82, in skipper return f args, kwargs) File Users chanley dev site packages lib python numpy core tests test print.py , line 164, in test locale single return test locale independance np.float32) File Users chanley dev site packages lib python numpy core tests test print.py , line 157, in test locale independance err msg= Failed locale test for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed locale test for type <type numpy.float32 > ACTUAL 1,2 DESIRED 1.2 ====================================================================== FAIL test print.test locale double Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy testing decorators.py , line 82, in skipper return f args, kwargs) File Users chanley dev site packages lib python numpy core tests test print.py , line 169, in test locale double return test locale independance np.double) File Users chanley dev site packages lib python numpy core tests test print.py , line 157, in test locale independance err msg= Failed locale test for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed locale test for type <type numpy.float64 > ACTUAL 1,2 DESIRED 1.2 ====================================================================== FAIL test print.test locale longdouble Traceback most recent call last) File Users chanley dev site packages lib python nose case.py , line 182, in runTest self.test self.arg) File Users chanley dev site packages lib python numpy testing decorators.py , line 82, in skipper return f args, kwargs) File Users chanley dev site packages lib python numpy core tests test print.py , line 174, in test locale longdouble return test locale independance np.longdouble) File Users chanley dev site packages lib python numpy core tests test print.py , line 157, in test locale independance err msg= Failed locale test for type s tp) File Users chanley dev site packages lib python numpy testing utils.py , line 183, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed locale test for type <type numpy.float128 > ACTUAL 1,2 DESIRED 1.2 Ran 1799 tests in 8.869s FAILED KNOWNFAIL=1, SKIP=1, failures=6) <nose.result.TextTestResult run=1799 errors=0 failures=6>","bug"
|
|
"6013","1577","broken arithmetic with masked arrays Trac 979).
|
|
Original ticket http projects.scipy.org numpy ticket 979 on 2009 01 07 by trac user jguyer, assigned to pierregm.
|
|
I m getting unexpected broadcasting and raveling of masked values in basic arithmetic.
|
|
I don t know if this is related to 1424 !python >>> import numpy >>> numpy.
|
|
version 1.3.0.dev6298 >>> A = numpy.ma.array [[1.
|
|
], ...
|
|
[2.
|
|
], ...
|
|
[3.
|
|
]], mask=[[False], ... [True], ... [True]]) >>> B = numpy.array [[2., 3.
|
|
], ... [4., 5.
|
|
], ... [6., 7.]])
|
|
>>> A B masked array data = [[2.0 ] [ 2.0] [ ]], mask = [[False True] [ True False] [ True True]], fill value = 1e 20) >>> A.data B array [[ 2., 3.
|
|
], [ 8., 10.
|
|
], [ 18., 21.]])
|
|
>>> A B).data array [[ 2., 3.
|
|
], [ 2., 2.
|
|
], [ 3., 3.]])
|
|
With NumPy 1.2, the mask is still broken, but the value is correct !python >>> import numpy >>> numpy.
|
|
version 1.2.0 >>> A = numpy.ma.array [[1.
|
|
], ...
|
|
[2.
|
|
], ...
|
|
[3.
|
|
]], mask=[[False], ... [True], ... [True]]) >>> B = numpy.array [[2., 3.
|
|
], ... [4., 5.
|
|
], ... [6., 7.]])
|
|
>>> A B masked array data = [[2.0 ] [ 10.0] [ ]], mask = [[False True] [ True False] [ True True]], fill value=1e 20) >>> A.data B array [[ 2., 3.
|
|
], [ 8., 10.
|
|
], [ 18., 21.]])
|
|
>>> A B).data array [[ 2., 3.
|
|
], [ 8., 10.
|
|
], [ 18., 21.]])
|
|
I thought I d seen cases where NumPy 1.1 returned both the correct value and mask, but I m not able to reproduce now.
|
|
NumPy 1.1 and 1.2 behave the same.","bug"
|
|
"6013","1578","NaN s are silenty converted to ints Trac 980).
|
|
Original ticket http projects.scipy.org numpy ticket 980 on 2009 01 07 by stefanv, assigned to cournape.
|
|
From a mailing list post by Josef Why does converting nan to an integer not throw an exception as with inf), instead numpy silently replaces nan by zero?
|
|
>>> inti = np.array [0,1,2]) >>> inti[1] = np.inf Traceback most recent call last) File <pyshell 34> , line 1, in ?
|
|
inti[1] = np.inf OverflowError cannot convert float infinity to long >>> inti[1] = np.nan >>> inti array [0, 0, 2])","bug"
|
|
"6013","1579","problem with locales Trac 981).
|
|
Original ticket http projects.scipy.org numpy ticket 981 on 2009 01 08 by chanley, assigned to unknown.
|
|
Seen in latest version of numpy svn on multiple platforms.
|
|
>>> import numpy as np >>> import locale >>> locale.format string f ,np.double 1.2)) Traceback most recent call last) File <stdin> , line 1, in <module> File usr stsci pyssgdev Python 2.5.1 lib python2.5 locale.py , line 188, in format string val[key] = format perc.group ), val[key], grouping) TypeError numpy.float64 object does not support item assignment >>>","bug"
|
|
"6013","1580","Problem accessing data attribute of ndarray object created with certain sizes Trac 982).
|
|
Original ticket http projects.scipy.org numpy ticket 982 on 2009 01 09 by trac user jtaylor2, assigned to unknown.
|
|
When creating a ndarray object on a 64 bit machine where the number of bytes of data in the array is described by the following rule size 0x80000000) != 0 AND size 0xFFFFFFFF) != 0xFFFFFFFF Accessing the data attribute the data buffer) of the ndarray object results in a ValueError exception.
|
|
Note the following examples thor> python Python 2.5.1 r251 54863, Jan 16 2008, 16 16 13) [GCC 3.4.6 20060404 Red Hat 3.4.6 9)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.3.0.dev6306 >>> a=numpy.zeros 0x7FFFFFFF,dtype= b1 ) >>> a.data <read write buffer for 0xa054a0, size 2147483647, offset 0 at 0x2a99e13e70> >>> a=numpy.zeros 0x80000000,dtype= b1 ) >>> a.data Traceback most recent call last) File <stdin> , line 1, in <module> ValueError size must be zero or positive >>> a=numpy.zeros 0xFFFFFFFE,dtype= b1 ) >>> a.data Traceback most recent call last) File <stdin> , line 1, in <module> ValueError size must be zero or positive >>> a=numpy.zeros 0xFFFFFFFF,dtype= b1 ) >>> a.data <read write buffer for 0xa2e410, size 1, offset 0 at 0x2a99e13e30> >>> a=numpy.zeros 0x180000000,dtype= b1 ) >>> a.data Traceback most recent call last) File <stdin> , line 1, in <module> ValueError size must be zero or positive >>> a=numpy.zeros 0x180000001,dtype= b1 ) >>> a.data Traceback most recent call last) File <stdin> , line 1, in <module> ValueError size must be zero or positive >>> a=numpy.zeros 0x1FFFFFFFF,dtype= b1 ) >>> a.data <read write buffer for 0xa2e590, size 1, offset 0 at 0x2a99e13e70>","bug"
|
|
"6013","1581","Cannot build numpy from svn Trac 983).
|
|
Original ticket http projects.scipy.org numpy ticket 983 on 2009 01 09 by nilswagner01, assigned to jarrodmillman.
|
|
python setup.py install prefix= HOME local Running from numpy source directory.
|
|
Traceback most recent call last) File setup.py , line 96, in <module> setup package ) File setup.py , line 68, in setup package from numpy.distutils.core import setup File home nwagner svn numpy numpy distutils core.py , line 25, in <module> from numpy.distutils.command import config, config compiler, \ File home nwagner svn numpy numpy distutils command build ext.py , line 16, in <module> from numpy.distutils.system info import combine paths File home nwagner svn numpy numpy distutils system info.py , line 185, in <module> opt local lib , sw lib ], platform bits) File home nwagner svn numpy numpy distutils system info.py , line 170, in libpaths out.extend [p bits, p]) TypeError cannot concatenate str and int objects","bug"
|
|
"6013","1582","machar single and machar double don t exist anymore Trac 984).
|
|
Original ticket http projects.scipy.org numpy ticket 984 on 2009 01 14 by lebedov, assigned to unknown.
|
|
There used to be two MachAr instances defined in numpy that stored single precision and double precision system parameters, respectively.
|
|
Although they are still listed in numpy lib info.py in numpy 1.2.1, they no longer appear to be defined in numpy lib machar.py.
|
|
They should either be redefined, or removed from numpy lib info.py.","bug"
|
|
"6013","1583","Remove need for site.cfg for installs in standard locations Trac 985).
|
|
Original ticket http projects.scipy.org numpy ticket 985 on 2009 01 14 by jarrodmillman, assigned to jarrodmillman.
|
|
Building numpy and scipy on the major UNIX like systems e.g., Fedora, Ubuntu) with libraries installed via their official packages shouldn t require a site.cfg.
|
|
On Fedora, used to require adding the following site.cfg to scipy [DEFAULT] library dirs = usr lib64 include dirs = usr include usr include suitesparse","bug"
|
|
"6013","1584","Von Mises draws not between pi and pi [patch] Trac 986).
|
|
Original ticket http projects.scipy.org numpy ticket 986 on 2009 01 16 by trac user anand, assigned to unknown.
|
|
As the following test shows, Von Mises draws are not currently staying between pi and pi at least on my Mac) from numpy import import numpy as np import pylab as pl for mu in arange 4,5) pi 2 q=pl.hist np.random.mtrand.vonmises mu,1,10000)),50) print mu raw input ) The attached patch makes them do so.","bug"
|
|
"6013","1585","random.exponential ) seems to use the reciprocal rate prarameter instead of the rate parameter probably just wrong documentation) Trac 987).
|
|
Original ticket http projects.scipy.org numpy ticket 987 on 2009 01 17 by trac user j.koehler, assigned to cournape.
|
|
hi, i hope i m doing this the right way.
|
|
i had some trouble using the random.exponential ) function.
|
|
i used the function as described in http docs.scipy.org doc numpy reference generated numpy.random.exponential.html i created some histograms of the random values and found out the distribution behaves like exp x L) L.","bug"
|
|
"6013","1586","Concatenate with sequence that contains empty sequences.
|
|
Trac 988).
|
|
Original ticket http projects.scipy.org numpy ticket 988 on 2009 01 22 by trac user kelson, assigned to unknown.
|
|
Howdy, After searching for this defect, perhaps nobody else has been bothered by this behaviour, but has anybody else noticed concatenate [[1],[2],[3],[4]] ).dtype = int32 concatenate [[1],[2],[3],[]] ).dtype = int32 concatenate [[],[2],[3],[4]] ).dtype = float64 I have a lot of old style Numeric stuff that likes to concatenate lists of indices which then were used with take [which does not like floats].
|
|
So below [in theory] I attach my attempt at a patch so that empty items in the sequence are not used to influence the common type that is used by PyArray ConvertToCommonType.
|
|
Maybe it is of some value, or maybe I didn t do this in a kosher manner and someone else has an idea of where things should be rethought for this defect?
|
|
Cheerio.","bug"
|
|
"6013","1587","masked array results differ between inplace and regular operators Trac 989).
|
|
Original ticket http projects.scipy.org numpy ticket 989 on 2009 01 26 by trac user danielmcdonald, assigned to pierregm.
|
|
The result of a masked array addition differs between inplace math operators and the regular math operators.
|
|
The unexpected behavior results in the regular operator, it appears that the mask variable is not being set in the same fashion.
|
|
Below is interpreter output.
|
|
NumPy 1.0.3 produces the expected result, NumPy 1.1.1 and NumPy 1.2.1 do not.
|
|
The following output was produced using Python 2.5.2, NumPy 1.2.1 on an OSX 10.5.6 x86 machine >>> from numpy import ma, array >>> a = ma.array [1,2,3],mask=array [False,False,False])) >>> a masked array data = [1 2 3], mask = [False False False], fill value=999999) >>> a = array [2,3,4]) >>> a masked array data = [2 6 12], mask = [False False False], fill value=999999) >>> a = ma.array [1,2,3],mask=array [False,False,False])) >>> a array [2,3,4]) masked array data = [ 2 6 12], mask = False, fill value=999999)","bug"
|
|
"6013","1588","SVD fails to converge Trac 990).
|
|
Original ticket http projects.scipy.org numpy ticket 990 on 2009 01 31 by trac user MikeTrumpis, assigned to charris.
|
|
I m running into a !LinAlgError SVD did not converge in both numpy and scipy for a given matrix npy file linked below).
|
|
The matrix is very badly conditioned but SVD is meant for that).
|
|
np version 1.3.0.dev6083 mac os 10.4, lapack lite built with Wl, framework Wl,Accelerate I can get results for this matrix using the SVD routines of either Matlab or Octave.
|
|
This is similar to ticket 1304, but the solution there r4914) can t help me, since my build doesn t use dlapack lite.c.
|
|
Since I can get solutions in Matlab and Octave, perhaps there is some post LAPACK solution for numpy too?
|
|
https cirl.berkeley.edu twiki pub User MikeTrumpis sinc operator.npy Mike In [156] [u,s,vt] = np.linalg.svd snc op, 1, 1) LinAlgError Traceback most recent call last) Users miket sandbox trunk testdata siemens <ipython console> in <module> ) opt local lib python2.5 site packages numpy linalg linalg.pyc in svd a, full matrices, compute uv) 1024 work, lwork, iwork, 0) 1025 if results[ info ] > 0 > 1026 raise LinAlgError, SVD did not converge 1027 s = s.astype realType result t)) 1028 if compute uv LinAlgError SVD did not converge In [157] [u,s,vt] = sp.linalg.svd snc op, full matrices=1, compute uv=1) LinAlgError Traceback most recent call last) Users miket sandbox trunk testdata siemens <ipython console> in <module> ) opt local lib python2.5 site packages scipy linalg decomp.pyc in svd a, full matrices, compute uv, overwrite a) 919 else clapack 920 raise NotImplementedError, calling gesdd from s gesdd.module name) > 921 if info>0 raise LinAlgError, SVD did not converge 922 if info<0 raise ValueError,\ 923 illegal value in th argument of internal gesdd info) LinAlgError SVD did not converge","bug"
|
|
"6013","1589","Make savez able to write ZIP64 files Trac 991).
|
|
Original ticket http projects.scipy.org numpy ticket 991 on 2009 02 02 by trac user phillies, assigned to stefanv.
|
|
When you try to write a zip file larger than 4 GB with numpy s savez you get an exception saying that you should enable ZIP64 extensions.
|
|
I did not find any way without altering the numpy code manually and reinstalling numpy) to make this possible.
|
|
From my point of view, this is a severe usability bug as I can not use savez anymore.
|
|
Please add a function, e.g.
|
|
savez64 ...) which uses ZIP64 by default or a function enableZip64 bool) which can turn ZIP64 on or off.","bug"
|
|
"6013","1590","Make savez able to write ZIP64 files Trac 992).
|
|
Original ticket http projects.scipy.org numpy ticket 992 on 2009 02 02 by trac user phillies, assigned to unknown.
|
|
When you try to write a zip file larger than 4 GB with numpy s savez you get an exception saying that you should enable ZIP64 extensions.
|
|
I did not find any way without altering the numpy code manually and reinstalling numpy) to make this possible.
|
|
From my point of view, this is a severe usability bug as I can not use savez anymore.
|
|
Please add a function, e.g.
|
|
savez64 ...) which uses ZIP64 by default or a function enableZip64 bool) which can turn ZIP64 on or off.","bug"
|
|
"6013","1591","[PATCH] Support for detecting libraries in several directories simultaneously Trac 993).
|
|
Original ticket http projects.scipy.org numpy ticket 993 on 2009 02 02 by FrancescAlted, assigned to rgommers.
|
|
I ve found a problem with `numpy.distutils` when I want to specify multiple libraries for a !NumPy component that are spread in two or more directories.
|
|
For example, for linking against MKL, a `site.cfg` with the next contents [mkl] library dirs = opt intel Compiler 11.0 074 mkl lib em64t opt intel Compiler 11.0 074 lib intel64 include dirs = opt intel Compiler 11.0 074 mkl include mkl libs = mkl gf lp64, mkl gnu thread, mkl core, iomp5 gives the next output mkl info libraries mkl gf lp64,mkl gnu thread,mkl core,iomp5 not found in opt intel Compiler 11.0 074 mkl lib em64t libraries mkl gf lp64,mkl gnu thread,mkl core,iomp5 not found in opt intel Compiler 11.0 074 lib intel64 NOT AVAILABLE I ve tracked down the problem in that `numpy.distutils` tries to find the complete set of libraries in one single directory, instead in the complete list.
|
|
I m attaching a fix that is able to find the set of required libraries in all the directories specified in `library dirs` and only in one of them.
|
|
With it, the output is mkl info FOUND libraries = [ mkl gf lp64 , mkl gnu thread , mkl core , iomp5 , pthread ] library dirs = [ opt intel Compiler 11.0 074 mkl lib em64t , opt intel Compiler 11.0 074 lib intel64 ] define macros = [ SCIPY MKL H , None)] include dirs = [ opt intel Compiler 11.0 074 mkl include ] So, it seems to work well.
|
|
Moreover, the fix simplifies quite a bit the logic in `system info.py`.
|
|
However, caveat emptor as I ve only tested the MKL section, more testing should be carried out with other libraries than MKL before to apply the patch.
|
|
Finally, attention should be payed to the next issue In `atlas info.calc info )` method, I ve not completely simplified the loop for d in lib dirs atlas = self.check libs2 d,atlas libs,[]) lapack atlas = self.check libs2 d,[ lapack atlas ],[]) if atlas is not None lib dirs2 = [d] self.combine paths d,[ atlas , ATLAS ]) lapack = self.check libs2 lib dirs2,lapack libs,[]) if lapack is not None break if atlas atlas 1 = atlas because I m not sure about how can affect the use of `self.combine paths )` in the previous loop, but I d say that it could be rewritten as atlas = self.check libs2 lib dirs,atlas libs,[]) lapack atlas = self.check libs2 lib dirs,[ lapack atlas ],[]) if atlas is not None lib dirs2 = [lib dirs] self.combine paths lib dirs,[ atlas , ATLAS ]) lapack = self.check libs2 lib dirs2,lapack libs,[]) atlas 1 = atlas So, if anybody with more knowledge than me can double check the patch, it would be great.
|
|
Thanks","bug"
|
|
"6013","1592","[PATCH] Support for better Core2 handling in numpy.distutils Trac 994).
|
|
Original ticket http projects.scipy.org numpy ticket 994 on 2009 02 02 by FrancescAlted, assigned to cournape.
|
|
I find that the next patch could be desirable in order to have a better handling of the Core2 architecture numpy distutils system info.py revision 6337) numpy distutils system info.py working copy) 838,7 838,7 if cpu.is Itanium ) plt = 64 l = mkl ipf elif cpu.is Xeon ) elif cpu.is Xeon ) or cpu.is Core2 ) plt = em64t l = mkl em64t else Thanks","bug"
|
|
"6013","1593","numpy.load can t handle gzip file handles Trac 995).
|
|
Original ticket http projects.scipy.org numpy ticket 995 on 2009 02 03 by trac user MatthewMiller, assigned to unknown.
|
|
I d like to log the state of my program as it progresses.
|
|
Using the numpy.save numpy.load functions on the same filehandle repeatedly works very well for this but ends up making a file which very quickly grows to gigabytes.
|
|
The data compresses well, though, so I thought I d use Python s built in gzip module underneath.
|
|
This works great for saving but when it comes time to play back, there s an issue >>> import numpy >>> import gzip >>> f=open test.gz ) >>> g=gzip.GzipFile None, rb ,9,f) >>> g <gzip open file test.gz , mode r at 0xbaad50 0xc0ab90> >>> numpy.load g) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib64 python2.5 site packages numpy lib io.py , line 133, in load fid.seek N,1) back up TypeError seek ) takes exactly 2 arguments 3 given) Turns out you can t rewind gzip file handles in Python.
|
|
Oops.
|
|
The offending code is that which distinguishes between npy and npz files.
|
|
Could there maybe be something added to just trust me that it s an npy?
|
|
If I use numpy.lib.format.read array ) directly, there s no problem.","enhancement"
|
|
"6013","1594","Re enable umath tests Trac 996).
|
|
Original ticket http projects.scipy.org numpy ticket 996 on 2009 02 03 by trac user hansres, assigned to unknown.
|
|
Currently building and running umath tests is disabled.
|
|
This module is used for testing the generalized ufunc framework see ticket 887), and it should be enabled again.
|
|
This includes the following steps.
|
|
Revert changeset 6019.
|
|
Timestamp 11 12 08 06 39 18 3 months ago) Author charris Message Temporarily remove umath test from build, it isn t linking on windows.
|
|
The problem is the blas gemm problem seen elsewhere and I want to get windows compiling first.
|
|
It s almost there.
|
|
Revert changeset 6001.
|
|
Enable umath tests when building with numscons apply attached patch).
|
|
I have asked Chuck Harris about this issue and he suggested opening a ticket to track it.","enhancement"
|
|
"6013","1595","[proposal][patch] bool array) should do the right thing in the unambiguous cases.
|
|
Trac 997).
|
|
Original ticket http projects.scipy.org numpy ticket 997 on 2009 02 04 by trac user alsuren, assigned to unknown.
|
|
After stumbling across http mail.python.org pipermail python list 2008 December 519124.html I thought I might as well make a concrete proposal.
|
|
The test case proposed in the email is >>>> import numpy >>>> y = numpy.zeros 3,)) >>>> y > array [ 0., 0., 0.])
|
|
>>>> bool y==y) > Traceback most recent call last) > File <stdin> , line 1, in <module> > ValueError The truth value of an array with more than one element is > ambiguous.
|
|
Use a.any ) or a.all ) >>>> ll1 = [y,1] >>>> y in ll1 > True >>>> ll2 = [1,y] >>>> y in ll2 > Traceback most recent call last) > File <stdin> , line 1, in <module> > ValueError The truth value of an array with more than one element is > ambiguous.
|
|
Use a.any ) or a.all ) May I suggest that bool y==y) > True bool zeros 5,))==ones 5,))) > False are not ambiguous in each case, any array)==all array), and this value should simply be returned.
|
|
I ll attach a patch against 1.1.1 which shows the general idea, but which I ve not tested.","enhancement"
|
|
"6013","1596","make latex failure in Numpy docs Trac 998).
|
|
Original ticket http projects.scipy.org numpy ticket 998 on 2009 02 04 by scottza, assigned to unknown.
|
|
Numpy PDF documentation fails to build with make latex , but make html succeeds.
|
|
From mailing list discussion http projects.scipy.org pipermail numpy discussion 2009 January 039917.html > 2009 1 27 Nils Wagner <nwagner iam.uni stuttgart.de> > a make latex in numpy doc failed with > > ... > > Intersphinx hit PyObject > http docs.python.org dev c api structures.html > writing... Sphinx error > too many nesting section levels for LaTeX, at heading > numpy.ma.MaskedArray.
|
|
lt > make [latex] Fehler 1 > > > I am using sphinxv0.5.1 > BTW, make html works fine here.
|
|
I see this problem too.
|
|
It used to work, and I don t think I ve changed anything on my system.
|
|
Python 2.5.2 r252 60911, Oct 5 2008, 19 24 49) [GCC 4.3.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.3.0.dev6335 >>> import sphinx >>> sphinx.
|
|
version 0.5.1","bug"
|
|
"6013","1597","NumPy does not easy install as a dependency Trac 999).
|
|
Original ticket http projects.scipy.org numpy ticket 999 on 2009 02 05 by trac user mjwakefield, assigned to dmcooke.
|
|
When NumPy is listed as a dependency for another package automatic installation of NumPy fails.
|
|
easy install numpy works successfully Best match numpy 1.2.1 Downloading http downloads.sourceforge.net numpy numpy 1.2.1.tar.gz?modtime=1225265011 big mirror=0 Processing numpy 1.2.1.tar.gz Running numpy 1.2.1 setup.py q bdist egg dist dir var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 egg dist tmp H5FSHT Running from numpy source directory.
|
|
Warning distutils distribution has been initialized, it may be too late to add a subpackage command Warning distutils distribution has been initialized, it may be too late to add a subpackage fcompiler non existing path in private var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils site.cfg Warning distutils distribution has been initialized, it may be too late to add a subpackage distutils Warning distutils distribution has been initialized, it may be too late to add a subpackage testing var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py 1343 DeprecationWarning os.popen4 is deprecated.
|
|
Use the subprocess module.
|
|
sin, sout = os.popen4 svnversion ) Traceback most recent call last) File Library Frameworks Python.framework Versions Current bin easy install , line 8, in <module> load entry point setuptools==0.6c9 , console scripts , easy install ) ) File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 1671, in main File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 1659, in with ei usage File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 1675, in <lambda> File Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils core.py , line 152, in setup dist.run commands ) File Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 975, in run commands self.run command cmd) File Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 211, in run File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 427, in easy install File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 478, in install item File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 519, in process distribution File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg pkg resources.py , line 522, in resolve File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg pkg resources.py , line 758, in best match File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg pkg resources.py , line 770, in obtain File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 446, in easy install File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 476, in install item File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 655, in install eggs File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 930, in build and install File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools command easy install.py , line 919, in run setup File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools sandbox.py , line 27, in run setup File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools sandbox.py , line 63, in run File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages setuptools 0.6c9 py2.6.egg setuptools sandbox.py , line 29, in <lambda> File setup.py , line 96, in <module> File setup.py , line 89, in setup package File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils core.py , line 150, in setup File setup.py , line 56, in configuration File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 851, in add subpackage File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 834, in get subpackage File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 781, in get configuration from setup py File numpy setup.py , line 8, in configuration import os File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 851, in add subpackage File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 834, in get subpackage File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 781, in get configuration from setup py File private var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy f2py setup.py , line 81, in configuration File var folders Kg KgodBp6HFYaMgVoWCkGln mNU Tmp easy install p3img4 numpy 1.2.1 numpy distutils misc util.py , line 1239, in add scripts AttributeError NoneType object has no attribute extend","bug"
|
|
"6013","1598","seg fault during ptp on MaskedArrays ; Solaris 9, NumPy 1.2.1 Trac 1000).
|
|
Original ticket http projects.scipy.org numpy ticket 1000 on 2009 02 05 by trac user gbburkhardt, assigned to cournape.
|
|
Built NumPy 1.2.1 on a Solaris 9 system, netlib lapack 3.2, fftw 3.2, UMFPACK 5.2.0, AMD 2.2.0, UFconfig 3.2.0.
|
|
Ran unit tests, and program seg fault d in the ptp on MaskedArrays test.
|
|
Log from test, backtrace, some variable queries attached.
|
|
site.cfg file attached.","bug"
|
|
"6013","1599","dype comparison with other types raise a TypeError Trac 1001).
|
|
Original ticket http projects.scipy.org numpy ticket 1001 on 2009 02 06 by trac user ashleywalsh, assigned to unknown.
|
|
If you compare a dtype with another object for equality a TypeError is raised if the second object cannot be converted to a dtype >>> import numpy >>> numpy.dtype float) == float True >>> numpy.dtype float) == i False >>> numpy.dtype float) == object ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError data type not understood >>> numpy.dtype float) == x Traceback most recent call last) File <stdin> , line 1, in <module> TypeError data type not understood In my case this caused epydoc to crash because it attempted to compare a dtype that was class attribute with I think) an object was an attribute of the same name on a super class.","bug"
|
|
"6013","1600","MaskedArray.resize method raises TypeError Trac 1002).
|
|
Original ticket http projects.scipy.org numpy ticket 1002 on 2009 02 06 by scottza, assigned to pierregm.
|
|
A TypeError is raised by the MaskedArray.resize ) method.
|
|
Python 2.5.2 r252 60911, Oct 5 2008, 19 24 49) [GCC 4.3.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.3.0.dev6344 >>> x = np.ma.array [0,1,2]) >>> x.resize 2,)) Traceback most recent call last) File <stdin> , line 1, in <module> File home scott .virtualenvs scipy dev lib python2.5 site packages numpy ma core.py , line 2721, in resize self.
|
|
data.resize newshape, refcheck, order) TypeError an integer is required >>> np.ma.resize x, 2,)) masked array data = [0 1], mask = False, fill value = 999999) >>> x = np.array [0,1,2]) >>> x.resize 2,)) >>> x array [0, 1]) Changing the default value for refcheck from True to 1 in numpy ma core.py doesn t resolve the problem.
|
|
Any fix, should also consider 1438, since the order keyword to the ndarray.resize ) method is broken.","bug"
|
|
"6013","1601","Add cyclic garbage collection support to object arrays Trac 1003).
|
|
Original ticket http projects.scipy.org numpy ticket 1003 on 2009 02 08 by stefanv, assigned to unknown.
|
|
Melt van Schoor reported a memory leak, illustrated by the following code from numpy import class Lattice def init self, size) self.size = size self.lat = empty self.size, self.size), dtype=object) for x in range self.size) for y in range self.size) self.lat[x,y]=Location self) compare with this.. self.lat=[[Location self) for y in xrange self.size)] for x in xrange self.size)] class Location list) pass def init self, parent) self.parent=parent This line is crucial to triggering the leak To trigger, run this it will consume increasing amounts of memory) for t in xrange 2000000) Lattice 80)","bug"
|
|
"6013","1602","round vs. numpy.round Trac 1004).
|
|
Original ticket http projects.scipy.org numpy ticket 1004 on 2009 02 09 by trac user lxkain, assigned to unknown.
|
|
I can t believe what I m seeing!
|
|
Version 1.2.1 >>> numpy.round 66873 0.5) 33436.0 >>> round 66873 0.5) 33437.0 >>> round 66875 0.5) 33438.0 >>> numpy.round 66875 0.5) 33438.0","bug"
|
|
"6013","1603","numpy not working with itertools.chain objects Trac 1005).
|
|
Original ticket http projects.scipy.org numpy ticket 1005 on 2009 02 11 by trac user rogerbrent, assigned to unknown.
|
|
numpy.array ), numpy.sum ), etc.
|
|
don t seem to work with itertools.chain objects.
|
|
e.g.
|
|
from itertools import chain iter = chain xrange 5)) array iter) returns array <itertools.chain object at 0x0172C828>, dtype=object).
|
|
list iter) would work as expected, returning [0, 1, 2, 3, 4].
|
|
However if I define my own chain ) function then array iter) works fine.
|
|
Python v2.5.4 and numpy v1.2.1","bug"
|
|
"6013","1604","docs document the feature added in r6358 Trac 1006).
|
|
Original ticket http projects.scipy.org numpy ticket 1006 on 2009 02 12 by pv, assigned to pv.
|
|
The field access pattern arr[[ field1 , field2 ]] should be documented in the reference guide.","bug"
|
|
"6013","1605","docs document the generalized ufunc C API Trac 1007).
|
|
Original ticket http projects.scipy.org numpy ticket 1007 on 2009 02 12 by pv, assigned to pv.
|
|
The generalized ufuncs see 1485) brought some C API changes.
|
|
The numpy reference manual should be updated accordingly.
|
|
There s some text under doc neps , some parts could perhaps be merged from there).","bug"
|
|
"6013","1606","Loss of precision in complex) arcsinh arctanh Trac 1008).
|
|
Original ticket http projects.scipy.org numpy ticket 1008 on 2009 02 15 by pv, assigned to unknown.
|
|
Numpy s complex `arcsinh` and `arctanh` suffer from loss of precision at small arguments >>> import numpy as np >>> np.
|
|
version 1.3.0.dev6366 >>> np.arcsinh 1e 50) 1e 50 >>> np.arcsinh 1e 50 0j) 0j >>> np.arctanh 1e 50) 1e 50 >>> np.arctanh 1e 50 0j) 0j However, complex `arcsin` and `arctan` are OK >>> np.arcsin 1e 50 0j) 1e 50 0j) >>> np.arctan 1e 50 0j) 1e 50 0j)","bug"
|
|
"6013","1607","strptime and python 2.4 Trac 1009).
|
|
Original ticket http projects.scipy.org numpy ticket 1009 on 2009 02 19 by cournape, assigned to unknown.
|
|
Recent change in numpy.lib introduced a problem on python 2.4 datetime.datetime.strptime is not available before 2.5.","bug"
|
|
"6013","1608","Can t test for None in array Trac 1010).
|
|
Original ticket http projects.scipy.org numpy ticket 1010 on 2009 02 19 by trac user anjiro, assigned to unknown.
|
|
Testing for `None` in an array always returns `False` >>> import numpy as np >>> 5 in np.array [None, 5, 23]) True >>> None in np.array [None, 5, 23]) False","bug"
|
|
"6013","1609","Can t test for None in array Trac 1011).
|
|
Original ticket http projects.scipy.org numpy ticket 1011 on 2009 02 19 by trac user anjiro, assigned to unknown.
|
|
Testing for `None` in an array always returns `False` >>> import numpy as np >>> 5 in np.array [None, 5, 23]) True >>> None in np.array [None, 5, 23]) False","bug"
|
|
"6013","1610","dtype attributes change after an arithmetic operation involving a scalar Trac 1012).
|
|
Original ticket http projects.scipy.org numpy ticket 1012 on 2009 02 19 by trac user hana, assigned to unknown.
|
|
If you add an integer scalar to an integer array, the resulting array changes the attributes num and char of dtype >>> from numpy import array >>> a = array [1,2]) >>> b = 0 a >>> a.dtype.num 7 >>> b.dtype.num 5 >>> a.dtype.char l >>> b.dtype.char i This has an effect on scipy.ndimage functions >>> from scipy import ndimage >>> ndimage.sum a) 3.0 >>> ndimage.sum b) Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.5 lib python2.5 site packages scipy ndimage measurements.py , line 126, in sum return nd image.statistics input, labels, index, 0) RuntimeError data type not supported Note that if you do b=a 0 instead of b=0 a), it works.
|
|
We are running python 2.5, numpy 1.2.1, scipy 0.7.0","bug"
|
|
"6013","1611","Ufunc generates unintuitive AttributeError on object arrays Trac 1013).
|
|
Original ticket http projects.scipy.org numpy ticket 1013 on 2009 02 20 by wesm, assigned to unknown.
|
|
This error is very unintuitive for end users, arrays formed from SQL query results can frequently end up as object arrays by accident.
|
|
In [15] arr = np.random.randn 100).astype object) In [16] np.log arr) AttributeError Traceback most recent call last) H \workspace\Python\src\<ipython console> in <module> ) AttributeError log Same AttributeError is raised for other ufuncs","bug"
|
|
"6013","1612","Ufunc generates AttributeError on object arrays Trac 1014).
|
|
Original ticket http projects.scipy.org numpy ticket 1014 on 2009 02 20 by wesm, assigned to unknown.
|
|
This error is very unintuitive for end users, arrays formed from SQL query results can frequently end up as object arrays by accident.
|
|
In [15] arr = np.random.randn 100).astype object) In [16] np.log arr) AttributeError Traceback most recent call last) H \workspace\Python\src\<ipython console> in <module> ) AttributeError log Same AttributeError is raised for other ufuncs","bug"
|
|
"6013","1613","Ufunc generates AttributeError on object arrays Trac 1015).
|
|
Original ticket http projects.scipy.org numpy ticket 1015 on 2009 02 20 by wesm, assigned to unknown.
|
|
This error is very unintuitive for end users, arrays formed from SQL query results can frequently end up as object arrays by accident.
|
|
In [15] arr = np.random.randn 100).astype object) In [16] np.log arr) AttributeError Traceback most recent call last) H \workspace\Python\src\<ipython console> in <module> ) AttributeError log Same AttributeError is raised for other ufuncs","bug"
|
|
"6013","1614","simple improvement to make testing utils work with ndarray subclasses Trac 1016).
|
|
Original ticket http projects.scipy.org numpy ticket 1016 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1615","simple improvement to make testing utils work with ndarray subclasses Trac 1017).
|
|
Original ticket http projects.scipy.org numpy ticket 1017 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1616","simple improvement to make testing utils work with ndarray subclasses Trac 1018).
|
|
Original ticket http projects.scipy.org numpy ticket 1018 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1617","simple improvement to make testing utils work with ndarray subclasses Trac 1019).
|
|
Original ticket http projects.scipy.org numpy ticket 1019 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1618","simple improvement to make testing utils work with ndarray subclasses Trac 1020).
|
|
Original ticket http projects.scipy.org numpy ticket 1020 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1619","simple improvement to make testing utils work with ndarray subclasses Trac 1021).
|
|
Original ticket http projects.scipy.org numpy ticket 1021 on 2009 02 22 by ddale, assigned to unknown.
|
|
I found a simple way to modify numpy s testing utils that makes it easier to use them for unittesting ndarray subclasses.
|
|
I tried running numpy s svn 6456) test suite both with and without the patch applied, and I got identical results 1 known failure and 11 skips on a 64 bit linux platform.
|
|
Its a really small patch, any chance of getting it in before 1.3 is released?","enhancement"
|
|
"6013","1620","numpy segfaults when ufunc is called with ndarray subclass and out=None Trac 1022).
|
|
Original ticket http projects.scipy.org numpy ticket 1022 on 2009 02 22 by ddale, assigned to unknown.
|
|
I am seeing some really strange behavior when I try to pass an ndarray subclass and out=None to numpy s ufuncs.
|
|
This example will reproduce the problem with svn numpy, the first print statement yields 1 as expected, the second yields <type builtin function or method > and the third yields a segmentation fault.
|
|
I m running this as a script, not with ipython import numpy as np class MyArray np.ndarray) array priority = 20 def new cls) return np.asarray 1).view cls).copy ) def repr self) return my array str = repr def mul self, other) return super MyArray, self).
|
|
mul other) def rmul self, other) return super MyArray, self).
|
|
rmul other) mine = MyArray ) print np.multiply 1, 1, None) x = np.multiply mine, mine, None) print type x) print x","bug"
|
|
"6013","1621","Subclassed ndarray fails with ValueError when assigning to a sliced array Trac 1023).
|
|
Original ticket http projects.scipy.org numpy ticket 1023 on 2009 02 23 by trac user psederberg, assigned to unknown.
|
|
Hi Folks I ve been working on a subclass of ndarray that has named dimensions that you can use to slice within a custom getitem .
|
|
For example, you can do something like this ` x[ time>0 ] ` to slice the array where the time dimension is greater than zero.
|
|
This is all working nicely, right up until I try to assign to a slice like this ` x[ time>0 ] = 1 ` or this ` x[ time>0 ] = x[ time>0 ] 1 ` giving me the following error ` ValueError field named time>0 not found.
|
|
` I m not sure what is raising this error, so I m scared it s actually in the interpreter or somewhere deep in the numpy code, but I figured I d try here first.
|
|
Any ideas for what s going on here?
|
|
All assignments with standard slices work just fine.
|
|
Is there some code in numpy that is checking the contents of getitem and causing it to fail here?
|
|
Perhaps something having to do with record arrays?
|
|
Thanks for any thoughts, Per PS > I m running this on NumPy 1.1.0 in Debian Lenny.","bug"
|
|
"6013","1622","Untested changes to numpy.lib.io in r6345 Trac 1024).
|
|
Original ticket http projects.scipy.org numpy ticket 1024 on 2009 02 23 by stefanv, assigned to unknown.
|
|
http projects.scipy.org scipy numpy changeset 6345","bug"
|
|
"6013","1623","Incorrect documentation for put function Trac 1025).
|
|
Original ticket http projects.scipy.org numpy ticket 1025 on 2009 02 24 by trac user afoglia, assigned to unknown.
|
|
The documentation for the put method is incorrect.
|
|
In particular the psuedocode should read ind = array indices, copy=False) vals = array values, copy=False).astype a.dtype) for i, v in zip ind,vals) a.flat[i] = v In particular, it s the last line.
|
|
The i th element of the array a is not filled with the i th element of the values array, but the value that is next in the values array.
|
|
If the documentation were correct as written, then a.put indices, values)[indices] == values would not be guaranteed up to nans).
|
|
http docs.scipy.org doc numpy reference generated numpy.put.html numpy.put","bug"
|
|
"6013","1624","ufuncs do not respect ndarray subclasses in some cases Trac 1026).
|
|
Original ticket http projects.scipy.org numpy ticket 1026 on 2009 02 25 by ddale, assigned to unknown.
|
|
The following script demonstrates that ufuncs do not respect ndarray subclasses when an iterable like list or tuple is passed to the second argument.
|
|
I have also seen this behavior with numpy s !MaskedArray.
|
|
import numpy as np class MyArray np.ndarray) array priority = 20 def new cls) return np.asarray 1).view cls).copy ) def array wrap self, obj, context=None) print array wrap , self, obj, context return obj.view type self)) def str self) return MyArray s) super MyArray,self).
|
|
str ) mine = MyArray ) print np.multiply 3,mine) yields MyArray print np.multiply mine,3) yields MyArray print np.multiply [1,2,3],mine) yields MyArray print np.multiply mine,[1,2,3]) yields ndarray print np.multiply mine, 1,2,3)) yields ndarray print np.multiply mine,np.array [1,2,3])) yields MyArray","bug"
|
|
"6013","1625","Document coremath Trac 1027).
|
|
Original ticket http projects.scipy.org numpy ticket 1027 on 2009 02 26 by cournape, assigned to unknown.
|
|
Document the core math library, if only mention it.","bug"
|
|
"6013","1626","Make coremath lib known to distutils, so other extensions can use it Trac 1028).
|
|
Original ticket http projects.scipy.org numpy ticket 1028 on 2009 02 26 by cournape, assigned to cournape.
|
|
COre math routines are now in separate lib, but extensions outside core cannot use it automatically.","enhancement"
|
|
"6013","1627","Align data on unpickling?
|
|
Trac 1029).
|
|
Original ticket http projects.scipy.org numpy ticket 1029 on 2009 02 26 by pv, assigned to unknown.
|
|
Unpickling numpy arrays sometimes produces surprising data alignment, cf.
|
|
1149 http scipy.org scipy scipy ticket 794 While this behavior is OK, and the two bugs were only exposed because of this, one can wonder If mis aligned data causes performance issues eg.
|
|
in arithmetic).
|
|
If the cost of aligning data on unpickling is small compared to the probability that aligned temporary copies need to be made repeatedly later on.
|
|
Eg.
|
|
calling LAPACK routines repeatedly with a mis aligned array > temporary copy each time.)
|
|
Answering these probably requires some benchmarks on how costly the copy on unpickling actually is.","enhancement"
|
|
"6013","1628","C function PyArray New seems to ignore the strides argument.
|
|
Trac 1030).
|
|
Original ticket http projects.scipy.org numpy ticket 1030 on 2009 02 27 by trac user alexis.roche, assigned to unknown.
|
|
I create the following two arrays as follows X = np.asarray [[0,1,2],[3,4,5]]).astype uint8 ) x = X[ , 0] x == array [0, 3]) I then pass X to a C routine that attempts to create another view on X[ ,0] using npy intp dim[1] = {2}; npy intp strides[1] = {3}; int itemsize = 1; PyArrayObject y = PyArrayObject ) PyArray New PyArray Type, 1, dim, type, strides, void )PyArray DATA X), itemsize, NPY CARRAY, NULL); While I expect y to be equal to x, the result is array [ 0, 1]), which would have been the expected result if I had set the strides argument to 1.
|
|
I am running numpy version 1.3.0.dev6468.
|
|
I have been using PyArray New for 3 years in similar context with devel versions of numpy, and only caught that weird behavior recently.","bug"
|
|
"6013","1629","feature request Add ViewVC access to repository Trac 1031).
|
|
Original ticket http projects.scipy.org numpy ticket 1031 on 2009 03 02 by trac user gbburkhardt, assigned to unknown.
|
|
Please add ViewVC access to the NumPy and SciPy repositories http www.viewvc.org ).
|
|
I ve found this facility really useful in the past, and I suspect it will allow me to pull snapshots from behind my company s proxy.
|
|
Thank you very much.","enhancement"
|
|
"6013","1630","Crash on fastputmask on win32 Trac 1032).
|
|
Original ticket http projects.scipy.org numpy ticket 1032 on 2009 03 02 by cournape, assigned to cournape.
|
|
The test suite crashes in PyArray FastPutMask, since revision r6528.
|
|
I don t see how an additional check could causes a segfault, so I guess it just show a previously undetected bug.","bug"
|
|
"6013","1631","build for r6530 on Solaris doesn t reference lm properly Trac 1033).
|
|
Original ticket http projects.scipy.org numpy ticket 1033 on 2009 03 02 by trac user gbburkhardt, assigned to unknown.
|
|
The umath.so library references fmodf, et al, resulting in ImportError ld.so.1 python fatal relocation error file opt gnu python 2.5.2 lib python2.5 site packages numpy core umath.so symbol fmodf referenced symbol not found Record of build and run of test is attached.","bug"
|
|
"6013","1632","Test new trac setup Trac 1034).
|
|
Original ticket http projects.scipy.org numpy ticket 1034 on 2009 03 03 by stefanv, assigned to cournape.
|
|
Just checking the ticket notification system.","bug"
|
|
"6013","1633","Fix up Nan handling for max amax argmax fmac min sort equivalent Trac 1035).
|
|
Original ticket http projects.scipy.org numpy ticket 1035 on 2009 03 03 by cournape, assigned to cournape.
|
|
This is a meta issue Extensive discussions on the ML http mail.scipy.org pipermail numpy discussion 2008 September 037413.html Also some other issues 1518, 1484, 1307, 1392,","bug"
|
|
"6013","1634","Add ismember ) to arraysetops.py Trac 1036).
|
|
Original ticket http projects.scipy.org numpy ticket 1036 on 2009 03 04 by rc, assigned to rc.
|
|
Neil Crighton wrote As far as I know there isn t an obvious way to get the functionality of setmember1d working on non unique inputs.
|
|
However, I ve needed this operation quite a lot, so here s a function I wrote that does it.
|
|
It s only a few times slower than numpy s setmember1d.
|
|
You re welcome to use it.
|
|
import numpy as np def ismember a1,a2) Test whether items from a2 are in a1.
|
|
This does the same thing as np.setmember1d, but works on non unique arrays.
|
|
Only a few 2 4) times slower than np.setmember1d, and a lot faster than [i in a2 for i in a1].
|
|
An example that np.setmember1d gets wrong >>> a1 = np.array [5,4,5,3,4,4,3,4,3,5,2,1,5,5]) >>> a2 = [2,3,4] >>> mask = ismember a1,a2) >>> a1[mask] array [4, 3, 4, 4, 3, 4, 3, 2]) a2 = set a2) a1 = np.asarray a1) ind = a1.argsort ) a1 = a1[ind] mask = [] need this bit because prev is not defined for first item item = a1[0] if item in a2 mask.append True) a2.remove item) else mask.append False) prev = item main loop for item in a1[1 ] if item == prev mask.append mask[ 1]) elif item in a2 mask.append True) prev = item a2.remove item) else mask.append False) prev = item restore mask to original ordering of a1 and return mask = np.array mask) return mask[ind.argsort )]","enhancement"
|
|
"6013","1635","Pickling unpickling empty unicode types does not work Trac 1037).
|
|
Original ticket http projects.scipy.org numpy ticket 1037 on 2009 03 04 by FrancescAlted, assigned to cournape.
|
|
This shows the problem In [18] cp = cPickle.dumps numpy.array [u ]), 0) In [19] cPickle.loads cp) TypeError Traceback most recent call last) home faltet PyTables pytables PyTablesPro trunk <ipython console> in <module> ) TypeError data type not understood , <type numpy.dtype >, U0 , 0, 1)) However, plain strings does work In [20] cp = cPickle.dumps numpy.array [ ]), 0) In [21] cPickle.loads cp) Out[21] array [ ], dtype= |S1 )","bug"
|
|
"6013","1636","test failing test umath.TestComplexFunctions.test loss of precision <type numpy.complex256 >,) Trac 1038).
|
|
Original ticket http projects.scipy.org numpy ticket 1038 on 2009 03 05 by stsci sienkiew, assigned to unknown.
|
|
numpy 1.3.0.dev6545 python 2.5.1 nose 0.10.4 This test fails on Mac OSX 10.4 tiger), PPC, 32 bit Solaris 8, SPARC, 32 bit This failure not reported on Mac OSX 10.4 tiger), x86, 32 bit Red Hat Enterprise Linux, x86, 32 bit Red Hat Enterprise Linux, x86 64, 64 bit ====================================================================== FAIL test umath.TestComplexFunctions.test loss of precision <type numpy.complex256 >,) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 nose case.py , line 182, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.1 numpy core tests test umath.py , line 455, in check loss of precision check x series, 2 eps) File usr stsci pyssgdev 2.5.1 numpy core tests test umath.py , line 437, in check assert np.all d < rtol), x[np.argmax d)], d.max )) AssertionError nan, nan)","bug"
|
|
"6013","1637","test fails Test find duplicates Trac 1039).
|
|
Original ticket http projects.scipy.org numpy ticket 1039 on 2009 03 05 by stsci sienkiew, assigned to unknown.
|
|
numpy 1.3.0.dev6545 python 2.5.1 nose 0.10.4 This test fails on Solaris 8, sparc, 32 bit No failure reported on Mac OSX 10.4 tiger), x86 or PPC Red Hat Enterprise Linux 3 or 4, x86 or x86 64 ====================================================================== FAIL Test find duplicates Traceback most recent call last) File usr stsci pyssgdev 2.5.1 numpy testing decorators.py , line 169, in knownfailer return f args, kwargs) File usr ra pyssg 2.5.1 numpy lib tests test recfunctions.py , line 152, in test find duplicates assert equal test[ 1], control) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 121, in assert equal return assert array equal actual, desired, err msg) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 193, in assert array equal header= Arrays are not equal ) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 186, in assert array compare verbose=verbose, header=header) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 295, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [2, 0]) y array [0, 2])","bug"
|
|
"6013","1638","64 bit arrayobject bugs Trac 1040).
|
|
Original ticket http projects.scipy.org numpy ticket 1040 on 2009 03 05 by trac user jmiller, assigned to cournape.
|
|
This ticket and patch addresses a number of 64 bit bugs in arrayobject.c PyArray ToString >>> a.tostring ) Traceback most recent call last) File <stdin> , line 1, in <module> SystemError Negative size passed to PyString FromStringAndSize PyArray CopyObject ?
|
|
>>> a = numpy.ones 0x80000000, dtype= c ) >>> a[ ] = 1 0x80000000 Traceback most recent call last) File <stdin> , line 1, in <module> SystemError error return without exception set >>> a.astype b1 ) Segmentation fault core dumped) The patch also fixes a number of error string formatting bugs caused by use of d for a Py ssize t sized value.","bug"
|
|
"6013","1639","gfortran win32 Trac 1041).
|
|
Original ticket http projects.scipy.org numpy ticket 1041 on 2009 03 08 by cournape, assigned to cournape.
|
|
Not all combinations will work document which one do work.","bug"
|
|
"6013","1640","clip, fill and co behavior when the fill clipping values cannot be casted meaningfully to the input Trac 1042).
|
|
Original ticket http projects.scipy.org numpy ticket 1042 on 2009 03 09 by cournape, assigned to unknown.
|
|
For example !python import numpy as np a = np.array [1, 2, 3]) np.fill a, 1.3) 1.3 converted to 1 Clip is another interesting case...","bug"
|
|
"6013","1641","adding new fortran compiler Trac 1043).
|
|
Original ticket http projects.scipy.org numpy ticket 1043 on 2009 03 09 by trac user avi, assigned to unknown.
|
|
Hi, I have added a new compiler pathscale and added it to numpy distutils fcompiler and edited init .py to add it to the list.
|
|
However when I run python setup.py config fc I get the pathscale compiler recognized but then the setup bypasses and continues and subsequently picks the gfortran, which I don t want.
|
|
When I added some verbosity to the above, I noticed this difference between pathscale and gfortran python setup.py v config fc Running os.system opt pathscale ice9 native 3.2n B sicortex bin pathf95 v ; echo ?
|
|
> tmp tmpoXv0XE VCZd1T ) > tmp tmpoXv0XE yNOdmp 2> 1 ) update environment ...) customize PGroupFCompiler with gfortran.. Running os.system .root0 usr bin gfortran version ; echo ?
|
|
> tmp tmpoXv0XE xX3wVg ) > tmp tmpoXv0XE tQhxpb 2> 1 ) update environment ...) exec command [ .root0 usr bin gfortran , Wall , ffixed form , fno second underscore , fPIC , O3 , funroll loops , print libgcc file name ],) Retaining cwd .root0 net home avi builds numpy 1.2.1 preserve environment []) So it seems to me that it is finding the pathscale compiler but skipping the exec command.
|
|
Any help in resolving this appreciated.
|
|
I am attaching the log file for the above script run and the pathscale.py files.
|
|
Other tidbits on system and os.. sc1 m3n6 builds scipy 0.7.0) avi uname a Linux sc1 m3n6.scsystem 2.6.18.8 sc lustre perfmon 1 SMP Thu Sep 25 13 19 45 EDT 2008 mips64 SiCortex ICE9B V1.0 FPU V0.1 SiCortex SC 1000 GNU Linux sc1 m3n6 builds scipy 0.7.0) avi python c import os,sys;print os.name,sys.platform posix linux2 Thank You Avi","enhancement"
|
|
"6013","1642","Ufuncs return scalars for 0D input Trac 1044).
|
|
Original ticket http projects.scipy.org numpy ticket 1044 on 2009 03 09 by pv, assigned to unknown.
|
|
Continuation of 1440.
|
|
Ufuncs return array scalars for 0D array input >>> import numpy as np >>> type np.conjugate np.array 1 2j))) <type numpy.complex128 > >>> type np.sum np.array 3.
|
|
), np.array 5.)))
|
|
<type numpy.float64 > Should they return 0D arrays instead?","bug"
|
|
"6013","1643","Invalid formatting of inf nan imaginary part Trac 1045).
|
|
Original ticket http projects.scipy.org numpy ticket 1045 on 2009 03 09 by pv, assigned to unknown.
|
|
Complex inf nan are formatted wrong >>> import numpy as np >>> np.
|
|
version 1.3.0.dev6603.05139b8de06aa0aa8f2574540876589cbc8820ee >>> print np.complex128 complex 1, np.inf)) 1infj) >>> print repr np.complex128 complex 1, np.inf))) 1infj) >>> print np.complex128 complex 1, np.nan)) 1nanj) >>> print np.complex128 complex 1, np.nan)) 1nanj) So `inf` and `nan` don t get prefixed by the positive sign.
|
|
Ordinary numbers do >>> print np.complex128 complex np.inf, 1)) inf 1j)","bug"
|
|
"6013","1644","SKIP option to doctests is not supported by python 2.4 Trac 1046).
|
|
Original ticket http projects.scipy.org numpy ticket 1046 on 2009 03 10 by cournape, assigned to pv.
|
|
Some doctests failures happen on 2.4 because of this.","bug"
|
|
"6013","1645","build failure r6609 Trac 1047).
|
|
Original ticket http projects.scipy.org numpy ticket 1047 on 2009 03 10 by cournape, assigned to unknown.
|
|
Syntax error in scalartypes.inc","bug"
|
|
"6013","1646","numpy build fails on Solaris Trac 1048).
|
|
Original ticket http projects.scipy.org numpy ticket 1048 on 2009 03 10 by stsci sienkiew, assigned to unknown.
|
|
numpy trunk, revision 6604 Python 2.5.1 Solaris 8 SUNWSpro 6.2 C compiler cc G build temp.solaris 2.8 sun4u 2.5 numpy core src multiarraymodule.o Lbuild temp.solaris 2.8 sun4u 2.5 lnpymath lm o build lib.solaris 2.8 sun4u 2.5 numpy core multiarray.so building numpy.core.umath extension compiling C sources C compiler cc DNDEBUG O xcode=pic32 compile options Ibuild src.solaris 2.8 sun4u 2.5 numpy core src Inumpy core include Ibuild src.solaris 2.8 sun4u 2.5 numpy core include numpy Inumpy core src Inumpy core include I usr ra pyssg Python 2.5.1 include python2.5 c cc build src.solaris 2.8 sun4u 2.5 numpy core src umathmodule.c numpy core include numpy ufuncobject.h , line 397 warning statement not reached numpy core src umath loops.inc.src , line 183 warning non constant initializer op SCONV numpy core src umath loops.inc.src , line 212 warning non constant initializer op SCONV numpy core src umath loops.inc.src , line 213 warning non constant initializer op SCONV numpy core src umath ufunc object.inc , line 3178 syntax error before or at double numpy core src umath ufunc object.inc , line 3183 undefined symbol maxpriority numpy core src umath ufunc object.inc , line 3239 cannot recover from previous errors cc acomp failed for build src.solaris 2.8 sun4u 2.5 numpy core src umathmodule.c numpy core include numpy ufuncobject.h , line 397 warning statement not reached numpy core src umath loops.inc.src , line 183 warning non constant initializer op SCONV numpy core src umath loops.inc.src , line 212 warning non constant initializer op SCONV numpy core src umath loops.inc.src , line 213 warning non constant initializer op SCONV numpy core src umath ufunc object.inc , line 3178 syntax error before or at double numpy core src umath ufunc object.inc , line 3183 undefined symbol maxpriority numpy core src umath ufunc object.inc , line 3239 cannot recover from previous errors cc acomp failed for build src.solaris 2.8 sun4u 2.5 numpy core src umathmodule.c error Command cc DNDEBUG O xcode=pic32 Ibuild src.solaris 2.8 sun4u 2.5 numpy core src Inumpy core include Ibuild src.solaris 2.8 sun4u 2.5 numpy core include numpy Inumpy core src Inumpy core include I usr ra pyssg Python 2.5.1 include python2.5 c build src.solaris 2.8 sun4u 2.5 numpy core src umathmodule.c o build temp.solaris 2.8 sun4u 2.5 build src.solaris 2.8 sun4u 2.5 numpy core src umathmodule.o failed with exit status 2 At line 183 of numpy core src umath loops.inc.src, this is a struct initializer, which can only contain constants.
|
|
From the context, it looks like maybe it should say cdouble tmp; cdouble out; tmp.real = double) in1[0]); tmp.imag = double) in1[1]); At line 3178 of umath ufunc object.inc, a declaration comes after an executable statement.","bug"
|
|
"6013","1647","Some tests fail with the new behavior of the sign ufunc.
|
|
Trac 1049).
|
|
Original ticket http projects.scipy.org numpy ticket 1049 on 2009 03 10 by charris, assigned to unknown.
|
|
Sign now returns nan as the sign of nan.
|
|
This breaks some tests and until we decide on the logical value of nan I will leave them broken.
|
|
Failing tests are File usr lib64 python2.5 site packages numpy lib tests test ufunclike.py , line 25, in test ufunclike Failed example nx.sign a) Expected array [ 1., 1., 0., 0., 1., 1.])
|
|
Got array [ 1., 1., NaN, 0., 1., 1.])
|
|
File usr lib64 python2.5 site packages numpy lib tests test ufunclike.py , line 40, in test ufunclike Failed example nx.sign a, y) Expected array [ True, True, False, False, True, True], dtype=bool) Got array [ True, True, True, False, True, True], dtype=bool) File usr lib64 python2.5 site packages numpy lib tests test ufunclike.py , line 43, in test ufunclike Failed example y Expected array [ True, True, False, False, True, True], dtype=bool) Got array [ True, True, True, False, True, True], dtype=bool) ..............................................................................................................................................................................................................................................................................................................................................................................................................................","bug"
|
|
"6013","1648","Make array printing consistent with our recent formatting work Trac 1050).
|
|
Original ticket http projects.scipy.org numpy ticket 1050 on 2009 03 12 by cournape, assigned to unknown.
|
|
Example !python import numpy as np a = np.array [np.nan]) print a print a[0] Returns array [ NaN]) nan","bug"
|
|
"6013","1649","Renaming and removing NaN and related IEEE754 special cases Trac 1051).
|
|
Original ticket http projects.scipy.org numpy ticket 1051 on 2009 03 12 by bsouthey, assigned to unknown.
|
|
Numpy defines multiple versions of certain IEEE 754 special cases Inf , inf , infty , Infinity nan , NaN In addition, NAN, INF, PINF, NINF, PZERO, and NZERO are also defined.
|
|
In a 2008 discussion on the numpy mailing list Travis suggested http mail.scipy.org pipermail numpy discussion 2008 April 032661.html to use numpy.inf numpy.nan I think I have found all the relevant files and created the appropriate patches for them.
|
|
The net result is that numpy.nan, numpy.inf, numpy.ninf, numpy.pzero, and numpy.nzero are defined.","enhancement"
|
|
"6013","1650","AttributeError keys Trac 1052).
|
|
Original ticket http projects.scipy.org numpy ticket 1052 on 2009 03 13 by trac user Jack, assigned to pierregm.
|
|
I have installed Gnuradio 3.1.3 on RedHat 4.The installation of gnuradio was successful.Even I have got USRP connected and have loaded the firmware.
|
|
But when I try to run the given examples of gr utils for example usrp fft.py i get the following error [root localhost python] .
|
|
usrp fft.py Traceback most recent call last) File .
|
|
usrp fft.py , line 23, in ?
|
|
from gnuradio import gr, gru File usr lib python2.3 site packages gnuradio gru init .py , line 37, in ?
|
|
exec from gnuradio.gruimpl.
|
|
s import f,) File <string> , line 1, in ?
|
|
File usr lib python2.3 site packages gnuradio gruimpl gnuplot freqz.py , line 28, in ?
|
|
import numpy File usr lib python2.3 site packages numpy init .py , line 107, in ?
|
|
import ma File usr lib python2.3 site packages numpy ma init .py , line 14, in ?
|
|
import core File usr lib python2.3 site packages numpy ma core.py , line 114, in ?
|
|
max filler.update [ k, numpy.inf) for k in [numpy.float32, numpy.float64]]) AttributeError keys What is the reason for such type of error.Can someone guide me through this problem.
|
|
Thanks.","bug"
|
|
"6013","1651","let memmap use offset keyword in mmap.mmap Trac 1053).
|
|
Original ticket http projects.scipy.org numpy ticket 1053 on 2009 03 14 by sturlamolden, assigned to unknown.
|
|
In Python 2.6, mmap has a offset keyword.
|
|
NumPy s memmap should use this to allow big files to be memory mapped on 32 bit systems.
|
|
Only a minor change is required.","enhancement"
|
|
"6013","1652","Feature request intent transpose) option in f2py Trac 1054).
|
|
Original ticket http projects.scipy.org numpy ticket 1054 on 2009 03 14 by trac user mscottshell, assigned to unknown.
|
|
I propose the addition of an intent transpose) option in f2py that will reverse the ordering of array axes as seen in Fortran relative to those of the calling routines in Python.
|
|
The advantage of this option is that it will permit simultaneous Python and Fortran access of array data in a way that is natural to each, and it will eliminate the need for array copying.
|
|
Consider the following example taken from a molecular simulation routine for making changes to the positions of N atoms in an N,3) array called Pos.
|
|
If we were to perform this operation in Python, it might look like for p in Pos do something with p, a length 3 array Now imagine we were to construct a Fortran routine, to be compiled with f2py, that were to contain a similar loop over atoms subroutine FortranFunc Pos, Dim, NAtom) implicit none integer, intent in) Dim, NAtom real 8), intent inout), dimension NAtom, Dim) Pos !f2py intent in,out) Pos integer i do i = 1, NAtom !do something with Pos i, ), a length 3 array enddo end subroutine Here, two factors will make this function inefficient 1) a copy is made of Pos to Fortran ordering by the f2py routines, and 2) the loop over the Pos vector clashes with Fortran ordering because the last axis is accessed the fastest, rather than the first.
|
|
On the other hand, an intent transpose) argument would reverse the order of the dimensions in the function when an array is passed to a Fortran routine, and the reverse them again upon exiting i.e., so that only Fortran saw the reversal).
|
|
That is, the above function might be written subroutine FortranFunc Pos, Dim, NAtom) implicit none integer, intent in) Dim, NAtom real 8), intent inout), dimension Dim, NAtom) Pos !f2py intent in,out,transpose) Pos integer i do i = 1, NAtom !do something with Pos ,i), a length 3 array enddo end subroutine By introducing such a transpose argument, one would avoid copying, and it would be possible to simultaneously achieve natural traversing of arrays in Python and Fortran for the same loop construct.
|
|
With current conventions, the closest workaround seems to be to use Fortran ordered arrays e.g., order= F ).
|
|
However, any numerical NumPy operation will back convert these to C ordered e.g., Pos = Pos 1), which will require subsequent copying and conversions when the f2py compiled Fortran routine is called.
|
|
Alternatively, the .T view of arrays might be used here upon sending arrays to Fortran routines, but this would seem to litter Python code unnecessarily.
|
|
For programmers who are concerned with the most efficient code, and thus aware of ordering issues, the intent transpose) option would come as a natural convention, rather than a point of confusion.
|
|
In this sense, the transposition of array indices in going to Fortran naturally reflects differences in storage order between the two languages; though the loops will be different between Python and Fortran, the actual manner in which the array data is accessed in memory would be the same.
|
|
Thus, an intent transpose) option might help create an awareness of efficiency issues and inherent storage differences between Fortran and Python C) that could guide faster code development.","enhancement"
|
|
"6013","1653","enhancements to fftpack Trac 1055).
|
|
Original ticket http projects.scipy.org numpy ticket 1055 on 2009 03 15 by sturlamolden, assigned to charris.
|
|
1) Make sure fftpack lite releases the GIL.
|
|
Functions in fftpack lite.c and fftpack.c are reentrant.
|
|
Currently NumPy s FFTs lock up the interpreter.
|
|
2) Use C99 restrict qualifier, GNU extension restrict , or MSVC extension restrict in fftpack.c to help compilers optimize better.
|
|
The code is derived from Fortran so arrays are not aliased.
|
|
3) Use OpenMP pragmas in fftpack lite for multiple FFTs in parallel e.g.
|
|
for multidimensional transforms).","enhancement"
|
|
"6013","1654","typedef npy intp to Py ssize t cf.
|
|
PEP353) Trac 1056).
|
|
Original ticket http projects.scipy.org numpy ticket 1056 on 2009 03 15 by sturlamolden, assigned to unknown.
|
|
arrayobject.h contains these typedefs typedef Py intptr t npy intp; typedef Py uintptr t npy uintp; From PEP 353 Conceptually, Py intptr t and Py ssize t are different things Py intptr t needs to be the same size as void , and Py ssize t the same size as size t. These could differ, e.g.
|
|
on machines where pointers have segment and offset.
|
|
On current flat address space machines, there is no difference, so for all practical purposes, Py intptr t would have worked as well.
|
|
This would thus be preferred, although the difference will not be noticeable on most platforms typedef Py ssize t npy intp; typedef size t npy uintp;","bug"
|
|
"6013","1655","numpy.array inplace exchange duplicates overwrites elements Trac 1057).
|
|
Original ticket http projects.scipy.org numpy ticket 1057 on 2009 03 18 by trac user praeteri, assigned to unknown.
|
|
simply check this In [1] from numpy import array In [2] a = array [ array [1,1]), array [2,2]), array [3,3]), array [4,4]), array [5,5]) ]) In [3] a[0], a[1] = a[1], a[0] In [4] a Out[4] array [[2, 2], [2, 2], [3, 3], [4, 4], [5, 5]]) This data format is used e.g.
|
|
by pylab.load.
|
|
The inplace exchange e.g.
|
|
by random.shuffle.
|
|
Hence, it is a very common situation...
|
|
I don t know if this is the way it should be.","bug"
|
|
"6013","1656","fromiter with dtype= b or B Trac 1058).
|
|
Original ticket http projects.scipy.org numpy ticket 1058 on 2009 03 19 by trac user russel, assigned to unknown.
|
|
As of SVN from this morning, fromiter cannot create an array whose dtype is a byte type In [1] np.fromiter range 10), dtype= b ) MemoryError Traceback most recent call last) Users baxter <ipython console> in <module> ) MemoryError cannot allocate array memory patch and test attached.","bug"
|
|
"6013","1657","Invalid value of longdouble finfo tiny eps on some platforms Trac 1059).
|
|
Original ticket http projects.scipy.org numpy ticket 1059 on 2009 03 19 by pv, assigned to unknown.
|
|
On some platforms, `finfo` apparently reports bogus limits for `longdouble` dual G5 Mac running OS X 10.5.6 and Enthought s EPD Python 2.5.2 |EPD Py25 4.1.30101| r252 60911, Dec 19 2008, 15 28 32) http permalink.gmane.org gmane.comp.python.numeric.general 28872 possibly also the FreeBSD 64 buildbot http buildbot.scipy.org builders FreeBSD 64 builds 874 steps shell 2 logs stdio On the G5 Mac, the user reported >>> import numpy as np >>> info = np.finfo np.longcomplex) >>> print eps , info.eps, info.eps.dtype >>> print tiny , info.tiny, info.tiny.dtype >>> print log10 , np.log10 info.tiny), np.log10 info.tiny info.eps) eps 1.3817869701e 76 float128 tiny 1.08420217274e 19 float128 log10 nan nan Which means that both `eps` and `tiny` are bogus.
|
|
The latter is also negative.","bug"
|
|
"6013","1658","patch to move ccompiler to types module Trac 1060).
|
|
Original ticket http projects.scipy.org numpy ticket 1060 on 2009 03 23 by trac user gaurdro, assigned to unknown.
|
|
new module is removed in Py3k, types should be used.
|
|
I updated the import statement and the first only?)
|
|
use of the module in the file to use the new types module.","enhancement"
|
|
"6013","1659","Unhelpful error message for big arrays creation Trac 1061).
|
|
Original ticket http projects.scipy.org numpy ticket 1061 on 2009 03 23 by cournape, assigned to unknown.
|
|
From Matthew Brett In [13] np.zeros n,)) OverflowError Traceback most recent call last) home mb312 tmp max speed.py in <module> ) > 1 2 3 4 5 OverflowError long int too large to convert to int","bug"
|
|
"6013","1660","np.arange does not fail when input size overflows Trac 1062).
|
|
Original ticket http projects.scipy.org numpy ticket 1062 on 2009 03 23 by cournape, assigned to unknown.
|
|
Reported by M. Brett import numpy as np np.arange 2 32) Returns an empty array on 32 bits machine).","bug"
|
|
"6013","1661","Add a format specifier for Py ssize t Trac 1063).
|
|
Original ticket http projects.scipy.org numpy ticket 1063 on 2009 03 23 by cournape, assigned to unknown.
|
|
Right now, we have no portable way to print a Py ssize t z).
|
|
Would be nice to fix a couple of warnings on some 64 bits arch.","bug"
|
|
"6013","1662","numpy.interp bug Trac 1064).
|
|
Original ticket http projects.scipy.org numpy ticket 1064 on 2009 03 23 by trac user changimeno, assigned to charris.
|
|
`interp x, xp, fp, left=None, right=None)` the right parameter option does not work properly.
|
|
Example In [2] x = arange 4) 2 2 In [3] y=x 2 In [4] xx = arange 11) In [5] x Out[5] array [2, 4, 6, 8]) In [6] y Out[6] array [ 4, 16, 36, 64]) In [7] xx Out[7] array [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]) In [14] numpy.interp xx,x,y) Out[14] array [ 4., 4., 4., 10., 16., 26., 36., 50., 64., 64., 64.])
|
|
>>>> RIGHT!!!!
|
|
In [15] numpy.interp xx,x,y,left=0.)
|
|
Out[15] array [ 0., 0., 4., 10., 16., 26., 36., 50., 64., 64., 64.])
|
|
>>>> RIGHT!!!!
|
|
In [16] numpy.interp xx,x,y,left=0.,right=0.)
|
|
Out[16] array [ 0., 0., 4., 10., 16., 26., 36., 50., 0., 0., 0.])
|
|
>>>> WRONG!!!!
|
|
numpy.interp set the vales of the interpolated array yy) for xx[i] >= x[len x) 1] to the given RIGHT value.
|
|
So, the last GOOD value xx[i] == x[len x) 1]) is wrongly set to the RIGHT value instead of the proper value yy[i] = y[len x) 1].
|
|
I edited the file numpy lib src compiled base.c and fixed this bug see attachment).
|
|
Now In [19] numpy.interp xx,x,y,0.,0.)
|
|
Out[19] array [ 0., 0., 4., 10., 16., 26., 36., 50., 64., 0., 0.])
|
|
>>>> RIGHT!!!!
|
|
My only worry is that I modified binary search together with arr interp and I do not know if this function is called by other routines and whether this modification can alter the results.
|
|
diff > sgg > else if dval > dlist [len 1]) > result = len ; > sgg 448c454,455 < else if indx >= lenxp 1) > sgg else if indx >= lenxp 1) > else if indx > lenxp 1) Cheers, Chan","bug"
|
|
"6013","1663","Unusual behavior of numpy.tile Invalid issue, please close.
|
|
Trac 1065).
|
|
Original ticket http projects.scipy.org numpy ticket 1065 on 2009 03 24 by trac user lawrennd, assigned to unknown.
|
|
Hi, a = np.array [[1], [2], [3], [4]]) a.shape 4, 1) b = np.tile a, 1, 1, 2)) b.shape 1, 4, 2) I m surprised at this behavior, shouldn t the shape be 4, 1, 2)?
|
|
This applies in Python 2.4 with numpy 1.2.1 Apologies if this is a known feature !","bug"
|
|
"6013","1664","Minor Trac workflow glitches Trac 1066).
|
|
Original ticket http projects.scipy.org numpy ticket 1066 on 2009 03 24 by pv, assigned to stefanv.
|
|
There are currently some workflow glitches in this trac There are both `apply` and `review positive` states for tickets The Severity field is present, and defaults to `blocker`.
|
|
I think the intention was to disable it altogether.
|
|
I can try to look at this eventually, if someone else does not manage to do it first.","bug"
|
|
"6013","1665","gfortran does not get necessary flags for universal builds Trac 1067).
|
|
Original ticket http projects.scipy.org numpy ticket 1067 on 2009 03 25 by cournape, assigned to dmcooke.
|
|
Reported by Chris Barker.
|
|
The problem is in numpy.distutils some gfortran versions are not correctly detected as universal","bug"
|
|
"6013","1666","numpy crashes on long double computation on windows for amd64 Trac 1068).
|
|
Original ticket http projects.scipy.org numpy ticket 1068 on 2009 03 26 by cournape, assigned to unknown.
|
|
See Bruce Southey report !python import numpy as np info = np.finfo np.longcomplex) Info the beta1 was built with the binaries from equation.com gcc 4.4 20090313).
|
|
I can t reproduce this with my own toolchain mingw rev 692 gcc snapshot 4.4 20090320).","bug"
|
|
"6013","1667","cholesky decomposition not working in win32 1.3.0rc superpack Trac 1069).
|
|
Original ticket http projects.scipy.org numpy ticket 1069 on 2009 03 29 by trac user chris.flesher, assigned to unknown.
|
|
So I tried using the cholesky decomposition using the following example from numpy import A = mat 2,1;3,4 ) linalg.cholesky A) This code gives the error that A is not positive semidefinate even though it is.
|
|
I tried running the same code using numpy 1.0.1 on ubuntu and the code produces the correct answer.
|
|
This is a pain for me since I am trying to reproduce some MATLAB code in Windows.","bug"
|
|
"6013","1668","doc sphinxext only directives incompatible w Sphinx 0.6.1 Trac 1070).
|
|
Original ticket http projects.scipy.org numpy ticket 1070 on 2009 03 29 by pierregm, assigned to pv.
|
|
The syntax has changed in Sphinx 0.6.1, and ``latexwriter`` is no longer available directly.
|
|
Solution 1 compatible w older versions of Sphinx) Use try from sphinx.latexwriter import LaTeXTranslator except ImportError from sphinx.writers.latex import LaTeXTranslator Solution 2 simplest) Just use from sphinx.writers.latex import LaTeXTranslator","bug"
|
|
"6013","1669","loadtxt fails if the last column contains empty value Trac 1071).
|
|
Original ticket http projects.scipy.org numpy ticket 1071 on 2009 03 30 by trac user Electrion, assigned to unknown.
|
|
Hello, I recently noticed that loadtxt ) fails when I try to load a file that contains no value on the last column when it is TAB delimited.
|
|
For example s = StringIO X\tY\tC\n1.\t1.\t1.\n2.\t2.\t\n ) loadtxt s) will fail with Exception ValueError setting an array element with a sequence.
|
|
The use of dtypes or converters like {2 lambda s float s or Nan )} ) doesn t help.
|
|
After investigation, I noticed that the loadtxt function strip all lines before splitting and then remove the \t\n .
|
|
If think the split line subfunction in numpy.io.loadtxt def split line line) Chop off comments, strip, and split at delimiter.
|
|
line = line.split comments)[0].strip ) [...] should be replaced by def split line line) Chop off comments, strip, and split at delimiter.
|
|
line = line.split comments)[0].strip \r\n ) [...] to remove any new line character but not the delimiter.","bug"
|
|
"6013","1670","Example conventions should be clarified in reference guide Trac 1072).
|
|
Original ticket http projects.scipy.org numpy ticket 1072 on 2009 03 30 by trac user dlary, assigned to pv.
|
|
When you try and follow the example on the page http docs.scipy.org doc numpy reference generated numpy.fft.fft.html numpy.fft.fft you get the following error NameError name np is not defined","bug"
|
|
"6013","1671","Merge autosummary to Sphinx Trac 1073).
|
|
Original ticket http projects.scipy.org numpy ticket 1073 on 2009 03 30 by pv, assigned to pv.
|
|
Sphinx 0.6 ships with the autosummary extension.
|
|
However, the version shipped does not contain all features contained in our sphinxext.autosummary, and does not play nice with Numpydoc.
|
|
This situation should be fixed the autosummary extension should be merged to Sphinx, and removed from Numpy s sphinxext.","bug"
|
|
"6013","1672","Merge plot directive back to Matplotlib Trac 1074).
|
|
Original ticket http projects.scipy.org numpy ticket 1074 on 2009 03 30 by pv, assigned to mdboom.
|
|
Numpy s sphinxext contains a modified plot directive, extracted from Matplotlib at one point.
|
|
It contains some features not present in Matplotlib s plot directive.
|
|
The new features should be merged back to matplotlib.","bug"
|
|
"6013","1673","Improving contiguity detection Trac 1075).
|
|
Original ticket http projects.scipy.org numpy ticket 1075 on 2009 03 31 by cournape, assigned to unknown.
|
|
The contiguous nature of arrays is more conservative than it should !python import numpy as np a = np.ones 10) b = a[None, ] print a.flags print b.flags b is just a view on a, and the data buffer is contiguous in both cases, but b is flagged as neither C or F contiguous","bug"
|
|
"6013","1674","fftfreq incompatible with fft for even n Trac 1076).
|
|
Original ticket http projects.scipy.org numpy ticket 1076 on 2009 04 03 by trac user alk, assigned to unknown.
|
|
If n is even for example n = 8, then fft is described, to return the frequencies at [0, 1, 2, 3, 4, 3, 2, 1].
|
|
But fftfreq 8) 8 returns [ 0., 1., 2., 3., 4., 3., 2., 1.]","bug"
|
|
"6013","1675","fftshift incompatible with fft for even n Trac 1077).
|
|
Original ticket http projects.scipy.org numpy ticket 1077 on 2009 04 03 by trac user alk, assigned to unknown.
|
|
If n is even for example n = 8, then fft is described, to return the frequencies at [0, 1, 2, 3, 4, 3, 2, 1].
|
|
But fftshift array [0,1,2,3,4, 3, 2, 1])) returns [ 4, 3, 2, 1, 0, 1, 2, 3] and not [ 3, 2, 1, 0, 1, 2, 3, 4]","bug"
|
|
"6013","1676","Bus error attempting to create ragged array Trac 1078).
|
|
Original ticket http projects.scipy.org numpy ticket 1078 on 2009 04 06 by trac user anand, assigned to unknown.
|
|
This has been observed as recently as 1.4.0.dev6728 In [1] import numpy as np In [2] a=np.ones 2) In [3] b=np.array 3) In [4] np.array a,b)) Segmentation fault","bug"
|
|
"6013","1677","add option for header to savetxt Trac 1079).
|
|
Original ticket http projects.scipy.org numpy ticket 1079 on 2009 04 06 by trac user kkuhlman, assigned to unknown.
|
|
To create symmetry with the skiprows option in loadtxt, it would make sense to have a header option in savetxt.
|
|
This would allow writing a string of some header info before writing the array.
|
|
This seems to be a simple change, so I will include a simple patch to add this.","enhancement"
|
|
"6013","1678","Bug in creating large arrays Trac 1080).
|
|
Original ticket http projects.scipy.org numpy ticket 1080 on 2009 04 08 by trac user anusr, assigned to unknown.
|
|
It seems that you can create really large arrays say Numpy.zeros) if the dimensions are in the power of 2, but otherwise it would give a ValueError dimensions too large .
|
|
For example, a = numpy.zeros [1048576, 1048576, 1048576, 1048576]) can be happily created without generating a single error with a regular 4 GB RAM.
|
|
1048576 = 2 20.
|
|
However, if you try to create an array of a considerably less size, say d = numpy.zeros [102480, 102840]), it will give you the above mentioned error.
|
|
It s probably a problem with 32 bit system, but it s a problem.","bug"
|
|
"6013","1679","Strange behavior of array ) in a corner case Trac 1081).
|
|
Original ticket http projects.scipy.org numpy ticket 1081 on 2009 04 09 by GaelVaroquaux, assigned to unknown.
|
|
In [1] from numpy import array In [2] array [array []), array 0, object)]) [1] 30002 segmentation fault ipython","bug"
|
|
"6013","1680","numpybook doesn t build Trac 1082).
|
|
Original ticket http projects.scipy.org numpy ticket 1082 on 2009 04 10 by charris, assigned to pv.
|
|
Running [charris f9 numpybook] lyx export pdf numpybook.lyx ... LayoutFile.cpp 278) Existing textclass mybook is redefined by home charris Workspace numpy.git doc numpybook mybook.layout This is with LyX 1.6.2 2009 03 14).
|
|
Running within lyx produces latex errors when trying to export dvi.
|
|
The eps extension is also unrecognized.
|
|
There may be a missing dependency somewhere...
|
|
I may be doing something wrong but the directory contains no documentation on making the document.","bug"
|
|
"6013","1681","numpy.array truncates to specified dtype Trac 1083).
|
|
Original ticket http projects.scipy.org numpy ticket 1083 on 2009 04 11 by inducer, assigned to unknown.
|
|
>>> numpy.array [324938], dtype=numpy.uint8) array [74], dtype=uint8) Errors should never pass silently.
|
|
Unless explicitly silenced.","bug"
|
|
"6013","1682","can t import numpy on Vista x64 system Trac 1084).
|
|
Original ticket http projects.scipy.org numpy ticket 1084 on 2009 04 14 by trac user ntv1534, assigned to unknown.
|
|
I recently upgraded my Python to version 2.6, and installed numpy via the sourceforge installer on my Windows Vista x64 laptop.
|
|
I got the following results SyntaxError invalid syntax >>> import numpy D \Python\lib\site packages\numpy\core\ init .py 5 Warning Windows 64 bits s upport is experimental, and only available for testing.
|
|
You are advised not to use it for production.
|
|
CRASHES ARE TO BE EXPECTED PLEASE REPORT THEM TO NUMPY DEVELOPERS import multiarray Traceback most recent call last) File <stdin> , line 1, in <module> File D \Python\lib\site packages\numpy\ init .py , line 130, in <module> import add newdocs File D \Python\lib\site packages\numpy\add newdocs.py , line 9, in <module> from lib import add newdoc File D \Python\lib\site packages\numpy\lib\ init .py , line 13, in <module> from polynomial import File D \Python\lib\site packages\numpy\lib\polynomial.py , line 18, in <modul e> from numpy.linalg import eigvals, lstsq File D \Python\lib\site packages\numpy\linalg\ init .py , line 47, in <modu le> from linalg import File D \Python\lib\site packages\numpy\linalg\linalg.py , line 22, in <module > from numpy.linalg import lapack lite ImportError DLL load failed The specified module could not be found.","bug"
|
|
"6013","1683","Surprising results from in place operations involving views Trac 1085).
|
|
Original ticket http projects.scipy.org numpy ticket 1085 on 2009 04 15 by trac user tillmann, assigned to charris.
|
|
>>> import numpy as np >>> x=np.array [[1,2],[3,4]]) >>> print x [[1 2] [3 4]] >>> x =x.T >>> print x [[2 5] [8 8]] when writing code on such a high level of abstraction this behavior is not expected and imho wrong.
|
|
At least a warning should be issued or better the in place operations should fall back to creating copies if the results are not correct otherwise.","bug"
|
|
"6013","1684","Cannot compile numpy > 1.1.1 using Visual Studio.net 2003 Trac 1086).
|
|
Original ticket http projects.scipy.org numpy ticket 1086 on 2009 04 15 by trac user salimfadhley, assigned to unknown.
|
|
When compiling numpy 1.3.0 the following compile error was observed using Microsoft Visual Studio.
|
|
Apparantly this is from the r5610.
|
|
It s a bug that shows up in all Microsoft compilers because they can t handle long strings.
|
|
As a quick workaround, we could break up the long strings in the mtrand.c file into string per line and use line continuations.
|
|
C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\link.exe DLL nologo INCREMENTAL NO LIBPATH D \Python24\libs LIBPATH D \Python24\PCBuild LIBPATH build\temp.win32 2.4 EXPORT initlapack lite buil d\temp.win32 2.4\Release\numpy\linalg\lapack litemodule.obj build\temp.win32 2.4\Release\numpy\linalg\pyt hon xerbla.obj build\temp.win32 2.4\Release\numpy\linalg\zlapack lite.obj build\temp.win32 2.4\Release\nu mpy\linalg\dlapack lite.obj build\temp.win32 2.4\Release\numpy\linalg\blas lite.obj build\temp.win32 2.4\ Release\numpy\linalg\dlamch.obj build\temp.win32 2.4\Release\numpy\linalg\f2c lite.obj OUT numpy\linalg\ lapack lite.pyd IMPLIB build\temp.win32 2.4\Release\numpy\linalg\lapack lite.lib building numpy.random.mtrand extension compiling C sources C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG In umpy\core\include Inumpy\core\include numpy Inumpy\core\src Inumpy\core\include ID \Python24\include ID \Python24\PC Tcnumpy\random\mtrand\mtrand.c Fobuild\temp.win32 2.4\Release\numpy\random\mtrand\mtra nd.obj mtrand.c numpy\random\mtrand\mtrand.c 2783) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 3136) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 3773) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 3975) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 4933) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5078) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5206) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5609) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5609) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5756) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5903) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 5903) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 6659) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 7291) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 7617) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 7965) error C2026 string too big, trailing characters truncated numpy\random\mtrand\mtrand.c 8158) error C2026 string too big, trailing characters truncated error Command C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG Inumpy\core\include Inumpy\core\include numpy Inumpy\core\src Inumpy\core\include ID \P ython24\include ID \Python24\PC Tcnumpy\random\mtrand\mtrand.c Fobuild\temp.win32 2.4\Release\numpy\ra ndom\mtrand\mtrand.obj failed with exit status 2","bug"
|
|
"6013","1685","Unsupported arch flag on Darwin Trac 1087).
|
|
Original ticket http projects.scipy.org numpy ticket 1087 on 2009 04 16 by trac user smaret, assigned to dmcooke.
|
|
Numpy 1.3.0 uses a arch flag for the gfortran compiler on Darwin.
|
|
This flag is only supported by universal capable compilers, which is not the case for the standard gfortran from gcc 4.x.
|
|
This breaks the compilation of modules that uses a fortran compiler, such as scipy.
|
|
There is a function can target) in numpy distutils fcompiler gnu.py that supposedly tests if the compiler support the arch flag, but apparently its is not working python c from numpy.f2py.diagnose import run; run ) | grep A 10 Gnu95FCompiler Gnu95FCompiler instance properties archiver = [ sw bin gfortran , cr ] compile switch = c compiler f77 = [ sw bin gfortran , Wall , ffixed form , fno second underscore , arch , i686 , fPIC , O3 , funroll loops ] compiler f90 = [ sw bin gfortran , Wall , fno second underscore , arch , i686 , fPIC , O3 , funroll loops ] compiler fix = [ sw bin gfortran , Wall , ffixed form , fno second underscore , Wall , fno second underscore , arch , i686 , fPIC , O3 , funroll loops ] I am using gfortran 4.3.3 and Python 2.5.4 installed with Fink on MacOS 10.5.6.","bug"
|
|
"6013","1686","Missing Cython Swig ... file in binary release delivery Numpy doc) Trac 1088).
|
|
Original ticket http projects.scipy.org numpy ticket 1088 on 2009 04 21 by trac user nroux, assigned to rgommers.
|
|
Hi, In the latest Numpy deliveries, the binary release does not have any more all what used to be in the doc folder.
|
|
The issue is that within this doc directory, there are important files required to use numpy with Cython, Pyrex, Swig... Could you please put back these files when delivering the binary releases ?
|
|
Thanks, Cheers, Nicolas.","bug"
|
|
"6013","1687","numpy.test does not find all tests Trac 1089).
|
|
Original ticket http projects.scipy.org numpy ticket 1089 on 2009 04 22 by cournape, assigned to unknown.
|
|
I noticed that nosetests runner runs a couple of more tests compared to numpy.test.
|
|
Incidentally, most of the undetected tests fail","bug"
|
|
"6013","1688","numscons build broken with new version.py scheme Trac 1090).
|
|
Original ticket http projects.scipy.org numpy ticket 1090 on 2009 04 22 by cournape, assigned to cournape.","bug"
|
|
"6013","1689","crackfortran patch needed Trac 1091).
|
|
Original ticket http projects.scipy.org numpy ticket 1091 on 2009 04 22 by trac user mesmith, assigned to unknown.
|
|
I want to re open ticket 1035.
|
|
Here is the information that I submitted.
|
|
I am re opening this issue because the crackfortran patch does not appear to be redundant to me.
|
|
The patch allows for setting the python callback argument as optional.
|
|
I use it by adding the follow f2py declarations to the FORTRAN file where the python callback is called.
|
|
[[BR]] c These lines are here to get the correct settings into flip.pyf cf2py intent callback, hide) py errmsg cf2py optional py errmsg cf2py external py errmsg cf2py use gp errmsg user routines [[BR]] where py errmsg is a python callback function.
|
|
Without the patch I get an error message from python when I attempt to execute the FORTRAN subroutine containing the python callback.
|
|
[[BR]] TypeError flip.gp setdef ) takes at least 1 argument 0 given) [[BR]] where gp setdef is the FORTRAN subroutine called from python and containing the cf2py lines above.
|
|
I have tried numpy version 1.3 and the patch is still required.
|
|
If there is another way to accomplish what I am trying to do without the patch and without requiring the python callback function as an argument I would like to know how to do it.
|
|
The user base for our software is growing and the necessity of applying the patch for each installation is becoming unmanagable.
|
|
I have updated the attached crackfortran.patch file for numpy 1.3.
|
|
Thanks","enhancement"
|
|
"6013","1690","Can t Import numpy in windows xp 64, python 2.6.2 64 bit Trac 1092).
|
|
Original ticket http projects.scipy.org numpy ticket 1092 on 2009 04 23 by trac user SevenThunders, assigned to rgommers.
|
|
The amd64 binary install of numpy fails using the binary install of the amd64 version of python 2.6.2.
|
|
Installation occurs without error but an attempt to import numpy crashes python in the cmd.exe window and returns with no effect in IDLE.
|
|
By no effect I mean numpy and it s methods are not visible.
|
|
I also compiled numpy using the latest 64 bit SDK from Microsoft v15), and this reproduced the same errors.
|
|
I don t know if this is the same problem in ticket 1084, but the behavior is a little different.","bug"
|
|
"6013","1691","Moore Penrose Inverse Error for NumPy Matrices Trac 1093).
|
|
Original ticket http projects.scipy.org numpy ticket 1093 on 2009 04 23 by trac user mcleane, assigned to unknown.
|
|
The inverse functionality in NumPy is useful, for instance A.I will properly calculate the Moore Penrose inverse in many cases of rectangular matrices.
|
|
However, this functionality is badly broken in at least one instance.
|
|
The singular matrix >>import numpy as np >>A = np.matrix [[2,1,1],[1,0, 1],[3,1,0]]) incorrectly reports the inverse to be >>print A.I [[ 4.50359963e 15 4.50359963e 15 4.50359963e 15] [ 1.35107989e 16 1.35107989e 16 1.35107989e 16] [ 4.50359963e 15 4.50359963e 15 4.50359963e 15]] whereas it should be [[0.0606060606061 0.151515151515 0.212121212121] [0.151515151515 0.121212121212 0.030303030303] [0.393939393939 0.515151515152 0.121212121212]] The correct value above can be calculated in NumPy by a sequence of commands, by first taking a full rank decomposition EG = A >>E = np.matrix [[2,1],[1,0],[3,1]]) >>G = np.matrix [[1,0, 1],[0,1,3]]) then the Moore Penrose inverse properly calculated by >>print G.I E.I I suspect the problem lies in using a technique to calculate inverses that fails badly when A is of neither full column rank nor full row rank, and this problem should be easy to rectify.","bug"
|
|
"6013","1692","masked array autotest fails with bus error Trac 1094).
|
|
Original ticket http projects.scipy.org numpy ticket 1094 on 2009 04 24 by trac user Gerard, assigned to charris.
|
|
Running the numpy version 1.3.0, but also occurs on 1.2.1) autotester fails on masked array test case with a bus error.
|
|
Used system is Solaris 5.8 uname a SunOS wsasd529 5.8 Generic 117350 18 sun4u sparc SUNW,Sun Blade 1500 output of numpy.test verbose=10) ... Tests how masks are initialized at the creation of Maskedarrays.
|
|
... ok Check the use of ndmin ... ok Make sure we re not losing the original mask w ndmin ... ok Tests creaating a masked array from alist of masked arrays.
|
|
... ok test deepcopy test core.TestMaskedArray) ... ok Test printing a masked array w fancy dtype.
|
|
... ok Test filled w flexible dtype ...
|
|
Bus error Bus error is caused by a misaligned memory write see output of purify below) MAW Misaligned memory write This is occurring while in DOUBLE fastputmask [arraytypes.inc.src 2282] s val = vals; for i = 0; i < ni; i ) { if mask[i]) { => in[i] = s val; } } } PyArray PutMask [multiarraymodule.c 4687] } } else { => func dest, mask >data, ni, values >data, nv); } } array putmask [multiarraymodule.c 4600] array, mask, values)) { return NULL; } => return PyArray PutMask PyArrayObject )array, values, mask); } NUMPY API PyCFunction Call [methodobject.c 77] PyEval EvalFrameEx [ceval.c 3564] PyEval EvalFrameEx [ceval.c 3650] PyEval EvalCodeEx [ceval.c 2831] PyEval EvalFrameEx [ceval.c 3659] ...
|
|
Commenting the execution of the fastputmask in multiarraymodule.c and thus running the slower version already implemented there) solves the problem multiarraymodule.c, line 4676 and further) func = self >descr >f >fastputmask; if func == NULL) { for i = 0; i < ni; i ) { tmp = Bool ) mask >data))[i]; if tmp) { src = values >data chunk i nv); memmove dest i chunk, src, chunk); } } } else { func dest, mask >data, ni, values >data, nv); } added here are the comment blocks such that func is not executed anymore.
|
|
Minimal piece of python code to reproduce from numpy.ma.core import f = array [ 1,1)], dtype=[ i ,int), f ,float)]) f.mask = [1,1] f.filled )","bug"
|
|
"6013","1693","polyint does not preserve object arrays Trac 1095).
|
|
Original ticket http projects.scipy.org numpy ticket 1095 on 2009 04 24 by pv, assigned to unknown.
|
|
`polyint` discards imaginary parts of complex variables >>> p = poly1d [1j,2j]) >>> p = poly1d [1j,2j,3j,4j]) >>> print p.coeffs [ 0.
|
|
1.j 0.
|
|
2.j 0.
|
|
3.j 0.
|
|
4.j] >>> print p.integ ).coeffs [ 0.]
|
|
It should retain the input dtype, and also work with object arrays for example, if one wants to use fractions.Fraction or mpmath.mpf objects to gain precision in the arithmetic).","bug"
|
|
"6013","1694","logaddexp inf, inf) should return inf, not nan Trac 1096).
|
|
Original ticket http projects.scipy.org numpy ticket 1096 on 2009 04 28 by trac user aswart, assigned to unknown.
|
|
logaddexp inf, inf) should return inf, not nan In [1] from numpy import logaddexp In [2] logaddexp inf, inf) Out[2] nan","bug"
|
|
"6013","1695","polydiv should be duck typing safe Trac 1097).
|
|
Original ticket http projects.scipy.org numpy ticket 1097 on 2009 04 28 by pv, assigned to unknown.
|
|
polydiv incorrectly casts objects to polynomials >>> import numpy as np >>> from sympy.mpmath import mpf >>> p = np.poly1d [mpf 1 )]) >>> p poly1d [1.0], dtype=object) >>> p mpf 2.0 ) Traceback most recent call last) File <stdin> , line 1, in <module> File ... numpy lib polynomial.py , line 1007, in div return polydiv self, other) File ... numpy lib polynomial.py , line 747, in polydiv while NX.allclose r[0], 0, rtol=1e 14) and r.shape[ 1] > 1) File ... numpy core numeric.py , line 1588, in allclose xinf = isinf x) TypeError function not supported for these types, and can t coerce safely to supported types I think it should first cast both its arguments first to ndarrays and treat 0 d arrays as scalars.
|
|
Similar things should be checked also in the other polynomial arithmetic operations.","bug"
|
|
"6013","1696","variance is inaccurate for arrays of identical, large values Trac 1098).
|
|
Original ticket http projects.scipy.org numpy ticket 1098 on 2009 04 29 by thouis, assigned to charris.
|
|
Variance calculation is inaccurate for arrays of large, identical values >>> from numpy import >>> ones 100000) 10.0 20).var ) 52756253943791624.0 There are more accurate algorithms for computing variance.
|
|
One example from Welford 1962) is in the attached file.","bug"
|
|
"6013","1697","vectorize incompatible with python 2.6 Trac 1099).
|
|
Original ticket http projects.scipy.org numpy ticket 1099 on 2009 04 30 by trac user edgarG, assigned to charris.
|
|
The message qualifying the TypeError exception raised when calling single argument python built in functions without any argument has changed in python 2.6 at least for some of them, e.g.
|
|
math.cos, math.radians...), from xxx takes exactly 1 argument 0 given) to xxx takes exactly one argument 0 given) .
|
|
This ruins vectorize s determination of the number of arguments required by its argument function, which is based on a regexp match of the previous TypeError message.
|
|
hence, in python 2.6 vcos = numpy.vectorize math.cos) logically raises ValueError, failed to determine the number of arguments for <built in function cos> ) Modifying the regexp to accomodate the one and convert it back to 1 is a quick and dirty workaround it will remain dangerous to base the feature on the interpretation of the error message text...
|
|
But I have no better idea to submit!","bug"
|
|
"6013","1698","svn version detection completely forgets about i18n Trac 1100).
|
|
Original ticket http projects.scipy.org numpy ticket 1100 on 2009 05 02 by trac user beb0s, assigned to dmcooke.
|
|
On my systen svn info outputs Revisione 6945 since subversion is correctly internationalized.
|
|
Therefore doing re.compile Revision [0 9] ) ) doesn t work and the install script keeps failing.
|
|
All this should be replaced by using some kind of python subversion interface.","bug"
|
|
"6013","1699","compilation with gfortran fails Trac 1101).
|
|
Original ticket http projects.scipy.org numpy ticket 1101 on 2009 05 04 by trac user panov, assigned to unknown.
|
|
Compilation of numpy 1.3.0 with gfortran gcc 4.3.3) fails, it does not add shared option on linking error Command usr bin gfortran Wall build temp.linux i686 2.6 numpy linalg lapack litemodule.o build temp.linux i686 2.6 numpy linalg python xerbla.o L usr local lib L usr lib Lbuild temp.linux i686 2.6 llapack lf77blas lcblas latlas lpython2.6 lgfortran o build lib.linux i686 2.6 numpy linalg lapack lite.so failed with exit status 1","bug"
|
|
"6013","1700","double definition of configuration macros in config.h numpyconfig.h Trac 1102).
|
|
Original ticket http projects.scipy.org numpy ticket 1102 on 2009 05 06 by cournape, assigned to unknown.
|
|
When running several times python setup.py build without removing the build directory, the configuration macros are appended to the existing config header, resulting in strange warnings or even compilation errors.","bug"
|
|
"6013","1701","f2py problem with .f90 files Trac 1103).
|
|
Original ticket http projects.scipy.org numpy ticket 1103 on 2009 05 08 by astrofrog, assigned to unknown.
|
|
I am having problems compiling even a basic f90 test provided in the default numpy 1.3.0 distribution If I go to numpy distutils tests f2py f90 ext in the numpy tar.gz file, and type python setup.py build what happens is that the .mod file produced is moved to a directory which is not part of the subsequent include path.
|
|
I found this problem when trying to write my own setup.py file, and found it also occurs with this standard example.
|
|
A log of the error is attached.","bug"
|
|
"6013","1702","Detect C API forward incompatibility Trac 1104).
|
|
Original ticket http projects.scipy.org numpy ticket 1104 on 2009 05 10 by cournape, assigned to unknown.
|
|
ATM, we only check for ABI compatibility when importing numpy.
|
|
But if an extension is built against version N, it will generally not run with numpy version M<N forward compatibility).
|
|
With normal shared libraries, this problem is easily detected because trying to run against an older library without the features of the new library will cause a link error.
|
|
With python extensions, no link error will occur.
|
|
We need to compare the C API version at runtime RUNTIME C API), compare it against the actual C API the extension was built against BUILTIN C API), and barfs if BUILTIN C API > RUNTIME C API, even if ABI is backward) compatible.
|
|
This will cause false alarms, but we can t detect them easily it would require scanning the extensions sources).","bug"
|
|
"6013","1703","[Ubuntu Jaunty]python numpy install SyntaxWarning Trac 1105).
|
|
Original ticket http projects.scipy.org numpy ticket 1105 on 2009 05 12 by trac user jhaitas, assigned to unknown.
|
|
This is reported on Launchpad at [https bugs.launchpad.net ubuntu source python numpy bug 335646 |https bugs.launchpad.net ubuntu source python numpy bug 335646] Setting up python numpy 1 1.2.1 1ubuntu1) ... usr lib python2.6 dist packages numpy lib tests test io.py 68 SyntaxWarning assertion is always true, perhaps remove parentheses?
|
|
assert c.readlines ), Install seems to be OK.","bug"
|
|
"6013","1704","Wrong exception type raised when getting a non existent dtype field Trac 1106).
|
|
Original ticket http projects.scipy.org numpy ticket 1106 on 2009 05 13 by jpaalasm, assigned to unknown.
|
|
A !ValueError is raised when getting a non existent dtype field.
|
|
It should be an !IndexError, of course.
|
|
In [14] dt = numpy.dtype [ f1 , numpy.uint)]) In [15] dt[ f1 ] Out[15] dtype uint32 ) In [16] dt[ f2 ] ValueError Traceback most recent call last) home jpaalasm <ipython console> in <module> ) ValueError only integers, strings or unicode values allowed for getting fields.","bug"
|
|
"6013","1705","loadtxt support for blocks of data Trac 1107).
|
|
Original ticket http projects.scipy.org numpy ticket 1107 on 2009 05 13 by trac user ctenney, assigned to unknown.
|
|
Problem I wanted to be able to easily read multiple blocks of array data into Numpy from a single text file e.g.
|
|
xyz positions of a collection of atoms at different times).
|
|
I was unable to find an existing tool to accomplish this.
|
|
Solution I modified the loadtxt function in numpy lib io.py diff pasted below).
|
|
Summary I added two new parameter options numcols and datablocks .
|
|
The defaults were chosen to reproduce previous behavior.
|
|
numcols is an optional int that defaults to None.
|
|
If set, numcols specifies the number of columns in a valid data line.
|
|
This is useful for filtering out non comment header info particularly from files with multiple blocks of data).
|
|
datablocks is a boolean that defaults to False.
|
|
If True, data blocks within the file are assumed to be separated by two or more non data lines e.g.
|
|
blank, comment, or invalid numcols ).
|
|
An additional leading dimension is added to each returned array if two or more data blocks are present.
|
|
e.g.
|
|
If X = np.loadtxt ..., datablocks=True), then X[0] is the first block, X[1] the second, et cetera; and X.ravel returns the same result as np.loadtxt ..., datablocks=False).)
|
|
The result is unchanged for single block data sets.
|
|
Notes I successfully ran the changes against numpy lib tests test io.py.
|
|
When faced with multi block datasets, use of either or both of these options makes plotting in matplotlib almost as easy as plotting in gnuplot, which is particularly handy for data exploration via ipython pylab.
|
|
diff for numpy lib io.py for v1.3.0 332,333c332 < skiprows=0, usecols=None, unpack=False, < numcols=None, datablocks=False) > skiprows=0, usecols=None, unpack=False) 368,379d366 < numcols int < If set, numcols specifies the number of columns in a valid data line.
|
|
< This is useful for filtering out non comment header info particularly < from files with multiple blocks of data).
|
|
< datablocks bool < If True, data blocks within the file are assumed to be separated by two < or more non data lines e.g.
|
|
blank, comment, or invalid numcols ).
|
|
An < additional leading dimension is added to each returned array if two or < more data blocks are present.
|
|
e.g.
|
|
If X = np.loadtxt ..., < datablocks=True), then X[0] is the first block, X[1] the second, et < cetera; and X.ravel returns the same result as np.loadtxt ..., < datablocks=False).)
|
|
The result is unchanged for single block data sets.
|
|
471,472d457 < if numcols and len first vals) != numcols < first vals = None 495,497d479 < X.append []) < block = 0 < blanks = 0 501d482 < blanks = 1 503,509d483 < elif numcols and len vals) != numcols < blanks = 1 < continue < elif datablocks and blanks >= 2 < block = 1 < X.append []) < blanks = 0 515,516c489 < X[block].append tuple [conv val) for conv, val) in zip converters, vals)])) < > X.append tuple [conv val) for conv, val) in zip converters, vals)])) 536c509 < return np.rollaxis X, 1, 0) > return X.T","enhancement"
|
|
"6013","1706","FAIL Test bug in reduceat with structured arrays copied for speed.
|
|
Trac 1108).
|
|
Original ticket http projects.scipy.org numpy ticket 1108 on 2009 05 13 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL Test bug in reduceat with structured arrays copied for speed.
|
|
Traceback most recent call last) File usr stsci pyssgdev Python 2.5.1 lib python2.5 site packages nose case.py , line 182, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.1 numpy core tests test umath.py , line 664, in test reduceat assert np.all h1 == h2) AssertionError Affects rev 6976 through 6983 latest I tried), on 64 bit Red Hat Enterprise Linux 4 and 32 bit PowerPC Macintosh.
|
|
Does not affect 32 bit Solaris, 32 bit x86 Mac, 32 bit RHEL 3 or 4.","bug"
|
|
"6013","1707","MultiArrayError does not inherit from Exception Trac 1109).
|
|
Original ticket http projects.scipy.org numpy ticket 1109 on 2009 05 14 by trac user jguyer, assigned to unknown.
|
|
Calling `argmin )` on an empty array throws an error that doesn t inherit from `Exception`, such that it s not caught in a generic `try` block.
|
|
[http docs.python.org library exceptions.html exceptions.Exception Python Library Reference] says All user dexefxacx81ned exceptions should also be derived from [`Exception`].
|
|
!python >>> import numpy as np >>> try ... np.array []).argmin ) ... except Exception, e ... print caught , e ... Traceback most recent call last) File <stdin> , line 2, in <module> multiarray.error attempt to get argmax argmin of an empty sequence","bug"
|
|
"6013","1708","Error packing Numpy scalars with struct.pack Trac 1110).
|
|
Original ticket http projects.scipy.org numpy ticket 1110 on 2009 05 19 by trac user bryancatanzaro, assigned to unknown.
|
|
Hello I m using NumPy with PyCUDA.
|
|
PyCUDA needs to convert NumPy scalars into C structs, and does so using the standard struct.pack.
|
|
This seemed to work with earlier versions of NumPy Python, but with NumPy version 1.3.0 on Python 2.6.2, it causes an error when packing a NumPy scalar into an unsigned struct element.
|
|
Here is a code snippet import numpy import struct struct.pack I , numpy.uint32 0)) And here is the error sys 1 DeprecationWarning struct integer overflow masking is deprecated Traceback most recent call last) File <stdin> , line 1, in <module> SystemError Objects longobject.c 336 bad argument to internal function The pack proceeds correctly if I don t pack to an unsigned struct element, e.g.
|
|
struct.pack i , numpy.uint32 0)) Thanks, bryan","bug"
|
|
"6013","1709","Fix endianness detection on ppc64 builds Trac 1111).
|
|
Original ticket http projects.scipy.org numpy ticket 1111 on 2009 05 20 by dwf, assigned to unknown.
|
|
Trying to build either a fat64 or a 4 way binary of NumPy on Mac OS X is failing in svn r7002 an endianness detection error.
|
|
This came up on the mailing list [http mail.scipy.org pipermail scipy user 2009 May 021066.html thread]) and I encountered it myself.
|
|
Some warnings show PPC64 builds are the problem gcc arch ppc64 arch x86 64 DNDEBUG g fwrapv O3 Wall Wstrict prototypes O3 Wall DNDEBUG g fwrapv Wstrict prototypes arch x86 64 arch ppc64 Inumpy core include Ibuild src.macosx 10.5 fat64 2.6 numpy core include numpy Inumpy core src Inumpy core src multiarray Inumpy core src umath Inumpy core include I Library Frameworks Python64.framework Versions 2.6 include python2.6 Ibuild src.macosx 10.5 fat64 2.6 numpy core src multiarray Ibuild src.macosx 10.5 fat64 2.6 numpy core src umath c build src.macosx 10.5 fat64 2.6 numpy core src sortmodule.c In file included from numpy core include numpy npy endian.h 22, from numpy core include numpy ndarrayobject.h 33, from numpy core include numpy noprefix.h 7, from numpy core src sortmodule.c.src 29 numpy core include numpy npy cpu.h 39 6 warning warning PPC64!
|
|
Hooray!
|
|
In file included from numpy core include numpy ndarrayobject.h 33, from numpy core include numpy noprefix.h 7, from numpy core src sortmodule.c.src 29 numpy core include numpy npy endian.h 33 10 error error Unknown CPU can not set endianness In file included from numpy core include numpy npy endian.h 22, from numpy core include numpy ndarrayobject.h 33, from numpy core include numpy noprefix.h 7, from numpy core src sortmodule.c.src 29 numpy core include numpy npy cpu.h 30 6 warning warning AMD64!
|
|
Hooray!
|
|
lipo can t figure out the architecture type of var folders jz jzMQLnV GOWCP55tPgPBl Tc Tmp ccjBE91A.out I ve attached a tiny patch that fixes this, adds PPC64 as one of the architectures detected as big endian.","bug"
|
|
"6013","1710","More crashes on Windows 64 Trac 1112).
|
|
Original ticket http projects.scipy.org numpy ticket 1112 on 2009 05 20 by trac user langstephen, assigned to unknown.
|
|
Similar but different from ticket 1682.
|
|
1.3.0 compiled on Xeon using Visual Studio 2005 compiler version 14.00.50727.762).
|
|
Can import numpy but get a hard crash if we do this from numpy import a=empty shape=4) print a this works print a[0] crashes Or this import numpy as np y=np.longdouble 4) print y also crashes on any operation on y, not just y 2 as in ticket 1068 One note if it helps Numpy 1.3.0 was compiled with one modification to msvccompiler.py in <PYTHON DIR> Lib distutils.
|
|
The change made was to embed manifest inside of a DLL.
|
|
If the manifest is NOT embedded inside of the generated DLL then doing an xe2x80x9cimport numpyxe2x80x9d from Python results in error stating that MSVCR80.dll is not found.
|
|
msvccompiler line 297) self.compile options = [ nologo , Ox , MD , W3 , GS was modified to include MT flag to embed manifest self.compile options = [ nologo , Ox , MT , W3 , GS","bug"
|
|
"6013","1711","ppc 32 and 64 bits) test failure test umath.TestComplexFunctions.test loss of precision <type numpy.complex64 >,) Trac 1113).
|
|
Original ticket http projects.scipy.org numpy ticket 1113 on 2009 05 20 by dwf, assigned to unknown.
|
|
I m seeing this test fail with a PPC64 build on with a 4 way universal binary Python 2.6.2, r7007.
|
|
Looks similar to 1636, but it s a complex64 rather than a complex256.
|
|
FAIL test umath.TestComplexFunctions.test loss of precision <type numpy.complex64 >,) Traceback most recent call last) File Library Frameworks Python64.framework Versions 2.6 lib python2.6 site packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File Library Frameworks Python64.framework Versions 2.6 lib python2.6 site packages numpy core tests test umath.py , line 524, in check loss of precision assert np.all d < 1e 15) AssertionError","bug"
|
|
"6013","1712","win amd64 lapack lite import fails Trac 1114).
|
|
Original ticket http projects.scipy.org numpy ticket 1114 on 2009 05 21 by trac user fuzion, assigned to rgommers.
|
|
The lapack lite.pyd is there, but it looks about 1MB smaller than the one included in numpy 1.3.0 sse3.
|
|
I m not sure what might be missing since I haven t tried to make a dist yet.
|
|
Here s the error from numpy import array, float32 File C \Python26x64\lib\site packages\numpy\ init .py , line 130, in <modul e> import add newdocs File C \Python26x64\lib\site packages\numpy\add newdocs.py , line 9, in <modu le> from lib import add newdoc File C \Python26x64\lib\site packages\numpy\lib\ init .py , line 13, in <mo dule> from polynomial import File C \Python26x64\lib\site packages\numpy\lib\polynomial.py , line 18, in < module> from numpy.linalg import eigvals, lstsq File C \Python26x64\lib\site packages\numpy\linalg\ init .py , line 47, in <module> from linalg import File C \Python26x64\lib\site packages\numpy\linalg\linalg.py , line 22, in <m odule> from numpy.linalg import lapack lite ImportError DLL load failed The specified module could not be found.
|
|
Also, forgive me for the terrible bug report...","bug"
|
|
"6013","1713","Crash in fromfile reading empty file Trac 1115).
|
|
Original ticket http projects.scipy.org numpy ticket 1115 on 2009 05 21 by pv, assigned to unknown.
|
|
Consider touch foo python c import numpy as np; np.fromfile foo , sep= ) glibc detected python double free or corruption top) 0x098870b0 I think `array fromfile` probably does a call like `realloc ptr, 0)` which in C89 is equivalent to `{free ptr); return NULL;}`.
|
|
Or some similar mess up.","bug"
|
|
"6013","1714","Fromstring returns wrong result for empty text strings.
|
|
Trac 1116).
|
|
Original ticket http projects.scipy.org numpy ticket 1116 on 2009 05 23 by charris, assigned to unknown.
|
|
The fromstring ctor operates incorrectly on empty text strings.
|
|
In [1] fromstring , sep= ) Out[1] array [ 0.])
|
|
In [2] fromstring ) Out[2] array [], dtype=float64)","bug"
|
|
"6013","1715","correlate not order dependent Trac 1117).
|
|
Original ticket http projects.scipy.org numpy ticket 1117 on 2009 05 25 by trac user robochat, assigned to unknown.
|
|
Currently, if v and a have different lengths, the current correlation function will internally swap the inputs so that the output always has the same orientation.
|
|
This is a bug, correlations should be order dependent eg.
|
|
correlate a,v) != correlate v,a) in general whereas convolutions are symmetric).
|
|
So >>> import numpy as N >>> x = N.array [1,0,0]) >>> y = N.array [0,0,1]) >>> N.correlate x,y, full ) array [1, 0, 0, 0, 0]) >>> N.correlate y,x, full ) array [0, 0, 0, 0, 1]) This works fine.
|
|
However, if the arrays have different lengths, we get a problem.
|
|
>>> y2=N.array [0,0,0,1]) >>> N.correlate x,y2, full ) array [0, 0, 0, 0, 0, 1]) >>> N.correlate y2,x, full ) array [0, 0, 0, 0, 0, 1]) scipy.correlate also shows the same problem but ndimage.filters.correlate does not.","bug"
|
|
"6013","1716","fromfile crashes on empty file Trac 1118).
|
|
Original ticket http projects.scipy.org numpy ticket 1118 on 2009 05 28 by cournape, assigned to unknown.
|
|
!
|
|
python import numpy as np f = open yo , w ) f.write ) f.close ) a = np.fromfile open yo ), sep= )","bug"
|
|
"6013","1717","NPZ files are not compressed Trac 1119).
|
|
Original ticket http projects.scipy.org numpy ticket 1119 on 2009 05 31 by trac user luis ..., assigned to unknown.
|
|
NPZ files are not compressed.
|
|
You need to pass compression=ZIP DEFLATED to the zipfile constructor to get compression.
|
|
By default, you get compression=ZIP STORED, i.e., no compression.
|
|
If this is the intended behaviour, then the documentation should be fixed.","bug"
|
|
"6013","1718","correlate should conjugate its second argument before multiplying its input Trac 1120).
|
|
Original ticket http projects.scipy.org numpy ticket 1120 on 2009 06 01 by cournape, assigned to unknown.
|
|
correlate x, y) computes sum x y[k ]) for every k, instead of the usual definition of sum x conj y[ k])).","bug"
|
|
"6013","1719","PyInt Check np.int32 1)) is 0 on windows python 2.6) Trac 1121).
|
|
Original ticket http projects.scipy.org numpy ticket 1121 on 2009 06 01 by cournape, assigned to unknown.
|
|
Find out why...","bug"
|
|
"6013","1720","Syntax Error numpy core src multiarray descriptor.c Trac 1122).
|
|
Original ticket http projects.scipy.org numpy ticket 1122 on 2009 06 01 by stsci sienkiew, assigned to unknown.
|
|
numpy core src multiarray descriptor.c contains a syntax error.
|
|
Line 2107 contains a variable declaration that is not at the beginning of a block.
|
|
if PyString Check op) || PyUnicode Check op)) { PyObject obj = PyDict GetItem self >fields, op); if obj == NULL) { PyErr Format PyExc KeyError, Field named \ s\ not found.
|
|
, PyString AsString op)); return NULL; } PyObject descr = PyTuple GET ITEM obj, 0); Py INCREF descr); retval = descr; } suggest if PyString Check op) || PyUnicode Check op)) { PyObject obj = PyDict GetItem self >fields, op); PyObject descr; if obj == NULL) { PyErr Format PyExc KeyError, Field named \ s\ not found.
|
|
, PyString AsString op)); return NULL; } descr = PyTuple GET ITEM obj, 0); Py INCREF descr); retval = descr; } observed on Solaris 8 and 10 with SUNWspro 6 compiler.","bug"
|
|
"6013","1721","nansum for empty or all nan arrays is not zero Trac 1123).
|
|
Original ticket http projects.scipy.org numpy ticket 1123 on 2009 06 02 by josef pkt, assigned to charris.
|
|
Michale Hearne numpy.nansum []) In my copy of numpy 1.1.1, I get 0.0.
|
|
This is what I would expect to see.
|
|
However, this behavior seems to have changed in 1.3.0, in which I get nan.
|
|
Keith Goodman >> np.nansum 9.0) IndexError 0 d arrays can t be indexed.
|
|
Josef >>> np.nansum [np.nan, np.nan]) 1.
|
|
QNAN shouldn t this be zero numpy\lib\function base.py this should fix it, def nanop op, fill, a, axis=None) y = array a,subok=True) mask = isnan a) mask = asarray isnan a)) if mask.all ) return np.nan if not issubclass y.dtype.type, np.integer) y[mask] = fill return op y, axis=axis)","bug"
|
|
"6013","1722","Remove string exceptions from c code.
|
|
Trac 1124).
|
|
Original ticket http projects.scipy.org numpy ticket 1124 on 2009 06 02 by charris, assigned to charris.
|
|
The string exceptions are gone from the core src c files but some remain in lib src compiled base.c numarray capi.c fft fftpack litemodule.c f2py tons of them) They should be removed.","bug"
|
|
"6013","1723","implement the rest of the dictionary interface for NpzFile Trac 1125).
|
|
Original ticket http projects.scipy.org numpy ticket 1125 on 2009 06 03 by dwf, assigned to unknown.
|
|
The patch attached implements a more complete dictionary interface for NpzFile.
|
|
That is to say, it provides in addition to the already present getitem )) contains tests for containment of a key in self.files iter returns an iterator over self.files keys ) returns a copy of self.files values ) returns a list of all arrays in the NpzFile iterkeys ) returns an iterator [really a generator] over all items in self.files itervalues ) returns an iterator over all arrays in the NpzFile items ) returns a list of key, value) pair tuples iteritems ) returns an iterator over key, value) pair tuples This allows the convenient for fname in mynpzfile syntax, as well as letting you write code that doesn t care whether its input is an in memory dictionary or a lazily loaded NpzFile.","enhancement"
|
|
"6013","1724","numpy.delete does not support boolean index arrays Trac 1126).
|
|
Original ticket http projects.scipy.org numpy ticket 1126 on 2009 06 04 by jpaalasm, assigned to unknown.
|
|
To be consistent with the behaviour of array indexing, numpy.delete should support boolean array indexing.
|
|
In [27] a = numpy.arange 10) In [28] less than 5 = a < 5 In [29] a, less than 5 Out[29] array [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]), array [ True, True, True, True, True, False, False, False, False, False], dtype=bool)) In [30] a[less than 5], a[numpy.where less than 5)[0]] Out[30] array [0, 1, 2, 3, 4]), array [0, 1, 2, 3, 4])) In [31] numpy.delete a, less than 5), numpy.delete a, numpy.where less than 5)[0]) Out[31] array [2, 3, 4, 5, 6, 7, 8, 9]), array [5, 6, 7, 8, 9])) The two arrays in Out[31] should be identical as they are also in Out[30].","bug"
|
|
"6013","1725","Make the dtype object immutable and not coerce other types when compared hashable requirements) Trac 1127).
|
|
Original ticket http projects.scipy.org numpy ticket 1127 on 2009 06 05 by FrancescAlted, assigned to unknown.
|
|
The dtype object seems to be shared after a copy of a structured array is done.
|
|
This is inconvenient when you change properties in one of arrays and these are propagated to the other.
|
|
The next snippet exposes the problem In [21] dt = np.dtype f4,f8 ) In [22] ra = np.empty 1, dt) In [23] ra.dtype Out[23] dtype [ f0 , <f4 ), f1 , <f8 )]) In [24] rb = ra[ ] In [25] ra.dtype.names = float , double ) In [26] ra.dtype Out[26] dtype [ float , <f4 ), double , <f8 )]) In [27] rb.dtype Out[27] dtype [ float , <f4 ), double , <f8 )]) not the original!","bug"
|
|
"6013","1726","str function on string improperly truncates trailing zeroes Trac 1128).
|
|
Original ticket http projects.scipy.org numpy ticket 1128 on 2009 06 05 by trac user nrook, assigned to unknown.
|
|
The ` str ` method of a `string ` instance will truncate trailing zeroes the ` \x00 ` character) in the conversion.
|
|
For instance, `str string \x00 )) == ` evaluates to true, while `str string \x00 )) == \x00 ` evalutes to false.
|
|
Here is a longer example in the interpreter http pastebin.com f70292a3c Apologies for the external link, but I could not get Trac to easily write out the whole example without modifying it.)
|
|
A similar effect seems to occur with arrays this expression, `numpy.array [ ab\x00\x00 ])[0] == ab ` evaluates to true.","bug"
|
|
"6013","1727","F CONTIGUOUS masked arrays don t pickle correctly Trac 1129).
|
|
Original ticket http projects.scipy.org numpy ticket 1129 on 2009 06 06 by trac user reggie, assigned to pierregm.
|
|
If a multidimensional masked array which isn t C CONTIGUOUS is pickled and then unpickled, the order of the data in the unpickled array is different from that in the original array.
|
|
For example Python 2.5.2 r252 60911, Jul 31 2008, 17 28 52) [GCC 4.2.3 Ubuntu 4.2.3 2ubuntu7)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import cPickle, numpy >>> numpy.
|
|
version 1.3.0 >>> a = numpy.ma.arange 10) >>> a.shape = 1, 2) >>> b = a.T >>> b.flags C CONTIGUOUS False F CONTIGUOUS True OWNDATA False WRITEABLE True ALIGNED True UPDATEIFCOPY False >>> b masked array data = [[0 2 4 6 8] [1 3 5 7 9]], mask = False, fill value = 999999) >>> cPickle.loads cPickle.dumps b)) masked array data = [[0 4 8 3 7] [2 6 1 5 9]], mask = [[False False False False False] [False False False False False]], fill value = 999999) >>> I ve attached a patch that seems to fix this problem.","bug"
|
|
"6013","1728","set string function reset incomplete Trac 1130).
|
|
Original ticket http projects.scipy.org numpy ticket 1130 on 2009 06 06 by rgommers, assigned to unknown.
|
|
!rst There seems to be a bug in set string function when resetting the formatting function to the default.
|
|
After doing that the dtype of the array that is printed is the character string, not the numpy type.
|
|
Example .. code block python In [1] a=arange 10, dtype=uint16) In [2] a Out[2] array [0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=uint16) In [3] set string function lambda x str x 2)) In [4] a Out[4] [ 0 2 4 6 8 10 12 14 16 18] In [5] set string function None) reset to default In [6] a Out[6] array [0, 1, 2, 3, 4, 5, 6, 7, 8, 9], H ) The ability to reset the function by using None as argument was introduced in 949.","bug"
|
|
"6013","1729","build clib command does not alway call build src when needed Trac 1131).
|
|
Original ticket http projects.scipy.org numpy ticket 1131 on 2009 06 07 by cournape, assigned to dmcooke.
|
|
For example python setup.py build clib will fail because the sources for core math lib are not generated.","bug"
|
|
"6013","1730","Add a set of indexing functions for diagonals ndim), upper and lower triangular 2d) Trac 1132).
|
|
Original ticket http projects.scipy.org numpy ticket 1132 on 2009 06 08 by fperez, assigned to unknown.
|
|
This patch, as per the discussion on list, adds a set of utility functions to generate indices for manipulating diagonals in n d) and upper lower triangular matrices.
|
|
On my system, the test count goes up by 12 using test generators), though the tests could be still more thorough.
|
|
See full thread here http mail.scipy.org pipermail numpy discussion 2009 June 043089.html","enhancement"
|
|
"6013","1731","Clean up of arraysetops Trac 1133).
|
|
Original ticket http projects.scipy.org numpy ticket 1133 on 2009 06 08 by rc, assigned to unknown.
|
|
This ticket sumarizes the discussion in the numpy mailing list [1], concerning allowing non unique arrays as function arguments, better naming conventions and documentation.
|
|
1. the functions with nu suffix should be merged with the versions expecting unique arrays as input, example def intersect1d ar1, ar2, assume unique=False) if not assume unique return intersect1d nu ar1, ar2) else ... the current code The name of the keyword argument is yet to be resolved, other proposition so far were isunique=False or just unique=False Josef) ar1 unique=False, ar2 unique=False rc) 2. deprecate the nu functions 3. alias setmember1d nu as `in1d` or `isin1d`, because the function is a in and not a set operation 4. guarantee that setdiff1d works for non unique arrays even when implementation changes), and change documentation 5. union1d, ediff1d and unique1d are defined for non unique arrays document it properly 6. define np.setxor1d np.unique a), np.unique b)) to become np.setxor1d a, b, assume unique=False) 7. module name > > rename arraysetops to something easier to read like setfun.
|
|
I think it > > would only affect internal changes since all functions are exported to > > the main numpy name space > > 1e 4 I got used to arrayse tops) 0 internal change only).
|
|
Other numpy scipy submodules containing a bunch of functions are called pack fftpack, arpack, lapack), alg linalg), utils.
|
|
fun is used comonly in the matlab world.
|
|
8. keep docs in sync with correct usage [1] numpy discussion scipy.org, thread extract elements of an array that are contained in another array?","enhancement"
|
|
"6013","1732","distutils barfs when building a library from a .c.src file if add library does not have depends argument Trac 1134).
|
|
Original ticket http projects.scipy.org numpy ticket 1134 on 2009 06 09 by cournape, assigned to dmcooke.
|
|
Example !python ... config.add library yo , sources=[ yo.c.src ]) ... Will return Traceback most recent call last) File setup.py , line 185, in <module> setup package ) File setup.py , line 178, in setup package configuration=configuration ) File usr media src dsp numpy numpy git numpy distutils core.py , line 184, in setup return old setup new attr) File usr lib python2.5 distutils core.py , line 151, in setup dist.run commands ) File usr lib python2.5 distutils dist.py , line 974, in run commands self.run command cmd) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr media src dsp numpy numpy git numpy distutils command install.py , line 49, in run r = old install.run self) File usr lib python2.5 distutils command install.py , line 506, in run self.run command build ) File usr lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr media src dsp numpy numpy git numpy distutils command build.py , line 37, in run old build.run self) File usr lib python2.5 distutils command build.py , line 113, in run self.run command cmd name) File usr lib python2.5 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.5 distutils dist.py , line 994, in run command cmd obj.run ) File usr media src dsp numpy numpy git numpy distutils command build src.py , line 130, in run self.build sources ) File usr media src dsp numpy numpy git numpy distutils command build src.py , line 141, in build sources self.build library sources libname info) File usr media src dsp numpy numpy git numpy distutils command build src.py , line 222, in build library sources sources = self.template sources sources, lib name, build info)) File usr media src dsp numpy numpy git numpy distutils command build src.py , line 353, in template sources if self.force or newer group [source] depends, target file)) TypeError can only concatenate list not NoneType ) to list","bug"
|
|
"6013","1733","Remove setup.py files from installed files Trac 1135).
|
|
Original ticket http projects.scipy.org numpy ticket 1135 on 2009 06 09 by cournape, assigned to dmcooke.
|
|
It does not make much sense to install them, and it confuses some people, especially beginners in numpy, who run setup.py inside the installed version.","enhancement"
|
|
"6013","1734","Multiplying Python float to numpy.array of objects works but fails with a numpy.float64 Trac 1136).
|
|
Original ticket http projects.scipy.org numpy ticket 1136 on 2009 06 10 by b45ch1, assigned to unknown.
|
|
This is best explained by the following example class adouble def init self,x) self.x = x def mul self,rhs) if isinstance rhs,adouble) return adouble self.x rhs.x) else return adouble self.x rhs) def str self) return str self.x) x = adouble 3.)
|
|
y = adouble 2.)
|
|
u = array [adouble 3.
|
|
), adouble 5.)])
|
|
v = array [adouble 2.
|
|
), adouble 7.)])
|
|
z = array [2.,3.])
|
|
print numpy.
|
|
version = ,numpy.
|
|
version print x y ok print u v ok print u z ok print u 3. ok print u z[0] NOT OK!
|
|
print u float64 3.)
|
|
NOT OK!
|
|
Output basti shlp tmp python numpy float64 issue.py numpy.
|
|
version = 1.4.0.dev7039 6.0 [6.0 35.0] [6.0 15.0] [9.0 15.0] Traceback most recent call last) File numpy float64 issue.py , line 26, in <module> print u z[0] NOT OK!
|
|
TypeError unsupported operand type s) for numpy.ndarray and numpy.float64","bug"
|
|
"6013","1735","repr ) of numpy floats is locale sensitive Trac 1137).
|
|
Original ticket http projects.scipy.org numpy ticket 1137 on 2009 06 10 by inducer, assigned to unknown.
|
|
The attached script gives this result for me 3,2999999999999998 Note that there s a comma in there, which probably shouldn t happen.","bug"
|
|
"6013","1736","mirr does not function correctly Trac 1138).
|
|
Original ticket http projects.scipy.org numpy ticket 1138 on 2009 06 11 by jseabold, assigned to unknown.
|
|
The tests happen to be right for the first value given in the tests.
|
|
However, in the second test the compare to value is wrong.
|
|
Try the second function in either Excel or OpenOffice Calc.
|
|
Also compare np.mirr 5000,1000,2000,3000), .05, .08) to OO calc answer or Excel.
|
|
There are two reasons for the discrepancy.
|
|
The first value in a MIRR problem is to be considered a sunk cost or present value at time zero and is not to be discounted or compounded hence the NET in net present value).
|
|
The second is that the current implementation drops the periods that are of the opposite sign in computing the NPV for the numerator and denominator.
|
|
However, the values should still be discounted over these periods.
|
|
Attached is a patch which works as expected along with correct tests.
|
|
The first test is from the OpenFormula document <http www.oasis open.org committees documents.php?wg abbrev=office formula>.
|
|
The second test is the one from the existing code.
|
|
This is my first patch, so please let me know if this isn t how I should go about doing this.","bug"
|
|
"6013","1737","record array passing formats as tuple raises error Trac 1139).
|
|
Original ticket http projects.scipy.org numpy ticket 1139 on 2009 06 17 by huard, assigned to unknown.
|
|
If a tuple is passed to the formats argument to rec.fromarrays for example, the following error is raised usr local lib python2.5 site packages numpy 1.4.0.dev7045 py2.5 linux x86 64.egg numpy core records.pyc in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 506 names = descr.names 507 else > 508 parsed = format parser formats, names, titles, aligned, byteorder) 509 names = parsed.
|
|
names 510 descr = parsed.
|
|
descr usr local lib python2.5 site packages numpy 1.4.0.dev7045 py2.5 linux x86 64.egg numpy core records.pyc in init self, formats, names, titles, aligned, byteorder) 110 111 def init self, formats, names, titles, aligned=False, byteorder=None) > 112 self.
|
|
parseFormats formats, aligned) 113 self.
|
|
setfieldnames names, titles) 114 self.
|
|
createdescr byteorder) usr local lib python2.5 site packages numpy 1.4.0.dev7045 py2.5 linux x86 64.egg numpy core records.pyc in parseFormats self, formats, aligned) 124 formats.append ) 125 formats = , .join formats) > 126 dtype = sb.dtype formats, aligned) 127 fields = dtype.fields 128 if fields is None TypeError data type not understood WARNING Failure executing file <load.py> If the tuple is converted into a list, then everything works fine.","bug"
|
|
"6013","1738","ndenumerate crash under win32 with python 2.5 Trac 1140).
|
|
Original ticket http projects.scipy.org numpy ticket 1140 on 2009 06 19 by trac user senyai, assigned to unknown.
|
|
Hi.
|
|
I ve got a bug.
|
|
Dunno what to write here ) I ve tested it at two PCs.
|
|
import numpy as np list np.ndenumerate np.array [[]])))","bug"
|
|
"6013","1739","lookfor is broken on Python2.6 Trac 1141).
|
|
Original ticket http projects.scipy.org numpy ticket 1141 on 2009 06 19 by pv, assigned to unknown.
|
|
Lookfor on Python2.6 is broken >>> import numpy as np >>> np.lookfor binary representation ) Traceback most recent call last) File <stdin> , line 1, in <module> File home pauli koodi proj scipy numpy dist linux lib python2.6 site packages numpy 1.4.0.dev7066.420e060cb4a1975bf5ccb6e0b56bc3750159c4a6 py2.6 linux i686.egg numpy lib utils.py , line 574, in lookfor cache = lookfor generate cache module, import modules, regenerate) File home pauli koodi proj scipy numpy dist linux lib python2.6 site packages numpy 1.4.0.dev7066.420e060cb4a1975bf5ccb6e0b56bc3750159c4a6 py2.6 linux i686.egg numpy lib utils.py , line 724, in lookfor generate cache for n, v in inspect.getmembers item) File usr lib python2.6 inspect.py , line 256, in getmembers value = getattr object, key) AttributeError abstractmethods","bug"
|
|
"6013","1740",".real doesn t work on object arrays Trac 1142).
|
|
Original ticket http projects.scipy.org numpy ticket 1142 on 2009 06 19 by inducer, assigned to unknown.
|
|
numpy could really do a better job here by propagating the .real to the elements of the object array Python 2.5.4 r254 67916, Feb 18 2009, 03 00 47) [GCC 4.3.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as n >>> n. version 1.3.0 >>> a = n.array 5 7j) >>> b = n.array [a], dtype=object) >>> b.real array [ 5 7j)], dtype=object) >>> n.real b) array [ 5 7j)], dtype=object)","bug"
|
|
"6013","1741","Improve performance of PyUFunc Reduce Trac 1143).
|
|
Original ticket http projects.scipy.org numpy ticket 1143 on 2009 06 20 by trac user brandyn, assigned to unknown.
|
|
I have a very large 2d array, m. I take a very small but still wide) slice, like s = m[i i 2] Now x = sum s[i] for i in xrange 2)) Is at least 10x as fast as x = numpy.sum s, 0) Even though both return the same results, and certainly intuitively one would expect the latter to be faster!
|
|
Note sum in the former case is python s built in sum.
|
|
The result x is a long 1d vector.)
|
|
I discovered this on accident out of laziness to look up how to make sum work over just one axis, so I used the python sum which worked just fine.
|
|
When cleaning things up, I replaced it with the numpy sum, and suddenly my runs slowed to a crawl 97 seconds became 1190, and there are a lot of other things happening in there besides this sum!)
|
|
I have verified that s.base is m, and that the issue is with sum, and not the original slice operation.
|
|
Always possible I ve done something stupid, of course, but for the moment it appears this is a bonified bug.
|
|
Brandyn","enhancement"
|
|
"6013","1742","unit test suite test fails after installing scipy and numpy Trac 1144).
|
|
Original ticket http projects.scipy.org numpy ticket 1144 on 2009 06 20 by trac user iftikhar.khan, assigned to unknown.
|
|
Hi, Just installed python 2.6 and the 2.6 versions of scipy numpy onto my Windows machine.
|
|
Get the following error when running scipy.test ) and numpy.test ) >>> numpy.test ) Running unit tests for numpy Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 242, in test self.
|
|
show system info ) File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 149, in show system info nose = import nose ) File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 49, in import nose raise ImportError msg) ImportError Need nose >= 0.10.0 for tests see http somethingaboutorange.com mrl projects nose >>> scipy.test ) Running unit tests for scipy Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 242, in test self.
|
|
show system info ) File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 149, in show system info nose = import nose ) File C \Python26\lib\site packages\numpy\testing\nosetester.py , line 49, in import nose raise ImportError msg) ImportError Need nose >= 0.10.0 for tests see http somethingaboutorange.com mrl projects nose >>> I went to the nosetest website but it is dedicated to just the UNIX OS.
|
|
Thanks Iftikhar","bug"
|
|
"6013","1743","Fail to compile fortran code with recent Intel Fortran compiler for ia64 Trac 1145).
|
|
Original ticket http projects.scipy.org numpy ticket 1145 on 2009 06 20 by cournape, assigned to unknown.
|
|
Recent fortran compiler verstring Intel R) Fortran IA 64 Compiler Professional for applications running on IA 64, Version 11.0 Build 20081105 Package ID l cprof p 11.0.074 Copyright C) 1985 2008 Intel Corporation.
|
|
All rights reserved.
|
|
Intel Fortran 11.0 1558","bug"
|
|
"6013","1744","Clarify the documentation standard vis a vis classes and their attributes Trac 1146).
|
|
Original ticket http projects.scipy.org numpy ticket 1146 on 2009 06 21 by trac user dgoldsmith, assigned to pv.
|
|
For example, it is not clear that, but for the optional Attributes and Methods sections, documenting a class is materially equivalent to documenting its constructor.
|
|
For other perceived deficiencies, please see http mail.scipy.org pipermail scipy dev 2009 June 012198.html and the thread of which it is a part.","bug"
|
|
"6013","1745","Feature Suggestion Ability to lock arrays so they are read only Trac 1147).
|
|
Original ticket http projects.scipy.org numpy ticket 1147 on 2009 06 24 by trac user brandyn, assigned to unknown.
|
|
This is most important for slices or views onto larger arrays that you don t want to accidentally alter.
|
|
Typical case is where some routine returns a view onto another array, but the calling routine doesn t expect or realize this and thinks it can safely alter the returned array.
|
|
Any array whether original or a view) should be lockable.
|
|
If a view is locked, it may still be changed via changes to the original locking would not promise the data will never change use a copy for that), but one would not be allowed to change the data via the view.
|
|
When a view is created, it should inherit locked state from the source.
|
|
Developers can use this either by locking an original when they know it shouldn t change any more no matter what, or by locking views any time they conceptually should be using a copy but aren t for efficiency reasons because they don t expect the view to be written.
|
|
Even if this is only partially implemented just the most common write operations) it would go a long way toward catching common bugs.
|
|
Brandyn","enhancement"
|
|
"6013","1746","f2py modules crash if RTLD GLOBAL in effect Trac 1148).
|
|
Original ticket http projects.scipy.org numpy ticket 1148 on 2009 06 25 by trac user hugh, assigned to unknown.
|
|
I was getting seg faults when the l bfgs b optimiser fortran code) in SciPy ran and discovered this ONLY occurred when the library openbabel was loaded.
|
|
The backtrace seems to suggest that the error is in numpy core multiarray.so.
|
|
The following code induces the crash ===================================== import openbabel comment out this line to stop seg fault import numpy from scipy.optimize.lbfgsb import fmin l bfgs b def f x) return x[0] 2 x[1] 2 def g x) return numpy.array [2 x[0], 2 x[1]]) x0 = numpy.array [3,1]) opt, energy, dict = fmin l bfgs b f, x0, fprime=g) print opt Here is the backtrace ====================== gdb) bt 0 0x00007f93e8635720 in PyArray API ) from usr local lib64 python2.5 site packages numpy core multiarray.so 1 0x00007f93e435afa3 in initminpack2 ) at usr local lib64 python2.5 site packages numpy core include numpy multiarray api.h 983 2 0x00007f93ea23e2f6 in PyImport LoadDynamicModule ) from usr lib64 libpython2.5.so.1.0 3 0x00007f93ea23c6e9 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 4 0x00007f93ea23cc79 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 5 0x00007f93ea23d200 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 6 0x00007f93ea23d335 in PyImport ImportModuleLevel ) from usr lib64 libpython2.5.so.1.0 7 0x00007f93ea221c5d in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 8 0x00007f93ea1ab683 in PyObject Call ) from usr lib64 libpython2.5.so.1.0 9 0x00007f93ea222251 in PyEval CallObjectWithKeywords ) from usr lib64 libpython2.5.so.1.0 10 0x00007f93ea226278 in PyEval EvalFrameEx ) from usr lib64 libpython2.5.so.1.0 11 0x00007f93ea229c4d in PyEval EvalCodeEx ) from usr lib64 libpython2.5.so.1.0 12 0x00007f93ea229e42 in PyEval EvalCode ) from usr lib64 libpython2.5.so.1.0 13 0x00007f93ea23ada0 in PyImport ExecCodeModuleEx ) from usr lib64 libpython2.5.so.1.0 14 0x00007f93ea23b028 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 15 0x00007f93ea23c6e9 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 16 0x00007f93ea23c982 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 17 0x00007f93ea23cfcf in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 18 0x00007f93ea23d335 in PyImport ImportModuleLevel ) from usr lib64 libpython2.5.so.1.0 19 0x00007f93ea221c5d in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 20 0x00007f93ea1ab683 in PyObject Call ) from usr lib64 libpython2.5.so.1.0 21 0x00007f93ea222251 in PyEval CallObjectWithKeywords ) from usr lib64 libpython2.5.so.1.0 22 0x00007f93ea226278 in PyEval EvalFrameEx ) from usr lib64 libpython2.5.so.1.0 23 0x00007f93ea229c4d in PyEval EvalCodeEx ) from usr lib64 libpython2.5.so.1.0 24 0x00007f93ea229e42 in PyEval EvalCode ) from usr lib64 libpython2.5.so.1.0 25 0x00007f93ea23ada0 in PyImport ExecCodeModuleEx ) from usr lib64 libpython2.5.so.1.0 26 0x00007f93ea23b028 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 27 0x00007f93ea23c6e9 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 28 0x00007f93ea23c982 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 29 0x00007f93ea23cfcf in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 30 0x00007f93ea23d335 in PyImport ImportModuleLevel ) from usr lib64 libpython2.5.so.1.0 31 0x00007f93ea221c5d in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 32 0x00007f93ea1ab683 in PyObject Call ) from usr lib64 libpython2.5.so.1.0 33 0x00007f93ea222251 in PyEval CallObjectWithKeywords ) from usr lib64 libpython2.5.so.1.0 34 0x00007f93ea226278 in PyEval EvalFrameEx ) from usr lib64 libpython2.5.so.1.0 35 0x00007f93ea229c4d in PyEval EvalCodeEx ) from usr lib64 libpython2.5.so.1.0 36 0x00007f93ea229e42 in PyEval EvalCode ) from usr lib64 libpython2.5.so.1.0 37 0x00007f93ea23ada0 in PyImport ExecCodeModuleEx ) from usr lib64 libpython2.5.so.1.0 38 0x00007f93ea23b028 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 39 0x00007f93ea23bfa7 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 40 0x00007f93ea23c6e9 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 41 0x00007f93ea23c982 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 42 0x00007f93ea23d016 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 43 0x00007f93ea23d335 in PyImport ImportModuleLevel ) from usr lib64 libpython2.5.so.1.0 44 0x00007f93ea221c5d in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 45 0x00007f93ea1ab683 in PyObject Call ) from usr lib64 libpython2.5.so.1.0 46 0x00007f93ea222251 in PyEval CallObjectWithKeywords ) from usr lib64 libpython2.5.so.1.0 47 0x00007f93ea226278 in PyEval EvalFrameEx ) from usr lib64 libpython2.5.so.1.0 48 0x00007f93ea229c4d in PyEval EvalCodeEx ) from usr lib64 libpython2.5.so.1.0 49 0x00007f93ea229e42 in PyEval EvalCode ) from usr lib64 libpython2.5.so.1.0 50 0x00007f93ea23ada0 in PyImport ExecCodeModuleEx ) from usr lib64 libpython2.5.so.1.0 51 0x00007f93ea23b028 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 52 0x00007f93ea23c6e9 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 53 0x00007f93ea23c982 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 54 0x00007f93ea23cfcf in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 55 0x00007f93ea23d335 in PyImport ImportModuleLevel ) from usr lib64 libpython2.5.so.1.0 56 0x00007f93ea221c5d in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 57 0x00007f93ea1ab683 in PyObject Call ) from usr lib64 libpython2.5.so.1.0 58 0x00007f93ea222251 in PyEval CallObjectWithKeywords ) from usr lib64 libpython2.5.so.1.0 59 0x00007f93ea226278 in PyEval EvalFrameEx ) from usr lib64 libpython2.5.so.1.0 60 0x00007f93ea229c4d in PyEval EvalCodeEx ) from usr lib64 libpython2.5.so.1.0 Type <return> to continue, or q <return> to quit 61 0x00007f93ea229e42 in PyEval EvalCode ) from usr lib64 libpython2.5.so.1.0 62 0x00007f93ea2447a1 in ?? )
|
|
from usr lib64 libpython2.5.so.1.0 63 0x00007f93ea24499c in PyRun StringFlags ) from usr lib64 libpython2.5.so.1.0 64 0x00007f93ea245c10 in PyRun SimpleStringFlags ) from usr lib64 libpython2.5.so.1.0 65 0x00007f93ea24f06a in Py Main ) from usr lib64 libpython2.5.so.1.0 66 0x00007f93e95b8436 in libc start main ) from lib64 libc.so.6 67 0x00000000004006c9 in start )","bug"
|
|
"6013","1747","argmin doesn t work with datetime objects Trac 1149).
|
|
Original ticket http projects.scipy.org numpy ticket 1149 on 2009 06 25 by trac user doylejg, assigned to unknown.
|
|
With a numpy array of datetime or datetime.timedelta objects, it appears that min will work, but argmin will not.
|
|
I have attached the traceback.
|
|
Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.5 site packages numpy core fromnumeric.py , line 632, in argmin return argmin axis) TypeError unsupported operand type s) for int and datetime.timedelta","bug"
|
|
"6013","1748","Object arrays can t be multiplied by numpy scalars Trac 1150).
|
|
Original ticket http projects.scipy.org numpy ticket 1150 on 2009 06 25 by inducer, assigned to unknown.
|
|
Python 2.5.4 r254 67916, Feb 18 2009, 03 00 47) [GCC 4.3.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.3.0 >>> numpy.zeros 5, dtype=object) 5. array [0.0, 0.0, 0.0, 0.0, 0.0], dtype=object) >>> numpy.zeros 5, dtype=object) numpy.float64 5) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError unsupported operand type s) for numpy.ndarray and numpy.float64","bug"
|
|
"6013","1749","signal.medfilt2d crashes under Windows for scipy 0.7 and above Trac 1151).
|
|
Original ticket http projects.scipy.org numpy ticket 1151 on 2009 06 26 by trac user PCA, assigned to unknown.
|
|
The following Python script below) causes a fatal error under Microsoft Windows XP Professional Version 2002 Service Pack 3 with ActivePython 2.5.2.2 and numpy 1.2.1 win32 superpack python2.5 with scipy 0.7.0 win32 superpack python2.5 or above but works with scipy 0.6.0.win32 p3 py2.5 The combination of numpy 1.1.0 win32 superpack python2.5 scipy 0.6.0.win32 p3 py2.5.exe works without any deprecation warnings.
|
|
import numpy as N import scipy as SP from scipy import signal import pylab as P import matplotlib as MPL def main ) lena = SP.lena ) print lena.shape P.imshow lena, interpolation= nearest , cmap=MPL.cm.gray) s = signal.medfilt2d lena.astype float), [15,15]) if name == main main ) P.show )","bug"
|
|
"6013","1750","fromfile ) returns MemoryError exception when attempting to read at EOF Trac 1152).
|
|
Original ticket http projects.scipy.org numpy ticket 1152 on 2009 06 26 by trac user jludlow, assigned to charris.
|
|
Consider the following brief example >>> fid = open temp.bin , wb ) >>> x = array [1.0, 2.0, 3.0, 4.0]) >>> x.tofile fid) >>> fid.close ) >>> fid = open temp.bin , rb ) >>> y = fromfile fid, count=5) >>> y array [ 1., 2., 3., 4.])
|
|
>>> z = fromfile fid, count=5) Traceback most recent call last) MemoryError >>> fromfile ) appears to behave as expected when more items are requested than are available in the file, nicely truncating the length of the returned array to only the available items.
|
|
However, a subsequent fromfile ) on the same file object returns the MemoryError exception.
|
|
This seems to happen only when a fromfile ) call is initiated right on the EOF.
|
|
Whether this is a defect or the intended behavior I m not sure.
|
|
It would be more consistent with other file access methods like fid.read bytes) to simply return an empty array when nothing more is available to read.","bug"
|
|
"6013","1751","performing operations inplace to save memory allocations Trac 1153).
|
|
Original ticket http projects.scipy.org numpy ticket 1153 on 2009 06 27 by trac user lucaciti, assigned to charris.
|
|
Hi!
|
|
From the Guide to !NumPy , pag 169 If your arrays are large, complicated expressions can take longer than absolutely necessary due to the creation and later) destruction of temporary calculation spaces.
|
|
For example, the expression G=a b c is equivalent to t1=A B; G=T1 C; del t1; It will be more quickly executed as G=A B; add G,C,G) which is the same as G=A B; G =C.
|
|
I have tried to overcome this limitation by using the ref count of the input objects.
|
|
After some time spent to understand how ufuncs work and where I should have tried to hook, I think I found a decent solution.
|
|
It is a starting point but kind of works.
|
|
Attached you can find a patch which enables this feature.
|
|
The basic idea is the following G=a b c is performed in two steps, the first is evaluating a b and storing the result in a temporary array; the second is summing this temporary array and c, and returning the result; as the first input of the second operation has ref count 1 i.e., besides the current function, nobody else cares about it if it is of the correct size and type, we can use this array as output of the second operation and save the allocation of an array.
|
|
The performance increase is substantial, in my pc I get for .5 r s) 100x100 108us instead of 109us 1000x1000 26.2ms instead of 31ms 10000x1000 269ms instead of 311ms for .5 r 2) 2) 100x100 135us instead of 138us 1000x1000 34.2ms instead of 43.3ms 10000x1000 352ms instead of 428ms The approach works in most cases it passes most of the numpy tests) but makes the assumption that ufuncs can work even if one of the inputs is used as output.
|
|
Is this an acceptable constraint on ufuncs?
|
|
It works with the base arithmetic functions, trigonometric functions, abs, sign, power, ...
|
|
I wasn t able to make complex operations work and I had to exclude them from this enhancement.
|
|
It should be easy to fix by replacing, for example out >real = a.real b.real a.imag b.imag; out >imag = a.real b.imag a.imag b.real; with real type tmp = a.real b.real a.imag b.imag; outp >imag = a.real b.imag a.imag b.real; outp >real = tmp; in the code inside the inner loop which performs the scalar operation.
|
|
Similarly, divmod does not work because it is internally evaluated in two steps and assumes that the inputs do not change between the two.
|
|
This too can be easily fixed.
|
|
Also masked arrays do not work but I never use them and I do not know how to fix it.
|
|
Probably the problem is still something similar to the above.
|
|
Of course there might be some other functions which does not work.
|
|
Thorough tests should be able to find them.
|
|
To conclude, I made some step towards optimizing compound operations in !NumPy.
|
|
This can have a substantial positive effect in terms of time 20 for big matrices) and reduces peaks of memory consumption.
|
|
But before going further trying to fix the open issues, I would like to have some feedback and some help if you think this is a direction in which it is worth working).
|
|
Thanks.
|
|
Luca","enhancement"
|
|
"6013","1752","NPY macros for endianness are broken if endian.h was not available Trac 1154).
|
|
Original ticket http projects.scipy.org numpy ticket 1154 on 2009 06 30 by cournape, assigned to unknown.
|
|
I managed to screw them up, NPY BIG ENDIAN and NPY LITTLE ENDIAN did not follow the BIG ENDIAN and LITTLE ENDIAN counterparts...","bug"
|
|
"6013","1753","Mac OSX installer sets file owner to UID 501 regardless of installing user s UID Trac 1155).
|
|
Original ticket http projects.scipy.org numpy ticket 1155 on 2009 06 30 by trac user gorm, assigned to rgommers.
|
|
It appears that the mac OSX installers .dmg) sets the file ownership of all installed files to UID 501 regardless of the actual UID of the installing user.
|
|
I have a UID different from 501 due to NFS related issues).
|
|
I suspect the correct behavior would be to set the owner to be the user actually running the installer.","bug"
|
|
"6013","1754","vectorizing an Unbound Method Trac 1156).
|
|
Original ticket http projects.scipy.org numpy ticket 1156 on 2009 06 30 by trac user barronh, assigned to unknown.
|
|
The current vectorize implementation cannot appropriately vectorize an UnboundMethodType function.
|
|
The root of the problem in is numpy.lib.function base.
|
|
get nargs line 1778.
|
|
The number of arguments that a function takes is decremented if the function is a MethodType.
|
|
MethodType is inclusive of bound and unbound methods.
|
|
To make the unbound method useful, it must take an instance as the first argument, but the check against the number of arguments prevents this.
|
|
reproduce with >>> from numpy import vectorize, arange >>> class foo >>> b=2 >>> def bar self, a) >>> return a self.b >>> >>> vectorize foo.bar) foo ),arange 9)) ValueError mismatch between python function inputs and received arguments Changing numpy.lib.function base line 1778 solves the problem < if instance obj, types.MethodType) > if instance obj, types.MethodType) and not instance obj, types.UnboundMethodType)","bug"
|
|
"6013","1755","Unexpected results with inf and complex numbers Trac 1157).
|
|
Original ticket http projects.scipy.org numpy ticket 1157 on 2009 07 01 by cournape, assigned to unknown.
|
|
!python import numpy as np a = np.complex 1, np.inf) b = 1 1j np.inf print a print b Outputs 1 infj nan infj","bug"
|
|
"6013","1756","handle nan inf signed zeros in comparison methods for testing Trac 1158).
|
|
Original ticket http projects.scipy.org numpy ticket 1158 on 2009 07 02 by cournape, assigned to unknown.","enhancement"
|
|
"6013","1757","Complex division overflows too early Trac 1159).
|
|
Original ticket http projects.scipy.org numpy ticket 1159 on 2009 07 06 by pv, assigned to unknown.
|
|
Complex division overflows early, and returns bogus results in some cases >>> asarray [1e210 0j]) 1e100 array [ Inf 0.j]) It s easy to see why loops.c.src C TYPE divide ... type d = in2r in2r in2i in2i; type )op1)[0] = in1r in2r in1i in2i) d; type )op1)[1] = in1i in2r in1r in2i) d;","bug"
|
|
"6013","1758","numscons Headers not installed with in place build Trac 1160).
|
|
Original ticket http projects.scipy.org numpy ticket 1160 on 2009 07 07 by stefanv, assigned to cournape.
|
|
The numpyconfig.h and multiarray api.h headers are not installed with in place builds using numscons.","bug"
|
|
"6013","1759","Errors and or wrong result with reverse slicing in numpy.delete Trac 1161).
|
|
Original ticket http projects.scipy.org numpy ticket 1161 on 2009 07 08 by trac user kgdunn, assigned to unknown.
|
|
If the slice object argument for np.delete is in reverse order, then then np.delete may fail with this error ValueError shape mismatch objects cannot be broadcast to a single shape.
|
|
In other cases it gives unexpected wrong) results.
|
|
Example A throws error >>> a = np.arange 18) >>> a.shape = 3, 6) >>> a array [[ 0, 1, 2, 3, 4, 5], [ 6, 7, 8, 9, 10, 11], [12, 13, 14, 15, 16, 17]]) >>> slice obj fwd = slice None, None, 2) >>> slice obj rev = slice None, None, 2) >>> np.delete a, obj=slice obj fwd, axis=1) case 1 succeeds >>> np.delete a, obj=slice obj rev, axis=1) case 2 fails The only difference between case 1 and 2 is whether one steps forwards or backwards.
|
|
I know it is a bit weird to slice in reverse, but slice obj rev is created programmatically it is intended to be similar to the behaviour expected with >>> a[ , slice obj rev] returns column 6, 4 and 2, in that order array [[ 5, 3, 1], [11, 9, 7], [17, 15, 13]]) Example B gives wrong result This deletes column 0 and 1, instead of 3 and 4 >>> np.delete a, obj=slice 4, 2, 1), axis=1) array [[ 2, 3, 4, 5], [ 8, 9, 10, 11], [14, 15, 16, 17]]) I was not able to find consistent rules for when it will fail, or when it will give a wrong result.
|
|
It seems to be caused by the way the code in numpy lib function base.py operates it takes the start chunk, end chunk and then middle pieces.
|
|
The slicing objects generated this way in function base.py are not of the correct shape.
|
|
I m not sure of the reason why that code is structured to extract the start, end and then middle, so I m reluctant to change it and provide a patch.
|
|
Why not just use xrange start, stop, step) to find the indices to delete, then take the setdiff1d, and copy over those over?
|
|
This is what is done already in the code region marked with the comment default behavior .
|
|
Hope this detail helps, Kevin","bug"
|
|
"6013","1760","linalg.matrix power crashes with nested list input Trac 1162).
|
|
Original ticket http projects.scipy.org numpy ticket 1162 on 2009 07 09 by trac user dgoldsmith, assigned to unknown.
|
|
linalg.matrix power raises an AttributeError when a nested list is passed in as the array argument Microsoft Windows [Version 6.0.6002] Copyright c) 2006 Microsoft Corporation.
|
|
All rights reserved.
|
|
C \Users\DG>python Python 2.5.4 r254 67916, Dec 23 2008, 15 10 54) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> from numpy import linalg as LA >>> LA.matrix power [[1, 2], [3, 4]], 3) Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python25\Lib\site packages\numpy\core\defmatrix.py , line 118, in matrix power if len M.shape) != 2 or M.shape[0] != M.shape[1] AttributeError list object has no attribute shape Yet the doc states matrix power M, n) Raise a square matrix to the integer) power n. Parameters M array like I ve corrected the doc, for now, but I assume this actually a code bug.","bug"
|
|
"6013","1761","Incorrect conversion to Int64 by loadtxt traced to getconv in numpy.lib.io) Trac 1163).
|
|
Original ticket http projects.scipy.org numpy ticket 1163 on 2009 07 09 by trac user onsi, assigned to unknown.
|
|
I m running version 1.2.1 but this error should also occur in 1.3.0 based on the source currently in the trunk.
|
|
I try importing the following ascii data stored in sample.csv 9007200000000000,670927001710,0.010190886[[BR]] 9007200000000001,660927001348,0.00976051[[BR]] 9007200000000002,650883003926,0.009154096 using maximal verbosity for clarity) import numpy arr=numpy.loadtxt sample.csv ,dtype=[ id0 ,numpy.int64), id1 ,numpy.int64), flt ,numpy.float32)],delimiter= , ,comments= ) I get [ 9007200000000000L, 670927001710L, 0.010190886445343494)[[BR]] 9007200000000000L , 660927001348L, 0.0097605101764202118)[[BR]] 9007200000000002L, 650883003926L, 0.009154096245765686)][[BR]] After some digging, i found the culprit to be the converter used by loadtxt to convert strings to dtypes.
|
|
lib.io.
|
|
getconv line 352 in trunk) returns lambda x int float x)) as the converter for any dtype that is a subclass of int, which int64 is.
|
|
Unfortunately, float does not faithfully reproduce long integers and so 9007200000000001 gets rounded to 9007200000000000.
|
|
This is fairly serious as int64s are often used as IDs in various numerical simulation contexts.
|
|
Changing the converter to int ) should resolve this problem though then some error checking needs to take place to ensure that int is fed an integer string.","bug"
|
|
"6013","1762","pg.py specifies module for all compilers, but this isn t valid for pgf77 Trac 1164).
|
|
Original ticket http projects.scipy.org numpy ticket 1164 on 2009 07 10 by trac user noam.bernstein, assigned to dmcooke.
|
|
numpy distutils fcompiler pg.py specifies module for searching for module files.
|
|
Since fortran 77 doesn t have modules, pgf77 complains.
|
|
This can be fixed by adding noswitcherror to the default set of flags for pgf77.
|
|
I m definitely seeing it in 1.2.1, but searching for module dir switch suggests that it s not been fixed in more recent versions.
|
|
A nearly trivial) diff is attached.","bug"
|
|
"6013","1763","Clarify the documentation standard on the issue of documenting class methods Trac 1165).
|
|
Original ticket http projects.scipy.org numpy ticket 1165 on 2009 07 10 by trac user dgoldsmith, assigned to pv.
|
|
Proposal regarding modification of the NumPy Docstring Standard Under the heading Method docstrings in the section Documenting classes Reword as follows Do not include self in the list of parameters.
|
|
Otherwise, follow the docstring standard for an ordinary function, except when the method duplicates a function in the package namespace.
|
|
In such cases, simply provide the brief summary and the See Also section, whose content should simply be <packagename>.<name of equivalent function>.
|
|
On those rare occasions when a method does some additional, significant processing e.g., parameter manipulation) before or after) handing off to a nominally equivalent function, then document only the added functionality in the method, using only the sections needed to do so, and for the rest refer to the function.
|
|
In general, the guiding principle is to document behavior where it is implemented in the source code, with ufuncs considered to live in their function implementation rather than any method implementation they may have.","bug"
|
|
"6013","1764","numpy reduce operation on masked array with no mask fails Trac 1166).
|
|
Original ticket http projects.scipy.org numpy ticket 1166 on 2009 07 11 by trac user reggie, assigned to pierregm.
|
|
If you have a MaskedArray with no masked values, but whose mask is not nomask, a numpy operation on that array that results in an array whose shape is different from the original eg.
|
|
numpy.maximum.reduce), will result in an exception being raised.
|
|
For example In [1] import numpy In [2] numpy.
|
|
version Out[2] 1.3.0 In [3] x = numpy.ma.arange 5) In [4] x.mask = numpy.ma.getmaskarray x) In [5] x Out[5] masked array data = [0 1 2 3 4], mask = [False False False False False], fill value = 999999) In [6] numpy.maximum.reduce x) ValueError Traceback most recent call last) home reggie src python <ipython console> in <module> ) usr local lib python2.6 site packages numpy ma core.py in array finalize self, obj) 2444 Finalize the mask ........... 2445 if self.
|
|
mask is not nomask > 2446 self.
|
|
mask.shape = self.shape 2447 return 2448 .................................. ValueError total size of new array must be unchanged In [7] I have attached a patch that seems to fix the problem.","bug"
|
|
"6013","1765","speed up np.diag Trac 1167).
|
|
Original ticket http projects.scipy.org numpy ticket 1167 on 2009 07 11 by trac user lucaciti, assigned to unknown.
|
|
I have re written diag and eye using a different approach.
|
|
I think it is cleaner but, more importantly, it gives a significant speed improvement.
|
|
I have also run some tests with square, large, tall matrices, k outside the matrix and so on, and the result is the same as the original versions.
|
|
Below some comparison of the execution time.
|
|
The improvement is often by a factor 3, with a maximum of 9 and a minimum of 2 except when diag has to create a big matrix in this case most of the time is spent in memory allocation and the two versions take the same time).
|
|
In [1] import numpy as np In [2] import qqq In [3] timeit qqq.eye 10) 10000 loops, best of 3 23.6 xc2xb5s per loop In [4] timeit qqq.eye2 10) 100000 loops, best of 3 12.7 xc2xb5s per loop In [5] timeit qqq.eye 10,1000) 10000 loops, best of 3 140 xc2xb5s per loop In [6] timeit qqq.eye2 10,1000) 10000 loops, best of 3 24.9 xc2xb5s per loop In [7] timeit qqq.eye 1000,10) 1000 loops, best of 3 212 xc2xb5s per loop In [8] timeit qqq.eye2 1000,10) 10000 loops, best of 3 24.7 xc2xb5s per loop In [9] timeit qqq.eye 1000,1000) 10 loops, best of 3 26.5 ms per loop In [10] timeit qqq.eye2 1000,1000) 100 loops, best of 3 10.9 ms per loop In [11] timeit qqq.eye 1000,1000,500) 10 loops, best of 3 27.1 ms per loop In [12] timeit qqq.eye2 1000,1000,500) 100 loops, best of 3 10.9 ms per loop In [13] x = np.random.rand 10,15) In [14] timeit qqq.diag x) 10000 loops, best of 3 25.9 xc2xb5s per loop In [15] timeit qqq.diag2 x) 100000 loops, best of 3 8.16 xc2xb5s per loop In [16] x = np.random.rand 1000,15) In [17] timeit qqq.diag x) 10000 loops, best of 3 26 xc2xb5s per loop In [18] timeit qqq.diag2 x) 100000 loops, best of 3 8.24 xc2xb5s per loop In [19] x = np.random.rand 1000,1500) In [20] timeit qqq.diag x) 10000 loops, best of 3 77.2 xc2xb5s per loop In [21] timeit qqq.diag2 x) 10000 loops, best of 3 31.9 xc2xb5s per loop In [22] timeit qqq.diag x, 79) 10000 loops, best of 3 75.7 xc2xb5s per loop In [23] timeit qqq.diag2 x, 79) 10000 loops, best of 3 31 xc2xb5s per loop In [24] x = np.random.rand 1000,15) In [25] timeit qqq.diag x, 5) 10000 loops, best of 3 27.4 xc2xb5s per loop In [26] timeit qqq.diag2 x, 5) 100000 loops, best of 3 7.88 xc2xb5s per loop In [27] x = np.random.rand 15) In [28] timeit qqq.diag x, 5) 10000 loops, best of 3 30.9 xc2xb5s per loop In [29] timeit qqq.diag2 x, 5) 100000 loops, best of 3 11.3 xc2xb5s per loop In [30] x = np.random.rand 1500) In [31] timeit qqq.diag x, 5) 10 loops, best of 3 25.4 ms per loop In [32] timeit qqq.diag2 x, 5) 10 loops, best of 3 25.4 ms per loop While the two versions of diag give the same results with strings as well, the new eye gives different results.
|
|
It should not be a problem because I do no think somebody is using it and, to be honest, I think the results of the new version are somehow more correct.
|
|
In fact, as zeros 3, dtype= S10 ) returns empty strings, I expect eye to have empty strings outside the chosen diagonal.
|
|
In [33] qqq.eye 3,dtype= S10 ) Out[33] array [[ True , False , False ], [ False , True , False ], [ False , False , True ]], dtype= |S10 ) In [34] qqq.eye2 3,dtype= S10 ) Out[34] array [[ 1 , , ], [ , 1 , ], [ , , 1 ]], dtype= |S10 )","enhancement"
|
|
"6013","1766","Make tester s package name detection work packages installed in a non standard location Trac 1168).
|
|
Original ticket http projects.scipy.org numpy ticket 1168 on 2009 07 11 by GaelVaroquaux, assigned to unknown.
|
|
Right now, numpy s tester package name detection used to do coverage analysis) fails for a package that is installed in a non standard location.
|
|
This is due to the following logics def get package name filepath) find the package name given a path name that s part of the package fullpath = filepath[ ] pkg name = [] while site packages in filepath filepath, p2 = os.path.split filepath) if p2 == site packages break pkg name.append p2) which fails if the package is installed outside a site package directory.
|
|
I attach a patch that fixes this problem works for me), and I believe is more robust.","bug"
|
|
"6013","1767","IndexError needed when accessing rows of 2D array that don t exist Trac 1169).
|
|
Original ticket http projects.scipy.org numpy ticket 1169 on 2009 07 14 by trac user markbakker, assigned to unknown.
|
|
When I create a 2D array with 1 row and 10 columns >>> a = ones 1,10)) I can access the first and only) row as a[0].
|
|
When I try to access row 5 I get an indexerror as well I should) >>> a[5] IndexError When I try to access row 5 through 8 I don t get an IndexError, but an emtpy matrix.
|
|
a[5 8] Out[105] array [], shape= 0, 10), dtype=float64) I think this is a bug, and numpy should return and IndexError.
|
|
Thanks, Mark","bug"
|
|
"6013","1768","Possible Bug in F2PY Fortran Compiler Detection Trac 1170).
|
|
Original ticket http projects.scipy.org numpy ticket 1170 on 2009 07 14 by trac user selevercin, assigned to dmcooke.
|
|
I believe there is a bug in the way f2py detects fortran compilers, where the executable name never gets entered.
|
|
I ve tried posting this to the mailing list, but my post wont show up.)
|
|
Everything discussed here was performed on Linux cluster1 2.6.17 1.2142 FC4smp 1 SMP Tue Jul 11 22 59 20 EDT 2006 x86 64 x86 64 x86 64 GNU Linux Trying to get f2py to recognize ifort, I started to output comments in some of the code to help trace down the mistake.
|
|
I started in numpy distutils fcompiler init .py , which appears to check for the existence of a compiler based on whether or not it can find a version number in numpy distutils fcompiler init .py .
|
|
I ve included my output, but I ll also add short summary 1. f2py successfully locates all compatible compilers with the platform init .py) 2. f2py begins trying to get versions for each possible compiler init .py) 3. f2py INCORRECTLY creates a version command, to test if the compiler exists ccompiler.py) Where the version command sticks with the wrong executables opt pgi linux86 64 6.2 bin pgf77 and opt pgi linux86 64 6.2 bin pgf90 ).
|
|
For example, to detect if ifort is installed, the following command is wrongly) run opt pgi linux86 64 6.2 bin pgf77 FI V c tmp tmplG4 3d ESXFqO.f o tmp tmplG4 3d ESXFqO.o This command is garbage, because it is trying pgf77 rather than ifort.
|
|
Running this command is silly but I ll go ahead and include it pgf77 Warning Unknown switch FI pgf77 6.2 5 64 bit target on x86 64 Linux Copyright 1989 2000, The Portland Group, Inc. All Rights Reserved.
|
|
Copyright 2000 2006, STMicroelectronics, Inc. All Rights Reserved.
|
|
PGFTN x86 64 Linux x86 64 6.2 5 Copyright 1989 2000, The Portland Group, Inc. All Rights Reserved.
|
|
Copyright 2000 2006, STMicroelectronics, Inc. All Rights Reserved.
|
|
Since ifort is in the path which ifort home user10 intel fce 9.1.041 bin ifort The following works well home user10 intel fce 9.1.041 bin ifort FI V c tmp tmplG4 3d ESXFqO.f o tmp tmplG4 3d ESXFqO.o Intel R) Fortran Compiler for Intel R) EM64T based applications, Version 9.1 Build 20070109 Package ID l fc c 9.1.041 Copyright C) 1985 2007 Intel Corporation.
|
|
All rights reserved.
|
|
FOR NON COMMERCIAL USE ONLY Intel Fortran 9.1 6565 I tried a quick hack to get the ifort executable in the version command, but this didn t cause ifort to be successfully detected.
|
|
My Test Output python setup.py install prefix= {HOME} numpy dir | grep NDR Running from numpy source directory.
|
|
home nreveles numpy 1.3.0 numpy distutils command config.py 361 DeprecationWarning Usage of get output is deprecated please do not use it anymore, and avoid configuration checks involving running executable on the target machine.
|
|
DeprecationWarning) NDR self= <numpy.distutils.fcompiler.gnu.GnuFCompiler instance at 0x2aaaaee5d950> NDR version cmd = [ opt pgi linux86 64 6.2 bin pgf77 , version ] NDR compiler = <numpy.distutils.fcompiler.gnu.GnuFCompiler instance at 0x2aaaaee5d950> NDR comp.
|
|
version = None NDR self= <numpy.distutils.fcompiler.intel.IntelFCompiler instance at 0x2aaaaee62200> NDR version cmd = [ opt pgi linux86 64 6.2 bin pgf77 , FI , V , c , tmp tmplG4 3d ESXFqO.f , o , tmp tmplG4 3d ESXFqO.o ] NDR compiler = <numpy.distutils.fcompiler.intel.IntelFCompiler instance at 0x2aaaaee62200> NDR comp.
|
|
version = None NDR compiler = <numpy.distutils.fcompiler.lahey.LaheyFCompiler instance at 0x2aaaaee62680> NDR self= <numpy.distutils.fcompiler.lahey.LaheyFCompiler instance at 0x2aaaaee62680> NDR version cmd = [ opt pgi linux86 64 6.2 bin pgf90 , version ] NDR comp.
|
|
version = None NDR compiler = <numpy.distutils.fcompiler.pg.PGroupFCompiler instance at 0x2aaaaee633b0> NDR self= <numpy.distutils.fcompiler.pg.PGroupFCompiler instance at 0x2aaaaee633b0> NDR version cmd = [ opt pgi linux86 64 6.2 bin pgf77 , V 2> dev null ] NDR comp.
|
|
version = 6.2 5","bug"
|
|
"6013","1769","Inconsistent indexing of arrays with no elements, depending on shape Trac 1171).
|
|
Original ticket http projects.scipy.org numpy ticket 1171 on 2009 07 14 by trac user jguyer, assigned to unknown.
|
|
An empty, rank 0 array can be indexed with an empty list !python >>> import numpy as np >>> np.
|
|
version 1.3.0.dev6304 >>> a = np.zeros 0,)) >>> a[..., []] array [], dtype=float64) but an empty array of higher rank throws an index error !python >>> b = np.zeros 1, 0)) >>> b[..., []] Traceback most recent call last) File <stdin> , line 1, in <module> IndexError invalid index We have a not completely insane reason for wanting the second case to return `array [], shape= 1, 0), dtype=float64)`.
|
|
At a minimum, it seems the two cases should be consistent.
|
|
I wonder if this is related to ticket 1169?
|
|
I see this on a 10.5.7 Mac, with Python 2.6.2 and NumPy 1.3.0dev and on a Debian etch machine running Python 2.4.4 and NumPy 1.2.1.","bug"
|
|
"6013","1770","abs does not work with maxint Trac 1172).
|
|
Original ticket http projects.scipy.org numpy ticket 1172 on 2009 07 15 by trac user etu, assigned to unknown.
|
|
>>> a = array [ 128], int8) >>> a array [ 128], dtype=int8) >>> abs a) array [ 128], dtype=int8)","bug"
|
|
"6013","1771","FAIL Tests new ufuncs on MaskedArrays.
|
|
Solaris) Trac 1173).
|
|
Original ticket http projects.scipy.org numpy ticket 1173 on 2009 07 15 by stsci sienkiew, assigned to pierregm.
|
|
====================================================================== FAIL Tests new ufuncs on MaskedArrays.
|
|
Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test core.py , line 1403, in test testUfuncRegression assert mask equal ur.mask, mr.mask) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 237, in assert mask equal assert array equal m1, m2) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 193, in assert array equal header= Arrays are not equal ) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 186, in assert array compare verbose=verbose, header=header) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 395, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 25.0 ) x array [False, True, False, True, False, False, False, True], dtype=bool) y array [False, True, False, False, False, False, False, False], dtype=bool) numpy 1.4.0.dev7132 solaris 8 python 2.5.1","bug"
|
|
"6013","1772","FAIL test testUfuncRegression test old ma.TestUfuncs) Trac 1174).
|
|
Original ticket http projects.scipy.org numpy ticket 1174 on 2009 07 15 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test testUfuncRegression test old ma.TestUfuncs) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test old ma.py , line 689, in test testUfuncRegression self.failUnless eqmask ur.mask, mr.mask)) AssertionError numpy 1.4.0.dev7132 solaris 8 python 2.5.1","bug"
|
|
"6013","1773","np.isfinite on structured arrays returns a NotImplemented type object Trac 1175).
|
|
Original ticket http projects.scipy.org numpy ticket 1175 on 2009 07 15 by pierregm, assigned to unknown.
|
|
Consider the current code >>> a = np.array zip np.arange 3)),dtype=[ a ,float)]) >>> np.isfinite a) NotImplemented Instead of a NotImplemented type object, shouldn t np.isfinite raise a NotImplementedError exception when called on a structured array ?","bug"
|
|
"6013","1774","Zero dimensional arrays are converted to scalars by deepcopy.
|
|
Trac 1176).
|
|
Original ticket http projects.scipy.org numpy ticket 1176 on 2009 07 17 by trac user mauro, assigned to unknown.
|
|
When making a deepcopy of a numpy.ndarray with dtype S a numpy.string is created.
|
|
I am not too familiar with the numpy types, but I think they are basically the same.
|
|
However, they have different repr and this can make this somewhat confusing it confused me).","bug"
|
|
"6013","1775","Using interp with numpy scalar raises ValueError Trac 1177).
|
|
Original ticket http projects.scipy.org numpy ticket 1177 on 2009 07 17 by tonysyu, assigned to charris.
|
|
A simple example that reproduces the error where `x0` below is the scalar of interest) >>> import numpy as np >>> x0 = np.array 0.1) >>> x = np.linspace 0, 1) >>> y = np.linspace 0, 1) >>> np.interp x0, x, y) ValueError object of too small depth for desired array Note that `interp` is able to handle python scalars, but not numpy scalars i.e.
|
|
zero rank arrays).
|
|
Below is a simple patch that fixes this issue.
|
|
Unfortunately, I m not using svn for numpy, so this is just a normal diff).
|
|
Note that this fix returns a numpy scalar.
|
|
lib function base.py ==================== 1143a1144,1145 > elif isinstance x, np.ndarray) and x.ndim == 0 > return compiled interp x[np.newaxis], xp, fp, left, right)[0]","bug"
|
|
"6013","1776","[PATCH] alpha hppa support Trac 1178).
|
|
Original ticket http projects.scipy.org numpy ticket 1178 on 2009 07 19 by trac user armin76, assigned to unknown.
|
|
The following patch enables alpha and hppa support.
|
|
The alpha part is done by me and the hppa fix is by Jeroen Roovers <jer at gentoo dot org> wrt https bugs.gentoo.org show bug.cgi?id=277438","bug"
|
|
"6013","1777","bug using array with a PIL Image misc.imread) Trac 1179).
|
|
Original ticket http projects.scipy.org numpy ticket 1179 on 2009 07 22 by trac user nicopinto, assigned to unknown.
|
|
For some reason I have a problem converting a specific png image using array ).
|
|
Here is what I am getting using numpy 1.3.0 and scipy 0.7.0).
|
|
wget http files.getdropbox.com u 167753 spiky adj 023.png python c import numpy as np; import Image; print np.array Image.open spiky adj 023.png )) <PngImagePlugin.PngImageFile instance at 0xd1a050> wget http files.getdropbox.com u 167753 spiky norm 001.png python c import numpy as np; import Image; print np.array Image.open spiky norm 001.png )) [[134 30 140 ..., 230 83 59] [ 99 202 233 ..., 160 63 133] [ 93 241 35 ..., 7 240 101] ..., [206 132 196 ..., 139 190 112] [218 21 217 ..., 121 152 109] [ 83 188 187 ..., 6 240 251]] I was initially using scipy s misc.imread when I found this bug.
|
|
I am currently using the following workaround python c import numpy as np; import Image; print np.array Image.open spiky adj 023.png ).convert RGB )) Let me know what you think.","bug"
|
|
"6013","1778","numpy.average add an option to skip None values from count Trac 1180).
|
|
Original ticket http projects.scipy.org numpy ticket 1180 on 2009 07 23 by dalloliogm, assigned to unknown.
|
|
If you calculate numpy.average the same goes with similar function) on a list in which one of the elements is None, you get an error.
|
|
This is silly, because the error message TypeError unsupported operand type for int and NoneType is not clear at all for newbies; you have to make use of temporary copies of your lists arrays and manually remove all the None values all the time.
|
|
Example >>> import numpy >>> numpy.average [1, 2, None]) TypeError ... >>> numpy.average numpy.array [1, 2, None])) TypeError ...
|
|
In other programming language this problems is solved by adding an option like ignore None to all the functions which suffer from this problem.","enhancement"
|
|
"6013","1779","lenght of a masked array?
|
|
Trac 1181).
|
|
Original ticket http projects.scipy.org numpy ticket 1181 on 2009 07 23 by dalloliogm, assigned to unknown.
|
|
What should len ) of a masked array return?
|
|
If I mask an array, I would expect that its lenght is equal to the number of not masked elements, or at least I would like to have a function to retrieve such information.
|
|
>>> import numpy >>> m = numpy.ma.masked equal [1, 2, None, 3, None, None, 4], None) I would expect len m) to be 4 >>> len m) 7 There is not an alternative function as for average, var, for masked arrays >>> numpy.ma.average m) 2 >>> numpy.average m) TypeError >>> numpy.ma.len m) Function doesn t exists","enhancement"
|
|
"6013","1780","fftshift and ifftshift axes argument should accept integer Trac 1182).
|
|
Original ticket http projects.scipy.org numpy ticket 1182 on 2009 07 26 by rgommers, assigned to charris.
|
|
The `axes` argument of both fftshift and ifftshift should accept either an int or a tuple of ints.
|
|
However, it fails for an integer In [2] freqs = [[ 0, 1, 2], [ 3, 4, 4], [ 3, 2, 1]] In [3] np.fft.fftshift freqs, axes=0) TypeError Traceback most recent call last) Users jingzou <ipython console> in <module> ) Library Frameworks Python.framework Versions 4.2.30201 lib python2.5 site packages numpy 1.2.1n1 py2.5 macosx 10.3 fat.egg numpy fft helper.pyc in fftshift x, axes) 34 axes = range ndim) 35 y = tmp > 36 for k in axes 37 n = tmp.shape[k] 38 p2 = n 1) 2 TypeError int object is not iterable A typecheck for an integer will fix this.
|
|
Patch with tests is attached.","bug"
|
|
"6013","1781","Using C99 complex type and functions when available Trac 1183).
|
|
Original ticket http projects.scipy.org numpy ticket 1183 on 2009 07 27 by cournape, assigned to unknown.
|
|
This is useful for the following reasons C extensions writers who do not care about portability to non C99 platforms can use the C99 type operator support, etc...) We can reuse the C99 cmath functions instead of rolling our own This is necessary for windows 64 support, to help writing a minimal fake gfortran runtime.","enhancement"
|
|
"6013","1782","assert array almost equal os bogus for nan arrays Trac 1184).
|
|
Original ticket http projects.scipy.org numpy ticket 1184 on 2009 07 27 by cournape, assigned to unknown.
|
|
!python import numpy as np np.testing.assert array almost equal np.array [np.nan]), np.array [1]))","bug"
|
|
"6013","1783","assert approx equal bogus for nan Trac 1185).
|
|
Original ticket http projects.scipy.org numpy ticket 1185 on 2009 07 27 by cournape, assigned to unknown.
|
|
!python import numpy as np np.testing.assert approx equal np.nan, 1) This does not raise any AssertionError","bug"
|
|
"6013","1784","numpy.ma.abs not working correctly?
|
|
Trac 1186).
|
|
Original ticket http projects.scipy.org numpy ticket 1186 on 2009 08 03 by dalloliogm, assigned to pierregm.
|
|
the standard function abs ) raises an error when applied to an array list containing a None value.
|
|
To circumvent this problem, I tought of using a masked array and the numpy.ma.abs function, but it doesn t work.
|
|
>>> import numpy >>> ma1 = numpy.ma.masked equal [1, 2, 3, None, 4, None, 5], None) >>> ma1 masked array data = [1 2 3 4 5], mask = [False False False True False True False], fill value = ?)
|
|
>>> numpy.ma.abs ma1) TypeError Traceback most recent call last) home gioby <ipython console> in <module> ) usr lib python2.5 site packages numpy 1.3.0 py2.5 linux i686.egg numpy ma core.pyc in call self, a, args, kwargs) 643 result = self.f d1, args, kwargs) 644 else > 645 result = np.where m, d1, self.f d1, args, kwargs)) 646 If result is not a scalar 647 if result.ndim TypeError bad operand type for abs ) NoneType >>> numpy.abs ma1) TypeError Traceback most recent call last) home gioby <ipython console> in <module> ) TypeError bad operand type for abs ) NoneType >>> abs ma1) TypeError Traceback most recent call last) home gioby <ipython console> in <module> ) TypeError bad operand type for abs ) NoneType I expected numpy.ma.abs to ignore the masked values, as it works for other functions in the .ma module.
|
|
Thanks for reviewing this.","bug"
|
|
"6013","1785","anom fails on non masked arrays Trac 1187).
|
|
Original ticket http projects.scipy.org numpy ticket 1187 on 2009 08 05 by huard, assigned to pierregm.
|
|
Anom is defined as a method of masked arrays, but has no equivalent in the numpy namespace.
|
|
The class np.ma from method which transforms ma methods into functions does not handle this case.","bug"
|
|
"6013","1786","np.ma.average cast result to int when a mask is present Trac 1188).
|
|
Original ticket http projects.scipy.org numpy ticket 1188 on 2009 08 05 by huard, assigned to pierregm.
|
|
Here is an example >>> np.ma.average np.ma.array [1,2])) 1.5 >>> np.ma.average np.ma.array [1,2,3,4], mask=[False, False, True, True])) 1","bug"
|
|
"6013","1787","ndarray.getfield raises ValueError w confusing explanation Trac 1189).
|
|
Original ticket http projects.scipy.org numpy ticket 1189 on 2009 08 09 by trac user dgoldsmith, assigned to unknown.
|
|
Example code x = np.eye 3) x.getfield np.complex) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError Need 0 <= offset <= 8 for requested type but received offset = 0 So the reason ValueError was raised is not because an incorrect value was received as implied by the error message), but rather because no correct value is possible for this particular pairing of array dtype and getfield dtype argument), i.e., the compound inequality which needs to be true has no solution?","bug"
|
|
"6013","1788","[PATCH] Add support for arm sh endian ness detection Trac 1190).
|
|
Original ticket http projects.scipy.org numpy ticket 1190 on 2009 08 09 by trac user ford prefect, assigned to unknown.
|
|
This patch adds fallback endianness checking for ARM SuperH based CPUs.
|
|
The only difference between these and other architectures is that the endian ness is not fixed for the architecture, so a second define needs to be checked.","enhancement"
|
|
"6013","1789","Fix sdist npy pkg config files Trac 1191).
|
|
Original ticket http projects.scipy.org numpy ticket 1191 on 2009 08 10 by cournape, assigned to dmcooke.","bug"
|
|
"6013","1790","integer dot product Trac 1192).
|
|
Original ticket http projects.scipy.org numpy ticket 1192 on 2009 08 11 by trac user jloper, assigned to unknown.
|
|
the dot function is designed to work ONLY WITH FLOATING POINT NUMBERS, but it doesn t raise an exception or even a warning when you feed it integer values.
|
|
This is going to confuse people.
|
|
In [239] a=require randint 0,2,5000),int8) In [240] b=require randint 0,2,5000),int8) In [241] dot a,b); sum a b) Out[241] 82 Out[241] 1198 Let s just say it is a very rare use case that people want sum a b) mod 128.","bug"
|
|
"6013","1791","A faster identity function Trac 1193).
|
|
Original ticket http projects.scipy.org numpy ticket 1193 on 2009 08 12 by kwgoodman, assigned to unknown.
|
|
Using a trick that Robert Kern recently posted to the numpy list makes the identity function much faster.
|
|
Current version def identity n, dtype=None) a = array [1] n [0],dtype=dtype) b = empty n,n),dtype=dtype) b.flat = a return b Proposed version def myidentity n, dtype=None) a = zeros n,n), dtype=dtype) a.flat[ n 1] = 1 return a >> timeit identity 1) 100000 loops, best of 3 14.9 xc2xb5s per loop >> timeit identity 10) 10000 loops, best of 3 20 xc2xb5s per loop >> timeit identity 100) 1000 loops, best of 3 696 xc2xb5s per loop >> timeit identity 1000) 10 loops, best of 3 73.6 ms per loop >> timeit myidentity 1) 100000 loops, best of 3 6.57 xc2xb5s per loop >> timeit myidentity 10) 100000 loops, best of 3 7.08 xc2xb5s per loop >> timeit myidentity 100) 100000 loops, best of 3 16.4 xc2xb5s per loop >> timeit myidentity 1000) 100 loops, best of 3 5.92 ms per loop It would also speed up the functions that use identity for example np.linalg.inv).","enhancement"
|
|
"6013","1792","lnpymath cannot find error during scipy installation Trac 1194).
|
|
Original ticket http projects.scipy.org numpy ticket 1194 on 2009 08 15 by trac user gsever, assigned to cournape.
|
|
An SVN check out and install of numpy with python setupegg.py develop following the similarly for scipy with develop fails due to usr bin ld cannot find lnpymath issue.
|
|
usr bin ld cannot find lnpymath collect2 ld returned 1 exit status error Command usr bin g77 g Wall g Wall shared build temp.linux i686 2.6 scipy special cephesmodule.o build temp.linux i686 2.6 scipy special amos wrappers.o build temp.linux i686 2.6 scipy special specfun wrappers.o build temp.linux i686 2.6 scipy special toms wrappers.o build temp.linux i686 2.6 scipy special cdf wrappers.o build temp.linux i686 2.6 scipy special ufunc extras.o L usr lib python2.6 site packages numpy core lib L usr lib Lbuild temp.linux i686 2.6 lsc amos lsc toms lsc c misc lsc cephes lsc mach lsc cdf lsc specfun lnpymath lm lpython2.6 lg2c o scipy special cephes.so failed with exit status 1 The temporary solution to this problem is to create a symbolic link for npy math.o file under numpy build temp.linux i686 2.6 numpy core src npymath ) into your usr lib directory as libnpymath.so.
|
|
SciPy with python setupegg.py develop should compile fine.","bug"
|
|
"6013","1793","charray.count ) breaks if kwarg end assigned but start isn t Trac 1195).
|
|
Original ticket http projects.scipy.org numpy ticket 1195 on 2009 08 17 by trac user dgoldsmith, assigned to unknown.
|
|
Example >>> c = np.array [ aAaAaA , aA , abBABba ]).view np.chararray) >>> c chararray [ aAaAaA , aA , abBABba ], dtype= |S7 ) >>> c.count aA , end=3) array [3, 1, 0] Problem as reported by Pierre GM) The input of the method aA ,start,end) are checked for consistency.
|
|
However, a problem is that as soon any value is 0 or None or False, the following values are not checked and just discarded that s line 171 in defchararray).
|
|
When you call the `count` method, `start` defaults to None, meaning that the `end` parameter is never taken into account.
|
|
That s clearly a bug, actually 2 bugs `start` should default to 0 in that case you need it in the `count` method of strings) when one of the parameter is not None, it should be taken into account.
|
|
>>> if not chk or chk.dtype is object and chk.item ) is None) should become >>> if chk is None) of chk.dtype is object and chk.item ) is None)","bug"
|
|
"6013","1794","IndexExpression negative index bug.
|
|
Trac 1196).
|
|
Original ticket http projects.scipy.org numpy ticket 1196 on 2009 08 17 by mforbes, assigned to unknown.
|
|
Negative indices do not work properly because the `IndexExpression` class has a ` len )` method that returns the maximum integer >>> np.s [ 1] slice 0, 9223372036854775806, None) >>> a = arange 2) >>> a[ 1] array [0]) >>> a[s [ 1]] array [0, 1]) This can be fixed by having ` len )` return 0, or perhaps better, by removing ` getslice ` and ` len ` altogether as suggested by Robert Kern.
|
|
http www.mail archive.com numpy discussion scipy.org msg18251.html","bug"
|
|
"6013","1795","Memory leak when creating ndarray from list of ndarray subclass instances Trac 1197).
|
|
Original ticket http projects.scipy.org numpy ticket 1197 on 2009 08 17 by trac user chicagofire, assigned to unknown.
|
|
The following code leaks horribly on python v2.6 numpy v1.3.0 import numpy import gc class Array numpy.ndarray) pass if name == main print running arr = [Array shape = 1,1)), Array shape = 1,1))] for i in xrange 10000000000) numpy.array arr) if i 1000 == 0 print collecting gc.collect ) Commenting out PyArray EnsureArray in setArrayFromSequence stops the leak.","bug"
|
|
"6013","1796","Segmentation fault with |O dtype Trac 1198).
|
|
Original ticket http projects.scipy.org numpy ticket 1198 on 2009 08 18 by astrofrog, assigned to unknown.
|
|
The following code causes a segmentation fault when using numpy 1.4.0.dev7282 import numpy as np array = np.zeros 10,), dtype=[ clon , col1 ), |O4 ), err ang , col2 ), <i4 )]) array[ clon ] = np.zeros 10, dtype = np.string ) array[ err ang ] = np.zeros 10, dtype = np.int32) print array The problem occurs when using Python 2.6 64 bit on MacOS X, and the MacOS X traceback is as follows Thread 0 Crashed 0 multiarray.so 0x000000010045eb4a OBJECT copyswapn 154 arraytypes.c.src 1354) 1 multiarray.so 0x000000010049cce8 broadcast cast 552 convert datatype.c 243) 2 multiarray.so 0x00000001004ab889 PyArray CopyObject 89 ctors.c 804) 3 multiarray.so 0x00000001004ad5a4 PyArray SetField 228 methods.c 356) 4 multiarray.so 0x00000001004bcef9 array ass sub 841 mapping.c 794) 5 org.python.python 0x00000001000bc5d8 PyEval EvalFrameEx 10824 6 org.python.python 0x00000001000c0e0b PyEval EvalCodeEx 1483 7 org.python.python 0x00000001000c1206 PyEval EvalCode 54 8 org.python.python 0x00000001000e599e PyRun FileExFlags 174 9 org.python.python 0x00000001000e6781 PyRun SimpleFileExFlags 817 10 org.python.python 0x00000001000f68be Py Main 2766 11 Python 64 0x0000000100000f54 0x100000000 3924","bug"
|
|
"6013","1797","chararray.expandtabs broken Trac 1199).
|
|
Original ticket http projects.scipy.org numpy ticket 1199 on 2009 08 18 by trac user dgoldsmith, assigned to unknown.
|
|
>>> c = np.array [ aAaAaA , \taA\t , abBABba ]).view np.chararray); c chararray [ aAaAaA , \taA , abBABba ], dtype= |S7 ) >>> c.expandtabs ) Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python25\lib\site packages\numpy\core\defchararray.py , line 223, in expandtabs return self.
|
|
generalmethod endswith , broadcast self, tabsize)) File C \Python25\lib\site packages\numpy\core\defchararray.py , line 158, in generalmethod newitem = getattr val[0],name) newval) TypeError endswith ) takes at least 1 argument 0 given) And yet >>> help c.expandtabs) Help on method expandtabs in module numpy.core.defchararray expandtabs self, tabsize=None) method of numpy.core.defchararray.chararray instance So expandtabs isn t supposed to require an argument, and thus must itself generate the argument endswith is looking for, but isn t doing so.","bug"
|
|
"6013","1798","chararray.rstrip inconsistency Trac 1200).
|
|
Original ticket http projects.scipy.org numpy ticket 1200 on 2009 08 19 by trac user dgoldsmith, assigned to unknown.
|
|
>>> c = np.array [ aA ], dtype= S6 ).view np.chararray); c chararray [ aA ], dtype= |S6 ) >>> len c[0]) is A the trailing character... 4 >>> c.rstrip A ) ...or not?
|
|
chararray [ aA ], dtype= |S6 ) Note >>> len c.base[0]); c.base[0] 6 aA Consequent hypothesis rstrip is operating on c.base, not c. >>> len c.base[0]); c.base[0] 6 aA","bug"
|
|
"6013","1799","FAIL test umath.test arctan2 special values 0.0, 3.1415926535897931) [ SOLARIS ] Trac 1201).
|
|
Original ticket http projects.scipy.org numpy ticket 1201 on 2009 08 20 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test umath.test arctan2 special values 0.0, 3.1415926535897931) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 nose case.py , line 183, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 421, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL 0.0 DESIRED 3.1415926535897931 1.4.0.dev7284 august 6) 1.4.0.dev7303 august 19) Python 2.5.1, Solaris 8, cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20","bug"
|
|
"6013","1800","FAIL test umath.test arctan2 special values 0.0, 3.1415926535897931) [ SOLARIS ] Trac 1202).
|
|
Original ticket http projects.scipy.org numpy ticket 1202 on 2009 08 20 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test umath.test arctan2 special values 0.0, 3.1415926535897931) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 nose case.py , line 183, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 421, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL 0.0 DESIRED 3.1415926535897931 1.4.0.dev7284 august 6) 1.4.0.dev7303 august 19) Python 2.5.1, Solaris 8, cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20","bug"
|
|
"6013","1801","test umath.test arctan2 special values 0.0, 0.0) Trac 1203).
|
|
Original ticket http projects.scipy.org numpy ticket 1203 on 2009 08 20 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test umath.test arctan2 special values 0.0, 0.0) Traceback most recent call last) File usr stsci pyssgdev 2.5.1 nose case.py , line 183, in runTest self.test self.arg) File usr ra pyssg 2.5.1 numpy core tests test umath.py , line 235, in assert arctan2 isnzero assert ncu.arctan2 x, y) == 0 and np.signbit ncu.arctan2 x, y)) AssertionError 1.4.0.dev7284 august 6) 1.4.0.dev7303 august 19) Python 2.5.1, Solaris 8, cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20 I put in three tickets for this one the other two are 1799 and 1800) before I noticed that the problem is probably that arctan2 doesn t work.","bug"
|
|
"6013","1802","Test failure new ufuncs on MaskedArrays.
|
|
[ SOLARIS ] Trac 1204).
|
|
Original ticket http projects.scipy.org numpy ticket 1204 on 2009 08 20 by stsci sienkiew, assigned to pierregm.
|
|
====================================================================== FAIL Tests new ufuncs on MaskedArrays.
|
|
Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test core.py , line 1419, in test testUfuncRegression assert mask equal ur.mask, mr.mask) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 237, in assert mask equal assert array equal m1, m2) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 193, in assert array equal header= Arrays are not equal ) File usr stsci pyssgdev 2.5.1 numpy ma testutils.py , line 186, in assert array compare verbose=verbose, header=header) File usr stsci pyssgdev 2.5.1 numpy testing utils.py , line 571, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 25.0 ) x array [False, True, False, True, False, False, False, True], dtype=bool) y array [False, True, False, False, False, False, False, False], dtype=bool) 1.4.0.dev7284 august 6) 1.4.0.dev7303 august 19) Python 2.5.1, Solaris 8, cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20","bug"
|
|
"6013","1803","FAIL test testUfuncRegression test old ma.TestUfuncs) [ SOLARIS ] Trac 1205).
|
|
Original ticket http projects.scipy.org numpy ticket 1205 on 2009 08 20 by stsci sienkiew, assigned to pierregm.
|
|
====================================================================== FAIL test testUfuncRegression test old ma.TestUfuncs) Traceback most recent call last) File usr ra pyssg 2.5.1 numpy ma tests test old ma.py , line 689, in test testUfuncRegression self.failUnless eqmask ur.mask, mr.mask)) AssertionError 1.4.0.dev7284 august 6) 1.4.0.dev7303 august 19) Python 2.5.1, Solaris 8, cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20","bug"
|
|
"6013","1804","rk state.gauss should always be initialized Trac 1206).
|
|
Original ticket http projects.scipy.org numpy ticket 1206 on 2009 08 21 by trac user zouave, assigned to unknown.
|
|
rk state.gauss is left non initialized when rk state.has gauss is set to 0.
|
|
Both has gauss and gauss are returned by RandomState.get state ) and are part of pickled RandomStates.
|
|
Thus, two generators which have the exact same behavior may have different states and pickles at least until a gaussian is generated).
|
|
Suggestion always set gauss to 0. when has gauss is set to 0.","bug"
|
|
"6013","1805","ma.where type handling incorrect Trac 1207).
|
|
Original ticket http projects.scipy.org numpy ticket 1207 on 2009 08 22 by rgommers, assigned to pierregm.
|
|
When both `x` and `y` arguments are not masked arrays, the type is determined by ``np.max type1, type2).
|
|
This does not follow the NumPy casting rules, the solution is to replace ``np.max`` by ``np.find common type``.
|
|
Example >>> x = np.arange 4, dtype=np.int32) >>> y = np.arange 4, dtype=np.float32) 2.2 >>> np.ma.where x > 1.5, y, x).dtype dtype float32 ) >>> np.find common type [np.int32, np.float32], []) dtype float64 )","bug"
|
|
"6013","1806","Implement .real protocol Trac 1208).
|
|
Original ticket http projects.scipy.org numpy ticket 1208 on 2009 08 26 by stefanv, assigned to unknown.
|
|
Considering x = np.array [np.array 3 1j), np.array 4 1j)], dtype=object) both In [4] x.real Out[4] array [ 3 1j), 4 1j)], dtype=object) and In [6] np.real x) Out[6] array [ 3 1j), 4 1j)], dtype=object) does nothing, so that you have to do In [8] [np.real e) for e in x] Out[8] [array 3.0), array 4.0)] or [e.real for e in x].
|
|
We should check for the real method of e.","enhancement"
|
|
"6013","1807","Docstring for numpy.numarray.random array.multinomial is out of date.
|
|
Trac 1209).
|
|
Original ticket http projects.scipy.org numpy ticket 1209 on 2009 08 27 by trac user throwaway, assigned to unknown.
|
|
It appears that the semantics of the multinomial function have changed, and the docstring has not kept pace with it.
|
|
Here is what the docstring says There are len prob) 1 events.
|
|
prob[i] is the probability of the i th event, 0<=i<len prob).
|
|
The probability of event len prob) is 1.
|
|
Numeric.sum prob).
|
|
Contrast that with this behavior >>> from numpy.numarray.random array import multinomial >>> multinomial 10000, [0.5, 0.5]) array [5010, 4990]) It used to be that you d have to leave off the last probability entry.
|
|
It s much more convenient that you don t have to do that anymore, but since it continues to silently accept input in the old form, it s caused me a great deal of confusion.
|
|
Here is the old behavior, for sake of contrast >>> from RandomArray import multinomial >>> multinomial 10000, [0.5,]) array [4982, 5018]) >>> multinomial 10000, [0.5,0.5]) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.5 site packages Numeric RandomArray.py , line 221, in multinomial x = ranlib.multinomial trials, probs.astype Numeric.Float32), Numeric.multiply.reduce shape)) RuntimeError Sum of P i) > 1 in GENMUL The new form still accepts this input >>> from numpy.numarray.random array import multinomial >>> multinomial 10000, [0.5,]) array [10000]) These examples come from numpy version 1.0.4 Ubuntu Hardy), but the same problems seem to apply to version 1.2.1 Ubuntu Jaunty).
|
|
I don t have access to the latest numpy version to check there.
|
|
Apologies for the redundant report, if this problem has already been corrected.
|
|
Incidentally, the reason I moved to the numarray multinomial function is that the one in oldnumeric has a memory leak.
|
|
If this is worth attention, I will be happy to submit a bug report about that, too.","bug"
|
|
"6013","1808","endianness detection incomplete Trac 1210).
|
|
Original ticket http projects.scipy.org numpy ticket 1210 on 2009 08 30 by trac user kumanna, assigned to unknown.
|
|
On systems where endian.h is present, the npy endian.h should use the fact that endian.h is present.
|
|
Unfortunately, I can t find any way to enforce this.
|
|
One suggestion provided in http bugs.debian.org 544291 is to patch npy endian.h as follows .
|
|
numpy core include numpy npy endian.h 2009 04 05 10 09 20.000000000 0200 .
|
|
numpy core include numpy npy endian.h 2009 08 30 17 42 02.000000000 0200 5,8 5,9 NPY BYTE ORDER is set to the same value as BYTE ORDER set by glibc in endian.h include <stddef.h> ifdef NPY HAVE ENDIAN H if defined NPY HAVE ENDIAN H) || defined GLIBC ) Use endian.h if available include <endian.h> define NPY BYTE ORDER BYTE ORDER HTH.
|
|
Thanks.","bug"
|
|
"6013","1809","numpy.linalg.eig crashes on Mandriva OpenSuse for some matrices Trac 1211).
|
|
Original ticket http projects.scipy.org numpy ticket 1211 on 2009 08 30 by pv, assigned to unknown.
|
|
See this discussion http article.gmane.org gmane.comp.python.scientific.user 21936 The attached script crashes with SIGSEGV on on Mandriva 2009.1, 64 bit and OpenSUSE.","bug"
|
|
"6013","1810","genfromtext shall print useful error messages Trac 1212).
|
|
Original ticket http projects.scipy.org numpy ticket 1212 on 2009 09 01 by timmie, assigned to pierregm.
|
|
Trying to read wrong formatted text lead to errors.
|
|
these should be more useful Example The file http www.ncdc.noaa.gov oa climate rcsg cdrom ismcs alphanum.html After having converted this into a text file I tried this np.genfromtxt alphanum to text.txt , dtype=np.str , delimiter= | , skiprows=1, missing= ) But I got an error 1043 dtype = np.dtype ttype) 1044 > 1045 output = np.array data, dtype) 1046 if usemask 1047 if dtype.names ValueError setting an array element with a sequence I sometimes experience this error message with text read in.
|
|
Could this message be made more helpful like telling in which line of the input file this occurs?
|
|
The file is attached.
|
|
The errors occur in the following lines 1635 2851 5538 Note this also applies to loadtxt Diskussion See genfromtext advice http thread.gmane.org gmane.comp.python.numeric.general 32511","enhancement"
|
|
"6013","1811","median in average O n) time Trac 1213).
|
|
Original ticket http projects.scipy.org numpy ticket 1213 on 2009 09 01 by sturlamolden, assigned to charris.
|
|
NumPy s median uses quicksort which is O n log n) on average.
|
|
It is possible to find the median in average O n) time using quickselect instead.","enhancement"
|
|
"6013","1812","np.hist needs to be enabled for masked arrays Trac 1214).
|
|
Original ticket http projects.scipy.org numpy ticket 1214 on 2009 09 01 by timmie, assigned to pierregm.
|
|
In R project Note values is my dataset masking zeros values mask=ifelse values==0, NA, values)) http stat.ethz.ch R manual R patched library stats html ecdf.html plot ecdf values mask), main= Emprical cumulative distribution function , do.points=F, xlab= Data Values ) There masking the zeros has an effect.
|
|
With numpy not.","enhancement"
|
|
"6013","1813","Add MIPS support Trac 1215).
|
|
Original ticket http projects.scipy.org numpy ticket 1215 on 2009 09 02 by r0bertz, assigned to unknown.
|
|
I made a patch to make numpy support MIPS.","enhancement"
|
|
"6013","1814","PyArray INCREF numpy core src multiarray refcount.c Trac 1216).
|
|
Original ticket http projects.scipy.org numpy ticket 1216 on 2009 09 02 by nilswagner01, assigned to unknown.
|
|
>>> numpy.
|
|
version 1.4.0.dev7353 >>> scipy.
|
|
version 0.8.0.dev5838 Getting factors of real matrix ... ok Prefactorize with UMFPACK) matrix for solving with multiple rhs ...
|
|
Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 47346072863120 LWP 4228)] PyArray INCREF mp=0xc) at numpy core src multiarray refcount.c 103 103 if !PyDataType REFCHK mp >descr)) { gdb) bt 0 PyArray INCREF mp=0xc) at numpy core src multiarray refcount.c 103 1 0x00002b0fab9a34df in helper getCArrayObject input=0x1b2c1f0, type=12, minDim=1, maxDim=1) at build src.linux x86 64 2.5 scikits umfpack umfpack wrap.c 2535 2 0x00002b0fab9a4eab in wrap umfpack di defaults self=<value optimized out>, args=<value optimized out>) at build src.linux x86 64 2.5 scikits umfpack umfpack wrap.c 3054","bug"
|
|
"6013","1815","Segfault in lexsort.
|
|
Trac 1217).
|
|
Original ticket http projects.scipy.org numpy ticket 1217 on 2009 09 03 by charris, assigned to unknown.
|
|
It seems to happen when sorting a reversed order array followed by an array with larger bitwise data type.
|
|
Reported by Jeremy Mayes.
|
|
In [1] a = np.array [ ] 100) In [2] b = np.array [1.0] 100) In [3] i = lexsort a,b)) In [4] i = lexsort a[ 1])) In [5] i = lexsort b[ 1])) In [6] i = lexsort a,b[ 1])) In [7] i = lexsort a[ 1],b)) Crash These also work In [3] i = lexsort b[ 1],a)) In [4] i = lexsort b[ 1],b[ 1])) In [5] i = lexsort a[ 1],a[ 1])) In [6] i = lexsort a,b[ 1]))","bug"
|
|
"6013","1816","There should be a way to format individual entries in array2string Trac 1218).
|
|
Original ticket http projects.scipy.org numpy ticket 1218 on 2009 09 06 by inducer, assigned to unknown.
|
|
>>> import numpy >>> a = numpy.array [1,2,3]) >>> numpy.array2string a, style=lambda x str x)) [1 2 3]","enhancement"
|
|
"6013","1817","Inconsistency in axis indexing between eg.
|
|
numpy.sum ) and numpy.ma.sum ) Trac 1219).
|
|
Original ticket http projects.scipy.org numpy ticket 1219 on 2009 09 08 by trac user ckov, assigned to unknown.
|
|
Hi, I ve noticed a weird inconsistency between the masked array versions of some functions and the standard versions.
|
|
For example, if you compare np.sum ) and np.ma.sum ) in the standard version, if you choose to sum over an axis, the axes are 1 based, but in the masked array versions, they are 0 based.
|
|
The 0 based seems the most correct to me, but in any case it seems like they should at least be consistent.
|
|
Any way to resolve this?
|
|
Thanks, Charlie","bug"
|
|
"6013","1818","dotblas.pyd missing in NOSSE numpy 1.3.0 win32 superpack python2.6.exe Trac 1220).
|
|
Original ticket http projects.scipy.org numpy ticket 1220 on 2009 09 11 by trac user sole, assigned to cournape.
|
|
Hello, dotblas.pyd is missing in the windows nosse architecture for python 2.6 It is present for python 2.5.
|
|
Of course, that leads to performance issues under python 2.6.
|
|
See [Numpy discussion] Dot product performance on python 2.6 windows) dated Sep. 11, 2009.
|
|
Best regards, Armando","bug"
|
|
"6013","1819","Can t import universal numpy on MacOS Snow Leopard when running 32 bit python Trac 1221).
|
|
Original ticket http projects.scipy.org numpy ticket 1221 on 2009 09 13 by trac user rwest, assigned to unknown.
|
|
After building 32 64 bit numpy from SVN on MacOS 10.6 Snow Leopard according to the instructions at http blog.hyperjeff.net ?p=160 I would expect to have a universal i386 and x86 64) version of numpy.
|
|
If, however, I start a 32 bit python shell and import numpy then I get a ValueError array is too big.
|
|
on line line 600 of numpy core numerictypes.py These are my steps Installing numpy sudo mv System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy \ System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy APPLE DEFAULT export MACOSX DEPLOYMENT TARGET=10.6 export CFLAGS= arch i386 arch x86 64 export FFLAGS= arch i386 arch x86 64 export LDFLAGS= Wall undefined dynamic lookup bundle arch i386 arch x86 64 cd tmp svn co http svn.scipy.org svn numpy trunk numpy cd numpy python setup.py build fcompiler=gnu95 sudo python setup.py install Loading 32 bit python export VERSIONER PYTHON PREFER 32 BIT=yes python Trying to import numpy Python 2.6.1 r261 67515, Jul 7 2009, 23 51 51) [GCC 4.2.1 Apple Inc. build 5646)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File Library Python 2.6 site packages numpy init .py , line 130, in <module> import add newdocs File Library Python 2.6 site packages numpy add newdocs.py , line 9, in <module> from lib import add newdoc File Library Python 2.6 site packages numpy lib init .py , line 4, in <module> from type check import File Library Python 2.6 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Python 2.6 site packages numpy core init .py , line 8, in <module> import numerictypes as nt File Library Python 2.6 site packages numpy core numerictypes.py , line 600, in <module> typestr[key] = empty 1,),key).dtype.str[1 ] ValueError array is too big.
|
|
This is mentioned at http mail.scipy.org pipermail scipy user 2009 May 021090.html but I have not found a solution.","bug"
|
|
"6013","1820","memory leak in frompyfunc Trac 1222).
|
|
Original ticket http projects.scipy.org numpy ticket 1222 on 2009 09 14 by trac user zegres, assigned to unknown.
|
|
The following code which is an accumulator that resets whenever there is a zero) has a memory leak.
|
|
from numpy import frompyfunc,zeros def customAcc x,y) if y return x y return 0 customAccUfunc=frompyfunc customAcc,2,1) def cumSumReset a) return customAccUfunc.accumulate a).astype float32 ) def killMemory ) print running a1=zeros 10000) for i in range 20) ACCF2=cumSumReset a1) for x in range 200) killMemory )","bug"
|
|
"6013","1821","numpy.random.multivariate normal accepts indefinite covariance matrices Trac 1223).
|
|
Original ticket http projects.scipy.org numpy ticket 1223 on 2009 09 15 by trac user zero79, assigned to unknown.
|
|
numpy.random.multivariate normal will happily accept non physical covariance matrices i.e.
|
|
non symetric and those with det C)<0), which results in erroneous distributions.
|
|
attached is code that demonstrates the problem for a covariance matrix with det C)<0 except when rhoxz=1, which produces a valid covariance matrix).
|
|
!
|
|
usr bin python import numpy , pylab nmcsamples = 2 10 6 xavg = 1.0 yavg = 2.0 zavg = 3.0 xdev = 0.1 ydev = 0.1 zdev = 0.1 rhoxy = 1.0 rhoxz = numpy.arange 0.0 , 1.05 , 0.1 ) rhoyz = 1.0 nsamples = len rhoxz ) xdevmc = numpy.zeros nsamples ) ydevmc = numpy.zeros nsamples ) zdevmc = numpy.zeros nsamples ) rhoxymc = numpy.zeros nsamples ) rhoxzmc = numpy.zeros nsamples ) rhoyzmc = numpy.zeros nsamples ) for n in range 0 , nsamples ) cxx = xdev 2 cyy = ydev 2 czz = zdev 2 cxy = rhoxy xdev ydev cxz = rhoxz[n] xdev zdev cyz = rhoyz ydev zdev c = numpy.array [ [ cxx , cxy , cxz ] , [ cxy , cyy , cyz ] , [ cxz , cyz , czz ] ] ) x,y,z = numpy.transpose numpy.random.multivariate normal [ xavg , yavg , zavg ] , c , nmcsamples ) ) print rhoxz[n],numpy.linalg.det c ) xavgmc = numpy.mean x ) yavgmc = numpy.mean y ) zavgmc = numpy.mean z ) xdevmc[n] = numpy.sqrt numpy.mean x xavgmc ) 2 ) ) ydevmc[n] = numpy.sqrt numpy.mean y yavgmc ) 2 ) ) zdevmc[n] = numpy.sqrt numpy.mean z zavgmc ) 2 ) ) rhoxymc[n] = numpy.mean x y ) xavgmc yavgmc ) xdevmc[n] ydevmc[n] rhoxzmc[n] = numpy.mean x z ) xavgmc zavgmc ) xdevmc[n] zdevmc[n] rhoyzmc[n] = numpy.mean y z ) yavgmc zavgmc ) ydevmc[n] zdevmc[n] pylab.subplots adjust wspace = 0.4 ) pylab.rc font , size = 9 ) pylab.rc legend , fontsize = 9 ) pylab.subplot 2 , 3 , 1 ) pylab.plot rhoxz , numpy.ones nsamples ) xdev , rhoxz , xdevmc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \Delta x ) pylab.axis [ 0.0 , 1.0 , 0.09 , 0.12 ] ) pylab.subplot 2 , 3 , 2 ) pylab.plot rhoxz , numpy.ones nsamples ) ydev , rhoxz , ydevmc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \Delta y ) pylab.axis [ 0.0 , 1.0 , 0.09 , 0.12 ] ) pylab.subplot 2 , 3 , 3 ) pylab.plot rhoxz , numpy.ones nsamples ) zdev , rhoxz , zdevmc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \Delta z ) pylab.legend input , monte carlo ) ) pylab.axis [ 0.0 , 1.0 , 0.09 , 0.12 ] ) pylab.subplot 2 , 3 , 4 ) pylab.plot rhoxz , numpy.ones nsamples ) rhoxy , rhoxz , rhoxymc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \\rho {xy} ) pylab.subplot 2 , 3 , 5 ) pylab.plot rhoxz , rhoxz , rhoxz , rhoxzmc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \\rho {xz} ) pylab.subplot 2 , 3 , 6 ) pylab.plot rhoxz , numpy.ones nsamples ) rhoyz , rhoxz , rhoyzmc ) pylab.xlabel \\rho {xz} ) pylab.ylabel \\rho {yz} ) pylab.savefig multivariate normal problem.pdf )","bug"
|
|
"6013","1822","discrepencies in found tests between numscons and distutils builds Trac 1224).
|
|
Original ticket http projects.scipy.org numpy ticket 1224 on 2009 09 16 by cournape, assigned to dmcooke.","bug"
|
|
"6013","1823","datetime64 and timedelta64 tests Trac 1225).
|
|
Original ticket http projects.scipy.org numpy ticket 1225 on 2009 09 16 by trac user as8ca ..., assigned to unknown.
|
|
Initial version of tests for datetime64 and timedelta64 types.
|
|
We need many more tests.","bug"
|
|
"6013","1824","no silent overflow option Trac 1226).
|
|
Original ticket http projects.scipy.org numpy ticket 1226 on 2009 09 18 by trac user jloper, assigned to unknown.
|
|
> array 2 32,int32) OverflowError > array 2 30,int32) 4 0 oh it would just make me so happy if these two commands COULD have the same output.
|
|
just a little flag somewhere.
|
|
a little i love overflow errors flag.
|
|
Just for me.","enhancement"
|
|
"6013","1825","IsAligned ) should respect the PyArrayDescr Type.alignment field.
|
|
Trac 1227).
|
|
Original ticket http projects.scipy.org numpy ticket 1227 on 2009 09 19 by trac user hansres, assigned to unknown.
|
|
Currently `IsAligned )` always returns `True` for the xe2x80x9cvoidxe2x80x9d dtype, i.e., it assumes that the alignment field is 1.
|
|
This is fine for dtypes, e.g., created with `PyArray DescrConverter )` they always have alignment 1); however those created via `PyArray DescrAlignConverter )` typically donxe2x80x99t.
|
|
Fixing this is particularly useful when unpickling a python array, since it will ensures that the return value is aligned it creates a copy if the data buffer is not aligned).
|
|
I have consulted with Travis regarding this, and he agreed that that we should probably get rid of the void) special case.
|
|
The proposed patch always respects the `alignment` field Simplify improve IsAligned ) Allow PyArray VOID to have arbitrary alignment instead of assuming 1.
|
|
Don t special case PyArray STRING its alignment is 1, which is efficiently handled anyway.","bug"
|
|
"6013","1826","error in parsing fortran comments in crackfortran.py Trac 1228).
|
|
Original ticket http projects.scipy.org numpy ticket 1228 on 2009 09 21 by trac user ohm314, assigned to pearu.
|
|
Hi there, the regular expression `commentline` source trunk numpy f2py crackfortran.py 7268 line 285) matches also !
|
|
inside literal strings.
|
|
example CALL my error function some var, Something went wrong!
|
|
,310,info) This matched the end of the second line as a comment and caused an error.
|
|
I think the regex should be modified to !python commentline=re.compile r ?P<line> [ ] [ ] [ !]
|
|
|[ \ ] \ [ \ ] \ [ \ !]
|
|
|[ !]
|
|
))!
|
|
{1} ?P<rest>.
|
|
[ ] [ ] [ ] |[ \ ] \ [ \ ] \ [ \ ] )) ) best, omar","bug"
|
|
"6013","1827","[Numpy discussion] Numpy large array bug Trac 1229).
|
|
Original ticket http projects.scipy.org numpy ticket 1229 on 2009 09 21 by trac user lucaciti, assigned to unknown.
|
|
Kashyap Ashwin raised this question in [Numpy discussion] Numpy large array bug .
|
|
The unexpected behaviour is the following >>> import numpy as np >>> a=np.zeros 2 1024 1024 1024 10000), dtype= uint8 ) >>> a[ ]=1 >>> print a) array [0, 0, 0, ..., 0, 0, 0], dtype=uint8) also, I noticed that >>> a.put [2 1024 1024 1024 100,], 8) IndexError index out of range for array and suggested by Kashyap) >>> np.putmask a, np.ones like a), 5) Segmentation fault","bug"
|
|
"6013","1828","use explicit integer division in numerictypes Trac 1230).
|
|
Original ticket http projects.scipy.org numpy ticket 1230 on 2009 09 22 by trac user cmattoon, assigned to unknown.
|
|
In numpy core numerictypes.py a bits to bytes conversion is made several times using classic integer division.
|
|
Using python2.6 and the W flag, we get a DeprecationWarning.
|
|
These should be replaced with explicit integer division a b becomes a b 2to3 conversion tool will not make this change.
|
|
Changes on lines 250, 290, 320, 321, 565 of numerictypes.py, where the original comes from git HEAD as of September 19 2009","enhancement"
|
|
"6013","1829","chararray methods ignore all arguments following the first argument that evaluates to False Trac 1231).
|
|
Original ticket http projects.scipy.org numpy ticket 1231 on 2009 09 22 by mdboom, assigned to unknown.
|
|
When chararray builds an argument list to delegate to an underlying string method, it does the following for chk in val[1 ] if not chk or chk.dtype is object and chk.item ) is None) break newval.append chk) Unfortunately, this results in bugs like In [19] a = np.char.array [ 12345d ]) In [20] a.find d , 0, 2) Out[20] array [5]) This should be 1, since find should only be looking in the first two characters).
|
|
I suspect the intention of this code was if chk is None or chk.dtype is object and chk.item ) is None) but it seems like a spurious optimization anyway, and could perhaps just be removed.","bug"
|
|
"6013","1830","is ndarray.base the closest base or the ultimate base?
|
|
Trac 1232).
|
|
Original ticket http projects.scipy.org numpy ticket 1232 on 2009 09 22 by trac user lucaciti, assigned to unknown.
|
|
I raised this question in [Numpy discussion] is ndarray.base the closest base or the ultimate base?
|
|
http article.gmane.org gmane.comp.python.numeric.general 32981 In [1] import numpy as np In [2] x = np.arange 12) In [3] y = x[ 2] In [4] z = y[2 ] ...
|
|
In [8] z.base Out[8] array [ 0, 2, 4, 6, 8, 10]) It looks like the base of z is y , i.e.
|
|
its closest base, the array from which the view z was created.
|
|
But the documentation says base ndarray If the array is a view on another array, that array is its `base` unless that array is also a view).
|
|
The `base` array is where the array data is ultimately stored.
|
|
and it looks like the base should be x , the array where the data is ultimately stored.","enhancement"
|
|
"6013","1831","Deserialized arrays with base mutate strings Trac 1233).
|
|
Original ticket http projects.scipy.org numpy ticket 1233 on 2009 09 23 by trac user hniksic, assigned to mwiebe.
|
|
Numpy arrays with the base property are deserialized as arrays pointing to a storage contained within a Python string.
|
|
This is a problem since such arrays are mutable and can mutate existing strings.
|
|
Here is how to create one >>> import numpy, cPickle as p >>> a = numpy.array [1, 2, 3]) create an array >>> b = a[ 1] create a view >>> b array [3, 2, 1]) >>> b.base view s base is the original array array [1, 2, 3]) >>> c = p.loads p.dumps b, 1)) roundtrip the view through pickle >>> c array [3, 2, 1]) >>> c.base base is now a simple string \x03\x00\x00\x00\x02\x00\x00\x00\x01\x00\x00\x00 >>> s = c.base >>> s \x03\x00\x00\x00\x02\x00\x00\x00\x01\x00\x00\x00 >>> type s) <type str > >>> c[0] = 4 when the array is mutated... >>> s ...the string changes value!
|
|
\x04\x00\x00\x00\x02\x00\x00\x00\x01\x00\x00\x00 This is somewhat disconcerting, as Python strings are supposed to be immutable.
|
|
In this case the string was created by numpy and is probably not shared by anyone, so it doesn t present a problem in practice.
|
|
But in corner cases it can lead to serious bugs.
|
|
Python has a cache of one character strings, which cannot be turned off.
|
|
This means that one byte array views can change existing Python strings used elsewhere in the code.
|
|
For example >>> a = numpy.array [65], int8 ) >>> b = a[ 1] >>> c = p.loads p.dumps b, 1)) >>> c array [65], dtype=int8) >>> c.base A >>> c[0] = 66 >>> c.base B >>> A B Note how changing a numpy array permanently changed the contents of all A strings in this python instance, rendering python unusable.
|
|
The fix should be straightforward use a string subclass which will skip the one letter cache), or an entirely separate type for storage of base memory referenced by deserialized arrays.","bug"
|
|
"6013","1832","axis keyword could take multiple axes for reduction operations Trac 1234).
|
|
Original ticket http projects.scipy.org numpy ticket 1234 on 2009 09 23 by teoliphant, assigned to unknown.
|
|
Sometimes, data is stored in such a way that a particular logical dimension is actually stored along two NumPy dimensions.
|
|
For example, if you have multiple time series stored on disk in time chunks for localized access, you may have a NumPy 3 d array when memory mapping the file where the first and third dimensions are really time.
|
|
Reductions could be done over the third axis i.e.
|
|
mean, max, etc.)
|
|
by passing in a tuple as the axis keyword argument indicating the dimensions to reduce over a.mean axis= 0,2)) indicates to perform the reduction over dimensions 0 and 2.
|
|
This is equivalent to a.mean axis=0).mean axis= 1) in this case.
|
|
For a.std it is not the same.","enhancement"
|
|
"6013","1833","Coercing object arrays to string arrays has surprising behaviour Trac 1235).
|
|
Original ticket http projects.scipy.org numpy ticket 1235 on 2009 09 24 by mdboom, assigned to unknown.
|
|
NOTE I m looking into this not sure how to assign it to myself), but creating a bug to record progress.
|
|
In [14] x = np.array [ abcdefgh , ijklmnop ], O ) Without specifying the length, it seems to default to sizeof int)...
|
|
???
|
|
In [15] np.array x, S ) Out[15] array [ abcd , ijkl ], dtype= |S4 ) In [21] np.array x, np.string ) Out[21] array [ abcd , ijkl ], dtype= |S4 ) Specifying a length gives strange results In [16] np.array x, S8 ) Out[16] array [ abcdijkl , mnop\xe0\x01\x85\x08 ], dtype= |S8 ) This is what I expected to happen above, but the cast to a list seems like it should be unnecessary In [17] np.array list x)) Out[17] array [ abcdefgh , ijklmnop ], dtype= |S8 ) Assignment also seems broken In [18] y = np.empty x.shape, dtype= S8 ) In [19] y[ ] = x[ ] In [20] y Out[20] array [ abcdijkl , mnop\xc05\xf9\xb7 ], dtype= |S8 )","bug"
|
|
"6013","1834","proposal add a header and footer function to numpy.savetxt Trac 1236).
|
|
Original ticket http projects.scipy.org numpy ticket 1236 on 2009 09 24 by timmie, assigned to unknown.
|
|
Hello, sometime scripts and programs create a lot of data output.
|
|
For the programmer and also others not involved in the scripting but in the evaluation of the output it would be very nice the output files could be prepended with a file header describing what is written in the columns below and to append a footer.
|
|
A good example has been developed by the scipy.scikits.timeseries developers http pytseries.sourceforge.net lib.report.html scikits.timeseries.lib.reportlib.Report These formatting flags are a convenient way to save additional meta information.
|
|
E. g. of it is important to state the physical units of the data saved.
|
|
I would be happy if such a thing could be added to np.savetxt ).
|
|
What is the current common way to save a header above the saved ascii array?
|
|
For structured arrays, this could support writing the field names as column headers in the ascii file as default.
|
|
Thanks in advance.","enhancement"
|
|
"6013","1835","np.any and np.all short circuiting Trac 1237).
|
|
Original ticket http projects.scipy.org numpy ticket 1237 on 2009 09 24 by trac user lucaciti, assigned to unknown.
|
|
I noticed that python s any can be faster than numpy s any and the similarly for all ).
|
|
Then I wondered why.
|
|
I realized that numpy implements any as logical or.reduce and all as logical and.reduce).
|
|
This means that numpy cannot take advantage of short circuiting.
|
|
Looking at the timings confirmed my suspects.
|
|
I think python fetches one element at the time from the array and as soon as any of them is true it returns true.
|
|
Instead, numpy goes on until the end of the array even if the very first element is already true.
|
|
Looking at the code I think I found a way to fix it.
|
|
In the following table any x) is python s version, np.any x) is numpy s, while np.any x) is mine.
|
|
1.4.0.dev7417 x = np.zeros 100000, dtype=bool) x[i] = True timeit any x) timeit np.any x) x = np.ones 100000, dtype=bool) x[i] = False timeit all x) timeit np.all x) ANY i any x) np.any x) np.any x) 6.84 ms 831 xc2xb5s 189 xc2xb5s 50000 3.41 ms 832 xc2xb5s 98 xc2xb5s 10000 683 xc2xb5s 831 xc2xb5s 24.7 xc2xb5s 1000 68.9 xc2xb5s 859 xc2xb5s 8.41 xc2xb5s 100 7.92 xc2xb5s 888 xc2xb5s 6.9 xc2xb5s 10 1.42 xc2xb5s 832 xc2xb5s 6.68 xc2xb5s 0 712 ns 831 xc2xb5s 6.65 xc2xb5s ALL i all x) np.all x) np.all x) 6.65 ms 676 xc2xb5s 300 xc2xb5s 50000 3.32 ms 677 xc2xb5s 154 xc2xb5s 10000 666 xc2xb5s 676 xc2xb5s 36.4 xc2xb5s 1000 67.9 xc2xb5s 686 xc2xb5s 9.86 xc2xb5s 100 7.53 xc2xb5s 677 xc2xb5s 7.26 xc2xb5s 10 1.39 xc2xb5s 676 xc2xb5s 7.06 xc2xb5s 0 716 ns 678 xc2xb5s 6.96 xc2xb5s","enhancement"
|
|
"6013","1836","Allow autostrip keyword in genfromtxt Trac 1238).
|
|
Original ticket http projects.scipy.org numpy ticket 1238 on 2009 09 24 by jseabold, assigned to unknown.
|
|
This allows the user to specify whether the whitespace should be stripped from strings in genfromtxt.
|
|
It does not change the default behavior.
|
|
This is how it was handled.
|
|
>>> from StringIO import StringIO >>> import numpy as np >>> s = StringIO 1 4 2003 ,1.3,abcde ) >>> data = np.genfromtxt s, dtype=None, delimiter= , ) >>> data array 1 4 2003 , 1.3, abcde ), dtype=[ f0 , |S10 ), f1 , <f8 ), f2 , |S5 )]) Now you can do >>> s.seek 0) >>> data = np.genfromtxt s, dtype=None, delimiter= , , autostrip=True) >>> data array 1 4 2003 , 1.3, abcde ), dtype=[ f0 , |S8 ), f1 , <f8 ), f2 , |S5 )]) Diff file attached.","enhancement"
|
|
"6013","1837","S0 and U0 typecodes fail with np.dtype ) Trac 1239).
|
|
Original ticket http projects.scipy.org numpy ticket 1239 on 2009 09 24 by dwf, assigned to charris.
|
|
I noticed this inconsistency In [30] dtype str) Out[30] dtype |S0 ) In [31] dtype |S0 ) TypeError Traceback most recent call last) Users dwf <ipython console> in <module> ) TypeError data type not understood In [32] dtype unicode) Out[32] dtype >U0 ) In [33] dtype >U0 ) TypeError Traceback most recent call last) Users dwf <ipython console> in <module> ) TypeError data type not understood This would present problems if you, say, np.save ) d an array with an S0 field and tried to np.load ) it back; dtype ) would choke in the deserialization process.
|
|
The attached patch resolves it in the simplest way I could think of for U and S typecodes, 0 length is okay note that >U0 and <U0 are still different types, technically, despite byte order being meaningless here.","bug"
|
|
"6013","1838","Casting from Unicode to String array ignores exception Trac 1240).
|
|
Original ticket http projects.scipy.org numpy ticket 1240 on 2009 09 25 by mdboom, assigned to unknown.
|
|
Element [0, 1] in the a array below should raise a UnicodeEncodeError exception when casting to a string.
|
|
However, the exception is lost and the resulting array ends up with garbage data.
|
|
The attached patch fixes this.
|
|
There are many other places I found where to setitem error codes were being ignored it s possible these should be fixed, but only this one is exercised by this example.
|
|
In [4] a = np.array [[u abc , u \u03a3 ], [u asdf , u erw ]], dtype= U ) In [5] np.array a, dtype= S4 ) Out[5] array [[ abc , hv\xb1\x08 ], [ asdf , erw ]], dtype= |S4 )","bug"
|
|
"6013","1839","Array constructed with mixture of str and unicode objects fails length detection Trac 1241).
|
|
Original ticket http projects.scipy.org numpy ticket 1241 on 2009 09 25 by mdboom, assigned to unknown.
|
|
There is a special case bit of code in array small type to handle the case when a str obj is followed by a unicode obj and compare their lengths correctly.
|
|
However, the objects can also appear in the reverse order and this case needs to be handled as well see attached patch).
|
|
In [2] np.array [ 1234 , u 123 ]) Out[2] array [u 123 , u 123 ], dtype= <U3 ) In [3] np.array [u 123 , 1234 ]) Out[3] array [u 123 , u 1234 ], dtype= <U4 )","bug"
|
|
"6013","1840","numpy.matlib has no tests Trac 1242).
|
|
Original ticket http projects.scipy.org numpy ticket 1242 on 2009 09 26 by pv, assigned to unknown.
|
|
Numpy.matlib has no tests.
|
|
We should at least check that it s possible to import the module.","bug"
|
|
"6013","1841","numpy.lib.utils.who ) breaks with 0 d arrays Trac 1243).
|
|
Original ticket http projects.scipy.org numpy ticket 1243 on 2009 09 27 by trac user gnutrino, assigned to unknown.
|
|
The function who ) in numpy lib utils.py gives a stak trace if the dictionary it works over contains a ndarray with shape ) i.e.
|
|
a zero dimensional aray).
|
|
For example import numpy as np d = { foo np.array 1)} np.who d) gives ValueError Traceback most recent call last) usr lib python2.6 site packages numpy lib utils.py in who vardict) 274 maxbyte = len val[2]) 275 if val[4] > 276 totalbytes = int val[2]) 277 278 if len sta) > 0 ValueError invalid literal for int ) with base 10 4.0 this is caused by the line bytestr = str var.itemsize product var.shape)) and the fact that product returns 1.0 i.e a float rather than an int) when called with an empty tuple.
|
|
A more sane way to do this would be to use the .nbytes attribute of the ndarray rather than computing this value by hand.
|
|
Attached is a patch which trivially implements this fix.","bug"
|
|
"6013","1842","numpy.linalg.qr crashes with empty arrays Trac 1244).
|
|
Original ticket http projects.scipy.org numpy ticket 1244 on 2009 09 29 by trac user chtito, assigned to unknown.
|
|
Simple test !python import numpy as np a = np.zeros [0,2]) from numpy.linalg import qr qr a) python crashes","bug"
|
|
"6013","1843","saving large matrices through filters fails Trac 1245).
|
|
Original ticket http projects.scipy.org numpy ticket 1245 on 2009 10 02 by trac user jap, assigned to unknown.
|
|
While trying to save a large matrix through gzip, I got the following error >>> import numpy >>> Large Matrix = numpy.zeros 30000, 30000), dtype=numpy.float32) >>> import gzip >>> numpy.save gzip.open tmp matrix.npy.gz , w ), Large Matrix) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.6 dist packages numpy lib io.py , line 178, in save format.write array fid, arr) File usr lib python2.6 dist packages numpy lib format.py , line 326, in write array fp.write array.tostring C )) SystemError Negative size passed to PyString FromStringAndSize Even though this is a python bug, I d like to see a workaround in numpy by writing chunks smaller than 2GB.
|
|
This might also be good for performance of filtered writes.. As I m going to need this, I ll code this up shouldn t change the on disk format), is there any demand for this to be integrated into numpy?","bug"
|
|
"6013","1844","two different numpy.int32 types?
|
|
Trac 1246).
|
|
Original ticket http projects.scipy.org numpy ticket 1246 on 2009 10 02 by trac user zmic, assigned to unknown.
|
|
>>> import numpy as np >>> a = np.int8 1) >>> b = np.int32 2) >>> type a b)==type b a) False >>> print type a b) <type numpy.int32 > >>> print type b a) <type numpy.int32 >","bug"
|
|
"6013","1845","vectorize fails with f2py and keyword functions Trac 1247).
|
|
Original ticket http projects.scipy.org numpy ticket 1247 on 2009 10 02 by sturlamolden, assigned to unknown.
|
|
vectorize infer the number of arguments by calling the function with zero arguments and parsing the error message.
|
|
The regex used by get nargs in numpy lib function base.py is terr = re.compile r .
|
|
?
|
|
takes exactly ?P<exargs>\d ) argument s|) \ ?P<gargs>\d ) given\) ) It does not fit the error message when using keyword arguments >>> def test a, b=1) pass >>> test ) Traceback most recent call last) File <pyshell 6> , line 1, in <module> test ) TypeError test ) takes at least 1 argument 0 given) The problem also affects f2py wrapped Fortran, as the f2py generated wrapper calls PyArg ParseTupleAndKeywords.
|
|
The regex should be re.compile r .
|
|
?
|
|
takes exactly|at least) ?P<exargs>\d ) argument s|) \ ?P<gargs>\d ) given\) )","bug"
|
|
"6013","1846","Sun Studio Fortran libraries Trac 1248).
|
|
Original ticket http projects.scipy.org numpy ticket 1248 on 2009 10 04 by trac user beffa, assigned to dmcooke.
|
|
Building SciPy on a x64 machine with OpenSolaris 2009.06 with Sun Studio 12u1 I get the following error pkgbuild opt SunStudioExpress bin f90 Bdynamic G lCrun lCstd build temp.solaris 2.11 i86pc 2.6 build src.solaris 2.11 i86pc 2.6 scipy fftpack fftpackmodule.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src drfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zrfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zfftnd.o build temp.solaris 2.11 i86pc 2.6 build src.solaris 2.11 i86pc 2.6 fortranobject.o Lbuild temp.solaris 2.11 i86pc 2.6 ldfftpack lpython2.6 lfsu lsunmath lmvec lf77compat o build lib.solaris 2.11 i86pc 2.6 scipy fftpack fftpack.so[[BR]] pkgbuild ld fatal library lf77compat not found[[BR]] pkgbuild ld fatal file processing errors.
|
|
No output written to build lib.solaris 2.11 i86pc 2.6 scipy fftpack fftpack.so[[BR]] pkgbuild ld fatal library lf77compat not found[[BR]] pkgbuild ld fatal file processing errors.
|
|
No output written to build lib.solaris 2.11 i86pc 2.6 scipy fftpack fftpack.so[[BR]] pkgbuild error Command opt SunStudioExpress bin f90 Bdynamic G lCrun lCstd build temp.solaris 2.11 i86pc 2.6 build src.solaris 2.11 i86pc 2.6 scipy fftpack fftpackmodule.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src drfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zrfft.o build temp.solaris 2.11 i86pc 2.6 scipy fftpack src zfftnd.o build temp.solaris 2.11 i86pc 2.6 build src.solaris 2.11 i86pc 2.6 fortranobject.o Lbuild temp.solaris 2.11 i86pc 2.6 ldfftpack lpython2.6 lfsu lsunmath lmvec lf77compat o build lib.solaris 2.11 i86pc 2.6 scipy fftpack fftpack.so failed with exit status 1[[BR]] pkgbuild Bad exit status from var tmp pkgbuild bld pkgbuild tmp 2.18117 build) command output ends finished at Mon Sep 28 21 01 48 BST 2009 The error can be corrected by modifying the sun.py file included in NumPy with the attached patch.
|
|
Essentially removing the f77compat library.
|
|
As far as I know this is a SPARC only library.
|
|
The sun.py file should provide an architecture dependent library list.","bug"
|
|
"6013","1847","polyder does not preserve poly1d object Trac 1249).
|
|
Original ticket http projects.scipy.org numpy ticket 1249 on 2009 10 06 by pbrod, assigned to charris.
|
|
When passing a poly1d object to polyder it returns a ndarray and poly1d object for m=0 and m=1, respectively, as exemplified here >>> pd = poly1d [1,1,1]) >>> polyder pd, m=0) array [1, 1, 1]) >>> pd.deriv m=0) poly1d [1, 1, 1]) >>> polyder pd, m=1) poly1d [2, 1]) In both cases it should return a poly1d object.
|
|
A fix that solves this problem is given here def polyder p, m=1) m = int m) if m < 0 raise ValueError, Order of derivative must be positive see polyint) truepoly = isinstance p, poly1d) p = asarray p) if m == 0 if truepoly return poly1d p) return p else n = len p) 1 ix = arange n, 0, 1) if p.ndim > 1 ix = ix[..., newaxis] y = ix p[ 1] val = polyder y, m 1) if truepoly val = poly1d val) return val","bug"
|
|
"6013","1848","Behavior of np.sign for complex number Trac 1250).
|
|
Original ticket http projects.scipy.org numpy ticket 1250 on 2009 10 06 by huard, assigned to pv.
|
|
For complex arguments, np.sign returns the following > [10] np.sign 1j) <[10] 1 0j) > [11] np.sign 1j) <[11] 1 0j) > [12] np.sign 1 1j) <[12] 1 0j) So the sign of the real part takes precedence on the sign of the complex part when there is a real part.
|
|
I m not sure if this qualifies as a bug, but it s not intuitive, nor documented.","bug"
|
|
"6013","1849","64 Bit Numpy DLL Load failure Trac 1251).
|
|
Original ticket http projects.scipy.org numpy ticket 1251 on 2009 10 07 by trac user altiod, assigned to unknown.
|
|
Hi, I m running windows xp with python 64 2.6 bit and 64 bit numpy version 1.3.0 for python 2.6 Program I am using runs correctly in 32 bit python numpy, but when I use the 64 bit versions I get the following error e \Environ\Python26 64bit\lib\site packages\numpy\core\ init .py 5 Warning Windows 64 bits suppo rt is experimental, and only available for testing.
|
|
You are advised not to use it for production.
|
|
CRASHES ARE TO BE EXPECTED PLEASE REPORT THEM TO NUMPY DEVELOPERS import multiarray Traceback most recent call last) File TestGridWorldChiArm.py , line 1, in <module> import numpy File e \Environ\Python26 64bit\lib\site packages\numpy\ init .py , line 130, in <module> import add newdocs File e \Environ\Python26 64bit\lib\site packages\numpy\add newdocs.py , line 9, in <module> from lib import add newdoc File e \Environ\Python26 64bit\lib\site packages\numpy\lib\ init .py , line 13, in <module> from polynomial import File e \Environ\Python26 64bit\lib\site packages\numpy\lib\polynomial.py , line 18, in <module> from numpy.linalg import eigvals, lstsq File e \Environ\Python26 64bit\lib\site packages\numpy\linalg\ init .py , line 47, in <module> from linalg import File e \Environ\Python26 64bit\lib\site packages\numpy\linalg\linalg.py , line 22, in <module> from numpy.linalg import lapack lite ImportError DLL load failed The specified module could not be found.","bug"
|
|
"6013","1850","genfromtxt dtype enhancement Trac 1252).
|
|
Original ticket http projects.scipy.org numpy ticket 1252 on 2009 10 07 by jseabold, assigned to unknown.
|
|
Following up on the genfromtxt discussion, it would be nice if we could do the following import numpy as np from StringIO import StringIO s = StringIO 1,2,3.0 ) data = np.genfromtxt s, dtype=float, names=[ var1 , var2 , var3 ]) notice though that data.dtype.names is None True s.seek 0) data2 = np.genfromtxt s, dtype= int,float,float), names=[ var1 , var2 , var3 ], delimiter= , )","enhancement"
|
|
"6013","1851","Request for improving the fill value attribute behaviour Trac 1253).
|
|
Original ticket http projects.scipy.org numpy ticket 1253 on 2009 10 07 by trac user gsever, assigned to pierregm.
|
|
Assuming the following case I[39] f = np.arange 5, dtype=float) I[40] f[3] = 999999.9999 I[41] fm = ma.masked values f, 999999.9999) I[42] fm O[42] masked array data = [0.0 1.0 2.0 4.0], mask = [False False False True False], fill value = 999999.9999) I[43] fm[3] O[43] masked array data = , mask = True, fill value = 1e 20) Having two different fill value causing confusion.
|
|
Furthermore doing I[44] fm[3].filled ) O[44] array 1e 20) yields with wrong result while full filling works correctly.
|
|
I[47] fm.filled ) O[47] array [ 0.00000000e 00, 1.00000000e 00, 2.00000000e 00, 1.00000000e 06, 4.00000000e 00]) It would be nice to see this behaviour addressed in a way to eliminate confusions.
|
|
See the full discussion in context at [http www.nabble.com Questions about masked arrays td25776257.html Questions about masked arrays]","enhancement"
|
|
"6013","1852","newbyteorder fails on structured arrays with both names and titles Trac 1254).
|
|
Original ticket http projects.scipy.org numpy ticket 1254 on 2009 10 07 by mdboom, assigned to teoliphant.
|
|
Creating a newbyteorder of a dtype which contains both a name and title returns the <err> dtype.
|
|
Using this dtype to create a new array segfaults.
|
|
>>> import numpy as np >>> x = np.array [ 1, 1)], dtype=[ a , a ), i ), b , i )]) >>> x.dtype.newbyteorder > ) dtype <err>) >>> y = np.array x, dtype=x.dtype.newbyteorder > )) Segmentation fault And here s the backtrace from the segfault [Switching to Thread 1208456544 LWP 6456)] 0x080d5957 in vgetargs1 args=0x0, format=0xb7db54 , p va=0xbfe05bd4, flags=Variable flags is not available. )
|
|
at Python getargs.c 286 286 if !PyTuple Check args)) { gdb) bt 0 0x080d5957 in vgetargs1 args=0x0, format=0xb7db54 , p va=0xbfe05bd4, flags=Variable flags is not available. )
|
|
at Python getargs.c 286 1 0x080d5ec5 in PyArg ParseTuple SizeT args=0xb7db55, format=0xb7db54 ) at Python getargs.c 96 2 0x00b73092 in VOID setitem op=0xb7a9090c, ip=0x8fc8980 , ap=0x8fa80e8) at numpy core src multiarray arraytypes.c.src 620 3 0x00b748ce in VOID to VOID ip=0x8fd4838 \001 , op=0x8fc8980 , n=1, aip=0x8ebacd8, aop=0x8fa80e8) at numpy core src multiarray arraytypes.c.src 1392 4 0x00b53958 in PyArray CastTo out=0x8fa80e8, mp=0x8ebacd8) at numpy core src multiarray convert datatype.c 316 5 0x00b4ae74 in array copy into dest=0x8fa80e8, src=0x8ebacd8, usecopy=1) at numpy core src multiarray ctors.c 855 6 0x00b4be98 in PyArray FromArray arr=0x8ebacd8, newtype=0xb7a982f0, flags=112) at numpy core src multiarray ctors.c 2421 7 0x00b4c3ef in PyArray FromAny op=0x8ebacd8, newtype=0xb7a982f0, min depth=0, max depth=0, flags=112, context=0x0) at numpy core src multiarray ctors.c 1598 8 0x00b6a60a in PyArray CheckFromAny op=0x8ebacd8, descr=0xb7a982f0, min depth=0, max depth=0, requires=112, context=0x0) at numpy core src multiarray ctors.c 1769 9 0x00b75fde in array fromobject NPY UNUSED TAGGEDignored=0x0, args=0xb7f0a2cc, kws=0xb7a91dfc) at numpy core src multiarray multiarraymodule.c 1478 10 0x0805d984 in PyObject Call func=0xb7db50, arg=0xb7f0a2cc, kw=0xb7a91dfc) at Objects abstract.c 1861 11 0x080bf804 in PyEval EvalFrameEx f=0x8e85dc4, throwflag=0) at Python ceval.c 3784 12 0x080c43bb in PyEval EvalCodeEx co=0xb7eabec0, globals=0xb7f5eacc, locals=0xb7f5eacc, args=0x0, argcount=0, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at Python ceval.c 2836 13 0x080c450a in PyEval EvalCode co=0xb7eabec0, globals=0xb7f5eacc, locals=0xb7f5eacc) at Python ceval.c 494 14 0x080e3e52 in PyRun InteractiveOneFlags fp=0xd16720, filename=0x811a058 <stdin> , flags=0xbfe065dc) at Python pythonrun.c 1273 15 0x080e4005 in PyRun InteractiveLoopFlags fp=0xd16720, filename=0x811a058 <stdin> , flags=0xbfe065dc) at Python pythonrun.c 723 16 0x080e4525 in PyRun AnyFileExFlags fp=0xd16720, filename=0x811a058 <stdin> , closeit=0, flags=0xbfe065dc) at Python pythonrun.c 692 17 0x08056d85 in Py Main argc=0, argv=0xbfe06714) at Modules main.c 523 18 0x08056323 in main argc=1, argv=0xbfe06714) at .
|
|
Modules python.c 23","bug"
|
|
"6013","1853","spurious exceptions with new datatypes Trac 1255).
|
|
Original ticket http projects.scipy.org numpy ticket 1255 on 2009 10 08 by trac user tecki, assigned to unknown.
|
|
I wrote a new datatype and registered it with `PyArray RegisterDataType`, and got spurious exceptions somewhere from inside numpy.
|
|
I tracked it down to the function `PyArray DescrConverter`, which calls `PyArray DescrFromType` just after the `finish ` label).
|
|
This function sets a python exception with `PyErr SetString` and returns `NULL`.
|
|
But the calling function ignores this, and returns successfully.
|
|
The exception then is still dangling, and at a later time will be risen in completely unrelated code.
|
|
I guess a `PyErr Clear` should be added at the beginning of the `if` clause to solve the problem.","bug"
|
|
"6013","1854","repr of masked array crashes with Unicode elements Trac 1256).
|
|
Original ticket http projects.scipy.org numpy ticket 1256 on 2009 10 08 by mdboom, assigned to unknown.
|
|
In [7] np.ma.array u Unicode ) Out[7] ERROR An unexpected error occurred while tokenizing input The following traceback may be corrupted or invalid The error message is EOF in multi line statement , 382, 0)) KeyError Traceback most recent call last) wonkabar data1 builds <ipython console> in <module> ) home mdroe usr lib python2.5 site packages IPython Prompts.pyc in call self, arg) 550 551 and now call a possibly user defined print mechanism > 552 manipulated val = self.display arg) 553 554 user display hooks can change the variable to be stored in home mdroe usr lib python2.5 site packages IPython Prompts.pyc in display self, arg) 576 return IPython.generics.result display arg) 577 except TryNext > 578 return self.shell.hooks.result display arg) 579 580 Assign the default display method home mdroe usr lib python2.5 site packages IPython hooks.pyc in call self, args, kw) 139 print prio ,prio, cmd ,cmd dbg 140 try > 141 ret = cmd args, kw) 142 return ret 143 except ipapi.TryNext, exc home mdroe usr lib python2.5 site packages IPython hooks.pyc in result display self, arg) 169 170 if self.rc.pprint > 171 out = pformat arg) 172 if \n in out 173 So that multi line strings line up with the left column of home mdroe usr lib python2.5 pprint.pyc in pformat self, object) 109 def pformat self, object) 110 sio = StringIO ) > 111 self.
|
|
format object, sio, 0, 0, {}, 0) 112 return sio.getvalue ) 113 home mdroe usr lib python2.5 pprint.pyc in format self, object, stream, indent, allowance, context, level) 127 self.
|
|
readable = False 128 return > 129 rep = self.
|
|
repr object, context, level 1) 130 typ = type object) 131 sepLines = len rep) > self.
|
|
width 1 indent allowance) home mdroe usr lib python2.5 pprint.pyc in repr self, object, context, level) 193 def repr self, object, context, level) 194 repr, readable, recursive = self.format object, context.copy ), > 195 self.
|
|
depth, level) 196 if not readable 197 self.
|
|
readable = False home mdroe usr lib python2.5 pprint.pyc in format self, object, context, maxlevels, level) 205 and whether the object represents a recursive construct.
|
|
206 > 207 return safe repr object, context, maxlevels, level) 208 209 home mdroe usr lib python2.5 pprint.pyc in safe repr object, context, maxlevels, level) 290 return format commajoin components), readable, recursive 291 > 292 rep = repr object) 293 return rep, rep and not rep.startswith < )), False 294 home mdroe usr lib python2.5 site packages numpy ma core.pyc in repr self) 3167 parameters = dict name=name, nlen= len name), 3168 data=str self), mask=str self.
|
|
mask), > 3169 fill=str self.fill value), dtype=str self.dtype)) 3170 if self.dtype.names 3171 if n <= 1 home mdroe usr lib python2.5 site packages numpy ma core.pyc in get fill value self) 2956 2957 if self.
|
|
fill value is None > 2958 self.
|
|
fill value = check fill value None, self.dtype) 2959 return self.
|
|
fill value 2960 home mdroe usr lib python2.5 site packages numpy ma core.pyc in check fill value fill value, ndtype) 316 dtype=ndtype,) 317 else > 318 fill value = default fill value ndtype) 319 elif fields 320 fdtype = [ [0], [1]) for in ndtype.descr] home mdroe usr lib python2.5 site packages numpy ma core.pyc in default fill value obj) 165 defval = default filler[obj.subdtype[0].kind] 166 else > 167 defval = default filler[obj.kind] 168 elif isinstance obj, float) 169 defval = default filler[ f ] KeyError U","bug"
|
|
"6013","1855","Failure in genfromtxt tests python 2.4 2.5 only) Trac 1257).
|
|
Original ticket http projects.scipy.org numpy ticket 1257 on 2009 10 08 by pv, assigned to pierregm.
|
|
Buildbots are showing this ====================================================================== FAIL test inconsistent dtype test io.TestFromTxt) Traceback most recent call last) File var lib buildslave numpy real b15 numpy install 2.4 lib python2.4 site packages numpy lib tests test io.py , line 888, in test inconsistent dtype assert raises TypeError, np.genfromtxt, mdata, kwargs) File .. numpy install 2.4 lib python2.4 site packages numpy testing utils.py , line 971, in assert raises AssertionError TypeError not raised Does not occur on python2.6, only on 2.4 and 2.5.","bug"
|
|
"6013","1856","numpy.array does not upconvert from float to store complex types Trac 1258).
|
|
Original ticket http projects.scipy.org numpy ticket 1258 on 2009 10 08 by trac user pkwill, assigned to unknown.
|
|
The documentation to numpy.array ) states of the dtype keyword that, If not given, then the type will be determined as the minimum type required to hold the objects in the sequence.
|
|
This argument can only be used to xe2x80x98upcastxe2x80x99 the array.
|
|
However, if I try to create an array with a complex value while specifying a lower dtype foo = numpy.array [1 1j], dtype=numpy.float) I get an exception ValueError setting an array element with a sequence.
|
|
I think the expected and desired behavior in this case would be to return an ndarray of some complex dtype.
|
|
Note that if I don t specify the dtype foo = numpy.array [1 1j], dtype=None) there s no problem.
|
|
I ran into this because I m trying to use numpy.asfarray ) to coerce arbitrary objects into arrays of at least floating point type, but complex type if that s what the user provides.","bug"
|
|
"6013","1857","byteswapping a complex scalar fails Trac 1259).
|
|
Original ticket http projects.scipy.org numpy ticket 1259 on 2009 10 09 by mdboom, assigned to unknown.
|
|
>>> x = np.array [ 1j], <c8 ) >>> x.tostring ).encode hex ) 00000000000080bf This is a little endian representation, in the order real, imag) When I swap the whole array, it swaps each of the real, imag) parts separately >>> y = x.byteswap ) >>> y.tostring ).encode hex ) 00000000bf800000 and this round trips fine >>> z = np.fromstring y.tostring ), dtype= >c8 ) >>> assert z[0] == 1j >>> When I swap the scalar, it seems to swap the entire 8 bytes >>> y = x[0].byteswap ) >>> y.tostring ).encode hex ) bf80000000000000 ...and this doesn t round trip >>> z = np.fromstring y.tostring ), dtype= >c8 ) >>> assert z[0] == 1j Traceback most recent call last) File <stdin> , line 1, in <module> AssertionError >>> Travis O. suggested a fix on the mailing list here http www.nabble.com byteswapping a complex scalar td25788857.html","bug"
|
|
"6013","1858","Use FFT in np.correlate convolve?
|
|
Trac 1260).
|
|
Original ticket http projects.scipy.org numpy ticket 1260 on 2009 10 12 by trac user roger, assigned to unknown.
|
|
The `convolve` and `correlate` functions appear to be much slower than their MATLAB 2009a equivalents.
|
|
The MATLAB command `xcorr randn 1e6,1))` takes about 0.35s to execute, while the Python equivalent `x = randn 1e6);correlate x, x)` takes more than a minute then killed).
|
|
MATLAB is also much faster for arrays of 1e5 elements.
|
|
`fftconvolve` in `scipy.signal` is even slower.
|
|
Tested with Python 2.6.3 and numpy 1.30 x86) under x64 Windows 7 and x64 Ubuntu on an i7.","enhancement"
|
|
"6013","1859","Remove D flags when compiling on AIX Trac 1261).
|
|
Original ticket http projects.scipy.org numpy ticket 1261 on 2009 10 13 by trac user dpeterson, assigned to rgommers.
|
|
Sent to me via e mail by Burkhard Neinhues I ve been trying to compile numpy and scipy on HP UX.
|
|
I succeeded in the end, but some patches were required.
|
|
I ve also attached the results for numpy.test ) to this email.
|
|
Please find below the steps required to compile them, if you require the patches as separate files, please let me know.
|
|
NUMPY to have native support for lapack and blas export NUMPY FCONFIG= config fc fcompiler=hpux export LAPACK= opt mlib lib pa20 64 liblapack.sl export BLAS= opt mlib lib pa20 64 libveclib.sl I had to add several missing information to this file.
|
|
Is there any option to configure for 32 or 64 bits support?
|
|
I ve now hard wired 64 bits The file `parisc numpy.patch` contains the patches.
|
|
There are also some notes as comments) in the file about why the various patches were done.
|
|
Note that some of the changes support builds on additional platforms such as Solaris and AIX.
|
|
The file `parisc test.out` is the test output Burkhard referred too.","bug"
|
|
"6013","1860","genfromtxt dtype should be None by default Trac 1262).
|
|
Original ticket http projects.scipy.org numpy ticket 1262 on 2009 10 15 by dalloliogm, assigned to unknown.
|
|
If I understand correctly, genfromtxt should be the standard function to read csv files now.
|
|
The problem is that by default it tries to read all the fields of the file as if they were floats because dtype=float by default).
|
|
So every time a newbie tries to read a csv file containing a column with strings, he obtains an array with many misterious NaN fields, unless he understand he has to set dtype to None and this is not documented anywhere except for some mailing list messages).
|
|
Yesterday I gave an introductory talk to pylab and everybody was complaining about this, it is really weird to have to set this parameter manually.
|
|
For example, you can see that all the R s read.delim csv etc.. infer the data type from the files, without having to specify it explicitely.
|
|
So, please change the default value to None.","bug"
|
|
"6013","1861","Strange numpy.float96 behaviour Trac 1263).
|
|
Original ticket http projects.scipy.org numpy ticket 1263 on 2009 10 15 by trac user bgbg, assigned to unknown.
|
|
This is what I see on my windows computer >>> import sys >>> import numpy as np >>> print sys.version 2.5.4 r254 67916, Dec 23 2008, 15 10 54) [MSC v.1310 32 bit Intel)] >>> print np.
|
|
version 1.2.1 >>> print np.float 4.3) 4.3 >>> print np.float32 4.3) 4.3 >>> print np.float96 4.3) 2.03246817962e 262 >>> make sure the problem isn t in string representation ... res = np.float96 4.3) 1 >>> print res 2.72509429761e 108 >>> Expected behaviour >>> print np.float96 4.3) 3.4 On the same machine, cygwin version of python and numpy note the same version of numpy) >>> import sys >>> import numpy as np >>> print sys.version 2.5.2 r252 60911, Dec 2 2008, 09 26 14) [GCC 3.4.4 cygming special, gdc 0.12, using dmd 0.125)] >>> print np.
|
|
version 1.2.1 >>> print np.float 4.3) 4.3 >>> print np.float32 4.3) 4.3 >>> print np.float96 4.3) 4.3 >>> make sure the problem isn t in string representation ... res = np.float96 4.3) 1 >>> print res 5.3","bug"
|
|
"6013","1862","numpy.rec.fromrecords on an empty list Trac 1264).
|
|
Original ticket http projects.scipy.org numpy ticket 1264 on 2009 10 16 by dalloliogm, assigned to charris.
|
|
numpy s fromrecords raises an IndexError list out of range when applied to an empty list.
|
|
I think the expected behaviour should be to generate an empty array, as happens for the array ) function.
|
|
>>> import numpy >>> numpy.rec.fromrecords []) Traceback IndexError list index out of range","bug"
|
|
"6013","1863","64bit System take raises TypeError when using int64 indices Trac 1265).
|
|
Original ticket http projects.scipy.org numpy ticket 1265 on 2009 10 18 by trac user crassus, assigned to unknown.
|
|
I m running the scipy package within a VM for sage http sagemath.org numpy.version.version 1.3.0 CPU Intel i7 Code to produce the bug import numpy as N dat= N.arange 10) idxs= N.array [3,1,4,1,5,9], dtype=N.int64) print dat[idxs] works fine print dat.take N.array idxs, dtype=N.int32)) work fine, too print dat.take idxs) raises ... TypeError array cannot be safely cast to required type So take checks whether the indices can be converted to some int32.
|
|
This is obviously not always possible.
|
|
But indexing using square braces does not check it.
|
|
Furthermore on a 64bit system one should be able to use int64 as indices.
|
|
So arr.take idxs) should have the same behaviour as arr[idxs]","bug"
|
|
"6013","1864","Extremely long runtimes in numpy.fft.fft Trac 1266).
|
|
Original ticket http projects.scipy.org numpy ticket 1266 on 2009 10 19 by trac user koehler, assigned to unknown.
|
|
Although the documenttation of numpy.fft.fft states that This is most efficient for n a power of two.
|
|
This also stores a cache of working memory for different sizes of fft s, so you could theoretically run into memory problems if you call this too many times with too many different n s. ,I think that it may be important to report this oddity in the fft runtime.
|
|
Dependent on the array length, the fft runtime varies really extreme [ipython shell, from numpy import ] In [1] time fft.fft zeros 119516)) CPU times user 22.83 s, sys 0.39 s, total 23.23 s Wall time 23.53 s In [3] time fft.fft zeros 119517)) CPU times user 36.33 s, sys 0.08 s, total 36.40 s Wall time 36.51 s In [5] time fft.fft zeros 119518)) CPU times user 4.88 s, sys 0.08 s, total 4.96 s Wall time 5.02 s In [7] time fft.fft zeros 119519)) CPU times user 0.45 s, sys 0.00 s, total 0.45 s Wall time 0.45 s In [9] time fft.fft zeros 119515)) CPU times user 0.07 s, sys 0.00 s, total 0.08 s Wall time 0.08 s In [11] time fft.fft zeros 119514)) CPU times user 15.84 s, sys 0.06 s, total 15.90 s Wall time 15.95 s In [13] time fft.fft zeros 119513)) CPU times user 272.75 s, sys 1.03 s, total 273.78 s Wall time 275.63 s","bug"
|
|
"6013","1865","Unaligned object pointers are not handled correctly on platforms where alignment matters Trac 1267).
|
|
Original ticket http projects.scipy.org numpy ticket 1267 on 2009 10 19 by mdboom, assigned to unknown.
|
|
Creating arrays containing unaligned PyObject s crashes with bus errors on Solaris.
|
|
For example, this is enough np.zeros 10,), dtype= O,c ) For more discussion of this issue, see this mailing list thread http www.nabble.com object array alignment issues tt25912008.html a25912008 Much of that discussion is about performance, but the purpose of this bug is just basic correctness on alignment requiring platforms.
|
|
I will be attaching a largish) patch to address this for review.","bug"
|
|
"6013","1866","scalar does not inherit byteorder from the array it s created from Trac 1268).
|
|
Original ticket http projects.scipy.org numpy ticket 1268 on 2009 10 20 by mdboom, assigned to unknown.
|
|
In [9] x = np.array [1], dtype= <i4 ) In [10] x[0].dtype.byteorder Out[10] = In [11] x = np.array [1], dtype= >i4 ) In [12] x[0].dtype.byteorder Out[12] = This may in fact be correct behavior, but it can lead to confusing use of byteswap, since the following is not guaranteed x.byteswap ).tostring ) == x[0].byteswap ).tostring ) and for that matter, neither is x.tostring ) == x[0].tostring ) This bug began as part of 1857, but is probably a different bug.)","bug"
|
|
"6013","1867","refcount decremented with record scalar Trac 1269).
|
|
Original ticket http projects.scipy.org numpy ticket 1269 on 2009 10 21 by trac user batripler, assigned to unknown.
|
|
The following function reduces the refcount of None by one) every time it is run.
|
|
If the dtype is expanded to include more objects, the refcount will be decremented by one for each additional object field added to the record.
|
|
import sys,numpy def numpybug ) print sys.getrefcount None) dt=numpy.dtype [ a ,object), b ,object)]) sc=numpy.empty 1, dt)[0]; a=numpy.empty 1, dt) a[ ]=sc del sc, a print sys.getrefcount None)","bug"
|
|
"6013","1868","Array support for binary repr and base repr Trac 1270).
|
|
Original ticket http projects.scipy.org numpy ticket 1270 on 2009 10 22 by trac user mproeller, assigned to unknown.
|
|
The input values of binary repr and base repr can t be numpy arrays.
|
|
It would be nice to have the possibility to use arrays as input values as well, for instance import numpy as np a = np.arange 10) b = np.binary repr a) < not supported A work around example def binary repr arr, width=None) binary list = map lambda foo np.binary repr foo, width)), arr.flatten )) str len max = len np.binary repr arr.max ), width=width)) str len min = len np.binary repr arr.min ), width=width)) if str len max > str len min str len = str len max else str len = str len min binary array = np.fromiter binary list, dtype= |S str str len)) return binary array.reshape arr.shape)","enhancement"
|
|
"6013","1869","Bug in pow for negative exponents of complex zero Trac 1271).
|
|
Original ticket http projects.scipy.org numpy ticket 1271 on 2009 10 23 by trac user jstenar, assigned to unknown.
|
|
pow gives the wrong result for taking 0j to negative powers.
|
|
Except when the exponent is 1.
|
|
I have seen this both on python 2.6 with numpy 1.3 and python 2.4 and numpy 1.2.1 Example In [1] from numpy import In [2] array 0j) 1 Out[2] nan nan j) In [3] array 0j) 0.5 Out[3] 0j In [4] array 0) 0.5 Out[4] inf In [5] array 0j) 2 Out[5] 0j In [6] array 0j) 2.4 Out[6] 0j In [7] array 0j) 0.5 0j) Out[7] 0j In [8] numpy.
|
|
version Out[8] 1.3.0 In [9] import sys;sys.version Out[9] 2.6.2 r262 71605, Apr 14 2009, 22 40 02) [MSC v.1500 32 bit Intel)]","bug"
|
|
"6013","1870","Building Numpy as a 4 way Mac binary from source Trac 1272).
|
|
Original ticket http projects.scipy.org numpy ticket 1272 on 2009 10 23 by trac user brian, assigned to unknown.
|
|
Hi, I m unable to build numpy 1.3.0 as a 4 way binary on a 10.5 Mac i386 using gcc 4.0.1.
|
|
I receive the following error message when arch ppc64 is included as a compiler flag.
|
|
numpy core include numpy npy endian.h 33 10 error error Unknown CPU can not set endianness I m able to build successfully after editing npy endian.h Including defined NPY CPU PPC64) to line 28 seems to do the trick.","bug"
|
|
"6013","1871","invalid npymath configuration, when installed as an egg Trac 1273).
|
|
Original ticket http projects.scipy.org numpy ticket 1273 on 2009 10 24 by pv, assigned to dmcooke.
|
|
Npymath library location is reported wrong when the package is installed as an egg site packages ls d numpy numpy 1.4.0.dev7536 0a6cc5 py2.6 linux i686.egg like so >>> from numpy.distutils.misc util import get numpy include dirs, get info >>> get info npymath ) { define macros [], libraries [ npymath , m ], library dirs [ home pauli koodi proj scipy numpy dist linux lib python2.6 site packages numpy core lib ], include dirs [ home pauli koodi proj scipy numpy dist linux lib python2.6 site packages numpy core include ]} >>> import numpy >>> numpy.
|
|
version 1.4.0.dev7536 0a6cc5 Also, the libnpymath.a static library doesn t seem to be included in the egg.","bug"
|
|
"6013","1872","remove objects from constants.py Trac 1274).
|
|
Original ticket http projects.scipy.org numpy ticket 1274 on 2009 10 24 by rgommers, assigned to pv.
|
|
These objects line, lines, etc) leak through and show up in the doc wiki.
|
|
They should not.
|
|
Fix is a one liner, patch attached.","bug"
|
|
"6013","1873","memory corruption by ndarray instance.data assignment data should be made readonly) Trac 1275).
|
|
Original ticket http projects.scipy.org numpy ticket 1275 on 2009 10 25 by trac user mcieslik, assigned to unknown.
|
|
It s probably not a bug, but I d like to ask for clarification.
|
|
I try to make two arrays share memory.
|
|
I create big array from small arrays this involves a copy e.g.
|
|
by vstack).
|
|
Create slice views of the big array corresponding to the small arrays and do a array.data assignment i.e.
|
|
small array.data = big array[some slice].data This works, both arrays can now modify the same memory, but is not safe as it can lead to memory corruption see below).
|
|
My first question is this considered very very bad practice and if so how to point an ndarray instance to a new memory safely.
|
|
In [1] import numpy as np In [2] zeros = np.zeros 2,2)) In [3] double zeros = np.vstack zeros, zeros)) In [4] double zeros Out[4] array [[ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.]])
|
|
In [5] zeros.data = double zeros[0 3].data In [6] double double zeros = np.vstack double zeros,double zeros)) In [7] double double zeros Out[7] array [[ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.
|
|
], [ 0., 0.]])
|
|
In [8] double zeros.data = double double zeros[0 5].data In [9] zeros Out[9] array [[ 1.25794259e 312, 9.42758674e 317], [ 1.58101007e 322, 2.37151510e 322]]) In [10] zeros[0][0] = 1.1 In [11] zeros Out[11] array [[ 9.88131292e 324, 3.95252517e 323], [ 4.94065646e 324, 2.42092166e 322]]) In [12] zeros[0][0] = 2 In [13] glibc detected usr bin python corrupted double linked list 0x000000000117dab0 this kills the interpreter)","bug"
|
|
"6013","1874","comparing string and unicode arrays broken Trac 1276).
|
|
Original ticket http projects.scipy.org numpy ticket 1276 on 2009 10 27 by mdboom, assigned to unknown.
|
|
In [1] import numpy as np In [2] x = np.array [ abc , def ], np.string ) In [3] y = np.array [ abc , def ], np.unicode ) In [4] x == y Out[4] array [ True, True], dtype=bool) In [5] y == x Out[5] array [ True, True], dtype=bool) In [6] x = np.array [ abcd , def ], np.string ) In [7] x == y TypeError Traceback most recent call last) wonkabar data1 builds numpy tools c coverage <ipython console> in <module> ) TypeError array cannot be safely cast to required type In [8] y == x TypeError Traceback most recent call last) wonkabar data1 builds numpy tools c coverage <ipython console> in <module> ) TypeError array cannot be safely cast to required type If string richcompare is given a mixed pair of string and unicode arrays, it first casts the string array to a unicode array before performing the comparison.
|
|
Unfortunately, this cast doesn t always work.
|
|
Let s say we are comparing string array s and unicode array u .
|
|
s will be converted to unicode, but with the same elsize as u .
|
|
If the elsize of u is not big enough to hold s , an array cannot be safely cast to required type exception is raised.
|
|
The solution is to, rather than using the type of u verbatim, to copy it, and then set elsize to 4 s >descr >elsize, so it will always be big enough to hold a unicode version of s .
|
|
Patch forthcoming).","bug"
|
|
"6013","1875","Constructing real arrays with complex numbers leads to funny errors Trac 1277).
|
|
Original ticket http projects.scipy.org numpy ticket 1277 on 2009 10 28 by inducer, assigned to unknown.
|
|
>>> numpy.array [1j], dtype=numpy.float64) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError setting an array element with a sequence.","bug"
|
|
"6013","1876","Assigning complex data to real array loses imaginary part Trac 1278).
|
|
Original ticket http projects.scipy.org numpy ticket 1278 on 2009 10 28 by inducer, assigned to unknown.
|
|
>>> x = numpy.array [1], dtype=numpy.float64) >>> x[0] = numpy.complex128 1j) >>> x[0] 0.0","bug"
|
|
"6013","1877","Memory leak in array list of matrices) Trac 1279).
|
|
Original ticket http projects.scipy.org numpy ticket 1279 on 2009 10 31 by pv, assigned to unknown.
|
|
The following code leaks memory Numpy 1.4.0.dev7592) import numpy as np print np.
|
|
version foo = [] for i in xrange 200) foo.append np.matrix [1, 1])) while True Leaks memory np.array foo)","bug"
|
|
"6013","1878","doc standard for modules Trac 1280).
|
|
Original ticket http projects.scipy.org numpy ticket 1280 on 2009 11 02 by rgommers, assigned to pv.
|
|
Add a description of how to document modules to HOWTO DOCUMENT.txt.","enhancement"
|
|
"6013","1879","fread failure causes unintelligible error message Trac 1281).
|
|
Original ticket http projects.scipy.org numpy ticket 1281 on 2009 11 03 by dwf, assigned to unknown.
|
|
I don t really know why this is happening I ve seen this 2028 value before, in an fwrite failure, on this same Ubuntu 6.06 version) but suffice it to say this error should probably be caught and an error message reflecting the true nature of the situation issued.
|
|
In [2] x = num.load hActs.npy ) 1451583488 items requested but only 2028 read ValueError Traceback most recent call last) ais gobi1 gdahl speech babble filtDatadebug <ipython console> in <module> ) nobackup murray bin pylab 02 x86 64 GenuineIntel 6.06 Python 2.5.4 lib python2.5 site packages numpy lib io.pyc in load file, mmap mode) 193 return format.open memmap file, mode=mmap mode) 194 else > 195 return format.read array fid) 196 else Try a pickle 197 try nobackup murray bin pylab 02 x86 64 GenuineIntel 6.06 Python 2.5.4 lib python2.5 site packages numpy lib format.pyc in read array fp) 373 374 if fortran order > 375 array.shape = shape[ 1] 376 array = array.transpose ) 377 else ValueError total size of new array must be unchanged As you can see the computation continues until it tries to assign array.shape.
|
|
Ideally, the fact that fread ) failed to return the correct number of items should raise an exception immediately or as soon as possible from the Python code, if not from the C extension code).","bug"
|
|
"6013","1880","non consequent dtype object identity Trac 1282).
|
|
Original ticket http projects.scipy.org numpy ticket 1282 on 2009 11 04 by trac user mcieslik, assigned to unknown.
|
|
dtype object identity is different for different strings used to define the dtype.
|
|
In [46] import numpy as np In [47] s1 1 = np.dtype S1 ) In [48] s1 2 = np.dtype S1 ) In [49] i 1 = np.dtype int32 ) In [50] i 2 = np.dtype int32 ) In [51] s1 1 is s1 2 Out[51] False In [52] i 1 is i 2 Out[52] True","bug"
|
|
"6013","1881","TypeError when using rec.fromrecords with an object field Trac 1283).
|
|
Original ticket http projects.scipy.org numpy ticket 1283 on 2009 11 04 by astrofrog, assigned to unknown.
|
|
The following code np.rec.fromrecords [ 1, hello ), 2, world )],dtype=[ a ,np.int8), b ,np.object )]) raises a TypeError see transcript attached)","bug"
|
|
"6013","1882","ImportError No module named multiarray when moving cPickle files between Trac 1284).
|
|
Original ticket http projects.scipy.org numpy ticket 1284 on 2009 11 04 by trac user reckoner, assigned to unknown.
|
|
On winxp 32 bit Python 2.5.2 r252 60911, Feb 21 2008, 13 11 45) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import cPickle >>> a = numpy.eye 10) >>> cPickle.dump a,open from32bitxp.pkl , w )) >>> import numpy.core.multiarray >>> numpy.
|
|
version 1.0.4 >>> On linux 64 bit Python 2.5.4 r254 67916, Feb 5 2009, 19 52 35) [GCC 4.1.2 20071124 Red Hat 4.1.2 42)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import cPickle >>> cPickle.load open from32bitxp.pkl )) Traceback most recent call last) File <stdin> , line 1, in <module> ImportError No module named multiarray >>> numpy.
|
|
version 1.2.1 >>> import numpy.core.multiarray >>> Note that I transfer the from32bitxp file from the winxp32 machine to the linux host.
|
|
Also, I ve tried this with version 1.3 on winxp and get the same problem on the linux host.
|
|
Here s more interesting info On linux >>> a = numpy.eye 10) >>> cPickle.dump a,open from64bitLinux.pkl , w )) upon transferring the file to winxp 32 and on winxp32 >>> cPickle.load open from64bitLinux.pkl )) See?
|
|
No problem going from linux to winxp32; but problems going the other way.
|
|
Please let me know if you need more info on this.
|
|
Any help appreciated.","bug"
|
|
"6013","1883","Missing semicolon in cfuncs.py causes compile errors in f2py with quad precision.
|
|
Trac 1285).
|
|
Original ticket http projects.scipy.org numpy ticket 1285 on 2009 11 05 by trac user cgranade, assigned to unknown.
|
|
When compiling Fortran code that uses quadruple precision with the f2py tool, I get the following errors tmp tmpZsJCOP src.linux i686 2.6 tiswe integratemodule.c In function xe2x80x98long double from pyobjxe2x80x99 tmp tmpZsJCOP src.linux i686 2.6 tiswe integratemodule.c 149 error expected xe2x80x98;xe2x80x99 before xe2x80x98returnxe2x80x99 Tracking down the problem, there is a single semi colon missing in cfuncs.py that causes the error \t\t\t v) = longdouble )PyArray DATA obj))","bug"
|
|
"6013","1884","numpy.min ) and max ) are confusing vs. Python behavior Trac 1286).
|
|
Original ticket http projects.scipy.org numpy ticket 1286 on 2009 11 06 by trac user wkornewald, assigned to unknown.
|
|
I had to search a very long time to find this bug.
|
|
When importing from numpy import max, min and calling max 2, 3) which would work in Python) the numpy version returns 2 always the first argument) without complaining that I didn t pass an array.
|
|
This causes a lot of bugs also for my colleague who thought, too, that min and max would work like Python s min and max functions).
|
|
Please fix this, so it either throws an error or falls back to Python behavior.
|
|
The latter would be better, of course.","bug"
|
|
"6013","1885","uint64 formatting does not work for large numbers Trac 1287).
|
|
Original ticket http projects.scipy.org numpy ticket 1287 on 2009 11 06 by astrofrog, assigned to unknown.
|
|
Formatting of large uint64 numbers does not work it s necessary to wrap the numpy number inside long ) before formatting In [3] 40i np.uint64 2 64 1) Out[3] 1 In [4] 40i long np.uint64 2 64 1)) Out[4] 18446744073709551615 So it looks like somewhere, numpy is converting the uint64 number to a python int ) instead of long ) which only works for numbers < 2 63 1.","bug"
|
|
"6013","1886","Add a bits attribute to np.finfo ) Trac 1288).
|
|
Original ticket http projects.scipy.org numpy ticket 1288 on 2009 11 06 by astrofrog, assigned to unknown.
|
|
The number of bits occupied by an integer type can be found from In [6] np.iinfo np.int16).bits Out[6] 16 but this is not currently possible for floats In [7] np.finfo np.float32).bits AttributeError Traceback most recent call last) Users tom <ipython console> in <module> ) AttributeError finfo object has no attribute bits Would it be possible to implement this?","enhancement"
|
|
"6013","1887","Solaris numpy core src npymath ieee754.c , line 172 error Needs nextafterl implementation for this platform Trac 1289).
|
|
Original ticket http projects.scipy.org numpy ticket 1289 on 2009 11 09 by stsci sienkiew, assigned to unknown.
|
|
compile options Inumpy core include Ibuild src.solaris 2.8 sun4u 2.5 numpy core include numpy Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr stsci pyssgdev Python 2.5.4 include python2.5 Ibuild src.solaris 2.8 sun4u 2.5 numpy core src multiarray Ibuild src.solaris 2.8 sun4u 2.5 numpy core src umath c cc numpy core src npymath ieee754.c numpy core src npymath ieee754.c , line 172 error Needs nextafterl implementation for this platform cc acomp failed for numpy core src npymath ieee754.c numpy core src npymath ieee754.c , line 172 error Needs nextafterl implementation for this platform cc acomp failed for numpy core src npymath ieee754.c error Command cc DNDEBUG I usr stsci pyssgdev Python 2.5.4 include xcode=pic32 Inumpy core include Ibuild src.solaris 2.8 sun4u 2.5 numpy core include numpy Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr stsci pyssgdev Python 2.5.4 include python2.5 Ibuild src.solaris 2.8 sun4u 2.5 numpy core src multiarray Ibuild src.solaris 2.8 sun4u 2.5 numpy core src umath c numpy core src npymath ieee754.c o build temp.solaris 2.8 sun4u 2.5 numpy core src npymath ieee754.o failed with exit status 2 Python 2.5.4 cc Sun WorkShop 6 update 2 C 5.3 Patch 111679 14 2004 02 20 numpy trunk rev 7699","bug"
|
|
"6013","1888","ERROR test simple2 test utils.TestArrayAlmostEqualNulp) bad format string) Trac 1290).
|
|
Original ticket http projects.scipy.org numpy ticket 1290 on 2009 11 09 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== ERROR test simple2 test utils.TestArrayAlmostEqualNulp) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy testing tests test utils.py , line 317, in test simple2 self.failUnlessRaises AssertionError, failure) File usr stsci pyssgdev Python 2.5.4 lib python2.5 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) File usr stsci pyssgdev 2.5.4 numpy testing tests test utils.py , line 316, in failure nulp=1000) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 1112, in assert array almost equal nulp max is d) nulp, max nulp)) TypeError int argument required Format string does not match args to be printed.
|
|
numpy trunk rev 7699 Python 2.5.4 seen on Red Hat Enterprise 4, 32 bit Mac OSX Leopard, 32 bit not seen on Red Hat Enterprise 4, 64 bit","bug"
|
|
"6013","1889","ERROR test nan test utils.TestULP) Trac 1291).
|
|
Original ticket http projects.scipy.org numpy ticket 1291 on 2009 11 09 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== ERROR test nan test utils.TestULP) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy testing tests test utils.py , line 409, in test nan lambda assert array max ulp nan, inf, File usr stsci pyssgdev Python 2.5.4 lib python2.5 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) File usr stsci pyssgdev 2.5.4 numpy testing tests test utils.py , line 410, in <lambda> maxulp=maxulp)) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 1121, in assert array max ulp maxulp) TypeError int argument required numpy trunk rev 7699 Python 2.5.4 seen on Red Hat Enterprise 4, 32 bit Mac OSX Leopard, 32 bit not seen on Red Hat Enterprise 4, 64 bit","bug"
|
|
"6013","1890","64 bit numpy problem Trac 1292).
|
|
Original ticket http projects.scipy.org numpy ticket 1292 on 2009 11 09 by trac user mnikhil, assigned to unknown.
|
|
Another 64 bit specific numpy bug 32 bit python 2.5.1 >>> import numpy >>> numpy.version.version 1.0.3.1 >>> a = numpy.arange 10, dtype= int32 ) >>> i = numpy.array [3, 5, 6], dtype= int32 ) >>> numpy.add.reduceat a, i) array [ 7, 5, 30]) 64 bit python 2.6.1 >>> import numpy >>> numpy.version.version 1.3.0 >>> a = numpy.arange 10, dtype= int32 ) >>> i = numpy.array [3, 5, 6], dtype= int32 ) >>> numpy.add.reduceat a, i) array [ 8589934594, 12884901888, 64424509443] Is this expected?
|
|
Why?
|
|
What would have changed this?","bug"
|
|
"6013","1891","Changeset 7712 seems to break f2py Trac 1293).
|
|
Original ticket http projects.scipy.org numpy ticket 1293 on 2009 11 09 by plaes, assigned to pearu.
|
|
I m getting following compile error when building scipy after revision 7712.
|
|
Please note that vodemodule.c is generated by f2py.
|
|
creating build 2.6 temp.linux x86 64 2.6 build 2.6 src.linux x86 64 2.6 scipy integrate compile options DNO ATLAS INFO=2 Ibuild 2.6 src.linux x86 64 2.6 I usr lib64 python2.6 site packages numpy core include I usr include python2.6 c x86 64 pc linux gnu gcc build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb f in dvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error Each undeclared identifier is reported only once build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error for each function it appears in.)
|
|
build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 377 error ydot Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb jac in dvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error jac Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error nrowpd undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb f in zvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 637 error ydot Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb jac in zvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error jac Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error nrowpd undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb f in dvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error Each undeclared identifier is reported only once build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error for each function it appears in.)
|
|
build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 377 error ydot Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb jac in dvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error jac Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error nrowpd undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb f in zvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 637 error ydot Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c In function cb jac in zvode user routines build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error y Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error n undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error jac Dims undeclared first use in this function) build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error nrowpd undeclared first use in this function) error Command x86 64 pc linux gnu gcc pthread DNDEBUG march=native ggdb O2 pipe fPIC DNO ATLAS INFO=2 Ibuild 2.6 src.linux x86 64 2.6 I usr lib64 python2.6 site packages numpy core include I usr include python2.6 c build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.c o build 2.6 temp.linux x86 64 2.6 build 2.6 src.linux x86 64 2.6 scipy integrate vodemodule.o failed with exit status 1","bug"
|
|
"6013","1892","scipy build failure wrt changes in r7712 of numpy Trac 1294).
|
|
Original ticket http projects.scipy.org numpy ticket 1294 on 2009 11 09 by nilswagner01, assigned to unknown.
|
|
The problem file is generated by f2py.
|
|
In file included from scipy integrate odepackmodule.c 6 scipy integrate odepack.h In function xe2x80x98odepack odeintxe2x80x99 scipy integrate odepack.h 328 warning call to function xe2x80x98lsoda xe2x80x99 without a real prototype scipy integrate odepack.h 31 note xe2x80x98lsoda xe2x80x99 was declared here scipy integrate odepackmodule.c At top level scipy integrate multipack.h 114 warning xe2x80x98my make numpy arrayxe2x80x99 defined but not used scipy integrate odepack.h In function xe2x80x98odepack odeintxe2x80x99 scipy integrate odepack.h 226 warning xe2x80x98tcritxe2x80x99 may be used uninitialized in this function scipy integrate odepack.h 227 warning xe2x80x98waxe2x80x99 may be used uninitialized in this function usr bin gfortran Wall Wall shared build temp.linux x86 64 2.6 scipy integrate odepackmodule.o L home nwagner src ATLAS3.8.2 mybuild lib L usr lib64 Lbuild temp.linux x86 64 2.6 lodepack llinpack lite lmach lptf77blas lptcblas latlas lpython2.6 lgfortran o build lib.linux x86 64 2.6 scipy integrate odepack.so building scipy.integrate.vode extension compiling C sources C compiler usr bin gcc fno strict aliasing DNDEBUG fmessage length=0 O2 Wall D FORTIFY SOURCE=2 fstack protector funwind tables fasynchronous unwind tables g fwrapv fPIC creating build temp.linux x86 64 2.6 build src.linux x86 64 2.6 scipy integrate compile options DATLAS INFO= \ 3.8.2\ I home nwagner src ATLAS3.8.2 include Ibuild src.linux x86 64 2.6 I home nwagner local lib64 python2.6 site packages numpy core include I usr include python2.6 c gcc build src.linux x86 64 2.6 scipy integrate vodemodule.c build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb f in dvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error Each undeclared identifier is reported only once build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error for each function it appears in.)
|
|
build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 377 error xe2x80x98ydot Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 335 warning unused variable xe2x80x98ipar 5xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 334 warning unused variable xe2x80x98rpar 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 332 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb jac in dvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error xe2x80x98jac Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error xe2x80x98nrowpdxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 warning left hand operand of comma expression has no effect build src.linux x86 64 2.6 scipy integrate vodemodule.c 467 warning unused variable xe2x80x98ipar 8xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 466 warning unused variable xe2x80x98rpar 7xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 465 warning unused variable xe2x80x98nrowpd 6xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 464 warning unused variable xe2x80x98mu 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 463 warning unused variable xe2x80x98ml 3xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 461 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb f in zvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 637 error xe2x80x98ydot Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 595 warning unused variable xe2x80x98ipar 5xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 594 warning unused variable xe2x80x98rpar 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 592 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb jac in zvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error xe2x80x98jac Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error xe2x80x98nrowpdxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 warning left hand operand of comma expression has no effect build src.linux x86 64 2.6 scipy integrate vodemodule.c 727 warning unused variable xe2x80x98ipar 8xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 726 warning unused variable xe2x80x98rpar 7xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 725 warning unused variable xe2x80x98nrowpd 6xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 724 warning unused variable xe2x80x98mu 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 723 warning unused variable xe2x80x98ml 3xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 721 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb f in dvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error Each undeclared identifier is reported only once build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error for each function it appears in.)
|
|
build src.linux x86 64 2.6 scipy integrate vodemodule.c 376 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 377 error xe2x80x98ydot Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 335 warning unused variable xe2x80x98ipar 5xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 334 warning unused variable xe2x80x98rpar 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 332 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb jac in dvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 508 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error xe2x80x98jac Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 error xe2x80x98nrowpdxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 509 warning left hand operand of comma expression has no effect build src.linux x86 64 2.6 scipy integrate vodemodule.c 467 warning unused variable xe2x80x98ipar 8xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 466 warning unused variable xe2x80x98rpar 7xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 465 warning unused variable xe2x80x98nrowpd 6xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 464 warning unused variable xe2x80x98mu 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 463 warning unused variable xe2x80x98ml 3xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 461 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb f in zvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 636 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 637 error xe2x80x98ydot Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 595 warning unused variable xe2x80x98ipar 5xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 594 warning unused variable xe2x80x98rpar 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 592 warning unused variable xe2x80x98n 0xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c In function xe2x80x98cb jac in zvode user routinesxe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error xe2x80x98y Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 768 error xe2x80x98nxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error xe2x80x98jac Dimsxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 error xe2x80x98nrowpdxe2x80x99 undeclared first use in this function) build src.linux x86 64 2.6 scipy integrate vodemodule.c 769 warning left hand operand of comma expression has no effect build src.linux x86 64 2.6 scipy integrate vodemodule.c 727 warning unused variable xe2x80x98ipar 8xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 726 warning unused variable xe2x80x98rpar 7xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 725 warning unused variable xe2x80x98nrowpd 6xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 724 warning unused variable xe2x80x98mu 4xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 723 warning unused variable xe2x80x98ml 3xe2x80x99 build src.linux x86 64 2.6 scipy integrate vodemodule.c 721 warning unused variable xe2x80x98n 0xe2x80x99 error Command usr bin gcc fno strict aliasing DNDEBUG fmessage length=0 O2 Wall D FORTIFY SOURCE=2 fstack protector funwind tables fasynchronous unwind tables g fwrapv fPIC DATLAS INFO= \ 3.8.2\ I home nwagner src ATLAS3.8.2 include Ibuild src.linux x86 64 2.6 I home nwagner local lib64 python2.6 site packages numpy core include I usr include python2.6 c build src.linux x86 64 2.6 scipy integrate vodemodule.c o build temp.linux x86 64 2.6 build src.linux x86 64 2.6 scipy integrate vodemodule.o failed with exit status 1","bug"
|
|
"6013","1893","Histogram plotting example in Tentative Numpy Tutorial is broken for 1.4 devel Trac 1295).
|
|
Original ticket http projects.scipy.org numpy ticket 1295 on 2009 11 12 by plaes, assigned to pv.
|
|
Histogram plotting example in Tentative Numpy Tutorial should be updated to use new return parameters.","bug"
|
|
"6013","1894","numpy.histogram ) not working as expected for discrete values Trac 1296).
|
|
Original ticket http projects.scipy.org numpy ticket 1296 on 2009 11 12 by plaes, assigned to unknown.
|
|
I have a dataset of discrete integer) values from 0 to n and want to create a histogram ie count the occurrences of each number) of the dataset with bin size 1, but numpy.histogram ) doesn t seem to work right due to the way it handles last bin [max data) 1, max data)] ie 8 and 9, when max data) = 9).
|
|
Here s the script that should demonstrate it import numpy, scipy from matplotlib import pylab Generate random discrete data data = numpy.random.poisson 3, 300) max num = max data) hist, bins = numpy.histogram data, bins=max num) matplotlibs histogram also has the same issue S hist, bins, = pylab.hist data, bins=max data)) print Histogram str hist)) print No of max str len data) len filter lambda x x != max data), data)))) print No of max 1 str len data) len filter lambda x x != max data) 1, data))))","bug"
|
|
"6013","1895","Unit test failures some nextafter tests fail on Python2.4 2.5 Trac 1297).
|
|
Original ticket http projects.scipy.org numpy ticket 1297 on 2009 11 12 by pv, assigned to cournape.
|
|
The following unit tests fail on Python2.4 and 2.5 but not on 2.6) ====================================================================== ERROR test simple2 test utils.TestArrayAlmostEqualNulp) Traceback most recent call last) File var lib buildslave numpy real b15 numpy install 2.5 lib python2.5 site packages numpy testing tests test utils.py , line 317, in test simple2 self.failUnlessRaises AssertionError, failure) File usr lib python2.5 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) File var lib buildslave numpy real b15 numpy install 2.5 lib python2.5 site packages numpy testing tests test utils.py , line 316, in failure nulp=1000) File .. numpy install 2.5 lib python2.5 site packages numpy testing utils.py , line 1112, in assert array almost equal nulp TypeError int argument required ====================================================================== ERROR test nan test utils.TestULP) Traceback most recent call last) File var lib buildslave numpy real b15 numpy install 2.5 lib python2.5 site packages numpy testing tests test utils.py , line 370, in test nan lambda assert array max ulp nan, inf, File usr lib python2.5 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) File var lib buildslave numpy real b15 numpy install 2.5 lib python2.5 site packages numpy testing tests test utils.py , line 371, in <lambda> maxulp=maxulp)) File .. numpy install 2.5 lib python2.5 site packages numpy testing utils.py , line 1121, in assert array max ulp TypeError int argument required","bug"
|
|
"6013","1896","setdiff1d does not work on structured arrays Trac 1298).
|
|
Original ticket http projects.scipy.org numpy ticket 1298 on 2009 11 13 by trac user mcieslik, assigned to unknown.
|
|
because mergesort is hardcoded in setmember1d In [121] a = np.array [ 0,1), 0,2), 0,3)], dtype= int8,int8 ) In [122] b = np.array [ 0,1), 0,2), 0,4)], dtype= int8,int8 ) In [123] np.setdiff1d a,b) TypeError Traceback most recent call last) <<snip>> python2.6 site packages numpy lib arraysetops.pyc in setmember1d ar1, ar2) 312 values from the first array should always come before the values from the 313 second array.
|
|
> 314 perm = ar.argsort kind= mergesort ) 315 aux = ar[perm] 316 aux2 = tt[perm] TypeError requested sort not available for type In [124] a.sort kind= quicksort ) In [126] a.sort kind= mergesort ) TypeError Traceback most recent call last) net home mcieslik <ipython console> in <module> ) TypeError desired sort not supported for this type","enhancement"
|
|
"6013","1897","setitem bug on structured arrays with objects Trac 1299).
|
|
Original ticket http projects.scipy.org numpy ticket 1299 on 2009 11 19 by trac user crobin, assigned to unknown.
|
|
I think there is a bug with ` setitem ` when using advanced selection on a structured array with objects when dimension is at least 2. x = numpy.rec.fromarrays [[0,1],[ a , b ]], formats= i8,O )[None] print x x[x.nonzero )] = x.ravel )[ 1] print x This gives [[ 0, a ) 1, b )]] [[ 0, a ) 0, b )]] WRONG The very last ` b ` should be an ` a `.
|
|
As a bonus, the reference count of the string `xe2x80x98bxe2x80x99` decreases, even though it still remains in the array `x`.
|
|
So the interpreter will crash if you repeat the above while True x[x.nonzero )] = x.ravel )[ 1] ...leads to Fatal Python error deletion of interned string failed zsh IOT instruction core dumped) python2.6 Note that this bug does not seem to occur when `x` is dimension 1, or when basic selection is used.
|
|
The offset of the object in the dtype does not seem to matter.","bug"
|
|
"6013","1898","OverflowError when nan[arg]min or nan[arg]max called on unsigned integers Trac 1300).
|
|
Original ticket http projects.scipy.org numpy ticket 1300 on 2009 11 20 by tonysyu, assigned to unknown.
|
|
Functions that call nanop i.e.
|
|
nan[arg]min, nan[arg]max) fail with unsigned integers.
|
|
For example >>> np.nanmin np.array [0, 1], dtype=np.uint8)) OverflowError cannot convert float infinity to integer It seems that unsigned integers don t get identified as integers in the nanop function.
|
|
Checking against `np.integer` instead of the built in `int` will catch signed and unsigned integers, as shown below.
|
|
Index numpy lib function base.py =================================================================== numpy lib function base.py revision 7749) numpy lib function base.py working copy) 1349,7 1351,7 mask = isnan a) We only need to take care of NaN s in floating point arrays if not np.issubdtype y.dtype, int) if not np.issubdtype y.dtype, np.integer) y[mask] = fill res = op y, axis=axis)","bug"
|
|
"6013","1899","array 1,ndmin=33) > segmentation fault Trac 1301).
|
|
Original ticket http projects.scipy.org numpy ticket 1301 on 2009 11 21 by endolith, assigned to unknown.
|
|
Obviously 33 dimensions is unrealistic, but shouldn t it create an exception rather than crashing?
|
|
Job 1, xe2x80x9cipython pylabxe2x80x9d terminated by signal SIGSEGV Address boundary error)","bug"
|
|
"6013","1900","make hstack and vstack promote 1 D argument to 2 D when necessary Trac 1302).
|
|
Original ticket http projects.scipy.org numpy ticket 1302 on 2009 11 24 by trac user pfeldman, assigned to unknown.
|
|
I d like to see hstack and vstack promote 1 D arguments to 2 D when this is necessary to make the dimensions match.
|
|
In the following example, c works as expected while hstack does not [ ]|8> x <8> array [[1, 2, 3], [4, 5, 6], [7, 8, 9]]) [ ]|9> y <9> array [10, 11, 12]) [ ]|10> c [x,y] <10> array [[ 1, 2, 3, 10], [ 4, 5, 6, 11], [ 7, 8, 9, 12]]) [ ]|11> x <11> array [[1, 2, 3], [4, 5, 6], [7, 8, 9]]) [ ]|12> y <12> array [10, 11, 12]) [ ]|13> hstack x,y)) ValueError Traceback most recent call last) \<ipython console> in <module> ) C \Program Files\Python25\lib\site packages\numpy\lib\shape base.pyc in hstack tup) 503 504 > 505 return nx.concatenate map atleast 1d,tup),1) 506 507 row stack = vstack ValueError arrays must have same number of dimensions","enhancement"
|
|
"6013","1901","Bug in rec.fromarrays in handling length 0 strings Trac 1303), now with segfault..
|
|
Original ticket http projects.scipy.org numpy ticket 1303 on 2009 11 25 by trac user dyamins, assigned to unknown.
|
|
I m using 1.4.0.dev7746, on Mac OSX 10.5.8.
|
|
The problems below occur both for python 2.5 compiled 32 bit as well as python2.6 compiled 64 bit.
|
|
A problem arises when you try to create a record array passing a type of |S0 .
|
|
.
|
|
When not passing any dtype, this is created into a recarray with no problem .
|
|
However, trouble arises when I try to pass a length 0 dtype explicitly.
|
|
.
|
|
The same thing occurs if I cast to np arrays before passing to np.rec.fromarrays .","bug"
|
|
"6013","1902","Bug in handling |S0 format string Trac 1304).
|
|
Original ticket http projects.scipy.org numpy ticket 1304 on 2009 11 25 by trac user dyamins, assigned to unknown.
|
|
np.dtype won t accept |S0 as a type argument.
|
|
>>> np.dtype |S0 ) TypeError data type not understood I have to do something like this >>> d = np.dtype |S ) >>> d dtype |S0 ) to get what I want.
|
|
Is this intended?
|
|
Regardless, this inconsistency also means that things like >>> np.dtype d.descr) can fail even when d is a properly constructed dtype object with a |S0 type, which seems a little perverse.","bug"
|
|
"6013","1903","masked arrays and views Trac 1305).
|
|
Original ticket http projects.scipy.org numpy ticket 1305 on 2009 11 26 by trac user tgirod, assigned to pierregm.
|
|
when creating a view on a masked array with mask=False, changing the mask in the view won t propagate to the mask in the original array.
|
|
However, when the mask is already an array before the view is created, altering the mask in the view alters the mask in the original array.
|
|
this difference of behaviour can lead to troubles import numpy as np import numpy.ma as ma x = ma.zeros 25) xv = x.reshape 5,5)) print x.mask > False print xv.mask > False xv.mask = np.identity 5) print x.mask > False print xv.mask > [[ True False False False False] [False True False False False] [False False True False False] [False False False True False] [False False False False True]] ...","bug"
|
|
"6013","1904","F2PY doesn t work without python dev package Trac 1306).
|
|
Original ticket http projects.scipy.org numpy ticket 1306 on 2009 11 27 by trac user alexm920, assigned to pearu.
|
|
As installed, python numpy, which contains F2PY, does not have python dev as a dependency.
|
|
However, without that package installed, no fortran files can be successfully F2PY d. Took a while to figure out what was missing in my new install.","bug"
|
|
"6013","1905","take with axis=1 from matrix with matrix indices gives row instead of column Trac 1307).
|
|
Original ticket http projects.scipy.org numpy ticket 1307 on 2009 11 28 by thouis, assigned to unknown.
|
|
>>> from numpy import >>> version 1.3.0.dev5934 >>> q = matrix arange 9).reshape 3,3))) >>> take q, matrix [[0]]), axis=0) matrix [[0, 1, 2]]) >>> take q, matrix [[0]]), axis=1) matrix [[0, 3, 6]]) Expected behavior for the last expression >>> take q, [0], axis=1) matrix [[0], [3], [6]])","bug"
|
|
"6013","1906","Make np.dot and np.inner fast for read only arrays.
|
|
Trac 1308).
|
|
Original ticket http projects.scipy.org numpy ticket 1308 on 2009 11 29 by trac user hansres, assigned to unknown.
|
|
Currently `np.dot` `np.inner` are much slower 3x) for read only arrays than for writeable arrays.
|
|
In [1] a = randn 1e6) In [2] timeit dot a, a) 100 loops, best of 3 5.42 ms per loop In [3] a.setflags write=False) In [4] timeit dot a, a) 100 loops, best of 3 15.9 ms per loop.
|
|
This slowdown for read only inputs is caused by the creation of array copies dot inner call `PyArray FromAny` and require the `WRITABLE` flag.
|
|
Instead, because read only input arrays are fine, the `WRITEABLE` flag should not be is not required; and the `ALIGNED` flag should be sufficient.","enhancement"
|
|
"6013","1907","dot ) returns negative number when fed with large integer inputs Trac 1309).
|
|
Original ticket http projects.scipy.org numpy ticket 1309 on 2009 11 29 by trac user dosh1965, assigned to unknown.
|
|
>>> a 10000000000000000L >>> dot a,a) 8814407033341083648","bug"
|
|
"6013","1908","Performance of vstack ) vs array ) Add an helper function to easily stack on new axis Trac 1310).
|
|
Original ticket http projects.scipy.org numpy ticket 1310 on 2009 12 03 by bthyreau, assigned to unknown.
|
|
Hi, In many cases, one may want to stack on a new dimension; eg.
|
|
say, some 2D array on a 3D array.
|
|
For years, i was using regularly code such as ` out = array [arr1, arr2, arr3]) ` until i realized this can be very inefficient with large input array ; Indeed compared to a simple reshaped vstack ), it can be an order of magnitude slower, as it happens that many `PySequence GetItem )` are called unneededly on every individual items of the arrays.
|
|
I m not aware of any obvious function to do this common operation.
|
|
If it is the case, then may i suggest to add a function for that, for example, similar to the following code def stacknew tup, lastdim=False) prepend or append a new axis, then stack arrays on it if not lastdim a = vstack [x[None] for x in tup]) else a = vstack [x.T[None] for x in tup]).T return a This may make a great addition to the other stack family function, as well as making more obvious that using array ) for that is not the best way.","enhancement"
|
|
"6013","1909","128 bit float not correctley converted to string Trac 1311).
|
|
Original ticket http projects.scipy.org numpy ticket 1311 on 2009 12 04 by trac user rgraham nz, assigned to unknown.
|
|
numpy seems to not use the full precision of a 128 bit float when converting to a string.
|
|
Interestingly ipython displays the whole number.
|
|
See this ipython session In [1] import numpy as np In [2] a = np.float128 1.0821318064579998897 ) In [3] a Out[3] 1.0821318064579998897 In [4] str a) Out[4] 1.08213180646 Platform 64 bit ubuntu linux.
|
|
numpy version 1.3.0 3, python 2.6.4","bug"
|
|
"6013","1910","Changes in behaviour due to buffer protocol in numpy 1.5 Trac 1312).
|
|
Original ticket http projects.scipy.org numpy ticket 1312 on 2009 12 06 by GaelVaroquaux, assigned to pv.
|
|
The following used to work in 1.4 hashlib.md5 x).hexdigest ) Now, you have to do hashlib.md5 np.getbuffer x)).hexdigest ) Also, with extension code I get the following failure if not rebuild against a new version of numpy) from routines import joint histogram, similarity, similarity measures File numpy.pxi , line 74, in nipy.neurospin.register.routines nipy neurospin register routines.c 6042) ValueError numpy.dtype does not appear to be the correct type object Where routines.c is a cython generated file, compiled with numpy 1.4.
|
|
I am unsure which line in routines.pyx is responsible for the problem.","bug"
|
|
"6013","1911","Complex POW not implemented correctly in 1.4.0rc1 and trunk) Trac 1313).
|
|
Original ticket http projects.scipy.org numpy ticket 1313 on 2009 12 07 by FrancescAlted, assigned to unknown.
|
|
After the restructuration of the complex functions in numpy, the `pow` function seems broken.
|
|
With numpy 1.3.0 In [5] 2 0.5 1.5j) Out[5] 0.35812203996480685 0.60971190286187238j) In [6] 2 np.array 0.5 1.5j)) Out[6] 0.35812203996480685 0.60971190286187238j) same as python With 1.4.0rc1 In [49] 2 0.5 1.5j) Out[49] 0.35812203996480685 0.60971190286187238j) In [50] 2 np.array 0.5 1.5j)) Out[50] 0.35182011479251107 1.3279606132136452j) different!
|
|
I ve tried the next fix Index numpy core src npymath npy math complex.c.src =================================================================== numpy core src npymath npy math complex.c.src revision 7975) numpy core src npymath npy math complex.c.src working copy) 199,7 199,7 type bx, by, px, py; b = npy clog c x); p = npy clog c y); p = y; bx = npy creal c b); by = npy cimag c b); px = npy creal c p); which I think is the correct one, but the error is still there.
|
|
That affects also to trunk.","bug"
|
|
"6013","1912",".all ) and .any ) fail for arrays of type S and U Trac 1314).
|
|
Original ticket http projects.scipy.org numpy ticket 1314 on 2009 12 07 by mdboom, assigned to unknown.
|
|
In [3] x = np.array [ abc , def , ], S ) In [4] x.all ) TypeError Traceback most recent call last) wonkabar data1 builds betadrizzle <ipython console> in <module> ) TypeError cannot perform reduce with flexible type This is a sensible operation.
|
|
I wonder if there s a way to make it work.","bug"
|
|
"6013","1913",".argmax ) works, but .argmin ) fails for arrays of type S and U Trac 1315).
|
|
Original ticket http projects.scipy.org numpy ticket 1315 on 2009 12 07 by mdboom, assigned to unknown.
|
|
In [6] x = np.array [ xyz , abc , def ], S ) In [7] x.argmax ) Out[7] 0 In [8] x.argmin ) TypeError Traceback most recent call last) wonkabar data1 builds betadrizzle <ipython console> in <module> ) TypeError argmax is unsupported for this type","bug"
|
|
"6013","1914",".max ) and .min ) don t work on arrays of type S and U Trac 1316).
|
|
Original ticket http projects.scipy.org numpy ticket 1316 on 2009 12 07 by mdboom, assigned to unknown.
|
|
This is a sensible operation.
|
|
It would be nice to make it work if possible.
|
|
In [25] x.max ) TypeError Traceback most recent call last) wonkabar data1 builds betadrizzle <ipython console> in <module> ) TypeError cannot perform reduce with flexible type In [26] x.min ) TypeError Traceback most recent call last) wonkabar data1 builds betadrizzle <ipython console> in <module> ) TypeError cannot perform reduce with flexible type In [27]","bug"
|
|
"6013","1915","Uncaught SIGFPE exception for overflowed integer division Trac 1317).
|
|
Original ticket http projects.scipy.org numpy ticket 1317 on 2009 12 07 by cournape, assigned to unknown.
|
|
The following raises a SIGFPE division by 0) import numpy as np x = np.array [ 2 63]) x = 1 It is because 2 63 is not representable as a 64 bits integers, and on at least x86, this raises a div by zero FPU exception.","bug"
|
|
"6013","1916","All FPU exception are ignored by default Trac 1318).
|
|
Original ticket http projects.scipy.org numpy ticket 1318 on 2009 12 07 by cournape, assigned to unknown.
|
|
We have import numpy as np x = np.array [1, 2, 3, 4]) x = 0 x is now zero, no warning printed This is a regression, which appeared in numpy 1.1.0.","bug"
|
|
"6013","1917","assignment of complex to float array causes loss of imaginary part Trac 1319).
|
|
Original ticket http projects.scipy.org numpy ticket 1319 on 2009 12 08 by trac user pfeldman, assigned to unknown.
|
|
As the following sample program `err.py` demonstrates, the assignment of an array of complex values to a float array causes the imaginary part to disappear.
|
|
The result is wrong answers without any warning message.
|
|
START OF CODE from numpy import q = ones 2,dtype=complex) 1 1J) r = zeros 2,dtype=float) r[ ] = q print q = ,q print r = ,r END OF CODE [ ]|9> run err q = [ 1.
|
|
1.j 1.
|
|
1.j] r = [ 1.
|
|
1.]
|
|
[ ]|10>","enhancement"
|
|
"6013","1918","Error in 1.4.0rc1 on Python 2.6.4 and Windows numpy.dtype does not appear to be the correct type object Trac 1320).
|
|
Original ticket http projects.scipy.org numpy ticket 1320 on 2009 12 08 by jonovik, assigned to unknown.
|
|
Using Numpy 1.4.0rc1 under Windows XP or 7) >>> from scipy.stats import chi2 Traceback most recent call last) ValueError numpy.dtype does not appear to be the correct type object I has this problem on 32 bit Windows XP SP3, whereas this poster had it on 64 bit Windows 7 http thread.gmane.org gmane.comp.python.scientific.user 23108 Another poster suggested to downgrade to Numpy 1.3.0, which fixed the problem in both cases.
|
|
I forgot to save the traceback before I downgraded, but the abovementioned Windows 7 user had this File <stdin> , line 1, in <module> File D \Python26\lib\site packages\scipy\stats\ init .py , line 7, in <module> from stats import File D \Python26\lib\site packages\scipy\stats\stats.py , line 203, in <module> from morestats import find repeats is only reference to scipy.stats File D \Python26\lib\site packages\scipy\stats\morestats.py , line 7, in <module> import distributions File D \Python26\lib\site packages\scipy\stats\distributions.py , line 27, in <module> import vonmises cython File numpy.pxd , line 30, in scipy.stats.vonmises cython scipy\stats\vonmises cython.c 2939)","bug"
|
|
"6013","1919","A dtype should be immutable Trac 1321).
|
|
Original ticket http projects.scipy.org numpy ticket 1321 on 2009 12 09 by FrancescAlted, assigned to unknown.
|
|
After some discussion in the list, we felt that the dtype should be immutable.
|
|
Currently this is not the case, as the `names` attribute can be changed In [19] t = np.dtype i4,f4 ) In [20] t Out[20] dtype [ f0 , <i4 ), f1 , <f4 )]) In [21] hash t) Out[21] 9041335829180134223 In [22] t.names = one , other ) In [23] t Out[23] dtype [ one , <i4 ), other , <f4 )]) In [24] hash t) Out[24] 8637734220020415106 [although the change of the hash seems to depend on the Python version] As a bonus, an immutable `dtype` would render this ticket 1725 without effect.","bug"
|
|
"6013","1920","Scipy build fails on Mac OS X 10.5.8, Numpy import fails on Mac OS X 10.5.8 Trac 1322).
|
|
Original ticket http projects.scipy.org numpy ticket 1322 on 2009 12 10 by trac user ndufour, assigned to cournape.
|
|
Mac OS X 10.5.8 32 bit machine The problem appears to lie with the file multiarray.so Attempting to import Numpy returns Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy init .py , line 132, in <module> import add newdocs File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy add newdocs.py , line 9, in <module> from lib import add newdoc File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib init .py , line 4, in <module> from type check import File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core init .py , line 5, in <module> import multiarray ImportError dlopen Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so unknown required load command 0x80000022 Attempting to build Scipy returns Traceback most recent call last) File setup.py , line 160, in <module> setup package ) File setup.py , line 127, in setup package from numpy.distutils.core import setup File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy init .py , line 132, in <module> import add newdocs File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy add newdocs.py , line 9, in <module> from lib import add newdoc File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib init .py , line 4, in <module> from type check import File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core init .py , line 5, in <module> import multiarray ImportError dlopen Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so unknown required load command 0x80000022","bug"
|
|
"6013","1921","FAIL Test generic loops.
|
|
Trac 1323).
|
|
Original ticket http projects.scipy.org numpy ticket 1323 on 2009 12 11 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL Test generic loops.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test ufunc.py , line 86, in test generic loops assert almost equal fone x), fone val, err msg=msg) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 434, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal PyUFunc F F ACTUAL array [ 0.
|
|
0.j, 0.
|
|
0.j, 0.
|
|
0.j, 0.
|
|
0.j, 0.
|
|
0.j], dtype=complex64) DESIRED 1 Python 2.5.4 Red Hat Enterprise Linux 4, 64 bit Not on RHE 4, 32 bit.
|
|
Not on Mac OSX Leopard, 32 bit python.
|
|
Not on RHE 5, 64 bit.
|
|
Not on Solaris 8, 32 bit python.
|
|
NumPy version 1.5.0.dev7996","bug"
|
|
"6013","1922","FAIL test umath.TestComplexFunctions.test loss of precision <type numpy.complex64 >,) Trac 1324).
|
|
Original ticket http projects.scipy.org numpy ticket 1324 on 2009 12 11 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test umath.TestComplexFunctions.test loss of precision <type numpy.complex64 >,) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 nose case.py , line 183, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.4 numpy core tests test umath.py , line 721, in check loss of precision check x basic, 2 eps 1e 3) File usr stsci pyssgdev 2.5.4 numpy core tests test umath.py , line 691, in check arcsinh ) AssertionError 0, 0.0010023052, 0.9987238, arcsinh ) Python 2.5.4 Red Hat Enterprise Linux 4, 64 bit Not on RHE 4, 32 bit.
|
|
Not on Mac OSX Leopard, 32 bit python.
|
|
Not on RHE 5, 64 bit.
|
|
!)
|
|
NumPy version 1.5.0.dev7996 This appears to be similar to 1711, but I m first seeing it this week and that ticket is 7 months old.","bug"
|
|
"6013","1923","FAIL test umath.TestComplexFunctions.test precisions consistent Trac 1325).
|
|
Original ticket http projects.scipy.org numpy ticket 1325 on 2009 12 11 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test umath.TestComplexFunctions.test precisions consistent Traceback most recent call last) File usr stsci pyssgdev 2.5.4 nose case.py , line 183, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.4 numpy core tests test umath.py , line 602, in test precisions consistent assert almost equal fcf, fcd, decimal=6, err msg= fch fcd s f) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 434, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal fch fcd <ufunc arcsin > ACTUAL 2.3561945j DESIRED 0.66623943249251527 1.0612750619050355j) Python 2.5.4 Red Hat Enterprise Linux 4, 64 bit Not on RHE 4, 32 bit.
|
|
Not on Mac OSX Leopard, 32 bit python.
|
|
Not on RHE 5, 64 bit.
|
|
Not on Solaris 8, 32 bit python.
|
|
NumPy version 1.5.0.dev7996","bug"
|
|
"6013","1924","FAIL test buffer hashlib test regression.TestRegression) Trac 1326).
|
|
Original ticket http projects.scipy.org numpy ticket 1326 on 2009 12 11 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== FAIL test buffer hashlib test regression.TestRegression) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test regression.py , line 1255, in test buffer hashlib assert equal md5 x).hexdigest ), 2a1dd1e1e59d0a384c26951e316cd7e6 ) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 305, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL 1264d4a9f74dc462700fd163e3ff09a6 DESIRED 2a1dd1e1e59d0a384c26951e316cd7e6 Python 2.5.4, 32 bit executable Solaris 8 Not on Red Hat Enterprise Linux, 4 32 bit.
|
|
Not on RHE 4 64 bit.
|
|
Not on RHE 5 64 bit.
|
|
Not on Mac OSX Leopard, 32 bit python.
|
|
That is, failing on the big endian SPARC, passing on a bunch of x86 systems.)
|
|
NumPy version 1.5.0.dev7996","bug"
|
|
"6013","1925","nan to num with bool arrays Trac 1327).
|
|
Original ticket http projects.scipy.org numpy ticket 1327 on 2009 12 11 by trac user Nicolas Rougier, assigned to unknown.
|
|
Using both numpy 1.3.0 and 1.4.0rc1 I got the following exception using nan to num on a bool array.
|
|
>>> import numpy >>> Z = numpy.zeros 3,3),dtype=bool) >>> numpy.nan to num Z) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.6 dist packages numpy lib type check.py , line 374, in nan to num maxf, minf = getmaxmin y.dtype.type) File usr lib python2.6 dist packages numpy lib type check.py , line 307, in getmaxmin f = getlimits.finfo t) File usr lib python2.6 dist packages numpy core getlimits.py , line 103, in new raise ValueError, data type r not inexact dtype) ValueError data type <type numpy.bool > not inexact","bug"
|
|
"6013","1926","Incorrect sorting on field Trac 1328).
|
|
Original ticket http projects.scipy.org numpy ticket 1328 on 2009 12 12 by stefanv, assigned to unknown.
|
|
In [53] cc Out[53] array [[ 0, 1)], [ 0, 0)], [ 1, 1)], [ 1, 0)]], dtype=[ x , <i8 ), y , <i8 )]) In [54] np.sort cc, order= y )[ y ] Out[54] array [[1], [0], [1], [0]])","bug"
|
|
"6013","1927","spacing broken for large values Trac 1329).
|
|
Original ticket http projects.scipy.org numpy ticket 1329 on 2009 12 13 by cournape, assigned to unknown.
|
|
For example np.spacing 1e30) returns 0","bug"
|
|
"6013","1928","tolist ) does not work for structured array row Trac 1330).
|
|
Original ticket http projects.scipy.org numpy ticket 1330 on 2009 12 14 by astrofrog, assigned to pierregm.
|
|
The following code import numpy.ma as ma a = ma.zeros 3,),dtype=[ a ,int), b ,int)]) a[ a ][ ] = 1 a[ b ][ ] = 2 a.mask[0][0] = True print a print a[0] print a[0].tolist ) produces the following output [ , 2) 1, 2) 1, 2)] , 2) None, None) Whereas I would expect [ , 2) 1, 2) 1, 2)] , 2) None, 2) This is using revision 8007","bug"
|
|
"6013","1929","Can t iterate over row of masked structured array Trac 1331).
|
|
Original ticket http projects.scipy.org numpy ticket 1331 on 2009 12 14 by astrofrog, assigned to pierregm.
|
|
The following code demonstrates the problem import numpy.ma as ma a = ma.zeros 3,),dtype=[ a ,int), b ,int)]) a[ a ][ ] = 1 a[ b ][ ] = 2 print a[0] for elem in a[0] print elem a[ a ].mask = True, False, False) print a[0] for elem in a[0] print elem produces 1, 2) 1 2 , 2) Traceback most recent call last) File bug.py , line 14, in <module> for elem in a[0] TypeError iteration over a 0 d array i.e.
|
|
once one element is masked, one can no longer iterate over a row.
|
|
This is using revision 8007.","bug"
|
|
"6013","1930","set fill value does not work properly for masked structured array Trac 1332).
|
|
Original ticket http projects.scipy.org numpy ticket 1332 on 2009 12 15 by astrofrog, assigned to pierregm.
|
|
The following code doesn t seem to work import numpy.ma as ma t = ma.array zip [1,2,3],[4,5,6]),dtype=[ a ,int), b ,int)]) print repr t[ a ]) t[ a ].set fill value 10) print repr t[ a ]) as the output is masked array data = [1 2 3], mask = [False False False], fill value = 999999) masked array data = [1 2 3], mask = [False False False], fill value = 999999) and no exception is raised.","bug"
|
|
"6013","1931","exclude ndarray operations mul , pow etc Trac 1333).
|
|
Original ticket http projects.scipy.org numpy ticket 1333 on 2009 12 15 by trac user dmitrey, assigned to unknown.
|
|
numpy ndarray operations !
|
|
mul!
|
|
, !
|
|
div!
|
|
, !
|
|
pow!
|
|
, !
|
|
lt!
|
|
, !
|
|
gt!
|
|
, !
|
|
eq!
|
|
etc all non unary ones) should be replaced by the lines like return asarray multiply self, i)) if not getattr i, !
|
|
exclude ndarray operations!
|
|
, False) else i.!
|
|
rmul!
|
|
self) This bug prevents correct stacking numpy ndarrays with matrices, polynomials, SAGE arrays, FuncDesigner types, etc.
|
|
See here for details http comments.gmane.org gmane.comp.python.scientific.devel 12436","bug"
|
|
"6013","1932","multiarray.so returns mysterious error on import in OSX 10.5.8 Trac 1334).
|
|
Original ticket http projects.scipy.org numpy ticket 1334 on 2009 12 17 by trac user Tazlor, assigned to cournape.
|
|
I have installed the NumPy 1.4.0rc1 package from sourceforge for python 2.6 on OSX 10.5.8.
|
|
2.6 is the version of my Macpython installation, the system python remains 2.5. import numpy in 2.6 returns an ImportError on the file multiarray.so.
|
|
Here are the last few lines of the error traceback ImportError dlopen Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so unknown required load command 0x80000022 The file multiarray.so does exist in that directory.
|
|
Strangely, the system python 2.5) can import NumPy without incident.
|
|
I have tried to receive help on this problem in a google groups conversation http groups.google.com group comp.lang.python browse thread thread 03620cbf3f202353) and another person has confirmed experiencing the same symptoms.
|
|
Below I have copied the full output of import numpy with gdb.
|
|
The traceback appears at the end.
|
|
By the way, I don t have any user called ronald .
|
|
Last login Thu Dec 17 02 18 15 on ttys000 Polaris kevin gdb python GNU gdb 6.3.50 20050815 Apple version gdb 962) Sat Jul 26 08 14 40 UTC 2008) Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and or distribute copies of it under certain conditions.
|
|
Type show copying to see the conditions.
|
|
There is absolutely no warranty for GDB.
|
|
Type show warranty for details.
|
|
This GDB was configured as i386 apple darwin ...Reading symbols for shared libraries .. done gdb) run c import numpy Starting program Library Frameworks Python.framework Versions 2.6 bin python c import numpy Reading symbols for shared libraries .
|
|
done Program received signal SIGTRAP, Trace breakpoint trap.
|
|
0x8fe01010 in dyld dyld start ) gdb) c Continuing.
|
|
Reading symbols for shared libraries warning Could not find object file private tmp py bld python libpython2.6.a getbuildinfo.o) no debug information available for Users ronald Projects python r263 Modules getbuildinfo.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a acceler.o) no debug information available for Users ronald Projects python r263 Parser acceler.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a grammar1.o) no debug information available for Users ronald Projects python r263 Parser grammar1.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a listnode.o) no debug information available for Users ronald Projects python r263 Parser listnode.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a node.o) no debug information available for Users ronald Projects python r263 Parser node.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a parser.o) no debug information available for Users ronald Projects python r263 Parser parser.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a parsetok.o) no debug information available for Users ronald Projects python r263 Parser parsetok.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a bitset.o) no debug information available for Users ronald Projects python r263 Parser bitset.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a metagrammar.o) no debug information available for Users ronald Projects python r263 Parser metagrammar.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a firstsets.o) no debug information available for Users ronald Projects python r263 Parser firstsets.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a grammar.o) no debug information available for Users ronald Projects python r263 Parser grammar.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pgen.o) no debug information available for Users ronald Projects python r263 Parser pgen.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a myreadline.o) no debug information available for Users ronald Projects python r263 Parser myreadline.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a tokenizer.o) no debug information available for Users ronald Projects python r263 Parser tokenizer.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a abstract.o) no debug information available for Users ronald Projects python r263 Objects abstract.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a boolobject.o) no debug information available for Users ronald Projects python r263 Objects boolobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a bufferobject.o) no debug information available for Users ronald Projects python r263 Objects bufferobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a bytes methods.o) no debug information available for Users ronald Projects python r263 Objects bytes methods.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a bytearrayobject.o) no debug information available for Users ronald Projects python r263 Objects bytearrayobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a cellobject.o) no debug information available for Users ronald Projects python r263 Objects cellobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a classobject.o) no debug information available for Users ronald Projects python r263 Objects classobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a cobject.o) no debug information available for Users ronald Projects python r263 Objects cobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a codeobject.o) no debug information available for Users ronald Projects python r263 Objects codeobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a complexobject.o) no debug information available for Users ronald Projects python r263 Objects complexobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a descrobject.o) no debug information available for Users ronald Projects python r263 Objects descrobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a enumobject.o) no debug information available for Users ronald Projects python r263 Objects enumobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a exceptions.o) no debug information available for Users ronald Projects python r263 Objects exceptions.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a genobject.o) no debug information available for Users ronald Projects python r263 Objects genobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a fileobject.o) no debug information available for Users ronald Projects python r263 Objects fileobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a floatobject.o) no debug information available for Users ronald Projects python r263 Objects floatobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a frameobject.o) no debug information available for Users ronald Projects python r263 Objects frameobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a funcobject.o) no debug information available for Users ronald Projects python r263 Objects funcobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a intobject.o) no debug information available for Users ronald Projects python r263 Objects intobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a iterobject.o) no debug information available for Users ronald Projects python r263 Objects iterobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a listobject.o) no debug information available for Users ronald Projects python r263 Objects listobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a longobject.o) no debug information available for Users ronald Projects python r263 Objects longobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a dictobject.o) no debug information available for Users ronald Projects python r263 Objects dictobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a methodobject.o) no debug information available for Users ronald Projects python r263 Objects methodobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a moduleobject.o) no debug information available for Users ronald Projects python r263 Objects moduleobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a object.o) no debug information available for Users ronald Projects python r263 Objects object.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a obmalloc.o) no debug information available for Users ronald Projects python r263 Objects obmalloc.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a rangeobject.o) no debug information available for Users ronald Projects python r263 Objects rangeobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a setobject.o) no debug information available for Users ronald Projects python r263 Objects setobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a sliceobject.o) no debug information available for Users ronald Projects python r263 Objects sliceobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a stringobject.o) no debug information available for Users ronald Projects python r263 Objects stringobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a structseq.o) no debug information available for Users ronald Projects python r263 Objects structseq.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a tupleobject.o) no debug information available for Users ronald Projects python r263 Objects tupleobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a typeobject.o) no debug information available for Users ronald Projects python r263 Objects typeobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a weakrefobject.o) no debug information available for Users ronald Projects python r263 Objects weakrefobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a unicodeobject.o) no debug information available for Users ronald Projects python r263 Objects unicodeobject.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a unicodectype.o) no debug information available for Users ronald Projects python r263 Objects unicodectype.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a warnings.o) no debug information available for Users ronald Projects python r263 Python warnings.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a Python ast.o) no debug information available for Users ronald Projects python r263 Python Python ast.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a asdl.o) no debug information available for Users ronald Projects python r263 Python asdl.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a ast.o) no debug information available for Users ronald Projects python r263 Python ast.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a bltinmodule.o) no debug information available for Users ronald Projects python r263 Python bltinmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a ceval.o) no debug information available for Users ronald Projects python r263 Python ceval.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a compile.o) no debug information available for Users ronald Projects python r263 Python compile.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a codecs.o) no debug information available for Users ronald Projects python r263 Python codecs.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a errors.o) no debug information available for Users ronald Projects python r263 Python errors.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a frozen.o) no debug information available for Users ronald Projects python r263 Python frozen.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a frozenmain.o) no debug information available for Users ronald Projects python r263 Python frozenmain.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a future.o) no debug information available for Users ronald Projects python r263 Python future.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getargs.o) no debug information available for Users ronald Projects python r263 Python getargs.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getcompiler.o) no debug information available for Users ronald Projects python r263 Python getcompiler.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getcopyright.o) no debug information available for Users ronald Projects python r263 Python getcopyright.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getmtime.o) no debug information available for Users ronald Projects python r263 Python getmtime.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getplatform.o) no debug information available for Users ronald Projects python r263 Python getplatform.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getversion.o) no debug information available for Users ronald Projects python r263 Python getversion.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a graminit.o) no debug information available for Users ronald Projects python r263 Python graminit.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a import.o) no debug information available for Users ronald Projects python r263 Python import.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a importdl.o) no debug information available for Users ronald Projects python r263 Python importdl.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a marshal.o) no debug information available for Users ronald Projects python r263 Python marshal.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a modsupport.o) no debug information available for Users ronald Projects python r263 Python modsupport.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a mystrtoul.o) no debug information available for Users ronald Projects python r263 Python mystrtoul.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a mysnprintf.o) no debug information available for Users ronald Projects python r263 Python mysnprintf.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a peephole.o) no debug information available for Users ronald Projects python r263 Python peephole.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pyarena.o) no debug information available for Users ronald Projects python r263 Python pyarena.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pyfpe.o) no debug information available for Users ronald Projects python r263 Python pyfpe.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pystate.o) no debug information available for Users ronald Projects python r263 Python pystate.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pythonrun.o) no debug information available for Users ronald Projects python r263 Python pythonrun.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a structmember.o) no debug information available for Users ronald Projects python r263 Python structmember.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a symtable.o) no debug information available for Users ronald Projects python r263 Python symtable.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a sysmodule.o) no debug information available for Users ronald Projects python r263 Python sysmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a traceback.o) no debug information available for Users ronald Projects python r263 Python traceback.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getopt.o) no debug information available for Users ronald Projects python r263 Python getopt.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pystrcmp.o) no debug information available for Users ronald Projects python r263 Python pystrcmp.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pystrtod.o) no debug information available for Users ronald Projects python r263 Python pystrtod.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a formatter unicode.o) no debug information available for Users ronald Projects python r263 Python formatter unicode.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a formatter string.o) no debug information available for Users ronald Projects python r263 Python formatter string.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a dynload shlib.o) no debug information available for Users ronald Projects python r263 Python dynload shlib.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a mactoolboxglue.o) no debug information available for Users ronald Projects python r263 Python mactoolboxglue.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a thread.o) no debug information available for Users ronald Projects python r263 Python thread.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a config.o) no debug information available for Modules config.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a getpath.o) no debug information available for Users ronald Projects python r263 Modules getpath.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a main.o) no debug information available for Users ronald Projects python r263 Modules main.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a gcmodule.o) no debug information available for Users ronald Projects python r263 Modules gcmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a threadmodule.o) no debug information available for Users ronald Projects python r263 Modules threadmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a signalmodule.o) no debug information available for Users ronald Projects python r263 Modules signalmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a posixmodule.o) no debug information available for Users ronald Projects python r263 Modules posixmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a errnomodule.o) no debug information available for Users ronald Projects python r263 Modules errnomodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a pwdmodule.o) no debug information available for Users ronald Projects python r263 Modules pwdmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a sre.o) no debug information available for Users ronald Projects python r263 Modules sre.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a codecsmodule.o) no debug information available for Users ronald Projects python r263 Modules codecsmodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a zipimport.o) no debug information available for Users ronald Projects python r263 Modules zipimport.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a symtablemodule.o) no debug information available for Users ronald Projects python r263 Modules symtablemodule.c .
|
|
warning Could not find object file private tmp py bld python libpython2.6.a xxsubtype.o) no debug information available for Users ronald Projects python r263 Modules xxsubtype.c .
|
|
warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a muldi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a negdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a lshrdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ashldi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ashrdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a cmpdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ucmpdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a floatdidf.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a floatdisf.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunsdfsi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunssfsi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunsdfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixdfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunssfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixsfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixxfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunsxfdi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a floatdixf.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a fixunsxfsi.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a clear cache.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a enable execute stack.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a main.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a absvsi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a absvdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a addvsi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a addvdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a subvsi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a subvdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a mulvsi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a mulvdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a negvsi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a negvdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ffssi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ffsdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a clz.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a clzsi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a clzdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ctzsi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a ctzdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a popcount tab.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a popcountsi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a popcountdi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a paritysi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a paritydi2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a powisf2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a powidf2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a powixf2.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a mulsc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a muldc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a mulxc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a divsc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a divdc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a divxc3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a eprintf.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a gcc bcmp.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a divdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a moddi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a udivdi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a umoddi3.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a udiv w sdiv.o) more recent than executable timestamp warning .o file Developer SDKs MacOSX10.4u.sdk usr lib gcc i686 apple darwin9 4.0.1 libgcc.a udivmoddi4.o) more recent than executable timestamp .
|
|
done Traceback most recent call last) File <string> , line 1, in <module> File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy init .py , line 132, in <module> import add newdocs File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy add newdocs.py , line 9, in <module> from lib import add newdoc File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib init .py , line 4, in <module> from type check import File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core init .py , line 5, in <module> import multiarray ImportError dlopen Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core multiarray.so unknown required load command 0x80000022 Program exited with code 01. gdb) bt No stack.
|
|
gdb) c The program is not being run.
|
|
gdb)","bug"
|
|
"6013","1933","segfault in vdot Trac 1335).
|
|
Original ticket http projects.scipy.org numpy ticket 1335 on 2009 12 17 by nbecker, assigned to unknown.
|
|
It appears vdot does not check the length of the input vectors.","bug"
|
|
"6013","1934","Fix for 1299 exposes Bus error on Sparc Trac 1336).
|
|
Original ticket http projects.scipy.org numpy ticket 1336 on 2009 12 18 by drnlm, assigned to unknown.
|
|
The test case for the fix for 1299 results in a bus error on sparc in nonzero.
|
|
I haven t looked closely at why yet, but it looks like a existing bug that s being exposed by the test, rather than anything introduced by the fix.","bug"
|
|
"6013","1935","error npy cdouble definition is not compatible with C99 complex definition Trac 1337).
|
|
Original ticket http projects.scipy.org numpy ticket 1337 on 2009 12 18 by trac user mayzel, assigned to unknown.
|
|
I ve got the following errors after manual installation of numpy 1.4.0rc1 home nmr5 local lib python2.6 site packages numpy core include numpy npy common.h 120 2 error npy cdouble definition is not compatible with C99 complex definition !
|
|
Please contact Numpy maintainers and give detailed information about your compiler and platform home nmr5 local lib python2.6 site packages numpy core include numpy npy common.h 127 2 error npy cfloat definition is not compatible with C99 complex definition !
|
|
Please contact Numpy maintainers and give detailed information about your compiler and platform home nmr5 local lib python2.6 site packages numpy core include numpy npy common.h 134 2 error npy clongdouble definition is not compatible with C99 complex definition !
|
|
Please contact Numpy maintainers and give detailed information about your compiler and platform In file included from home nmr5 local lib python2.6 site packages numpy core include numpy ndarrayobject.h 1258, from home nmr5 local lib python2.6 site packages numpy core include numpy arrayobject.h 14, from cov.c 144 home nmr5 local lib python2.6 site packages numpy core include numpy multiarray api.h In function ` import array home nmr5 local lib python2.6 site packages numpy core include numpy multiarray api.h 1202 error `NPY ABI VERSION undeclared first use in this function) home nmr5 local lib python2.6 site packages numpy core include numpy multiarray api.h 1202 error Each undeclared identifier is reported only once home nmr5 local lib python2.6 site packages numpy core include numpy multiarray api.h 1202 error for each function it appears in.)
|
|
home nmr5 local lib python2.6 site packages numpy core include numpy multiarray api.h 1208 error `NPY API VERSION undeclared first use in this function Python version and numpy configuration are Python 2.6.4 r264 75706, Dec 14 2009, 16 29 30) [GCC 3.4.6 20060404 Red Hat 3.4.6 8)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.show config ) atlas threads info libraries = [ lapack , clapack , f77blas , cblas , atlas ] library dirs = [ home nmr5 local lib ] language = f77 include dirs = [ home nmr5 local include ] blas opt info libraries = [ clapack , f77blas , cblas , atlas ] library dirs = [ home nmr5 local lib ] define macros = [ ATLAS INFO , \\ 3.9.19\\ )] language = c include dirs = [ home nmr5 local include ] atlas blas threads info libraries = [ clapack , f77blas , cblas , atlas ] library dirs = [ home nmr5 local lib ] language = c include dirs = [ home nmr5 local include ] lapack opt info libraries = [ lapack , clapack , f77blas , cblas , atlas ] library dirs = [ home nmr5 local lib ] define macros = [ ATLAS INFO , \\ 3.9.19\\ )] language = f77 include dirs = [ home nmr5 local include ] lapack mkl info NOT AVAILABLE blas mkl info NOT AVAILABLE mkl info NOT AVAILABLE gcc is nmr5 gly.nmr.gu.se home nmr5 sampler> gcc v Reading specs from usr lib gcc x86 64 redhat linux 3.4.6 specs Configured with .. configure prefix= usr mandir= usr share man infodir= usr share info enable shared enable threads=posix disable checking with system zlib enable cxa atexit disable libunwind exceptions enable java awt=gtk host=x86 64 redhat linux Thread model posix gcc version 3.4.6 20060404 Red Hat 3.4.6 8) I m using CentOS release 4.5 Final) on cluster with Linux version 2.6.9 55.ELsmp mockbuild builder6.centos.org) gcc version 3.4.6 20060404 Red Hat 3.4.6 8)) 1 SMP Wed May 2 14 04 42 EDT 2007","bug"
|
|
"6013","1936","importing scipy.stats distributions throws an error Trac 1338).
|
|
Original ticket http projects.scipy.org numpy ticket 1338 on 2009 12 18 by trac user island quant, assigned to unknown.
|
|
Python 2.6.4 Numpy 1.4.0rc1 Scipy 0.7.1 See the first command from scipy.stats import distributions.
|
|
It fails.
|
|
But I can immediately run it again and it works.
|
|
So I am wrapping the call in a try except clause then importing again in the except segment.
|
|
Here is the simplest session in pylab.
|
|
Python 2.6.4 r264 75708, Oct 26 2009, 08 23 19) [MSC v.1500 32 bit Intel)] Type copyright , credits or license for more information.
|
|
IPython 0.10 An enhanced Interactive Python.
|
|
?
|
|
> Introduction and overview of IPython s features.
|
|
quickref > Quick reference.
|
|
help > Python s own help system.
|
|
object?
|
|
> Details about object .
|
|
?object also works, ??
|
|
prints more.
|
|
Welcome to pylab, a matplotlib based Python environment.
|
|
For more information, type help pylab) .
|
|
In [1] from scipy.stats import distributions ValueError Traceback most recent call last) M \<ipython console> in <module> ) C \Python26\lib\site packages\scipy\stats\ init .py in <module> ) 5 from info import doc 6 > 7 from stats import 8 from distributions import 9 from rv import C \Python26\lib\site packages\scipy\stats\stats.py in <module> ) 201 202 import scipy.stats is this a circular import ?
|
|
> 203 from morestats import find repeats is only reference to scipy.stats 204 import distributions 205 C \Python26\lib\site packages\scipy\stats\morestats.py in <module> ) 5 import statlib 6 import stats > 7 import distributions 8 from numpy import isscalar, r , log, sum, around, unique, asarray 9 from numpy import zeros, arange, sort, amin, amax, any, where, \ C \Python26\lib\site packages\scipy\stats\distributions.py in <module> ) 25 from scipy.special import gammaln as gamln 26 from copy import copy > 27 import vonmises cython 28 import textwrap 29 M \numpy.pxd in scipy.stats.vonmises cython scipy\stats\vonmises cython.c 2939) ) ValueError numpy.dtype does not appear to be the correct type object In [2]","bug"
|
|
"6013","1937","segfault in argsort, sort Trac 1339).
|
|
Original ticket http projects.scipy.org numpy ticket 1339 on 2009 12 18 by kwgoodman, assigned to unknown.
|
|
np.array 121).argsort 0).argsort 0) Segmentation fault","bug"
|
|
"6013","1938","Searchsorted doesn t work with scalars.
|
|
Trac 1340).
|
|
Original ticket http projects.scipy.org numpy ticket 1340 on 2009 12 19 by charris, assigned to unknown.
|
|
In [1] int64 0).searchsorted 0) ValueError Traceback most recent call last) home charris <ipython console> in <module> ) ValueError object of too small depth for desired array","bug"
|
|
"6013","1939","remove doc strings for matrix attributes .T, .A, ... Trac 1341).
|
|
Original ticket http projects.scipy.org numpy ticket 1341 on 2009 12 19 by josef pkt, assigned to pv.
|
|
The properties T, A, ... of matrices, do not pick up the docstrings form the corresponding getxxx functions.
|
|
as reported on mailing list) this is a documentation improvement that needs source edit in defmatrix.py T = property getT, None, doc= transpose ) A = property getA, None, doc= base array ) A1 = property getA1, None, doc= 1 d base array ) H = property getH, None, doc= hermitian conjugate) transpose ) I = property getI, None, doc= inverse ) could take the docstrings from the getxxx function instead ?
|
|
Ralf That seems like a good idea.
|
|
All that is needed is to delete the doc arguments for those five properties.","bug"
|
|
"6013","1940","pickling fails on some masked arrays Trac 1342).
|
|
Original ticket http projects.scipy.org numpy ticket 1342 on 2009 12 19 by pierregm, assigned to pierregm.
|
|
The current pickling process fails sometimes >>> base = ma.array [{ a 1}]) >>> target = cPickle.loads cPickle.dumps base)) ... numpy ma core.pyc in setstate self, state) 5408 5409 , shp, typ, isf, raw, msk, flv) = state > 5410 ndarray.
|
|
setstate self, shp, typ, isf, raw)) 5411 self.
|
|
mask.
|
|
setstate shp, make mask descr typ), isf, msk)) 5412 self.fill value = flv Using >>> self.
|
|
data.tolist ), instead of >>> self.
|
|
data.tostring cf) in MaskedArray.
|
|
getstate solves that particular issue, but breaks the other pickle.
|
|
There s a need for a whole new approach for pickling masked arrays","bug"
|
|
"6013","1941","ImportError from umath.so on OSX 10.6 Trac 1343).
|
|
Original ticket http projects.scipy.org numpy ticket 1343 on 2009 12 21 by fonnesbeck, assigned to unknown.
|
|
For the past several weeks, I have been unable to build a working version of Numpy from svn checkout on 64 bit Python on OSX 10.6.
|
|
The build itself is successful, but the resulting package gives an import error In [1] import numpy Traceback most recent call last) File <ipython console> , line 1, in <module> File Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy init .py , line 132, in <module> import add newdocs File Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy lib init .py , line 4, in <module> from type check import File Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy core init .py , line 6, in <module> import umath ImportError dlopen Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy core umath.so, 2) Symbol not found npy cexp Referenced from Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy core umath.so Expected in flat namespace in Library Python 2.6 site packages numpy 1.5.0.dev py2.6 macosx 10.6 universal.egg numpy core umath.so This has been occurring since at least rev 7726.
|
|
Here is my build script !
|
|
bin sh export MACOSX DEPLOYMENT TARGET=10.6 export CFLAGS= arch i386 arch x86 64 export FFLAGS= m32 m64 export LDFLAGS= Wall undefined dynamic lookup bundle arch i386 arch x86 64 export PYTHONPATH= Library Python 2.6 site packages python setupegg.py config fc fcompiler gfortran config L Users fonnesbeck Code libs build bdist egg Following a suggestion, I have also tried without setting LDFLAGS, but the same result occurs.
|
|
Here is the config.h that is generated, if that helps http files.me.com fonnesbeck d9eyxi","bug"
|
|
"6013","1942","reshape should be able to figure our remainder dimensions Trac 1344).
|
|
Original ticket http projects.scipy.org numpy ticket 1344 on 2009 12 22 by inducer, assigned to unknown.
|
|
arange 25).reshape 5, None) should do the right thing by default, namely pick the second dimension as 5 .","enhancement"
|
|
"6013","1943","test multiarray.TestIO.test ascii segmentation fault with Python 2.7 Trac 1345).
|
|
Original ticket http projects.scipy.org numpy ticket 1345 on 2009 12 23 by bsouthey, assigned to unknown.
|
|
Initially this was found with the SVN 1.5 series but also occurs in the 1.4 rc branch.
|
|
The discussion about the SVN issues is [http mail.scipy.org pipermail numpy discussion 2009 December 047481.html] From the test, the following should reproduce the crash import numpy as np from numpy.compat import asbytes, getexception np.fromstring asbytes 1 , 2 , 3 , 4 ),sep= , )","bug"
|
|
"6013","1944","ERROR Test filled w mvoid Trac 1346).
|
|
Original ticket http projects.scipy.org numpy ticket 1346 on 2009 12 31 by stsci sienkiew, assigned to unknown.
|
|
ERROR Test filled w mvoid Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy ma tests test core.py , line 506, in test filled w mvoid a = mvoid np.array 1, 2)), mask=[ 0, 1)], dtype=ndtype) File usr stsci pyssgdev 2.5.4 numpy ma core.py , line 5454, in new data = ndarray.
|
|
new self, ), dtype=dtype, buffer=data.data) TypeError buffer is too small for requested array NumPy version 1.5.0.dev8039 Python 2.5.4 happens on Red Hat Enterprise 4 32 bit, Solaris 32 bit, Mac OSX Leopard 32 bit does not happen on RHE 4 64 bit, RHE 5 64 bit","bug"
|
|
"6013","1945","Update the Numpy documentation site to reflect the 1.4 release Trac 1347).
|
|
Original ticket http projects.scipy.org numpy ticket 1347 on 2010 01 05 by trac user petar, assigned to pv.
|
|
http docs.scipy.org doc hosts documentation only for 1.3 and 1.5 dev versions.","bug"
|
|
"6013","1946","CHM of Numpy Reference Guide development version) is outdated Trac 1348).
|
|
Original ticket http projects.scipy.org numpy ticket 1348 on 2010 01 05 by trac user petar, assigned to pv.
|
|
[http docs.scipy.org doc numpy numpy chm.zip CHM of Numpy Reference Guide development version] was last updated on March 7th, 2009.
|
|
There s probably a problem with the automatic doc builds.","bug"
|
|
"6013","1947","Error when calling argsort on masked array Trac 1349).
|
|
Original ticket http projects.scipy.org numpy ticket 1349 on 2010 01 06 by trac user pschella, assigned to pierregm.
|
|
When calling argsort on a masked array the following error is produced Traceback most recent call last) File argsort error.py , line 13, in <module> print np.argsort x) error File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core fromnumeric.py , line 573, in argsort return argsort axis, kind, order) File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy ma core.py , line 4034, in argsort return d.argsort axis=axis, kind=kind, order=order) TypeError expected string or Unicode object, NoneType found","bug"
|
|
"6013","1948","Pickling error with numpy.void object Trac 1350).
|
|
Original ticket http projects.scipy.org numpy ticket 1350 on 2010 01 06 by wesm, assigned to unknown.
|
|
I encountered this error after pickling a record from a structured array with NumPy 1.3.
|
|
I was able to reproduce it with NumPy 1.4 Please see the attached npy file with the data x = np.load record.npy ) arr = np.array [x.item )], dtype= i8,f4,i4 ) cPickle.loads cPickle.dumps arr[0]))","bug"
|
|
"6013","1949","Pickling error with numpy.void object Trac 1351).
|
|
Original ticket http projects.scipy.org numpy ticket 1351 on 2010 01 06 by wesm, assigned to unknown.
|
|
I encountered this error after pickling a record from a structured array with NumPy 1.3.
|
|
I was able to reproduce it with NumPy 1.4 Please see the attached npy file with the data x = np.load record.npy ) arr = np.array [x.item )], dtype= i8,f4,i4 ) cPickle.loads cPickle.dumps arr[0]))","bug"
|
|
"6013","1950","clip out=...) doesn t respect write=False Trac 1352).
|
|
Original ticket http projects.scipy.org numpy ticket 1352 on 2010 01 06 by trac user quarl, assigned to unknown.
|
|
x = arange 5) y = arange 5) y.setflags write=False) x.clip 2, 3, out=y) Bad result wrote to y which should have been read only Wanted result `ValueError cannot copy back to a read only array` This does work as expected when x and y have different dtypes.)","bug"
|
|
"6013","1951","doc server error for numpy.var Trac 1353).
|
|
Original ticket http projects.scipy.org numpy ticket 1353 on 2010 01 08 by timmie, assigned to pv.
|
|
The following URL shows an 500 internal server error http docs.scipy.org doc numpy reference generated numpy.var.html","bug"
|
|
"6013","1952","numpy.savez should support zip compression Trac 1354).
|
|
Original ticket http projects.scipy.org numpy ticket 1354 on 2010 01 09 by jpaalasm, assigned to unknown.
|
|
It would be useful if the compression argument for zipfile.
|
|
!ZipFile could be specified when creating an npz file.","enhancement"
|
|
"6013","1953","Patch for pyton 2.7 distutils Trac 1355).
|
|
Original ticket http projects.scipy.org numpy ticket 1355 on 2010 01 10 by trac user bicatali, assigned to cournape.
|
|
Here is a small patch from our python team at Gentoo to install numpy with python 2.7 http sources.gentoo.org viewcvs.py checkout gentoo x86 dev python numpy files numpy 1.4.0 python 2.7.patch","bug"
|
|
"6013","1954","loadtxt ) leaks memory uncollectible gzip object) Trac 1356).
|
|
Original ticket http projects.scipy.org numpy ticket 1356 on 2010 01 10 by trac user maxy, assigned to unknown.
|
|
loadtxt ) leaves objects behind that the python garbage collector can t clean up.
|
|
Probably not a big issue, but very annoying when hunting other memory leaks.
|
|
In [1] from numpy import In [2] import gc In [3] gc.collect ); gc.garbage Out[3] 12 Out[3] [] In [4] data = loadtxt painting30sec.dat.gz ) In [5] gc.collect ); gc.garbage Out[5] 10 Out[5] [<gzip on 0x7f81c8578d88>] In [6] data = loadtxt painting30sec.dat.gz ) In [7] gc.collect ); gc.garbage Out[7] 10 Out[7] [<gzip on 0x7f81c8578d88>, <gzip on 0x7f81c8583320>]","bug"
|
|
"6013","1955","initialization of large array produces strange result Trac 1357).
|
|
Original ticket http projects.scipy.org numpy ticket 1357 on 2010 01 12 by trac user kilian, assigned to unknown.
|
|
Initialization of a ndarray with shape 2 16,2 16) produces a strange result it doesn t seem to contain any data) > In [1] np.zeros 2 16,2 16)) > Out[1] array [], shape= 65536, 65536), dtype=float64) Initialization of a much smaller array throws an exception > In [2] np.zeros 2 14,2 14)) > ValueError dimensions too large.
|
|
array seems to be fixed in version 1.4.0","bug"
|
|
"6013","1956","broken link Trac 1358).
|
|
Original ticket http projects.scipy.org numpy ticket 1358 on 2010 01 12 by trac user kilian, assigned to unknown.
|
|
the bug report link http numpy.scipy.org bug report.html on the numpy home page http numpy.scipy.org is broken.","bug"
|
|
"6013","1957","Python crashes in matplotlib with numpy r7465 Trac 1359).
|
|
Original ticket http projects.scipy.org numpy ticket 1359 on 2010 01 12 by trac user jaraco, assigned to unknown.
|
|
I ve run into an issue with numpy and matplotlib which crashes Python.
|
|
I tracked the problem down to r7465 in subversion.
|
|
If I build and install r7464, then run python c import matplotlib.
|
|
path , everything is fine.
|
|
If I update to r7465 or later, including the 1.4.0 release, and run the same Python command, Python crashes.
|
|
I have not compiled trunk r8040) Because it has something to do with the C api, it s not clear to me where it is failing or what I can do to track down the problem.
|
|
I m using the 0.99.1 release of matplotlib, Python 2.6.4 64 bit on Windows, and numpy.
|
|
I have Visual Studio 2008 Professional installed and that s what Python is using to build numpy).
|
|
I have confirmed the problem does not exhibit itself in 32 bit Python with numpy 1.4.0 and matplotlib 0.99.1, so the problem appears to be specific to the 64 bit build.
|
|
Can someone with experience in numpy development specifically around the changes relevant to r7465) review the code.
|
|
I hope the cause is apparent and is an easy fix.
|
|
If not, please let me know how I can help track down the cause.
|
|
Can you suggest a better test case to reproduce the problem that does not involve invoking matplotlib, perhaps using ctypes or a native numpy call?","bug"
|
|
"6013","1958","numpy.seterr not working for numpy.log10 on windows Trac 1360).
|
|
Original ticket http projects.scipy.org numpy ticket 1360 on 2010 01 12 by trac user Andrew MacKeith, assigned to unknown.
|
|
On Windows, seterr all= raise ) does not raise exceptions when Overflow occurs in numpy.log10.
|
|
The following uses numpy 1.4.0.
|
|
Python 2.6.4 r264 75708, Oct 26 2009, 08 23 19) [MSC v.1500 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.seterr ) { over ignore , divide ignore , invalid ignore , under ignore } >>> numpy.int16 32000) numpy.int16 3) 30464 >>> numpy.log10 0.0) inf >>> numpy.seterr all= raise ) { over ignore , divide ignore , invalid ignore , under ignore } >>> numpy.int16 32000) numpy.int16 3) Traceback most recent call last) File <stdin> , line 1, in<module> FloatingPointError overflow encountered in short scalars >>> numpy.log10 0.0) inf >>> numpy.log10 1.0) nan >>> numpy.seterr ) { over raise , divide raise , invalid raise , under raise } >>>","bug"
|
|
"6013","1959","repr for string item within array of dtype object is missing quotes Trac 1361).
|
|
Original ticket http projects.scipy.org numpy ticket 1361 on 2010 01 13 by yarikoptic, assigned to unknown.
|
|
In numpy version shipped with Debian > python c from numpy import ; print version ; r=repr array [ s , None])); print r; eval r) 1.3.0 array [s, None], dtype=object) Traceback most recent call last) File <string> , line 1, in <module> File <string> , line 1, in <module> NameError name s is not defined as well as current trunk > python c from numpy import ; print version ; r=repr array [ s , None])); print r; eval r) 1.5.0.dev array [s, None], dtype=object) Traceback most recent call last) File <string> , line 1, in <module> File <string> , line 1, in <module> NameError name s is not defined","bug"
|
|
"6013","1960","comparison operators e.g.
|
|
==) on array with dtype object do not work Trac 1362).
|
|
Original ticket http projects.scipy.org numpy ticket 1362 on 2010 01 13 by yarikoptic, assigned to unknown.
|
|
python c import numpy as N; print N. version ; a=N.array [1, 0,1)],dtype=object); print a==1; print a == 0,1), a[1] == 0,1) results in 1.5.0.dev [ True False] [False False] True I expected last line to be [False True] True So, it works for int but doesn t work for tuple...
|
|
I guess it doesn t try to compare element by element but does smth else.","bug"
|
|
"6013","1961","Test failures due to strange pylab numpy interactions Trac 1363).
|
|
Original ticket http projects.scipy.org numpy ticket 1363 on 2010 01 14 by cournape, assigned to unknown.
|
|
numpy.test ) pass every test on 1.4.0, but if pylab is imported, numpy is modified and some unicode related tests are failing import numpy works numpy.test ) import pylab 3 unicode failures numpy.test )","bug"
|
|
"6013","1962","segfault in numpy.dot for large arrays Trac 1364).
|
|
Original ticket http projects.scipy.org numpy ticket 1364 on 2010 01 17 by trac user QuestionMark, assigned to unknown.
|
|
Hello, I ve encountered a segfault in numpy when trying to compute a dot product for two arrays see code below.
|
|
The problem only seems to occur when the arrays reach a certain size.
|
|
I m using Numpy version 1.3.0, installed via macports, on a 2.33 GHz Intel Core2 Duo Macbook Pro.
|
|
Example code import numpy as N print Demonstration of Numpy Bug print loading X random numbers) SzList = [10,20,30,40,50,60,70,80,90,100] for Sz in SzList print X size = d, d 300,Sz) X = N.random.rand 300,Sz) Y = N.random.rand 300,3) print Attempting dot product of X and Y N.dot X.T,Y) print Finished without bug.
|
|
Result run through gdb) There were a number of warnings like this so many that they went off the top of the screen and I couldn t copy them all.
|
|
This was typical of the warnings.)
|
|
Reading symbols for shared libraries warning Could not find object file opt local var macports build opt local var macports sources rsync.macports.org release ports lang python26 work Python 2.6.4 build temp.macosx 10.5 i386 2.6 opt local var macports build opt local var macports sources rsync.macports.org release ports lang python26 work Python 2.6.4 Modules collectionsmodule.o no debug information available for opt local var macports build opt local var macports sources rsync.macports.org release ports lang python26 work Python 2.6.4 Modules collectionsmodule.c .
|
|
.
|
|
done Demonstration of Numpy Bug loading X random numbers) X size = 300,10 Attempting dot product of X and Y Finished without bug.
|
|
X size = 300,20 Attempting dot product of X and Y Finished without bug.
|
|
X size = 300,30 Attempting dot product of X and Y Finished without bug.
|
|
X size = 300,40 Attempting dot product of X and Y Finished without bug.
|
|
X size = 300,50 Attempting dot product of X and Y Finished without bug.
|
|
X size = 300,60 Attempting dot product of X and Y Program received signal EXC BAD ACCESS, Could not access memory.
|
|
Reason 13 at address 0x00000000 [Switching to process 55933 thread 0x117] 0x01038884 in ATL dupMBmm0 2 0 b0 ) Help please and thanks!
|
|
), Mark","bug"
|
|
"6013","1963","numpy build on interix Trac 1365).
|
|
Original ticket http projects.scipy.org numpy ticket 1365 on 2010 01 18 by trac user mduft, assigned to cournape.
|
|
Hey!
|
|
i have a small patch to make numpy build on interix.
|
|
maybe it s not perfect, but you ll have the better solution ;) The problem here is, that there is a complex.h, but still one cannot use complex... this i have to fool setup.py into thinking there is no complex.h this makes the build go through...","enhancement"
|
|
"6013","1964","Sort fails for big endian) structured arrays Trac 1366).
|
|
Original ticket http projects.scipy.org numpy ticket 1366 on 2010 01 19 by astrofrog, assigned to unknown.
|
|
The problem was originally demonstrated on numpy discussion with the following example Python 2.6.1 r261 67515, Jul 7 2009, 23 51 51) [GCC 4.2.1 Apple Inc. build 5646)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> data = np.array [ a , 2.
|
|
), b , 4.
|
|
), d , 3.
|
|
), c , 1.
|
|
)], ... dtype=[ col1 , |S5 ), col2 , >f8 )]) >>> >>> data array [ a , 2.0), b , 4.0), d , 3.0), c , 1.0)], dtype=[ col1 , |S5 ), col2 , >f8 )]) >>> data.sort order=[ col1 ]) >>> data array [ a , 2.0), b , 4.0), c , 1.0), d , 3.0)], dtype=[ col1 , |S5 ), col2 , >f8 )]) >>> data.sort order=[ col2 ]) >>> data array [ a , 2.0), d , 3.0), b , 4.0), c , 1.0)], dtype=[ col1 , |S5 ), col2 , >f8 )]) In the case of sorting by col2 , the order is incorrect.
|
|
Another example was demonstrated by Warren Weckesser, which shows that this also fails for big endian integer columns, but that the sort is correct if the column is changed to little endian In [137] z = np.array [ 0,), 258,), 256,), 1,), 255,)], dtype=[ num , <i2 )]) In [138] z.sort order= num ) In [139] z Out[139] array [ 0,), 1,), 255,), 256,), 258,)], dtype=[ num , <i2 )]) In [140] z = np.array [ 0,), 258,), 256,), 1,), 255,)], dtype=[ num , >i2 )]) In [141] z.sort order= num ) In [142] z Out[142] array [ 255,), 0,), 256,), 1,), 258,)], dtype=[ num , >i2 )]) In [143] np.
|
|
version Out[143] 1.3.0 The error was confirmed on two separate numpy installations, one on MacOS 10.5 with Python 2.5.4, and one on MacOS 10.6 with Python 2.6.1.","bug"
|
|
"6013","1965","test core.TestMaskedView fails with python2.4 Trac 1367).
|
|
Original ticket http projects.scipy.org numpy ticket 1367 on 2010 01 21 by drnlm, assigned to unknown.
|
|
Seen on both Linux i386 with python 2.4.6 and Linux sparc with python 2.4.4.
|
|
Not seen with python 2.5 ====================================================================== ERROR test view to flexible dtype test core.TestMaskedView) Traceback most recent call last) File home neil Trees Numpy svn install2.4 lib python2.4 site packages numpy ma tests test core.py , line 3313, in test view to flexible dtype test = a[0].view [ A , float), B , float)]) File .. install2.4 lib python2.4 site packages numpy ma core.py , line 2862, in view File .. install2.4 lib python2.4 site packages numpy ma core.py , line 2786, in array finalize TypeError attribute shape of numpy.generic objects is not writable ====================================================================== ERROR test view to subdtype test core.TestMaskedView) Traceback most recent call last) File home neil Trees Numpy svn install2.4 lib python2.4 site packages numpy ma tests test core.py , line 3334, in test view to subdtype test = a[0].view float, 2)) File .. install2.4 lib python2.4 site packages numpy ma core.py , line 2862, in view File .. install2.4 lib python2.4 site packages numpy ma core.py , line 2786, in array finalize TypeError attribute shape of numpy.generic objects is not writable","bug"
|
|
"6013","1966","numpy.ma sort method sorts in place but dosctring says it returns a sorted copy Trac 1368).
|
|
Original ticket http projects.scipy.org numpy ticket 1368 on 2010 01 21 by trac user bsulman, assigned to pierregm.
|
|
The dosctring to the sort method in numpy.ma says Return a sorted copy of an array, the same behavior as the numpy sort method.
|
|
However, the numpy.ma sort method sorts arrays in place, unlike the numpy sort method and in contradiction to the documentation.
|
|
The code should be changed to match the numpy code, or the docstring should be changed to accurately reflect the behavior.","bug"
|
|
"6013","1967","do not veil proper exception with bogus one in load library Trac 1369).
|
|
Original ticket http projects.scipy.org numpy ticket 1369 on 2010 01 21 by yarikoptic, assigned to unknown.
|
|
I ve got an exception thrown at me on a Linux system OSError home yhalchen proj pymvpa pymvpa build install lib python2.6 site packages mvpa clfs libsmlrc smlrc.pyd cannot open shared object file No such file or directory I was surprised to see that it looks for dynamic windows library... looking at code for ln in libname ext try libpath = os.path.join libdir, ln) return ctypes.cdll[libpath] except OSError, e pass shows the issue... actual sensible exception was OSError home yhalchen proj pymvpa pymvpa build install lib python2.6 site packages mvpa clfs libsmlrc smlrc.so undefined symbol Py InitModule4 64 so it might be sensible to make it more platform specific and or check for present files","enhancement"
|
|
"6013","1968","numpy.load ) crashes on invalid data Trac 1370).
|
|
Original ticket http projects.scipy.org numpy ticket 1370 on 2010 01 21 by trac user dhelfman, assigned to unknown.
|
|
>>> import cStringIO as io >>> import numpy as np >>> np.load io.StringIO 0 ) ) When I run this, the Python interpreter crashes and exits unceremoniously.
|
|
Note that if I change the 0 to a 1 , I get a proper IOError exception and no crash.
|
|
This is in MSYS on Windows XP with Python 2.6.4 from python.org, running numpy 1.4.0 built with MSVC 9.0 gdb python GNU gdb 5.2.1 Copyright 2002 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and or distribute copies of it under certain conditions.
|
|
Type show copying to see the conditions.
|
|
There is absolutely no warranty for GDB.
|
|
Type show warranty for details.
|
|
This GDB was configured as i686 pc mingw32 ... no debugging symbols found)... gdb) r c import cStringIO as IO ; import numpy as np ; np.load IO.StringIO 0 ) ) Starting program c \python26 python.exe c import cStringIO as io ; import numpy as np ; np.load io.StringIO 0 ) ) Program received signal SIGSEGV, Segmentation fault.
|
|
0x1e09ada3 in ?? )
|
|
gdb) bt 0 0x1e09ada3 in ?? )
|
|
1 0x1e0e1fd5 in ?? )
|
|
2 0x00a4e530 in ?? )
|
|
3 0x1e1d1d30 in ?? )
|
|
Cannot access memory at address 0x5 And this is on Mac OS X 10.4 with Python 2.6.4 from python.org, running numpy 1.4.0 built with gcc 4.0.1 from Apple gdb python GNU gdb 5.2.1 Copyright 2002 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and or distribute copies of it under certain conditions.
|
|
Type show copying to see the conditions.
|
|
There is absolutely no warranty for GDB.
|
|
Type show warranty for details.
|
|
This GDB was configured as i686 pc mingw32 ... no debugging symbols found)... gdb) r c import cStringIO as IO ; import numpy as np ; np.load IO.StringIO 0 ) ) Starting program c \python26 python.exe c import cStringIO as IO ; import numpy as np ; np.load IO.StringIO 0 ) ) Program received signal SIGSEGV, Segmentation fault.
|
|
0x1e09ada3 in ?? )
|
|
gdb) bt 0 0x1e09ada3 in ?? )
|
|
1 0x1e0e1fd5 in ?? )
|
|
2 0x00a4e530 in ?? )
|
|
3 0x1e1d1d30 in ?? )
|
|
Cannot access memory at address 0x5","bug"
|
|
"6013","1969","compilation fails on non Intel CPUs Trac 1371).
|
|
Original ticket http projects.scipy.org numpy ticket 1371 on 2010 01 22 by trac user tanner, assigned to unknown.
|
|
Numpy 1.4 fails to compile on non Intel architectures here ARM on Maemo Nokia N900) compile options Inumpy core include I home tanner tmp numpy 1.4.0 .
|
|
build src.linux armv5tel 2.5 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr include python2.5 I home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src multiarray I home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src umath c gcc numpy core src multiarray multiarraymodule onefile.c In file included from numpy core src multiarray multiarraymodule onefile.c 11 numpy core src multiarray scalartypes.c.src In function voidtype setfield numpy core src multiarray scalartypes.c.src 1208 warning implicit declaration of function NPY COPY PYOBJECT PTR NPY COPY PYOBJECT PTR is not defined because NPY SIZEOF PY INTPTR T was not defined before including numpy core include numpy npy cpu.h my patch fixes the includes and triggers an error if NPY SIZEOF PY INTPTR T is not defined.
|
|
with this I get compile options Inumpy core include I home tanner tmp numpy 1.4.0 .
|
|
build src.linux armv5tel 2.5 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr include python2.5 I home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src multiarray I home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src umath c gcc home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src npymath npy math complex.c gcc home tanner tmp numpy 1.4.0 build src.linux armv5tel 2.5 numpy core src npymath npy math.c In file included from numpy core src private npy fpmath.h 7, from numpy core src npymath npy math private.h 25, from numpy core src npymath npy math.c.src 56 numpy core include numpy npy cpu.h 99 3 error error Unknown CPU which my patch to numpy core src private npy fpmath.h fixes.","bug"
|
|
"6013","1970","Intel Visual Fortran 11.1 not recognized by numpy.distutils on Windows Trac 1372).
|
|
Original ticket http projects.scipy.org numpy ticket 1372 on 2010 01 24 by cgohlke, assigned to cournape.
|
|
On Windows with Intel Visual Fortran Compiler Pro 11.1 installed, the distutils module of numpy 1.4.0 fails to compile Fortran Python extensions error extension ... has Fortran sources but no Fortran compiler found warning build ext f77 compiler=... is not available.
|
|
The attached patch contains fixes for using the 32 and 64 bit Intel Visual Fortran 11.1 compilers.
|
|
It allows building many scikits and the PyMC packages on Windows.","bug"
|
|
"6013","1971","mvoid.
|
|
str fails with python2.4 Trac 1373).
|
|
Original ticket http projects.scipy.org numpy ticket 1373 on 2010 01 25 by drnlm, assigned to unknown.
|
|
The following code fragment import numpy as np from numpy.ma.core import array, mvoid ndtype = [ a , int), b , float)] a = mvoid np.array 1.0, 2.0, 3.0, 4.0)), mask= 0, 1), dtype=ndtype) str a) works fine with python2.5, but fails with python2.4 Traceback most recent call last) File .
|
|
demo2.py , line 6, in ?
|
|
str a) File home neil Trees Numpy svn install2.4 lib python2.4 site packages numpy ma core.py , line 5470, in str r = [ str ), p)[ m] for , m) in zip self.
|
|
data.tolist ), tuple m))] TypeError tuple indices must be integers In 2.4, numpy.bool isn t converted to an appropriate type for use as an index.
|
|
The attached patch adds an explicit cast to int to solve this.","bug"
|
|
"6013","1972","Ticket 628 not fixed for Solaris polyfit uses 100 CPU and does not stop) Trac 1374).
|
|
Original ticket http projects.scipy.org numpy ticket 1374 on 2010 01 26 by trac user michelkeijzers, assigned to unknown.
|
|
The exact contents of ticket 628 are not fixed for the Solaris version Python 2.5.1 r251 54863, Aug 25 2008, 09 23 26) [GCC 3.4.3] on sunos5 >>> import numpy >>> numpy.version.version 1.2.1 >>> numpy.polyfit [0.0], [ numpy.Inf], 2) This command results in 100 cpu load and the process cannot be stopped.","bug"
|
|
"6013","1973","min ) and max ) accept and return dictionaries Trac 1375).
|
|
Original ticket http projects.scipy.org numpy ticket 1375 on 2010 01 26 by trac user gabriel, assigned to unknown.
|
|
This is probably not desired, or is it?
|
|
In [1] numpy.min { a 45})[[BR]] Out[1] { a 45} The same goes for max ) However I see that Python s built in min ) also happily accepts dictionaries; it escapes me why.
|
|
Numpy s behavior is not restricted to dictionaries, but min ) and max ) will also happily accept esoteric user defined class instances and return them.","bug"
|
|
"6013","1974","Suspecting indentation error!
|
|
Trac 1376).
|
|
Original ticket http projects.scipy.org numpy ticket 1376 on 2010 01 27 by trac user lenard.palko, assigned to cournape.
|
|
In the file numpy distutils fcompiler gnu.py lines 303 to 308 303 if is win64 ) 304 c compiler = self.c compiler 305 if c compiler and c compiler.compiler type == msvc 306 return [] 307 else 308 raise NotImplementedError Only MS compiler supported with gfortran on win64 ) The else seems to have wrong indentation and from this reason if the platform is a win32 platform it detects it as a win64 platform and raises the nasty exception.
|
|
I found this error in the numpy 1.4.0 win32 superpack python2.5.exe distribution found for download on http sourceforge.net projects numpy files","bug"
|
|
"6013","1975","HP Fortran compiler flags incomplete Trac 1377).
|
|
Original ticket http projects.scipy.org numpy ticket 1377 on 2010 01 28 by trac user neinhues, assigned to cournape.
|
|
The flags for the HP Fortran compiler should be changed as described below numpy 1.4.0 numpy distutils fcompiler hpux.py 9,17 9,17 version pattern = r HP F90 ?P<version>[ \s ,] ) executables = { version cmd [ <F90> , version ], version cmd [ f90 , version ], compiler f77 [ f90 ], compiler fix [ f90 ], compiler f90 [ f90 ], linker so None, linker so [ ld , b ], archiver [ ar , cr ], ranlib [ ranlib ] } module dir switch = None XXX fix me module include switch = None XXX fix me pic flags = [ pic=long ] pic flags = [ Z ] def get flags self) return self.pic flags [ ppu , DD64 ] def get flags opt self)","bug"
|
|
"6013","1976","Change of Intel Fortran compiler flags Trac 1378).
|
|
Original ticket http projects.scipy.org numpy ticket 1378 on 2010 01 28 by trac user neinhues, assigned to cournape.
|
|
The flags for the Intel compiler v 11 require a change, numpy got broken on our machines with O3.
|
|
A patch is supplied.","bug"
|
|
"6013","1977","CPU ID not set correctly on PARISC Trac 1379).
|
|
Original ticket http projects.scipy.org numpy ticket 1379 on 2010 01 28 by trac user neinhues, assigned to unknown.
|
|
In numpy 1.4.0 numpy core include numpy npy cpu.h the code for setting the CPU ID does not work as intended, patch is attached","bug"
|
|
"6013","1978","endianess is not set correctly for HP UX Itanium Trac 1380).
|
|
Original ticket http projects.scipy.org numpy ticket 1380 on 2010 01 28 by trac user neinhues, assigned to unknown.
|
|
Please note that the Itanium can be configured to work in big or little endian mode, a check is required in order to set this correctly.
|
|
numpy 1.4.0 numpy core include numpy npy endian.h Patch is attached.","bug"
|
|
"6013","1979","Linker flags for MKL library are not set correctly Trac 1381).
|
|
Original ticket http projects.scipy.org numpy ticket 1381 on 2010 01 28 by trac user neinhues, assigned to cournape.
|
|
In numpy 1.4.0 numpy distutils system info.py somebody tried to link both against 32 and 64bit libraries which does not work.
|
|
A patch is attached.
|
|
Please note Using the MKL libraries with numpy only works if Python is also linked against the MKL libraries.
|
|
The reason is that the MKL libraries don t work for dl opened libraries like Python extensions","bug"
|
|
"6013","1980","Sun compiler flags not correct for 64 bits compile Trac 1382).
|
|
Original ticket http projects.scipy.org numpy ticket 1382 on 2010 01 28 by trac user neinhues, assigned to cournape.
|
|
The supplied compiler flags do not work at least they don t with the Studio 9 compilers) Suggestion Please implement a switch to configure for 32 bit or 64 bit builds.
|
|
A patch is attached.","bug"
|
|
"6013","1981","Handling of HP C compiler not correct Trac 1383).
|
|
Original ticket http projects.scipy.org numpy ticket 1383 on 2010 01 28 by trac user neinhues, assigned to cournape.
|
|
Patch is attached.
|
|
I don t know whether it is possible to distinguish between C and C in distutils as it is not safe to assume that all compiler flags valid for C are also valid for C","bug"
|
|
"6013","1982","C API to get human readable dtype description from a PyArray Descr Trac 1384).
|
|
Original ticket http projects.scipy.org numpy ticket 1384 on 2010 01 29 by cournape, assigned to unknown.
|
|
This is often useful for error messages there may already be one way, but did not find any doc.","bug"
|
|
"6013","1983","please provide objects.inv for intersphinx Trac 1385).
|
|
Original ticket http projects.scipy.org numpy ticket 1385 on 2010 01 29 by yarikoptic, assigned to pv.
|
|
intersphinx module expects to find objects.inv in the root of the documentation, ie I guess here `http docs.scipy.org doc numpy reference objects.inv` but it is not there.
|
|
SciPy reference has it available for instance under analogous url `http docs.scipy.org doc scipy reference objects.inv`","enhancement"
|
|
"6013","1984","100 times poorer performance indexing scalar record array Trac 1386).
|
|
Original ticket http projects.scipy.org numpy ticket 1386 on 2010 01 31 by trac user oscar.bristol, assigned to unknown.
|
|
Assigning to a field in a scalar record array like this B[ a ] = A[ a ] takes around 50 times as long as assigning like this B[ a ][ ] = A[ a ] and 100 times as long as assigning to a size 1 record array.
|
|
I ve attached a script test.py) that compares scalars and size 1 arrays and the two different methods of indexing above.
|
|
On my machine, the results are python test.py 100000 array, B[ a ][ ] = A[ a ] 0.25 seconds array, B[ a ] = A[ a ] 0.213 seconds scalar, B[ a ][ ] = A[ a ] 0.395 seconds scalar, B[ a ] = A[ a ] 22.0 seconds","bug"
|
|
"6013","1985","Undefined behavior for bincount for empty array Trac 1387).
|
|
Original ticket http projects.scipy.org numpy ticket 1387 on 2010 02 02 by cournape, assigned to unknown.
|
|
E.g.
|
|
import numpy as np x = np.bincount np.array [], dtype=int))","bug"
|
|
"6013","1986","numpy.frombuffer calls PyErr Format without actually throwing the error, causing bizarre errors later Trac 1388).
|
|
Original ticket http projects.scipy.org numpy ticket 1388 on 2010 02 03 by trac user baffe, assigned to unknown.
|
|
After calling numpy.frombuffer with a zero length str or array.array, numpy starts behaving very strange.
|
|
Example Python 2.5.2 r252 60911, Feb 21 2008, 13 11 45) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> buf = numpy.frombuffer , B ) >>> print numpy.array [42], B ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError data type not understood >>> print numpy.array [42], B ) Exact same line again!
|
|
[1] Leftover error that should have been raised by the previous operation?","bug"
|
|
"6013","1987","Bad documentation for np.seterr Trac 1389).
|
|
Original ticket http projects.scipy.org numpy ticket 1389 on 2010 02 04 by astrofrog, assigned to pv.
|
|
There seems to be some debate in the np.seterr documentation page, which isn t great for end users to see.
|
|
Check out the bottom of http docs.scipy.org doc numpy reference generated numpy.seterr.html around lies!!!
|
|
Also, is the problem mentioned in there a bug?
|
|
The person who wrote that should have filed a bug report and corrected the documentation...","bug"
|
|
"6013","1988","dtype.itemsize not working correctly Trac 1390).
|
|
Original ticket http projects.scipy.org numpy ticket 1390 on 2010 02 08 by astrofrog, assigned to unknown.
|
|
Should the following np.float32.itemsize not return an integer which is the number of bytes in the type?
|
|
At the moment, it is just returning In [3] np.float32.itemsize Out[3] <attribute itemsize of numpy.generic objects> but in older version of NumPy, an integer was returned.
|
|
using r8088)","bug"
|
|
"6013","1989","Fill value and empty arrays Trac 1391).
|
|
Original ticket http projects.scipy.org numpy ticket 1391 on 2010 02 08 by astrofrog, assigned to pierregm.
|
|
I m not sure if this is a bug report or feature request since I m not sure what the intended behavior is.
|
|
When a masked structured array field is copied from one array to another, the masks are copied, but I would expect the fill value to be copied too which it isn t).
|
|
Furthermore, ma.empty and ma.zeros don t accept a fill value= argument.
|
|
Therefore, one is forced to call set fill value after copying the field.
|
|
For example import numpy.ma as ma import numpy as np a = ma.array [ 1.,), 2.,), 3.,), 4.,)], mask = [ False,), False,), False,), False,)], fill value = np.nan,), dtype = [ ra , <f8 )]) newdtype = np.dtype [ ra , <f8 ), dec , <f8 )]) b = ma.zeros a.shape, dtype=newdtype) b[ ra ] = a[ ra ] print repr b) produces masked array data = [ 1.0, 0.0) 2.0, 0.0) 3.0, 0.0) 4.0, 0.0)], mask = [ False, False) False, False) False, False) False, False)], fill value = 1e 20, 1e 20), dtype = [ ra , <f8 ), dec , <f8 )]) when, as a user, I would expect masked array data = [ 1.0, 0.0) 2.0, 0.0) 3.0, 0.0) 4.0, 0.0)], mask = [ False, False) False, False) False, False) False, False)], fill value = nan, 1e 20), dtype = [ ra , <f8 ), dec , <f8 )]) And the only way to get the fill value correct is to explicitly add b[ ra ].set fill value a[ ra )) When copying whole fields, should the fill value for that field not be copied too?","bug"
|
|
"6013","1990","numpy has a directory under python site packages named core linux removes core Trac 1392).
|
|
Original ticket http projects.scipy.org numpy ticket 1392 on 2010 02 09 by trac user schaefer0, assigned to unknown.
|
|
I install numpy 1.4.0) in the day, every day.
|
|
I log out in the evening, every evening, and by next morning numpy s core directory has been deleted.
|
|
So I get clever and chmod on core to write protect it.
|
|
No go.
|
|
Still deleted.
|
|
Arggghhh!
|
|
Red hat linux uname a returns Linux kronos 2.6.18 164.11.1.el5xen 1 SMP Wed Jan 6 14 01 18 EST 2010 i686 i686 i386 GNU Linux","bug"
|
|
"6013","1991","assignment of float to int array causes loss of decimal part Trac 1393).
|
|
Original ticket http projects.scipy.org numpy ticket 1393 on 2010 02 09 by lebigot, assigned to unknown.
|
|
An assignment such as my array[0] = 5.4 does not produce a warning when my array is an array of integers.
|
|
This is different from what my array[0] = 5.4j does when assigning to an array of floats a warning is issued, in this case which I think is very much appropriate, as explicit is better than implicit ).
|
|
It would be better if both behaviors were consistent with each other issue a warning or not, for both complex in float array, and float in integer array).
|
|
And I second issuing a warning, to be consistent with import this . )","enhancement"
|
|
"6013","1992","np.isnan not working correctly with vector columns in masked structure arrays Trac 1394).
|
|
Original ticket http projects.scipy.org numpy ticket 1394 on 2010 02 09 by astrofrog, assigned to pierregm.
|
|
The following example fails import numpy as np import numpy.ma as ma arr = ma.zeros 5),dtype=[ a ,np.float32,3)]) np.isnan arr[ a ]) giving ValueError Traceback most recent call last) Users tom tmp test masked.py in <module> ) 3 4 arr = ma.zeros 5),dtype=[ a ,np.float32,3)]) > 5 np.isnan arr[ a ]) 6 7 Users tom Library Python 2.6 site packages numpy ma core.pyc in array wrap self, obj, context) 2806 Wraps the numpy array and sets the mask according to context.
|
|
2807 > 2808 result = obj.view type self)) 2809 result.
|
|
update from self) 2810 .......... Users tom Library Python 2.6 site packages numpy ma core.pyc in view self, dtype, type) 2879 pass 2880 Make sure to reset the fill value if needed > 2881 if getattr output, fill value , None) 2882 output.
|
|
fill value = None 2883 return output ValueError The truth value of an array with more than one element is ambiguous.
|
|
Use a.any ) or a.all ) This is in NumPy r8095","bug"
|
|
"6013","1993","incorrect conversion from float128 to long Trac 1395).
|
|
Original ticket http projects.scipy.org numpy ticket 1395 on 2010 02 09 by trac user amcnabb, assigned to unknown.
|
|
I created a float128 and confirmed that it indeed holds a 20 digit number.
|
|
However, when I converted it to a Python long which should be able to hold an arbitrarily large integer), the result was only as accurate as a float64.
|
|
>>> a = numpy.float128 1234567890) 10 10 1234567890) >>> a 12345678901234567890.0 >>> long a) 12345678901234567168L >>> The correct result for long a) would be 12345678901234567890L.
|
|
I am using numpy 1.3.0 6.fc12.x86 64 and python 2.6.2 2.fc12.x86 64.","bug"
|
|
"6013","1994","Numpy product of zero length array returns 1.0 Trac 1396).
|
|
Original ticket http projects.scipy.org numpy ticket 1396 on 2010 02 10 by trac user xdylanm, assigned to unknown.
|
|
>>> numpy.prod []) 1.0 or >>> a = numpy.array []) >>> numpy.prod a) 1.0 This behaviour seems counterintuitive.
|
|
Either an error should be raised nothing to multiply) or it should return zero.
|
|
Workaround test the length of the array before passing to prod.
|
|
numpy version 1.3.0","enhancement"
|
|
"6013","1995","scientific notation with a D instead of E on mac OSX Trac 1397).
|
|
Original ticket http projects.scipy.org numpy ticket 1397 on 2010 02 12 by trac user ricitron, assigned to unknown.
|
|
In numpy.float64 ) please convert [dD] to [eE] When running python 2.5) and numpy 1.4) on mac I get >>> numpy.float64 1.23D 04 ) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError setting an array element with a sequence.","bug"
|
|
"6013","1996","augmented assignment statements do not handle exceptions in object.
|
|
float ) correctly Trac 1398).
|
|
Original ticket http projects.scipy.org numpy ticket 1398 on 2010 02 13 by b45ch1, assigned to unknown.
|
|
Best explained at an example import numpy print numpy.
|
|
version = ,numpy.
|
|
version class adouble def init self,x) self.x = x def mul self,other) if not isinstance other, self.
|
|
class ) return self.
|
|
class self.x other) return self.
|
|
class self.x other.x) def rmul self,other) return self other def float self) raise Exception this is not possible ) def str self) return str self.x) x = numpy.array [adouble 1.
|
|
), adouble 2.
|
|
), adouble 3.)])
|
|
y = numpy.array [4.,5.,6.])
|
|
this should call the float method of adouble which would raise an Exception but it does not!
|
|
y = x Output b45ch1 shlp workspace numpy bugs python bug casting from object to dtype issue.py numpy.
|
|
version = 1.5.0.dev","bug"
|
|
"6013","1997","distutils fails to build ppc64 support on Mac OS X when requested Trac 1399).
|
|
Original ticket http projects.scipy.org numpy ticket 1399 on 2010 02 18 by dwf, assigned to rgommers.
|
|
Stuff including Fortran code ends up missing the ppc64 architecture even when Python supports it, due to a small oversight in the detection code.
|
|
Patch attached.","bug"
|
|
"6013","1998","atan2 regression with Sun Studio Compilers on Linux Trac 1400).
|
|
Original ticket http projects.scipy.org numpy ticket 1400 on 2010 02 18 by trac user ChristianMarquardt, assigned to unknown.
|
|
When compiling numpy 1.4.0 with the Sun Studio Compilers v12 Update 1) on Linux an OpenSUSE 11.1 in my case), about 30 tests in numpy.test ) fail; all failures are related to the arctan2 function.
|
|
I ve found that in r7732 a patch was applied to trunk numpy core src private npy config.h in response to 1799, 1800, and 1801, undef ing the HAVE ATAN2 variable in order to fix a broken atan2 ) implementation on Solaris.
|
|
This does no good with the SunStudio 12.1 compiler on Linux, though; it apparently only works without failures with HAVE ATAN2 defined.
|
|
The attached patch ensures that the original patch is only applied on Sun platforms i.e., not on Linux).
|
|
If applied, all tests are completed successfully under Linux.
|
|
BTW, I did not observe 1802 or 1803...
|
|
I have no access to a Solaris machine, so I don t know if the original patch which was for Sun workshop 6 I think that s not available any more) is required with Sun Studio 12.1 at all.","bug"
|
|
"6013","1999","Scipy Build on AIX 5.3 error Unknown CPU, please report this to numpy maintainers Trac 1401).
|
|
Original ticket http projects.scipy.org numpy ticket 1401 on 2010 02 18 by trac user A LARAS, assigned to unknown.
|
|
Hello, I m trying to setup Scipy on AIX system with error error Unknown CPU, please report this to numpy maintainers with information about your platform OS, CPU and compiler) usr local python lib python2.4 site packages numpy core include numpy npy endian.h 41 10 error Unknown CPU can not set endianness error Command c DNDEBUG O Iscipy interpolate src I usr local python lib python2.4 site packages numpy core include I usr local python include python2.4 c scipy interpolate src interpolate.cpp o build temp.aix 5.3 2.4 scipy interpolate src interpolate.o failed with exit status 1 python setup.py config fc fcompiler=gnu95 install prefix= usr local python python 2.4 numpy 1.4.0 OS AIX 5300 7 Processor Type PowerPC POWER3 Number Of Processors 1 CPU Type 64 bit Kernel Type 64 bit F compiler gfortran gcc c 4.2.4 2 gcc cpp 4.2.4 2 gcc gfortran 4.2.4 2 gcc 4.2.4 2 libgcc 4.2.4 2 Thanks for your assistance.
|
|
Atmane.","bug"
|
|
"6013","2000","Provide a function to compute the log determinant Trac 1402).
|
|
Original ticket http projects.scipy.org numpy ticket 1402 on 2010 02 19 by njsmith, assigned to unknown.
|
|
For large matrices, computing the determinant directly may cause overflow or underflow.
|
|
But in many cases e.g., doing maximum likelihood calculations with multivariate normals), all one really wants to calculate is the log determinant, which is not subject to these numerical issues.
|
|
Attached patch adds a function sign log det to np.linalg.
|
|
The idea is that it returns the sign of the determinant as a number with absolute value 1) and the log of the absolute value.
|
|
This avoids introducing complex numbers when calculating the determinant of a real matrix and anyway the absolute value is often what s of interest anyway).
|
|
It also changes the ordinary det function to be a simple wrapper around sign log det ; as a side effect, it becomes somewhat more robust against ill conditioned matrices.
|
|
For instance, without this patch, np.linalg.det [[1e300, 0], [0, 1e 300]]) returns 0; with this patch, it correctly) returns 1.
|
|
Tests and docs included; patch is against r8128.
|
|
I d like to get this API settled, because I m adding similar functions for sparse matrices to my scikits.sparse library, and I d like to be consistent with whatever numpy does.","enhancement"
|
|
"6013","2001","long double check is failing on GNU Linux PPC Trac 1403).
|
|
Original ticket http projects.scipy.org numpy ticket 1403 on 2010 02 21 by trac user mounir.lamouri, assigned to unknown.
|
|
Since 1.4.0, numpy doesn t work anymore on GNU Linux PowerPC systems.
|
|
Indeed, numpy.core is checking the long double type and PowerPC isn t supported.
|
|
Please, look at the attached build log.","bug"
|
|
"6013","2002","complex log doesn t raise warning for invalid value Trac 1404).
|
|
Original ticket http projects.scipy.org numpy ticket 1404 on 2010 02 21 by charris, assigned to unknown.
|
|
Compare In [14] np.exp array np.nan 1j)) Warning invalid value encountered in exp Out[14] nan nan j) In [15] np.log array np.nan 1j)) Out[15] nan nan j) Note that none of the non complex versions raise warnings, I assume that is a gcc library thing.
|
|
Maybe there is a flag...","bug"
|
|
"6013","2003","NameError global name numpy is not defined Trac 1405).
|
|
Original ticket http projects.scipy.org numpy ticket 1405 on 2010 02 22 by stsci sienkiew, assigned to chanley.
|
|
File usr stsci pyssgdev 2.5.4 numpy lib utils.py , line 79, in get numarray include include dirs = get numarray include dirs ) File usr stsci pyssgdev 2.5.4 numpy numarray util.py , line 45, in get numarray include dirs base = os.path.dirname numpy.
|
|
file ) NameError global name numpy is not defined Reproduce by numpy.get numarray include ) The problem is import numpy as np ... base = os.path.dirname numpy.
|
|
file ) NumPy version 2.0.0.dev8251","bug"
|
|
"6013","2004","Should complex number have an ordering at all?
|
|
Trac 1406).
|
|
Original ticket http projects.scipy.org numpy ticket 1406 on 2010 02 23 by trac user malev, assigned to unknown.
|
|
Hi there!
|
|
If you have and array with complex values, and try to execute numpy.amax array), amax, brings a result the max number of the array, but it only search in the real part.
|
|
I think that there shouldn t be result, like it happens when you call the max methods on a simple list.
|
|
In [1] import numpy as np In [2] a list = [1 2j, 1 3j, 2 2j,1 100j, 5 1j] In [3] a np = np.array [1 2j, 1 3j, 2 2j,1 100j, 5 1j]) In [4] np.amax a np) Out[4] 5 1j) In [5] max a list) TypeError Traceback most recent call last) home malev <ipython console> in <module> ) TypeError no ordering relation is defined for complex numbers The code here http pastebin.com b8JgMNhm finally, I m using Python 2.6.4 r264 75706, Dec 7 2009, 18 45 15) and numpy 1.3.0 Regards malev","bug"
|
|
"6013","2005","rec append fields does not work for recarray of length 1 Trac 1407).
|
|
Original ticket http projects.scipy.org numpy ticket 1407 on 2010 02 24 by trac user koert, assigned to unknown.
|
|
Hello, I have a program that works with recarrays where i append a column to them using rec append fields from numpy.lib.recfunctions.
|
|
I noticed this works fine for recarrays of any length including zero) except for length 1.
|
|
I assume this is a bug?
|
|
Best, Koert Python 2.5.4 on Windows XP","bug"
|
|
"6013","2006","alter code1.changeimports does not recognize certain import strings Trac 1408).
|
|
Original ticket http projects.scipy.org numpy ticket 1408 on 2010 02 24 by trac user Andrew MacKeith, assigned to unknown.
|
|
numpy.oldnumeric.alter code1.changeimports will not recognize certain import strings that contain several names to import.
|
|
[[BR]] For example, it does not recognize import math, time, sys, os, Numeric I looked at the re pattern, and I think it needs this change to fix it.
|
|
From r import\s [ ,\n\r] ,\s ) s) name To r import\s [ ;\n\r] ,\s ) s) name AFAICT the reason for the complemented set is to ensure the name is part of the import statement, which normally ends with a ; or new line.
|
|
The pattern was previously changed in ticket 3201 by Travis.","bug"
|
|
"6013","2007","Removal of unused init posix distutils function Trac 1409).
|
|
Original ticket http projects.scipy.org numpy ticket 1409 on 2010 02 24 by bsouthey, assigned to cournape.
|
|
Numpy calls the function distutils.sysconfig.
|
|
init posix in numpy distutils ccompiler.py.
|
|
This function has been removed in the Python 2.7 alpha 3.
|
|
Also, this call is no longer used within the numpy code as per Robert Kern s response in the thread distutils problem with NumPy 1.4 Py 2.7a3 Snow Leopard) http mail.scipy.org pipermail numpy discussion 2010 February 048882.html","bug"
|
|
"6013","2008","Python2.7 requires DistutilsExecError Trac 1410).
|
|
Original ticket http projects.scipy.org numpy ticket 1410 on 2010 02 24 by bsouthey, assigned to cournape.
|
|
DistutilsExecError is not defined by Python 2.7 alpha 3 in numpy distutils ccompiler.py Traceback most recent call last) File setup.py , line 210, in <module> setup package ) File setup.py , line 203, in setup package configuration=configuration ) File home bsouthey python numpy svn numpy numpy distutils core.py , line 186, in setup return old setup new attr) File usr local lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File usr local lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File usr local lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File home bsouthey python numpy svn numpy numpy distutils command build.py , line 37, in run old build.run self) File usr local lib python2.7 distutils command build.py , line 127, in run self.run command cmd name) File usr local lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File usr local lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File home bsouthey python numpy svn numpy numpy distutils command build src.py , line 152, in run self.build sources ) File home bsouthey python numpy svn numpy numpy distutils command build src.py , line 163, in build sources self.build library sources libname info) File home bsouthey python numpy svn numpy numpy distutils command build src.py , line 298, in build library sources sources = self.generate sources sources, lib name, build info)) File home bsouthey python numpy svn numpy numpy distutils command build src.py , line 385, in generate sources source = func extension, build dir) File numpy core setup.py , line 671, in get mathlib info mlibs = check mathlib config cmd) File numpy core setup.py , line 337, in check mathlib if config cmd.check func exp , libraries=libs, decl=True, call=True) File home bsouthey python numpy svn numpy numpy distutils command config.py , line 313, in check func libraries, library dirs) File usr local lib python2.7 distutils command config.py , line 251, in try link libraries, library dirs, lang) File home bsouthey python numpy svn numpy numpy distutils command config.py , line 149, in link libraries, library dirs, lang)) File home bsouthey python numpy svn numpy numpy distutils command config.py , line 89, in wrap method ret = mth self,) args)) File usr local lib python2.7 distutils command config.py , line 148, in link target lang=lang) File usr local lib python2.7 distutils ccompiler.py , line 750, in link executable debug, extra preargs, extra postargs, None, target lang) File usr local lib python2.7 distutils unixccompiler.py , line 256, in link self.spawn linker ld args) File home bsouthey python numpy svn numpy numpy distutils ccompiler.py , line 69, in CCompiler spawn raise DistutilsExecError Command s failed with exit status d s cmd, s, msg)) NameError global name DistutilsExecError is not defined","bug"
|
|
"6013","2009","Array reshape bug [obsolete] Trac 1411).
|
|
Original ticket http projects.scipy.org numpy ticket 1411 on 2010 02 25 by trac user Han, assigned to unknown.
|
|
See also http bugs.python.org issue6562 might be related?)
|
|
When reshaping a flat ndarray with a mistaken) 64 bits long, an OverflowError will be generated in the C layer, but not propagated to the Python layer.
|
|
This error can then be caught later on, in a totally unrelated piece of code.
|
|
>>> from future import with statement >>> import sys; sys.version 2.5.1 r251 54863, Apr 18 2007, 08 51 08) [MSC v.1310 32 bit Intel)] >>> import threading >>> import numpy as np >>> needing this to reproduce the error, but it can ... also be reproduced in other ways ... lock = threading.RLock ) >>> creating a flat array ... arr = np.arange 10) >>> reshaping the array, but mistakenly using a 64 bit integer ... arr shaped = arr.reshape 1<<63) >>> all seems to go well up to this point.. ... with lock ... print done!
|
|
... C \Python25\lib\threading.py 697 RuntimeWarning tp compare didn t return 1 or 2 for exception Traceback most recent call last) File <stdin> , line 2, in <module> File c \python25\lib\threading.py , line 94, in acquire me = currentThread ) File c \python25\lib\threading.py , line 697, in currentThread return active[ get ident )] OverflowError long int too large to convert to int Other ways to reproduce the OverflowError after creating arr shaped >>> print arr shaped >>> import <anything> >>> arr shaped 10 or other operations I also tested it on Python 2.6, with similar results.","bug"
|
|
"6013","2010","Build scipy on IBM AIX with fcompiler=gnu95 Trac 1412).
|
|
Original ticket http projects.scipy.org numpy ticket 1412 on 2010 02 25 by trac user A LARAS, assigned to cournape.
|
|
Hi There[[BR]] I m trying to build scipy from the trunk latest release) [[BR]] python setup.py config fc fcompiler=gnu95 install prefix= usr local python[[BR]] but stuck with the following error [[BR]] C compiler c r qlanglvl=extc89 DNDEBUG O[[BR]] compile options Iscipy interpolate src I usr local python lib python2.4 site packages numpy core include I usr local python include python2.4 c [[BR]] c r scipy interpolate src interpolate.cpp[[BR]] sh c r not found.
|
|
[[BR]] sh c r not found.
|
|
[[BR]] error Command c r qlanglvl=extc89 DNDEBUG O Iscipy interpolate src I usr local python lib python2.4 site packages numpy core include I usr local python include python2.4 c scipy interpolate src interpolate.cpp o build temp.aix 5.3 2.4 scipy interpolate src interpolate.o failed with exit status 127[[BR]] whith export CXX= c , I hit another error syntax error)[[BR]] c c qlanglvl=extc89 bI usr local Python.2.4 lib python2.4 config python.exp build temp.aix 5.3 2.4 scipy interpolate src interpolate.o Lbuild temp.aix 5.3 2.4 o build lib.aix 5.3 2.4 scipy interpolate interpolate.so[[BR]] c c No such file or directory[[BR]] c unrecognized option qlanglvl=extc89 [[BR]] c unrecognized option bI usr local Python.2.4 lib python2.4 config python.exp [[BR]] error Command c c qlanglvl=extc89 bI usr local Python.2.4 lib python2.4 config python.exp build temp.aix 5.3 2.4 scipy interpolate src interpolate.o Lbuild temp.aix 5.3 2.4 o build lib.aix 5.3 2.4 scipy interpolate interpolate.so failed with exit status 1[[BR]] the second c is passed as an options to CXX compiler, and I don t know how to get rid of this.
|
|
[[BR]] OS AIX 5.3 64 bits[[BR]] CC gcc GCC) 4.2.4[[BR]] CXX c GCC) 4.2.4[[BR]] Python Python 2.4 2, Dec 20 2004, 10 36 59) [C] on aix4[[BR]] Numpy 1.4[[BR]]","bug"
|
|
"6013","2011","np.random.binomial ) fails with int64 n argument on x86 32 bit Trac 1413).
|
|
Original ticket http projects.scipy.org numpy ticket 1413 on 2010 02 26 by dwf, assigned to unknown.
|
|
Reported to the list by James Bergstra, confirmed by me >>> numpy.random.binomial n=numpy.asarray [2,3,4], dtype= int64 ), p=numpy.asarray [.1, .2, .3], dtype= float64 )) produces TypeError array cannot be safely cast to required type It seems to be not only 32 bit specific but x86 specific.
|
|
On a ppc machine, 32 bit mode, it behaves as expected dwf morrislab python 32 Python 2.6.4 r264 75706, Feb 16 2010, 21 03 46) [GCC 4.0.1 Apple Inc. build 5493)] on darwin Type help , copyright , credits or license for more information.
|
|
import numpy numpy.
|
|
version 1.3.0 numpy.random.binomial n=numpy.asarray [2,3,4], dtype= int64 ), p=numpy.asarray [.1,.2,.3], dtype= float64 )) array [1, 1, 2]) So it smells a bit like an endianness bug problem with the definition of `NPY LONG`.
|
|
I can confirm the bug on OS X Intel 32 bit, and Linux x86 32 both 1.3.0 and most recent svn trunk), as well as its absence on Linux x86 64.
|
|
The problem seems to be with this line in [source trunk numpy random mtrand mtrand.pyx mtrand.pyx], line 3306 in the trunk on = <ndarray>PyArray FROM OTF n, NPY LONG, NPY ALIGNED)","bug"
|
|
"6013","2012","Possible mistake in distribution.c rk binomial btpe Trac 1414).
|
|
Original ticket http projects.scipy.org numpy ticket 1414 on 2010 02 26 by trac user batz900, assigned to unknown.
|
|
I believe there is a small type in rk binomial btpe in distribution.c.
|
|
On line 337 there is an else, but in the Binomial Random Variate Generation by Kachitvichyanukul And Schmeiser their Step 5.1 ends the if else if block before the testing of the if v > F) goto Step10; goto Step60; basically, I believe the else enclosure should be deleted at the end of the Step50 section).
|
|
Also I have a cleaner version written without goto) if you re interested.","bug"
|
|
"6013","2013","data corruption unpickling a numpy.void Trac 1415).
|
|
Original ticket http projects.scipy.org numpy ticket 1415 on 2010 02 27 by trac user mspacek, assigned to unknown.
|
|
See thread at http article.gmane.org gmane.comp.python.numeric.general 36611 Pickling and then unpickling an entry of <type numpy.void > or <type numpy.record > from a structured array or recarray results in data corruption on the round trip >>> import numpy as np >>> x = np.zeros 2,), dtype= i4,f4,a10 )) >>> x[ ] = [ 1,2., Hello ), 2,3., World )] >>> x array [ 1, 2.0, Hello ), 2, 3.0, World )], dtype=[ f0 , <i4 ), f1 , <f4 ), f2 , |S10 )]) >>> x[0] 1, 2.0, Hello ) >>> type x[0]) <type numpy.void > >>> import pickle >>> s = pickle.dumps x[0]) >>> newx0 = pickle.loads s) >>> newx0 30917960, 1.6904535998413144e 38, \xd0\xef\x1c\x1eZ\x03\x00d ) >>> s cnumpy.core.multiarray\nscalar\np0\n cnumpy\ndtype\np1\n S V18 \np2\nI0\nI1\ntp3\nRp4\n I4\nS | \np5\nN S f0 \np6\nS f1 \np7\nS f2 \np8\ntp9\n dp10\ng6\n g1\n S i4 \np11\nI0\nI1\ntp12\nRp13 \n I4\nS < \np14\nNNNI 1\nI 1\nI0\nNtp15\nbI0\ntp16\nsg7\n g1 \n S f4 \np17\nI0\nI1\ntp18\nRp19\n I4\nS < \np20\nNNNI 1\nI 1 \nI0\nNtp21\nbI4\ntp22\nsg8\n g1\n S S10 \np23\nI0\nI1\ntp24\nRp25 \n I4\nS | \np26\nNNNI10\nI1\nI0\nNtp27\nbI8\ntp28\nsI18\nI1\nI0 \nNtp29\nbS \\x01\\x00\\x00\\x00\\x00\\x00\\x00 Hello\\x00\\x00 \\x00\\x00\\x00 \np30\ntp31\nRp32\n.
|
|
>>> type newx0) <type numpy.void > >>> newx0.dtype dtype [ f0 , <i4 ), f1 , <f4 ), f2 , |S10 )]) >>> x[0].dtype dtype [ f0 , <i4 ), f1 , <f4 ), f2 , |S10 )]) >>> np.version.version 1.4.0 It seems that the pickling works fine, but there s corruption on unpickling.
|
|
Robert Kern says The implementation of numpy.core.multiarray.scalar is doing something wrong.
|
|
Pauli Virtanen says The relevant code path to look at is multiarraymodule array scalar > scalarapi.c PyArray Scalar.
|
|
Needs some cgdb ing to find out what s going on there.
|
|
A nice temporary workaround is to convert the original structured array x into a list of scalar arrays, each of whose entries then pickle unpickle correctly, and still allow dict style named field access >>> x2 = map np.asarray, x)","bug"
|
|
"6013","2014","functions returned by frompyfunc ) cause segmentation faults Trac 1416).
|
|
Original ticket http projects.scipy.org numpy ticket 1416 on 2010 02 28 by trac user ernest ac, assigned to unknown.
|
|
When called with a last argument equal to zero, frompyfunc produces a function that segfaults on execution.
|
|
Example In [3] f = np.frompyfunc lambda x x, 1, 0) In [4] f [1,2,3]) <segmentation fault> Observed in Numpy 1.3.0 and in more recent versions.","bug"
|
|
"6013","2015","diag returns ndarray Trac 1417).
|
|
Original ticket http projects.scipy.org numpy ticket 1417 on 2010 02 28 by trac user uhrm, assigned to pierregm.
|
|
When called with a `matrix` object, `diag` return an `ndarray`.
|
|
Example >>> B = numpy.matrix 1.7 2.2 0.3; 1.2 0.2 0.0; 1.0 1.0 0.7; 0.3 0.3 0.2 ) >>> B matrix [[ 1.7, 2.2, 0.3], [ 1.2, 0.2, 0.
|
|
], [ 1. , 1. , 0.7], [ 0.3, 0.3, 0.2]]) >>> d = numpy.diag B) >>> d array [ 1.7, 0.2, 0.7]) If I understand correctly, the return type should also be a `matrix`.
|
|
Since `matrix` is always 2 dimensional, question is whether it s a column or row vector I d prefer a column vector).","bug"
|
|
"6013","2016","Using MaskedArray as index to np.ma.take should create a MaskedArray Trac 1418).
|
|
Original ticket http projects.scipy.org numpy ticket 1418 on 2010 03 02 by trac user petershinners, assigned to unknown.
|
|
Given an index array where illegal indices are masked.
|
|
It should be possible to use that array for np.ma.take, and have it return a MaskedArray.
|
|
Indices that are masked should not be looked up.
|
|
The mask on the original indices and the new result should be equal.","enhancement"
|
|
"6013","2017","Building Numpy Scipy on AIX system Trac 1419).
|
|
Original ticket http projects.scipy.org numpy ticket 1419 on 2010 03 02 by trac user A LARAS, assigned to unknown.
|
|
Hi,[[BR]] I wounder if somebody have success to build Numpy Scipy on AIX system with any of the IBM compilers XLC, XLF) or Gnu gcc, g [[BR]] I have lot of issues make it installed on AIX 5.3, 64 bits machine[[BR]] vac 8.0[[BR]] xlf 10.1[[BR]] gcc 4.2.4 2[[BR]] AIX 5.3[[BR]]","bug"
|
|
"6013","2018","irfft2 handles axes argument improperly Trac 1420).
|
|
Original ticket http projects.scipy.org numpy ticket 1420 on 2010 03 04 by trac user sivaldimarsson, assigned to unknown.
|
|
The following output of an interactive session shows how numpy.fft.irfft handles axes argument badly.
|
|
The workaround is to use irfftn with a shape argument as shown.
|
|
siv siv desktop Documents ipython Python 2.6.4 r264 75706, Dec 7 2009, 18 43 55) Type copyright , credits or license for more information.
|
|
IPython 0.10 An enhanced Interactive Python.
|
|
?
|
|
> Introduction and overview of IPython s features.
|
|
quickref > Quick reference.
|
|
help > Python s own help system.
|
|
object?
|
|
> Details about object .
|
|
?object also works, ??
|
|
prints more.
|
|
In [1] import numpy as np In [3] import numpy.fft as npfft In [4] a = np.zeros 256,256)) In [5] a.shape Out[5] 256, 256) In [9] a[0 10,0 100] = 1 In [11] ahat = npfft.rfft2 a,axes= 1,0)) In [12] ahat.shape Out[12] 129, 256) In [14] ahatinvhat = npfft.irfft2 ahat,axes= 1,0)) In [15] ahatinvhat.shape Out[15] 129, 510) In [16] ahatinvhat a ValueError Traceback most recent call last) home siv Documents <ipython console> in <module> ) ValueError shape mismatch objects cannot be broadcast to a single shape In [22] chat = npfft.rfftn a,s=a.shape,axes= 1,0)) In [23] chat.shape Out[23] 129, 256) In [24] np.max np.abs ahat chat)) Out[24] 0.0 In [25] chatinvhat = npfft.irfftn chat,s=a.shape,axes= 1,0)) In [26] chatinvhat.shape Out[26] 256, 256) In [27] np.max np.abs a chatinvhat)) Out[27] 6.6613381477509392e 16","bug"
|
|
"6013","2019","Why does np.nan{min, max} clobber my array mask?
|
|
Trac 1421).
|
|
Original ticket http projects.scipy.org numpy ticket 1421 on 2010 03 05 by bsouthey, assigned to pierregm.
|
|
David Carmean provided the following to the numpy list Why does np.nan{min, max} clobber my array mask?
|
|
[http mail.scipy.org pipermail numpy discussion 2010 February 048628.html] The following code reproduces it under Linux Python2.6 with numpy version 2.0.0.dev8272 .
|
|
import numpy as np from numpy.ma.testutils import uut = np.array [[2, 1, 3, np.nan], [5, 2, 3, np.nan]]) msk = np.ma.masked invalid uut) oldmsk=msk.copy ) mvalue=np.nanmin msk, axis=1) assert array equal msk.mask, oldmsk.mask)","bug"
|
|
"6013","2020","ma.
|
|
extrema operation subclasses reduce method raise ValueError... Trac 1422).
|
|
Original ticket http projects.scipy.org numpy ticket 1422 on 2010 03 05 by trac user dgoldsmith, assigned to pierregm.
|
|
...if passed zero size array.
|
|
>>> mn = numpy.ma.core.
|
|
minimum operation ) >>> mn.reduce N.array ))) Traceback most recent call last) File <input> , line 1, in <module> File C \Python254\Lib\site packages\numpy\ma\core.py , line 5570, in reduce t = self.ufunc.reduce target, kargs) ValueError zero size array to ufunc.reduce without identity Notes 0) Problem discovered via a call from matplotlib.imshow; here s that traceback Traceback most recent call last) File <my code.py> , line 108, in <module> ax.imshow part2plot, cmap name, extent = extent) File C \Python254\lib\site packages\matplotlib\axes.py , line 6261, in imshow im.autoscale None ) File C \Python254\lib\site packages\matplotlib\cm.py , line 236, in autoscale None self.norm.autoscale None self.
|
|
A) File C \Python254\lib\site packages\matplotlib\colors.py , line 792, in autoscale None if self.vmin is None self.vmin = ma.minimum A) File C \Python254\Lib\site packages\numpy\ma\core.py , line 5555, in call return self.reduce a) File C \Python254\Lib\site packages\numpy\ma\core.py , line 5570, in reduce t = self.ufunc.reduce target, kargs) ValueError zero size array to ufunc.reduce without identity Script terminated.
|
|
1) Platform details numpy 1.4.0, MPL 0.99.0, python 2.5.4, 32 bit Vista Home Premium SP2","bug"
|
|
"6013","2021","Error running easy install numpy Trac 1423).
|
|
Original ticket http projects.scipy.org numpy ticket 1423 on 2010 03 05 by trac user rjain, assigned to unknown.
|
|
demoenv)jain jain desktop sandbox demoenv easy install numpy Searching for numpy Reading http pypi.python.org simple numpy Reading http numpy.scipy.org Reading http sourceforge.net project showfiles.php?group id=1369 package id=175103 Reading http numeric.scipy.org Best match numpy 1.3.0 Downloading http sourceforge.net projects numpy files NumPy 1.3.0 numpy 1.3.0.zip download Processing download Running numpy 1.3.0 setup.py q bdist egg dist dir tmp easy install zpQwWp numpy 1.3.0 egg dist tmp mmTb5S Running from numpy source directory.
|
|
non existing path in numpy distutils site.cfg tmp easy install zpQwWp numpy 1.3.0 numpy distutils command config.py 361 DeprecationWarning Usage of get output is deprecated please do not use it anymore, and avoid configuration checks involving running executable on the target machine.
|
|
DeprecationWarning) Found executable usr bin g77 gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found compiling configtest.c This file is generated from numpy distutils system info.py void ATL buildinfo void); int main void) { ATL buildinfo ); return 0; } usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran st write done usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran transfer integer usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran stop numeric usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran st write usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran transfer character collect2 ld returned 1 exit status usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran st write done usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran transfer integer usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran stop numeric usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran st write usr lib gcc x86 64 linux gnu 4.2.4 .. .. .. .. lib libf77blas.so undefined reference to ` gfortran transfer character collect2 ld returned 1 exit status gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found compiling configtest.c This file is generated from numpy distutils system info.py void ATL buildinfo void); int main void) { ATL buildinfo ); return 0; } ATLAS version 3.6.0 built by root on Thu Jun 3 21 13 07 UTC 2004 UNAME Linux ravel 2.6.5 3 SMP Tue Apr 13 13 41 54 UTC 2004 x86 64 GNU Linux INSTFLG MMDEF home camm atlas3 3.6.0 CONFIG ARCHS HAMMER64SSE2 gcc gemm ARCHDEF home camm atlas3 3.6.0 CONFIG ARCHS HAMMER64SSE2 gcc misc F2CDEFS DAdd DStringSunStyle CACHEEDGE 720896 F77 home camm usr bin g77, version GNU Fortran GCC) 3.3.3 Debian 20040422) F77FLAGS fomit frame pointer O m64 CC usr bin gcc, version gcc GCC) 3.3.3 Debian 20040422) CC FLAGS fomit frame pointer O mfpmath=387 m64 MCC usr bin gcc, version gcc GCC) 3.3.3 Debian 20040422) MCCFLAGS fomit frame pointer O mfpmath=387 m64 gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 5 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 5 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 5 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 5 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 7 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c In function xe2x80x98mainxe2x80x99 configtest.c 7 error size of array xe2x80x98test arrayxe2x80x99 is negative configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 5 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 4 warning function declaration isnxe2x80x99t a prototype configtest.c 5 warning function declaration isnxe2x80x99t a prototype tmp easy install zpQwWp numpy 1.3.0 numpy distutils command config.py 39 DeprecationWarning Usage of try run is deprecated please do not use it anymore, and avoid configuration checks involving running executable on the target machine.
|
|
DeprecationWarning) configtest.o In function `main tmp easy install zpQwWp numpy 1.3.0 configtest.c 5 undefined reference to `exp collect2 ld returned 1 exit status configtest.o In function `main tmp easy install zpQwWp numpy 1.3.0 configtest.c 5 undefined reference to `exp collect2 ld returned 1 exit status configtest.c 1 warning conflicting types for built in function xe2x80x98asinxe2x80x99 configtest.c 2 warning conflicting types for built in function xe2x80x98cosxe2x80x99 configtest.c 3 warning conflicting types for built in function xe2x80x98logxe2x80x99 configtest.c 4 warning conflicting types for built in function xe2x80x98fabsxe2x80x99 configtest.c 5 warning conflicting types for built in function xe2x80x98tanhxe2x80x99 configtest.c 6 warning conflicting types for built in function xe2x80x98atanxe2x80x99 configtest.c 7 warning conflicting types for built in function xe2x80x98acosxe2x80x99 configtest.c 8 warning conflicting types for built in function xe2x80x98floorxe2x80x99 configtest.c 9 warning conflicting types for built in function xe2x80x98fmodxe2x80x99 configtest.c 10 warning conflicting types for built in function xe2x80x98sqrtxe2x80x99 configtest.c 11 warning conflicting types for built in function xe2x80x98coshxe2x80x99 configtest.c 12 warning conflicting types for built in function xe2x80x98modfxe2x80x99 configtest.c 13 warning conflicting types for built in function xe2x80x98sinhxe2x80x99 configtest.c 14 warning conflicting types for built in function xe2x80x98frexpxe2x80x99 configtest.c 15 warning conflicting types for built in function xe2x80x98expxe2x80x99 configtest.c 16 warning conflicting types for built in function xe2x80x98tanxe2x80x99 configtest.c 17 warning conflicting types for built in function xe2x80x98ceilxe2x80x99 configtest.c 18 warning conflicting types for built in function xe2x80x98log10xe2x80x99 configtest.c 19 warning conflicting types for built in function xe2x80x98sinxe2x80x99 configtest.c 20 warning conflicting types for built in function xe2x80x98ldexpxe2x80x99 configtest.c 1 warning conflicting types for built in function xe2x80x98expm1xe2x80x99 configtest.c 2 warning conflicting types for built in function xe2x80x98log2xe2x80x99 configtest.c 3 warning conflicting types for built in function xe2x80x98rintxe2x80x99 configtest.c 4 warning conflicting types for built in function xe2x80x98atanhxe2x80x99 configtest.c 5 warning conflicting types for built in function xe2x80x98asinhxe2x80x99 configtest.c 6 warning conflicting types for built in function xe2x80x98acoshxe2x80x99 configtest.c 7 warning conflicting types for built in function xe2x80x98exp2xe2x80x99 configtest.c 8 warning conflicting types for built in function xe2x80x98log1pxe2x80x99 configtest.c 9 warning conflicting types for built in function xe2x80x98truncxe2x80x99 configtest.c 1 warning conflicting types for built in function xe2x80x98cosfxe2x80x99 configtest.c 2 warning conflicting types for built in function xe2x80x98coshfxe2x80x99 configtest.c 3 warning conflicting types for built in function xe2x80x98rintfxe2x80x99 configtest.c 4 warning conflicting types for built in function xe2x80x98fabsfxe2x80x99 configtest.c 5 warning conflicting types for built in function xe2x80x98floorfxe2x80x99 configtest.c 6 warning conflicting types for built in function xe2x80x98tanhfxe2x80x99 configtest.c 7 warning conflicting types for built in function xe2x80x98log10fxe2x80x99 configtest.c 8 warning conflicting types for built in function xe2x80x98logfxe2x80x99 configtest.c 9 warning conflicting types for built in function xe2x80x98sinhfxe2x80x99 configtest.c 10 warning conflicting types for built in function xe2x80x98acosfxe2x80x99 configtest.c 11 warning conflicting types for built in function xe2x80x98sqrtfxe2x80x99 configtest.c 12 warning conflicting types for built in function xe2x80x98ldexpfxe2x80x99 configtest.c 13 warning conflicting types for built in function xe2x80x98hypotfxe2x80x99 configtest.c 14 warning conflicting types for built in function xe2x80x98log2fxe2x80x99 configtest.c 15 warning conflicting types for built in function xe2x80x98exp2fxe2x80x99 configtest.c 16 warning conflicting types for built in function xe2x80x98atanfxe2x80x99 configtest.c 17 warning conflicting types for built in function xe2x80x98fmodfxe2x80x99 configtest.c 18 warning conflicting types for built in function xe2x80x98atan2fxe2x80x99 configtest.c 19 warning conflicting types for built in function xe2x80x98modffxe2x80x99 configtest.c 20 warning conflicting types for built in function xe2x80x98ceilfxe2x80x99 configtest.c 21 warning conflicting types for built in function xe2x80x98log1pfxe2x80x99 configtest.c 22 warning conflicting types for built in function xe2x80x98asinfxe2x80x99 configtest.c 23 warning conflicting types for built in function xe2x80x98acoshfxe2x80x99 configtest.c 24 warning conflicting types for built in function xe2x80x98sinfxe2x80x99 configtest.c 25 warning conflicting types for built in function xe2x80x98tanfxe2x80x99 configtest.c 26 warning conflicting types for built in function xe2x80x98atanhfxe2x80x99 configtest.c 27 warning conflicting types for built in function xe2x80x98truncfxe2x80x99 configtest.c 28 warning conflicting types for built in function xe2x80x98asinhfxe2x80x99 configtest.c 29 warning conflicting types for built in function xe2x80x98frexpfxe2x80x99 configtest.c 30 warning conflicting types for built in function xe2x80x98powfxe2x80x99 configtest.c 31 warning conflicting types for built in function xe2x80x98expfxe2x80x99 configtest.c 32 warning conflicting types for built in function xe2x80x98expm1fxe2x80x99 configtest.c 1 warning conflicting types for built in function xe2x80x98tanhlxe2x80x99 configtest.c 2 warning conflicting types for built in function xe2x80x98log10lxe2x80x99 configtest.c 3 warning conflicting types for built in function xe2x80x98coshlxe2x80x99 configtest.c 4 warning conflicting types for built in function xe2x80x98coslxe2x80x99 configtest.c 5 warning conflicting types for built in function xe2x80x98floorlxe2x80x99 configtest.c 6 warning conflicting types for built in function xe2x80x98rintlxe2x80x99 configtest.c 7 warning conflicting types for built in function xe2x80x98fabslxe2x80x99 configtest.c 8 warning conflicting types for built in function xe2x80x98acoslxe2x80x99 configtest.c 9 warning conflicting types for built in function xe2x80x98ldexplxe2x80x99 configtest.c 10 warning conflicting types for built in function xe2x80x98sqrtlxe2x80x99 configtest.c 11 warning conflicting types for built in function xe2x80x98loglxe2x80x99 configtest.c 12 warning conflicting types for built in function xe2x80x98expm1lxe2x80x99 configtest.c 13 warning conflicting types for built in function xe2x80x98hypotlxe2x80x99 configtest.c 14 warning conflicting types for built in function xe2x80x98log2lxe2x80x99 configtest.c 15 warning conflicting types for built in function xe2x80x98exp2lxe2x80x99 configtest.c 16 warning conflicting types for built in function xe2x80x98atanlxe2x80x99 configtest.c 17 warning conflicting types for built in function xe2x80x98frexplxe2x80x99 configtest.c 18 warning conflicting types for built in function xe2x80x98atan2lxe2x80x99 configtest.c 19 warning conflicting types for built in function xe2x80x98sinhlxe2x80x99 configtest.c 20 warning conflicting types for built in function xe2x80x98fmodlxe2x80x99 configtest.c 21 warning conflicting types for built in function xe2x80x98log1plxe2x80x99 configtest.c 22 warning conflicting types for built in function xe2x80x98asinlxe2x80x99 configtest.c 23 warning conflicting types for built in function xe2x80x98ceillxe2x80x99 configtest.c 24 warning conflicting types for built in function xe2x80x98sinlxe2x80x99 configtest.c 25 warning conflicting types for built in function xe2x80x98acoshlxe2x80x99 configtest.c 26 warning conflicting types for built in function xe2x80x98atanhlxe2x80x99 configtest.c 27 warning conflicting types for built in function xe2x80x98tanlxe2x80x99 configtest.c 28 warning conflicting types for built in function xe2x80x98trunclxe2x80x99 configtest.c 29 warning conflicting types for built in function xe2x80x98powlxe2x80x99 configtest.c 30 warning conflicting types for built in function xe2x80x98explxe2x80x99 configtest.c 31 warning conflicting types for built in function xe2x80x98modflxe2x80x99 configtest.c 32 warning conflicting types for built in function xe2x80x98asinhlxe2x80x99 configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype configtest.c 6 warning function declaration isnxe2x80x99t a prototype File build src.linux x86 64 2.5 numpy core include numpy config.h define SIZEOF SHORT 2 define SIZEOF INT 4 define SIZEOF LONG 8 define SIZEOF FLOAT 4 define SIZEOF DOUBLE 8 define SIZEOF PY INTPTR T 8 define SIZEOF PY LONG LONG 8 define MATHLIB m define HAVE SIN define HAVE COS define HAVE TAN define HAVE SINH define HAVE COSH define HAVE TANH define HAVE FABS define HAVE FLOOR define HAVE CEIL define HAVE SQRT define HAVE LOG10 define HAVE LOG define HAVE EXP define HAVE ASIN define HAVE ACOS define HAVE ATAN define HAVE FMOD define HAVE MODF define HAVE FREXP define HAVE LDEXP define HAVE EXPM1 define HAVE LOG1P define HAVE ACOSH define HAVE ASINH define HAVE ATANH define HAVE RINT define HAVE TRUNC define HAVE EXP2 define HAVE LOG2 define HAVE SINF define HAVE COSF define HAVE TANF define HAVE SINHF define HAVE COSHF define HAVE TANHF define HAVE FABSF define HAVE FLOORF define HAVE CEILF define HAVE RINTF define HAVE TRUNCF define HAVE SQRTF define HAVE LOG10F define HAVE LOGF define HAVE LOG1PF define HAVE EXPF define HAVE EXPM1F define HAVE ASINF define HAVE ACOSF define HAVE ATANF define HAVE ASINHF define HAVE ACOSHF define HAVE ATANHF define HAVE HYPOTF define HAVE ATAN2F define HAVE POWF define HAVE FMODF define HAVE MODFF define HAVE FREXPF define HAVE LDEXPF define HAVE EXP2F define HAVE LOG2F define HAVE SINL define HAVE COSL define HAVE TANL define HAVE SINHL define HAVE COSHL define HAVE TANHL define HAVE FABSL define HAVE FLOORL define HAVE CEILL define HAVE RINTL define HAVE TRUNCL define HAVE SQRTL define HAVE LOG10L define HAVE LOGL define HAVE LOG1PL define HAVE EXPL define HAVE EXPM1L define HAVE ASINL define HAVE ACOSL define HAVE ATANL define HAVE ASINHL define HAVE ACOSHL define HAVE ATANHL define HAVE HYPOTL define HAVE ATAN2L define HAVE POWL define HAVE FMODL define HAVE MODFL define HAVE FREXPL define HAVE LDEXPL define HAVE EXP2L define HAVE LOG2L define HAVE DECL ISNAN define HAVE DECL ISINF define HAVE DECL SIGNBIT define HAVE DECL ISFINITE ifndef cplusplus undef inline endif EOF configtest.c 5 warning function declaration isnxe2x80x99t a prototype File build src.linux x86 64 2.5 numpy core include numpy numpyconfig.h define NPY SIZEOF SHORT 2 define NPY SIZEOF INT 4 define NPY SIZEOF LONG 8 define NPY SIZEOF FLOAT 4 define NPY SIZEOF DOUBLE 8 define NPY SIZEOF LONGDOUBLE 16 define NPY SIZEOF PY INTPTR T 8 define NPY SIZEOF PY LONG LONG 8 define NPY SIZEOF LONGLONG 8 define NPY NO SMP 0 define NPY HAVE DECL ISNAN define NPY HAVE DECL ISINF define NPY HAVE DECL SIGNBIT define NPY HAVE DECL ISFINITE define NPY USE C99 FORMATS 1 define NPY INLINE inline ifndef STDC FORMAT MACROS define STDC FORMAT MACROS 1 endif EOF gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found gnu no Fortran 90 compiler found zip safe flag not set; analyzing archive contents... numpy.
|
|
import tools module references file numpy.
|
|
import tools module references path numpy.version module references file numpy.core.scons support module references file numpy.core.setup module references file numpy.core.generate numpy api module references file numpy.doc.
|
|
init module references file numpy.ma.core module MAY be using inspect.trace numpy.numarray.util module references file numpy.distutils.system info module references file numpy.distutils.exec command module references file numpy.distutils.misc util module references file numpy.distutils.fcompiler.compaq module references file numpy.distutils.fcompiler.
|
|
init module references file numpy.distutils.command.build src module references file numpy.distutils.command.scons module references file numpy.lib.utils module references file numpy.lib.utils module references path numpy.lib.utils module MAY be using inspect.getsource numpy.lib.utils module MAY be using inspect.getsourcefile numpy.testing.numpytest module references file numpy.testing.utils module references file numpy.testing.noseclasses module references file numpy.testing.nosetester module references file numpy.f2py.setup module references file numpy.f2py.f2py2e module references file numpy.f2py.setupscons module references file numpy.f2py.diagnose module references file Adding numpy 1.3.0 to easy install.pth file Installing f2py script to sandbox demoenv bin Installed sandbox demoenv lib python2.5 site packages numpy 1.3.0 py2.5 linux x86 64.egg Processing dependencies for numpy Finished processing dependencies for numpy Error in atexit.
|
|
run exitfuncs Traceback most recent call last) File usr lib python2.5 atexit.py , line 24, in run exitfuncs func targs, kargs) File tmp easy install zpQwWp numpy 1.3.0 numpy distutils misc util.py , line 219, in clean up temporary directory SystemError Parent module numpy.distutils not loaded Error in sys.exitfunc Traceback most recent call last) File usr lib python2.5 atexit.py , line 24, in run exitfuncs func targs, kargs) File tmp easy install zpQwWp numpy 1.3.0 numpy distutils misc util.py , line 219, in clean up temporary directory SystemError Parent module numpy.distutils not loaded","bug"
|
|
"6013","2022","Results of array.tofile and fromfile are different Trac 1424).
|
|
Original ticket http projects.scipy.org numpy ticket 1424 on 2010 03 07 by trac user HyunchulKim, assigned to unknown.
|
|
How to reproduce this bug.
|
|
[[BR]] Environment [[BR]] Python 2.6.4 at Ubuntu 9.10[[BR]] I tested both of Ubuntu providing version and source compiled version.
|
|
Both of them give same error.
|
|
[[BR]] from numpy import Regardless of specific array, it always give a false result.
|
|
x = array [0,1,2,3]) x.tofile test.output ) y = fromfile test.output ) print x == y The result is False print y The result is array [ 4.24399158e 314, 8.48798317e 314]) Interestingly, results of repeated tofile and fromfile give same results y.tofile test.output2 ) z = fromfile test.output2 ) print y == z The result is True !!!
|
|
print z The result is array [ 4.24399158e 314, 8.48798317e 314]) [[BR]] I believe that this is a must and urgent bug that give false results.","bug"
|
|
"6013","2023","importing umath tests and then CTRL D crashes with core dump Trac 1425).
|
|
Original ticket http projects.scipy.org numpy ticket 1425 on 2010 03 10 by johannct, assigned to unknown.
|
|
[cohen jarrett core] python > Python 2.6.2 r262 71600, Jan 25 2010, 18 46 45) [GCC 4.4.2 20091222 > Red Hat 4.4.2 20)] on linux2 Type help , copyright , credits or > license for more information.
|
|
> >>> import umath tests > >>> > python Modules gcmodule.c 277 visit decref Assertion `gc >gc.gc refs > != 0 failed.
|
|
> Aborted core dumped) > This occurs on my Fedora 12 box with python 2.6.2 and numpy 2.0.0.dev8291","bug"
|
|
"6013","2024","subclasses of ndarray have problems with reflected operations in python 3 Trac 1426).
|
|
Original ticket http projects.scipy.org numpy ticket 1426 on 2010 03 11 by ddale, assigned to unknown.
|
|
Quoting a discussion on the mailing list, where Pauli V. explained the reasons for the reported behavior >> Now that the trunk has some support for python3, I am working on making Quantities work with python3 as well.
|
|
I m running into some problems related to subclassing ndarray that can be illustrated with a simple script, reproduced below.
|
|
It looks like there is a problem with the reflected operations, I see problems with rmul and radd , but not with mul and add > >Thanks for testing.
|
|
I wish the test suite was more complete hint!
|
|
hint! )
|
|
> >Yes, Python 3 introduced some semantic changes in how subclasses of builtin classes = written in C) inherit the r operations.
|
|
> >Below I ll try to explain what is going on.
|
|
We probably need to change some things to make things work better on Py3, within the bounds we are able to.
|
|
> >Suggestions are welcome.
|
|
The most obvious one could be to explicitly implement rmul etc.
|
|
on Python 3. import numpy as np class A np.ndarray) def new cls, args, kwargs) return np.ndarray.
|
|
new cls, args, kwargs) class B A) def mul self, other) return self.view A).
|
|
mul other) def rmul self, other) return self.view A).
|
|
rmul other) def add self, other) return self.view A).
|
|
add other) def radd self, other) return self.view A).
|
|
radd other) a = A 10,)) b = B 10,)) print A rmul ) print a. rmul 2)) yields NotImplemented print a.view np.ndarray).
|
|
rmul 2)) yields NotImplemented >Correct.
|
|
ndarray does not implement rmul , but relies on an automatic wrapper generated by Python.
|
|
> >The automatic wrapper wrap binaryfunc r) does the following > >1.
|
|
Is `type other)` a subclass of `type self)`?
|
|
If yes, call mul with swapped arguments.
|
|
>2.
|
|
If not, bail out with NotImplemented.
|
|
> >So it bails out.
|
|
> >Previously, the ndarray type had a flag that made Python to skip the subclass check.
|
|
That does not exist any more on Python 3, and is the root of this issue.
|
|
print 2 a) ok !!??
|
|
>Here, Python checks > >1.
|
|
Does nb multiply from the left op succeed?
|
|
Nope, since floats don t know how to multiply ndarrays.
|
|
> >2.
|
|
Does nb multiply from the right op succeed?
|
|
Here the execution passes directly to array multiply, completely skipping the rmul wrapper.
|
|
> >Note also that in the C level number protocol there is only a single multiplication function for both left and right multiplication.
|
|
> print B rmul ) print b. rmul 2)) yields NotImplemented print b.view A).
|
|
rmul 2)) yields NotImplemented print b.view np.ndarray).
|
|
rmul 2)) yields NotImplemented print 2 b) yields TypeError unsupported operand type s) for int and B >But here, the subclass calls the wrapper ndarray.
|
|
rmul , which wants to be careful with types, and hence fails.
|
|
> >Yes, probably explicitly defining rmul for ndarray could be the >right solution.
|
|
Please file a bug report on this.","bug"
|
|
"6013","2025","Memory leak caused by calling astype float ) on structured array of dtype object Trac 1427).
|
|
Original ticket http projects.scipy.org numpy ticket 1427 on 2010 03 15 by trac user nathanielpeterson, assigned to unknown.
|
|
The attached code demonstrates the problem.
|
|
My manner of monitoring memory usage is specific to Linux.
|
|
Sorry, I don t know how to make it more general.","bug"
|
|
"6013","2026","genfromtxt problem with long integers Trac 1428).
|
|
Original ticket http projects.scipy.org numpy ticket 1428 on 2010 03 15 by josef pkt, assigned to unknown.
|
|
It seems genfromtxt has problems identifying long integers at least on Windows 32) I have a csv file with large integers >>> np.array 4160680000,int) Traceback most recent call last) File <pyshell 4> , line 1, in <module> np.array 4160680000,int) OverflowError long int too large to convert to int genfromtxt fails with non informative exception >>> s = Date,Open,High,Low,Close,Volume,Adj Close ... 2010 02 12,1075.95,1077.81,1062.97,1075.51,4160680000,1075.51 ... 2010 02 11,1067.10,1080.04,1060.59,1078.47,4400870000,1078.47 >>> sh = StringIO s) >>> data = np.genfromtxt sh, delimiter= , , dtype=None, names=True) Traceback most recent call last) File C \Programs\Python25\Lib\site packages\numpy\lib\io.py , line 1367, in genfromtxt output = np.array data, dtype=ddtype) TypeError expected a readable buffer object same with explicit dtypes using int for the long integer >>> dt= [ , S10 ), ,float), ,float), ,float), ,float), ,int), ,float)] >>> sh = StringIO s) >>> data = np.genfromtxt sh, delimiter= , , dtype=dt, names=True) Traceback most recent call last) File C \Programs\Python25\Lib\site packages\numpy\lib\io.py , line 1388, in genfromtxt rows = np.array data, dtype=[ , ) for in dtype flat]) TypeError expected a readable buffer object using float works >>> dt= [ , S10 ), ,float), ,float), ,float), ,float), ,float), ,float)] >>> sh = StringIO s) >>> data = np.genfromtxt sh, delimiter= , , dtype=dt, names=True) >>> data array [ 2010 02 12 , 1075.95, 1077.8099999999999, 1062.97, 1075.51, 4160680000.0, 1075.51), 2010 02 11 , 1067.0999999999999, 1080.04, 1060.5899999999999, 1078.47, 4400870000.0, 1078.47)], dtype=[ Date , |S10 ), Open , <f8 ), High , <f8 ), Low , <f8 ), Close , <f8 ), Volume , <f8 ), Adj Close , <f8 )]) >>>","bug"
|
|
"6013","2027","inconsistent behavior of argmin argmax with nan Trac 1429).
|
|
Original ticket http projects.scipy.org numpy ticket 1429 on 2010 03 15 by trac user jgregoire, assigned to unknown.
|
|
This is related to Ticket 1633 but I believe is an unreported issue.
|
|
When the array contains nan, argmin and argmax ignore the nan.
|
|
The problem is that if the zeroth element of the array is nan then argmin and argmax return 0.
|
|
Having the behavior of a function depend on the value of the zeroth element is very undesirable.
|
|
In [1] import numpy as np In [2] a=np.array [1., 2., 0., np.nan]) In [3] print np.argmin a), np.argmax a) > print np.argmin a), np.argmax a)) 2, 1) In [4] print np.argmin a[ 1]), np.argmax a[ 1]) > print np.argmin a[ 1]), np.argmax a[ 1])) 0, 0)","bug"
|
|
"6013","2028","numpy s documentation missing in Ubuntu Trac 1430).
|
|
Original ticket http projects.scipy.org numpy ticket 1430 on 2010 03 16 by trac user malev, assigned to unknown.
|
|
Hi there!
|
|
I looked after the numpy s doc in Ubuntu usr share doc python numpy) and there are only a list of .py files, but no documentation.
|
|
There is a but about it in launchpad.
|
|
https bugs.launchpad.net ubuntu source python numpy bug 531917","bug"
|
|
"6013","2029","accessing multiple fields in a recarray gives fields in wrong order Trac 1431).
|
|
Original ticket http projects.scipy.org numpy ticket 1431 on 2010 03 16 by samtygier, assigned to unknown.
|
|
when using the list in list notation to access multiple columns of a record array the columns are returned in the order that they are stored in the array, not in the order asked for.
|
|
>>> x = np.array [ 1.5,2.5, 1.0,2.0)), 3.,4., 4.,5.
|
|
)), 1.,3., 2.,6.
|
|
))], dtype=[ x , f4 ), y ,np.float32), value , f4 , 2,2))]) >>> x[[ x , y ]] array [ 1.5, 2.5), 3.0, 4.0), 1.0, 3.0)], dtype=[ x , <f4 ), y , <f4 )]) >>> x[[ y , x ]] array [ 1.5, 2.5), 3.0, 4.0), 1.0, 3.0)], dtype=[ x , <f4 ), y , <f4 )]) this different to what happens with a 2d array, where they are returned in the order asked for >>> a = array [0,0.1,0.2,0.3,0.4]) then >>> a[[0,1,4]] array [ 0. , 0.1, 0.4]) >>> a[[4,1,0]] array [ 0.4, 0.1, 0. ])
|
|
I suggest that the it would be more consistent and predictable for the order to be the one asked for in both cases.
|
|
The current behaviour for the recarray was undocumented until today, so there can t be many people using it.
|
|
therefore it might be worth break backwards compatibility to make the interface better.
|
|
Please see discussion at http thread.gmane.org gmane.comp.python.numeric.general 36933 a simple patch is attached.","bug"
|
|
"6013","2030","some submodules are missing from sphinx s objects.inv Trac 1432).
|
|
Original ticket http projects.scipy.org numpy ticket 1432 on 2010 03 16 by yarikoptic, assigned to pv.
|
|
hopefully this time I would not screw up and file a valid bug report ;) I found that intersphinx failed to link to mod `numpy.random` ... so, I got objects.inv and saw which mod references seems to be missing > for d in ; do grep q numpy\.
|
|
{d } mod objects.inv || echo d missing ; done compat missing core missing doc missing f2py missing lib missing linalg missing ma missing matrixlib missing polynomial missing random missing testing missing tests missing","bug"
|
|
"6013","2031","numpy array captures in statement when it shouldn t Trac 1433).
|
|
Original ticket http projects.scipy.org numpy ticket 1433 on 2010 03 16 by trac user graik, assigned to unknown.
|
|
The following code used to work perfectly well with older numpy versions >>> a = arange 10) >>> a in 0,None) FALSE Now it gives an error ValueError The truth value of an array with more than one element is ambiguous.
|
|
Use a.any ) or a.all ) This comes unexpected.
|
|
For example my library contains several situations like this def calculate a=None ) if a in 0,None) a = zeros 10 ) ... calculate a=arange 10) ) So I use the in to quickly check for identity with a couple of non valid values.
|
|
Apparently, the numpy array code now captures the in statement via contains and expects an array on the other side not even a list will do).
|
|
This is not logical.
|
|
While there might be some reason for treating the array in array situation differently, this should not break standard python behavior.
|
|
Simple python lists do accept this code >>> l = [1,2,3] >>> l in None,0) FALSE I would expect numpy arrays to do the same as it did before).
|
|
At the very least, the new contains should check for non array arguments and fall back to the standard behavior rather than raising an exception.
|
|
Thanks in advance!
|
|
Greetings Raik","enhancement"
|
|
"6013","2032","out argument ignored by var and std Trac 1434).
|
|
Original ticket http projects.scipy.org numpy ticket 1434 on 2010 03 18 by trac user a.naldi, assigned to unknown.
|
|
when providing an axis, the mean, var and std functions return an array of results.
|
|
The out argument enables to provide an already allocated array to store the result, which works fine for mean but not for var and std exemple ipython session In [1] import numpy as np In [2] data = np.array 1,2,3), 4,5,6), 7,8,9))) In [3] result = np.zeros 3,)) In [4] data.mean axis=1, out=result) == result Out[4] array [ True, True, True], dtype=bool) In [5] data.var axis=1, out=result) == result Out[5] array [False, False, False], dtype=bool) In [6] data.var axis=1, out=result) Out[6] array [ 0.66666667, 0.66666667, 0.66666667]) In [7] result Out[7] array [ 2., 5., 8.])","bug"
|
|
"6013","2033","numpy.gradient ) does not support array subclasses Trac 1435).
|
|
Original ticket http projects.scipy.org numpy ticket 1435 on 2010 03 18 by dopplershift, assigned to unknown.
|
|
For no good reason, numpy.gradient does not support ndarray subclasses, like masked arrays.
|
|
This is due to the use of numpy.zeros on line 822, instead of using something like numpy.zeros like ).
|
|
Changing zeros to zeros like allows the function to return arrays of the proper subclass.
|
|
Also, while the function makes an effort to determine a floating point type for the output array 805 807), it was not actually using this type.
|
|
I ve corrected this as well.","bug"
|
|
"6013","2034","Regression creating array from character Trac 1436).
|
|
Original ticket http projects.scipy.org numpy ticket 1436 on 2010 03 20 by dopplershift, assigned to unknown.
|
|
The following code, which works with numpy 1.4.0, results in an error In [1] import numpy as np In [2] v = m In [3] dt = np.dtype >c ) In [4] a = np.asarray v, dt) On 1.4.0 In [5] a Out[5] array m , dtype= |S1 ) In [6] np.
|
|
version Out[6] 1.4.0 On SVN trunk home rmay .local lib python2.6 site packages numpy core numeric.pyc in asarray a, dtype, order) 282 283 > 284 return array a, dtype, copy=False, order=order) 285 286 def asanyarray a, dtype=None, order=None) ValueError assignment to 0 d array In [5] np.
|
|
version Out[5] 2.0.0.dev8297","bug"
|
|
"6013","2035","Copy a complex array into a slice of a real array takes the real part.
|
|
Trac 1437).
|
|
Original ticket http projects.scipy.org numpy ticket 1437 on 2010 03 21 by trac user SevenThunders, assigned to unknown.
|
|
Numpy gives surprising results when copying complex data into a portion of a predefined real data array.
|
|
It takes the real part of the complex data, and does not retype the real array to be complex.
|
|
Here is some example code import numpy as np z = np.zeros 3,4)) u = np.random.randn 3,2) 1j np.random.randn 3,2) z[ ,1 3] = u print z) One can contrast this behavior, with say Matlab, or with numpy itself under different contexts, e.g.
|
|
adding to the above code u4 = np.random.randn 3,4) 1j np.random.randn 3,4) z = u4 print z) Thus when the right hand side has the same size as the left, numpy will promote the left hand side, z, to be complex.
|
|
I can only presume that the reason for this behavior is performance, probably due to maintaining a complex array as interleaved real and imaginary.
|
|
If so, then for consistency and to catch the obvious sorts of bugs this creates, it would be far preferable to throw an exception for the complex to real copy for a sliced array.
|
|
After all it is pretty easy to get the default behavior by taking the real part of the right hand side.","bug"
|
|
"6013","2036","numpy.trapz ) doesn t respect subclass Trac 1438).
|
|
Original ticket http projects.scipy.org numpy ticket 1438 on 2010 03 22 by dopplershift, assigned to unknown.
|
|
numpy.trapz uses asarray ) inside the code, which converts arrays that were ndarray subclasses.
|
|
Otherwise, the code can work fine with subclasses.
|
|
I ve attached a patch that fixes this.
|
|
No test case, because for some reason I can t get masked arrays to fail without my fix.
|
|
However, this does fix the problem for the python quantities package.","bug"
|
|
"6013","2037","numpy.array arr.flat) reorders arr s memory if arr.flags.fortran and arr.flags.owndata Trac 1439).
|
|
Original ticket http projects.scipy.org numpy ticket 1439 on 2010 03 27 by zachrahan, assigned to unknown.
|
|
import numpy a = numpy.array [[1,2],[3,4]], order= F ) print a l = list a.flat) print a f = numpy.fromiter a.flat, a.dtype) print a a2 = numpy.array a.flat) print a yields array [[1, 2], [3, 4]]) array [[1, 2], [3, 4]]) array [[1, 2], [3, 4]]) array [[1, 3], [2, 4]]) If a is C contiguous, this doesn t happen; likewise no problem I think) if a is a F contiguous view on another array.","bug"
|
|
"6013","2038","Numpy where cannot handle masked arrays with boolean arrays for masks Trac 1440).
|
|
Original ticket http projects.scipy.org numpy ticket 1440 on 2010 03 27 by trac user brendanarnold, assigned to unknown.
|
|
To reproduce ...
|
|
In [1] a = np.ma.arange 5) In [2] a Out[2] masked array data = [0 1 2 3 4], mask = False, fill value = 999999) In [3] np.where a == 1) Out[3] masked array data = [1], mask = False, fill value = 999999) ,) In [4] a.mask = a == 1 In [5] a Out[5] masked array data = [0 2 3 4], mask = [False True False False False], fill value = 999999) In [6] np.where a == 1) ValueError Traceback most recent call last) C \Documents and Settings\ba1224\My Documents\Python\<ipython console> in <module> ) C \Python26\lib\site packages\numpy\ma\core.pyc in array finalize self, obj) 2445 Finalize the mask ........... 2446 if self.
|
|
mask is not nomask > 2447 self.
|
|
mask.shape = self.shape 2448 return 2449 .................................. ValueError total size of new array must be unchanged","bug"
|
|
"6013","2039","rollaxis ) has confusing error messages and should maybe interpret negative start argument differently Trac 1441).
|
|
Original ticket http projects.scipy.org numpy ticket 1441 on 2010 03 31 by trac user kbasye, assigned to unknown.
|
|
The rollaxis ) function in numeric.py allows negative arguments for both the axis and the start.
|
|
For the axis, the usual Python interpretation of a negative value applies; the axis is chosen by counting from the back.
|
|
For the start, however, 1 is not the last possible position, but the second to last.
|
|
I understand that fixing this is a change in the function s behavior, but at least the existing behavior is undocumented ).
|
|
In both cases, the error messages can be confusing if too negative values are given because the error raised uses modified values of the arguments and suggests a legal range that s smaller than what the function can actually handle.
|
|
Since it s short and I ve changed many lines, I m just going to paste a suggested replacement here rather than a diff hope that s OK. def rollaxis a, axis, start=0) Roll the specified axis until it lies in a given position.
|
|
Parameters a ndarray Input array.
|
|
axis int The axis to roll.
|
|
The positions of the other axes do not change relative to one another.
|
|
start int, optional The axis is rolled until it lies before this position.
|
|
Returns res ndarray Output array.
|
|
See Also roll Roll the elements of an array by a number of positions along a given axis.
|
|
Examples >>> a = np.ones 3,4,5,6)) >>> rollaxis a, 3, 1).shape 3, 6, 4, 5) >>> rollaxis a, 2).shape 5, 3, 4, 6) >>> rollaxis a, 1, 4).shape 3, 5, 6, 4) First axis becomes the last >>> rollaxis a, 0, 1).shape 4, 5, 6, 3) Last axis becomes the first >>> rollaxis a, 1).shape 6, 3, 4, 5) n = a.ndim orig axis, orig start = axis, start if axis < 0 axis = n if start < 0 start = n 1 msg = rollaxis s d) must be >= d and < d if not 0 <= axis < n) raise ValueError, msg axis , orig axis, n, n) if not 0 <= start < n 1) raise ValueError, msg start , orig start, n 1, n 1) if axis < start) removing axis below will shift the start position start = 1 if axis == start return a axes = range 0,n) axes.remove axis) axes.insert start, axis) return a.transpose axes)","bug"
|
|
"6013","2040","warning directives are not portable Trac 1443).
|
|
Original ticket http projects.scipy.org numpy ticket 1443 on 2010 03 31 by cgohlke, assigned to unknown.
|
|
numpy trunk r8313 contains 14 warning statements and therefore can not be built with Microsoft Visual C compilers.
|
|
Please consider using pragma directives instead.
|
|
numpy\core\src\multiarray\scalartypes.c.src 1965) warning XXX the format is not implemented!
|
|
this needs more work numpy\core\src\scalarmathmodule.c.src 1310) warning XXX is it important to do something with the integers here?
|
|
numpy\core\src\scalarmathmodule.c.src 1354) warning XXX is it important to do something with the integers here?
|
|
numpy\core\src\scalarmathmodule.c.src 1397) warning XXX is it important to do something with the integers here?
|
|
numpy\core\src\scalarmathmodule.c.src 1440) warning XXX is it important to do something with the integers here?
|
|
numpy\core\src\scalarmathmodule.c.src 1524) warning XXX is it important to do something with the integers here?
|
|
numpy\core\src\multiarray\buffer.c 251) warning XXX should it use UTF 8 here?
|
|
numpy\core\src\multiarray\common.c 311) warning XXX what is the correct thing to do here?
|
|
numpy\core\src\multiarray\ctors.c 1099) warning XXX what is the correct thing to do here?
|
|
numpy\f2py\cfuncs.py 269) warning HAVE LONG LONG is not available.
|
|
Redefining pyobj from long long.
|
|
numpy\lib\src\ compiled base.c 577) warning XXX is this correct at all!?
|
|
numpy\numarray\ capi.c 93) warning XXX needs implementation numpy\numarray\ capi.c 130) warning XXX needs implementation numpy\numarray\ capi.c 153) warning XXX needs implementation See also [http stackoverflow.com questions 171435 portability of warning preprocessor directive]","bug"
|
|
"6013","2041","numpy.array fails to change a list of tuples with float and string, whith to large string of number Trac 1444).
|
|
Original ticket http projects.scipy.org numpy ticket 1444 on 2010 04 01 by trac user mario, assigned to unknown.
|
|
In [39] c Out[39] [ 0.017527029320000002, TD C3 0.440sec ), 9.0204180442999995e 06, TD C3 0.920sec )] In [40] numpy.array c) Out[40] array [[ 0.01752702932 , TD C3 0.440sec ], [ 9.0204180443e , TD C3 0.920sec ]], dtype= |S14 ) So it seems that the length of the number is restricted by the lenght of the string.
|
|
Proof In [71] e=[] In [72] e.append a, TD C3 0.920second )) In [73] e.append b, TD C3 0.920sec )) In [74] numpy.array e) Out[74] array [[ 9.0204180443e 06 , TD C3 0.920second ], [ 0.01752702932 , TD C3 0.920sec ]], dtype= |S17 ) A way out was to set the dtype to |S17 but thats no real solution.","bug"
|
|
"6013","2042","add loadtxt support for multiple datasets Trac 1445).
|
|
Original ticket http projects.scipy.org numpy ticket 1445 on 2010 04 01 by trac user Onanymous, assigned to unknown.
|
|
Just like it is implemented in gnuplot, different datasets are separated by two empty lines, when reading the file I can choose which dataset to read.
|
|
Probably another parameter, like loadtxt fname, dataset=0).
|
|
Meanwhile there is no easy way to do this, especially when datasets have different shapes","enhancement"
|
|
"6013","2043","vstack not equivalent to concatenate x, axis=0) Trac 1446).
|
|
Original ticket http projects.scipy.org numpy ticket 1446 on 2010 04 04 by mforbes, assigned to unknown.
|
|
The documentation states that `vstack tup) is equivalent to concatenate tup, axis=0)`.
|
|
This fails for 1D arrays because of the promotion 1D >2D.
|
|
>>> tup = [np.zeros 1), np.zeros 5)] >>> np.concatenate tup, axis=0) array [ 0., 0., 0., 0., 0., 0.])
|
|
>>> vstack tup) Traceback most recent call last) ... ValueError array dimensions must agree except for d 0 If the current behaviour of automatic promotion is desired, then this is a documentation bug and a bug with the exception message all the dimensions do agree except for d 0 in this case they don t exist!)
|
|
I will add a note about this on the documentation page too.","bug"
|
|
"6013","2044","Inconsistent behavior when indexing a masked array with a bool array Trac 1447).
|
|
Original ticket http projects.scipy.org numpy ticket 1447 on 2010 04 04 by trac user nathanielpeterson, assigned to unknown.
|
|
This part works as I would expected import numpy as np a=np.ma.fix invalid np.array [np.nan, 1,0,1])) b=np.ma.fix invalid np.array [np.nan, 1,0,1])) idx= a==b) print a[idx][3]) 1.0 Note that a[idx] has shape 4,).
|
|
But if I change the first element of b from np.nan to 2.0 then a[idx2] has shape 3,) despite np.alltrue idx==idx2) being True c=np.ma.fix invalid np.array [2.0, 1,0,1])) idx2= a==c) assert np.alltrue idx==idx2)) a[idx2][3] Traceback most recent call last) File home np test.py , line 12, in <module> a[idx2][3] File usr lib python2.6 dist packages numpy ma core.py , line 2578, in getitem dout = ndarray.
|
|
getitem ndarray.view self, ndarray), indx) IndexError index out of bounds I looked at repr idx) and repr idx2) and they appear to have equal values in all respects.
|
|
Therefore, it seems inconsistent and puzzling) why a[idx2] does not behave the same as a[idx].","bug"
|
|
"6013","2045","Intel Fortran compiler detection fails on Mac OS X 10.6 Trac 1448).
|
|
Original ticket http projects.scipy.org numpy ticket 1448 on 2010 04 08 by trac user themiwi, assigned to unknown.
|
|
Running the version cmd as in numpy distutils ccompiler.py) results in usr bin ifort FI V c test.f o test.o Intel R) Fortran Intel R) 64 Compiler Professional for applications running on Intel R) 64, Version 11.1 Build 20090827 Package ID m cprof p 11.1.067 Copyright C) 1985 2009 Intel Corporation.
|
|
All rights reserved.
|
|
Intel R) Fortran 11.1 2582 Using the following patch fixes things for me, but probably will break other platforms, so somebody with a better knowledge of the whole compiler detection code should take care of this.
|
|
Index numpy distutils fcompiler intel.py =================================================================== numpy distutils fcompiler intel.py revision 8324) numpy distutils fcompiler intel.py working copy) 24,8 24,8 compiler type = intel compiler aliases = ifort ,) description = Intel Fortran Compiler for 32 bit apps version match = intel version match 32 bit|IA 32 ) description = Intel Fortran Compiler for 32 64 bit apps version match = intel version match 32 bit|64)|IA 32|64) ) possible executables = [ ifort , ifc ]","bug"
|
|
"6013","2046","uint64 hash function bug Trac 1449).
|
|
Original ticket http projects.scipy.org numpy ticket 1449 on 2010 04 09 by trac user erian, assigned to unknown.
|
|
>>> nbr = 1<< 50L >>> numpy.uint64 nbr) == nbr True >>> hash numpy.uint64 nbr)) == hash nbr) False Affects 2.0.0.dev8324 as well","bug"
|
|
"6013","2047","Patch with Ziggurat method for Normal distribution Trac 1450).
|
|
Original ticket http projects.scipy.org numpy ticket 1450 on 2010 04 10 by trac user ilan, assigned to charris.
|
|
I ve written a patch which replaces the Box Muller transform in numpy random mtrand randomkit.c with the faster Ziggurat method.
|
|
The patch also includes updates to doc strings which include the relevant references Doornik, J.A.
|
|
2005), An Improved Ziggurat Method to Generate Normal Random Samples , mimeo, Nuffield College, University of Oxford, and www.doornik.com research.
|
|
Marsaglia, G. and Tsang, W. W. 2000), The Ziggurat Method for Generating Random Variables , Journal of Statistcal Software 5, http www.jstatsoft.org v05 i08 Marsaglia, G. 1964) Generating a variable from the tail of the normal distribution , Technometrics 6, 101 102.
|
|
My patch is based on the support code which J. Doornik has made available on his web page first reference).
|
|
I ve contacted him, and he has given permission to include his code in numpy on the condition that his paper is referenced which the patch does).
|
|
The patch was written against http svn.scipy.org svn numpy tags 1.4.0, and I ve just checked to make sure it can still be applied to the trunk revision 8324).
|
|
I ve tested the patch on Windows, MacOSX, Linux and Solaris.
|
|
The speedup is a bit over 2 times, when creating a large number of Standard Normal distributed random numbers.
|
|
To verify that the method works and produces good random numbers, I have created a large array, using randn ) transformed using the cumulative distribution function) it back to uniform distributed random numbers written the data to a large file performed the diehard tests I used the dieharder program) on the data Feel free to hammer the patch.
|
|
I hope this patch will make its way into numpy soon.","enhancement"
|
|
"6013","2048","add filename attribute to numpy.memmap patch tests included) Trac 1451).
|
|
Original ticket http projects.scipy.org numpy ticket 1451 on 2010 04 12 by trac user brentp, assigned to unknown.
|
|
the numpy memmap does not have an associated filename.
|
|
this patch adds that.","enhancement"
|
|
"6013","2049","pickle memmap d arrays patch tests) Trac 1452).
|
|
Original ticket http projects.scipy.org numpy ticket 1452 on 2010 04 14 by trac user brentp, assigned to unknown.
|
|
currently cant pickle a np.memmap object.
|
|
this patch adds that functionality.","enhancement"
|
|
"6013","2050","deprecated decorator does not work as advertized Trac 1453).
|
|
Original ticket http projects.scipy.org numpy ticket 1453 on 2010 04 14 by cournape, assigned to cournape.
|
|
The deprecated decorator effectively disables the test","bug"
|
|
"6013","2051","numpy.linalg.eig computes wrong eigenvalues Trac 1454).
|
|
Original ticket http projects.scipy.org numpy ticket 1454 on 2010 04 16 by trac user skavhaug, assigned to unknown.
|
|
When using numpy.linalg.eig to compute the eigenvector eigenvalue pairs of a non symmetric and non Hessian) matrix, the errors defined as np.linalg.norm np.dot a, v[ ,i]) w[i] v[ ,i]) is off by several orders of magnitude up to 14 orders) compared to when performing the same calculation in octave.","bug"
|
|
"6013","2052","numpy.cov should accept ddof keyword argument Trac 1455).
|
|
Original ticket http projects.scipy.org numpy ticket 1455 on 2010 04 16 by trac user aisaac, assigned to charris.
|
|
The ``bias`` keyword has been replaced with ``ddof`` on related functions e.g., np.var and np.std) but not yet for np.cov.","bug"
|
|
"6013","2053","add ``dot`` as array method Trac 1456).
|
|
Original ticket http projects.scipy.org numpy ticket 1456 on 2010 04 16 by trac user aisaac, assigned to unknown.
|
|
Having ``dot`` as an array method would make many expressions much more readable.
|
|
E.g., a .dot b) .dot c) instead of np.dot np.dot a,b),c)","enhancement"
|
|
"6013","2054","TestClog failure on Windows Trac 1457).
|
|
Original ticket http projects.scipy.org numpy ticket 1457 on 2010 04 18 by rgommers, assigned to unknown.
|
|
Reported against 1.4.1rc2, and according to Josef has been failing for a long time on Windows FAIL test special values test umath complex.TestClog) Traceback most recent call last) File C \Python26\lib\site packages\numpy\core\tests\test umath complex.py , l ine 179, in test special values assert almost equal np.log x), y) File C \Python26\lib\site packages\numpy\testing\utils.py , line 437, in asse rt almost equal DESIRED s\n str actual), str desired))) AssertionError Items are not equal ACTUAL [ NaN 2.35619449j] DESIRED inf 2.35619449019j) Marking as knownfail for the 1.4.1 release.","bug"
|
|
"6013","2055","loadtxt doesn t unpack structred arrays with unpack=True Trac 1458).
|
|
Original ticket http projects.scipy.org numpy ticket 1458 on 2010 04 18 by trac user gdub, assigned to unknown.
|
|
The following should work !python import numpy; from StringIO import StringIO d = StringIO M 21 72\nF 35 58 ) a,b,c = numpy.loadtxt d, dtype={ names gender , age , weight ), formats S1 , i4 , f4 )}, unpack=True)","bug"
|
|
"6013","2056","mtrand.beta does not handle small parameters well Trac 1459).
|
|
Original ticket http projects.scipy.org numpy ticket 1459 on 2010 04 20 by trac user JohnReid, assigned to unknown.
|
|
I ve been using scipy.stats.beta and I ve noticed it returns NaN when the parameters are small.
|
|
I ve tried using the same parameters in R code and it handles them just fine.
|
|
R also handles parameters up to 17 orders of magnitude smaller.
|
|
Is there any documentation on which parameter ranges are acceptable?
|
|
Can I expect similar results with other distributions?
|
|
Please see my recent post on scipy user scipy.org with Josef s follow up.
|
|
Here s the code I used import scipy.stats as S, numpy as N from rpy2.robjects import r alpha, beta = 0.0710, 0.4222 for i in xrange 20) x from scipy = S.beta.rvs alpha, beta) x from R = r.rbeta 1, alpha, beta) print Alpha= .2e; Beta= .2e; scipy.stats.beta.rvs= .2e; R.rbeta= .2e alpha, beta, x from scipy, x from R[0]) alpha = 10. beta = 10. and the output from it Alpha=7.10e 02; Beta=4.22e 01; scipy.stats.beta.rvs=2.75e 11; R.rbeta=6.60e 02 Alpha=7.10e 03; Beta=4.22e 02; scipy.stats.beta.rvs=3.73e 84; R.rbeta=4.50e 124 Alpha=7.10e 04; Beta=4.22e 03; scipy.stats.beta.rvs=1.00e 00; R.rbeta=1.00e 00 Alpha=7.10e 05; Beta=4.22e 04; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 313 Alpha=7.10e 06; Beta=4.22e 05; scipy.stats.beta.rvs=nan; R.rbeta=1.00e 00 Alpha=7.10e 07; Beta=4.22e 06; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 315 Alpha=7.10e 08; Beta=4.22e 07; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 316 Alpha=7.10e 09; Beta=4.22e 08; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 317 Alpha=7.10e 10; Beta=4.22e 09; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 318 Alpha=7.10e 11; Beta=4.22e 10; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 319 Alpha=7.10e 12; Beta=4.22e 11; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 320 Alpha=7.10e 13; Beta=4.22e 12; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 321 Alpha=7.10e 14; Beta=4.22e 13; scipy.stats.beta.rvs=nan; R.rbeta=3.95e 322 Alpha=7.10e 15; Beta=4.22e 14; scipy.stats.beta.rvs=nan; R.rbeta=1.00e 00 Alpha=7.10e 16; Beta=4.22e 15; scipy.stats.beta.rvs=nan; R.rbeta=1.00e 00 Alpha=7.10e 17; Beta=4.22e 16; scipy.stats.beta.rvs=nan; R.rbeta=0.00e 00 Alpha=7.10e 18; Beta=4.22e 17; scipy.stats.beta.rvs=nan; R.rbeta=1.00e 00 Alpha=7.10e 19; Beta=4.22e 18; scipy.stats.beta.rvs=nan; R.rbeta=0.00e 00 Alpha=7.10e 20; Beta=4.22e 19; scipy.stats.beta.rvs=nan; R.rbeta=0.00e 00 Alpha=7.10e 21; Beta=4.22e 20; scipy.stats.beta.rvs=nan; R.rbeta=0.00e 00","bug"
|
|
"6013","2057","test polynomial.TestDocs overly sensitive to output formatting Trac 1460).
|
|
Original ticket http projects.scipy.org numpy ticket 1460 on 2010 04 23 by drnlm, assigned to unknown.
|
|
Seen on the windows buildbot.
|
|
Using python 2.5, Windows XP, Mingw32 build.
|
|
FAIL test doctests test polynomial.TestDocs) Traceback most recent call last) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\lib\tests\test polynomial.py , line 90, in test doctests return rundocs ) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 962, in rundocs AssertionError Some doctests failed File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\lib\tests\test polynomial.py , line 20, in test polynomial Failed example print poly1d [100e 90, 1.234567e 9j 3, 1234.999e8])) Expected 2 1e 88 x 3 1.235e 09j) x 1.235e 11 Got 2 1e 088 x 3 1.235e 009j) x 1.235e 011 Not seen using python 2.6","bug"
|
|
"6013","2058","test umath complex failures with Windows XP Migw32 Trac 1461).
|
|
Original ticket http projects.scipy.org numpy ticket 1461 on 2010 04 23 by drnlm, assigned to unknown.
|
|
From the windows buildbot Windows XP, python 2.5, mingw32 ====================================================================== FAIL test umath complex.TestCexp.test special values <ufunc exp >, 1.
|
|
INF, 0, 1.
|
|
INF, 0) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ Inf NaNj]) y array 1.
|
|
INF 0j)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 60, in check ninf inf raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, inf) is 1.
|
|
IND00, 1.
|
|
IND00), expected 0, 0) >> raise AssertionError cexp inf, inf) is f, f), expected 0, 0) nan nanj).real, nan nanj).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 73, in check inf inf raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, inf) is 1.
|
|
IND00, 1.
|
|
IND00), expected inf, nan) >> raise AssertionError cexp inf, inf) is f, f), expected inf, nan) nan nanj).real, nan nanj).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 86, in check ninf nan raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, nan) is 1.
|
|
QNAN0, 1.
|
|
QNAN0), expected 0, 0) >> raise AssertionError cexp inf, nan) is f, f), expected 0, 0) nan nanj).real, nan nanj).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 99, in check inf nan raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, nan) is 1.
|
|
QNAN0, 1.
|
|
QNAN0), expected inf, nan) >> raise AssertionError cexp inf, nan) is f, f), expected inf, nan) nan nanj).real, nan nanj).imag)) ====================================================================== FAIL test special values test umath complex.TestClog) Traceback most recent call last) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 144, in test special values self.assertRaises FloatingPointError, np.log, x) AssertionError FloatingPointError not raised if hasattr excClass, name ) FloatingPointError = excClass.
|
|
unittest else FloatingPointError = str excClass) >> raise self.failureException, s not raised FloatingPointError ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ Inf NaNj]) y array 1.
|
|
INF 1.
|
|
INFj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ NaN Infj]) y array 1.
|
|
INF 1.
|
|
INFj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ Inf NaNj]) y array 1.
|
|
INF 1.
|
|
INFj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ NaN Infj]) y array 1.
|
|
INF 1.
|
|
INFj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1.
|
|
IND, 1.
|
|
INF, 1.
|
|
INF, 1.
|
|
INF) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest self.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 539, in check complex value assert equal f z1), z2) File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 256, in assert equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 686, in assert array equal File ..\numpy install25\Lib\site packages\numpy\testing\utils.py , line 596, in assert array compare AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ NaN NaNj]) y array 1.
|
|
INF 1.
|
|
INFj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values None,) Traceback most recent call last) File c \Python25\Lib\site packages\nose\case.py , line 186, in runTest sRunning unit tests for numpy NumPy version 2.0.0.dev8351 NumPy is installed in C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy Python version 2.5.4 r254 67916, Dec 23 2008, 15 10 54) [MSC v.1310 32 bit Intel)] nose version 0.11.2 elf.test self.arg) File C \buildbot\numpy\b11\numpy install25\Lib\site packages\numpy\core\tests\test umath complex.py , line 340, in check ninf nan raise AssertionError msgform z.real, z.imag)) AssertionError csqrt inf, nan) is 1.
|
|
QNAN0, 1.
|
|
QNAN0), expected nan, inf) >> raise AssertionError csqrt inf, nan) is f, f), expected nan, inf) nan nanj).real, nan nanj).imag)) The weird output formatting appears to be a long standing issue, see 1463 for example), but the results are clearly not what was expected.
|
|
Similar failures seen when using python 2.6","bug"
|
|
"6013","2059","importing umath results in glibc detected python corrupted double linked list error Trac 1462).
|
|
Original ticket http projects.scipy.org numpy ticket 1462 on 2010 04 23 by drnlm, assigned to unknown.
|
|
As pointed out by Hodgestar on scipy PYTHONPATH= usr lib python2.5 site packages numpy core python Python 2.5.5 r255 77872, Feb 1 2010, 19 53 42) [GCC 4.4.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import umath >>> quit ) glibc detected python corrupted double linked list 0x08cdc570 However PYTHONPATH= usr lib python2.5 site packages numpy core python Python 2.5.5 r255 77872, Feb 1 2010, 19 53 42) [GCC 4.4.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> import umath >>> quit ) The bug is present in numpy 1.3, but is reproducible using trunk on the sparc buildbot.","enhancement"
|
|
"6013","2060","Allow disabling balancing in `linalg.eig` Trac 1463).
|
|
Original ticket http projects.scipy.org numpy ticket 1463 on 2010 04 24 by pv, assigned to pv.
|
|
Eigenvalue computations via `eig` should perhaps allow the user to optionally disable matrix balancing.
|
|
The LAPACK expert driver DGEEVX supports such additional options, so adding the interface should be possible.
|
|
Balancing occasionally may cause some problems, such as in 2051, and having an option to turn it off could be useful in some special cases...","enhancement"
|
|
"6013","2061","numpy.set numeric ops causes double free Trac 1464).
|
|
Original ticket http projects.scipy.org numpy ticket 1464 on 2010 04 25 by trac user mras, assigned to unknown.
|
|
The following snippet bug.py) causes a double free on exit tested with Numpy v. 1.3.0 and the latest source code from SVN on Python 2.6.4) def binfunc x,y) return None import numpy print numpy.
|
|
version oldopts = numpy.set numeric ops add=binfunc) a = numpy.arange 16) print a a Output from gdb Starting program usr bin python bug.py [Thread debugging using libthread db enabled] 2.0.0.dev8355 None glibc detected usr bin python corrupted double linked list 0x082d8540 C Program received signal SIGINT, Interrupt.
|
|
0x0012d422 in kernel vsyscall ) gdb) bt 0 0x0012d422 in kernel vsyscall ) 1 0x00264bd3 in lll lock wait private ) at .. nptl sysdeps unix sysv linux i386 i686 .. i486 lowlevellock.S 95 2 0x001fbf57 in L lock 9510 ) from lib tls i686 cmov libc.so.6 3 0x001fa7c6 in GI libc free mem=0x8302050) at malloc.c 3714 4 0x0012160f in dl scope free old=0x8302050) at dl open.c 175 5 0x0011c327 in dl map object deps map=<value optimized out>, preloads=<value optimized out>, npreloads=<value optimized out>, trace mode=0, open mode= 2147483648) at dl deps.c 668 6 0x0012184e in dl open worker a=0xbfffed38) at dl open.c 326 7 0x0011d4e6 in dl catch error objname=<value optimized out>, errstring=<value optimized out>, mallocedp=<value optimized out>, operate=0x1216b0 <dl open worker>, args=0xbfffed38) at dl error.c 178 8 0x00121200 in dl open file=0x2aa85d libgcc s.so.1 , mode=<value optimized out>, caller dlopen=0x0, nsid= 2, argc=2, argv=0xbffff8c4, env=0xbffff8d0) at dl open.c 615 9 0x0028df92 in do dlopen ptr=0xbfffeec8) at dl libc.c 86 10 0x0011d4e6 in dl catch error objname=<value optimized out>, errstring=<value optimized out>, mallocedp=<value optimized out>, operate=0x28df30 <do dlopen>, args=0xbfffeec8) at dl error.c 178 11 0x0028e091 in dlerror run operate=<value optimized out>, args=0xbfffeec8) at dl libc.c 47 12 0x0028e1ab in GI libc dlopen mode name=0x2aa85d libgcc s.so.1 , mode= 2147483647) at dl libc.c 160 13 0x0026bfb8 in init ) at .. sysdeps i386 backtrace.c 44 14 0x00139160 in pthread once ) at .. nptl sysdeps unix sysv linux i386 pthread once.S 122 15 0x0026c1ad in GI backtrace array=0xbffff49c, size=64) at .. sysdeps i386 backtrace.c 121 16 0x001ebefb in libc message do abort=2, fmt=0x2af578 glibc detected s s 0x s \n ) at .. sysdeps unix sysv linux libc fatal.c 168 17 0x001f5ff1 in malloc printerr action=<value optimized out>, str=0x8302048 \300\207\371\267!
|
|
, ptr=0x82d8540) at malloc.c 6217 18 0x001f7a89 in int free av=<value optimized out>, p=0x82d8398) at malloc.c 4912 19 0x001fa7cd in GI libc free mem=0x82d83a0) at malloc.c 3716 20 0x081581b3 in code dealloc co=0xb7fd6020) at ..
|
|
Objects codeobject.c 270 21 0x081603d4 in func dealloc op=0xb7f82d84) at ..
|
|
Objects funcobject.c 454 22 0x0807fe6e in list dealloc op=0xb7f75dac) at ..
|
|
Objects listobject.c 306 23 0x080f8223 in PyInterpreterState Clear interp=0x8273008) at .. Python pystate.c 107 24 0x080fa312 in Py Finalize ) at .. Python pythonrun.c 476 25 0x0805c34b in Py Main argc=2, argv=0xbffff8c4) at .. Modules main.c 625 26 0x0805baeb in main argc=2, argv=0xbffff8c4) at .. Modules python.c 23 Earlier on I got another trace that I unfortunately neglected to save) where the double free was in Py Finalize > ... > dict dealloc > dict dealloc > ufunc dealloc.
|
|
I believe that was caused by me calling numpy.set numeric ops oldopts) at the end and thus causing something else to be doublefree d. I ve investigated to the best of my knowledge and I think the cause of the problem is in the SET macro in numpy core src multiarray number.c.
|
|
If I am reading the docs for [http docs.python.org c api dict.html PyDict GetItemString] correctly it returns a borrowed reference [http docs.python.org release 2.5.2 ext refcountsInPython.html] ) and the it s reference count should thus be incremented before being stored in the numerical operations array.
|
|
I ve tried to check other call sites of PyArray SetNumericOps to see if I had misunderstood the semantics, but I think it s simply a bug.
|
|
The below code seems to fix the problem, but as I am inexperienced at writing python extensions in C and with NumPy in particular I would appreciate some feedback.
|
|
Index numpy core src multiarray number.c =================================================================== numpy core src multiarray number.c revision 8355) numpy core src multiarray number.c working copy) 30,6 30,7 if !
|
|
PyCallable Check temp))) { \ return 1; \ } \ Py INCREF temp); \ Py XDECREF n ops.op); \ n ops.op = temp; \ }","bug"
|
|
"6013","2062","PyCObject is deprecated in Python2.7 Trac 1465).
|
|
Original ticket http projects.scipy.org numpy ticket 1465 on 2010 04 26 by charris, assigned to charris.
|
|
Just a heads up to myself that the version checks need to be updated.","bug"
|
|
"6013","2063","Uninstall in Windows does not remove some directories Trac 1466).
|
|
Original ticket http projects.scipy.org numpy ticket 1466 on 2010 04 27 by trac user threexk, assigned to cournape.
|
|
After uninstalling 1.4.0 numpy 1.4.0 win32 superpack python2.6.exe), the following directory remains dev Python26 Lib site packages numpy The first time I tried uninstalling it, a dialog popped up with counts of files and directories removed, but said that one file directory was not removed.
|
|
It gave no indication which.
|
|
I moved away the aforementioned directory, reinstalled 1.4.0, and uninstalled again.
|
|
This time, it did not report that any files directories could not be removed, but dev Python26 Lib site packages numpy.
|
|
I believe it is a bug that this directory is not removed at uninstall.
|
|
This is seen on Windows 7.","bug"
|
|
"6013","2064","MACOSX DEPLOYMENT TARGET should not have default Trac 1467).
|
|
Original ticket http projects.scipy.org numpy ticket 1467 on 2010 04 27 by trac user mallorn, assigned to cournape.
|
|
PROBLEM numpy distutils fcompiler gnu.py sets MACOSX DEPLOYMENT TARGET to a default of 10.3 which breaks distutils generating cross compiling errors under Snow Leopard or any non 10.3 OS version.
|
|
SOLUTION Do not set a default value of MACOSX DEPLOYMENT TARGET only set it if defined in the makefile or the environment.
|
|
Alternatively use the existing distutils.util.get platform ) function to get the build platform and extract the OS version from that if needed.","bug"
|
|
"6013","2065","Datetime test creation overflow fails Trac 1468).
|
|
Original ticket http projects.scipy.org numpy ticket 1468 on 2010 04 29 by pv, assigned to stefanv.
|
|
On a 32 bit system, r8372 ====================================================================== FAIL test creation overflow test datetime.TestDateTime) Traceback most recent call last) File home pauli wrk scipy numpy dist linux lib python2.6 site packages numpy core tests test datetime.py , line 68, in test creation overflow err msg= Datetime conversion error for unit s unit) File home pauli wrk scipy numpy dist linux lib python2.6 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal Datetime conversion error for unit ms ACTUAL 567052800 DESIRED 322689600000 >> raise AssertionError \nItems are not equal Datetime conversion error for unit ms\n ACTUAL 567052800\n DESIRED 322689600000 ) Any ideas what s wrong?","bug"
|
|
"6013","2066","Matrix multiplication dot) fails silently for large integers v1.3.0rc2 Trac 1469).
|
|
Original ticket http projects.scipy.org numpy ticket 1469 on 2010 04 30 by trac user gtpitt, assigned to unknown.
|
|
When multiplying 2 matrices with large integer values the following error was observed.
|
|
Please note that no warning or error message was received.
|
|
This is running in Eclipse Version 3.4.1 Build id M20080911 1700.
|
|
Pydev v 1.4.6.2788. version 1.3.0rc2 gm1 [[ 0 1] [ 1 157249]] gm2 [[ 8762 28349] [ 54857 177487]] gm = dot gm2,gm1) note the order, gm2 is first) [[ 28349 162893367] [ 177487 2139904344]] correct answer [[ 28349 4457860663] [ 177487 27909708120]] 4457860663 34 bits) 100001001101101011000111000110111","bug"
|
|
"6013","2067","Improve docs for writing new ufuncs e.g.
|
|
explain GIL issues) Trac 1470).
|
|
Original ticket http projects.scipy.org numpy ticket 1470 on 2010 05 02 by trac user kishor.iitr, assigned to pv.
|
|
I am trying to raise an exception in ufunc, I get segmentation fault.
|
|
The data type for the functions is PyArray DOUBLE.
|
|
Please note that if we change the type to PyArray OBJECT then it is working fine.","enhancement"
|
|
"6013","2068","Access to private attributes of Py objects without using the official C API Trac 1471).
|
|
Original ticket http projects.scipy.org numpy ticket 1471 on 2010 05 03 by stefanv, assigned to unknown.
|
|
PyPy generates proxy objects to many of the standard Python objects Py Slice, etc.).
|
|
These proxy objects only expose the official C API, which NumPy doesn t always respect.
|
|
Here is a general log which highlights some of the problems not all of the errors below need to be fixed, but many of them do) numpy core src multiarray common.c 616 error expected xe2x80x98 xe2x80x99 before xe2x80x98PyPyString Checkxe2x80x99 numpy core src multiarray scalartypes.c.src 513 error request for member xe2x80x98tp reprxe2x80x99 in something not a structure or union numpy core src multiarray scalartypes.c.src 513 error request for member xe2x80x98tp strxe2x80x99 in something not a structure or union numpy core src multiarray scalartypes.c.src 513 error request for member xe2x80x98tp reprxe2x80x99 in something not a structure or union numpy core src multiarray scalartypes.c.src 513 error request for member xe2x80x98tp strxe2x80x99 in something not a structure or union numpy core src multiarray scalartypes.c.src 591 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 591 error Each undeclared identifier is reported only once numpy core src multiarray scalartypes.c.src 591 error for each function it appears in.)
|
|
numpy core src multiarray scalartypes.c.src 606 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 591 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 606 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 591 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 606 error xe2x80x98Py PRINT RAWxe2x80x99 undeclared first use in this function) numpy core src multiarray scalartypes.c.src 2216 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98lengthxe2x80x99 numpy core src multiarray scalartypes.c.src 2216 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98lengthxe2x80x99 numpy core src multiarray scalarapi.c 375 error xe2x80x98PyComplexObjectxe2x80x99 undeclared first use in this function) numpy core src multiarray scalarapi.c 375 error expected expression before xe2x80x98)xe2x80x99 token numpy core src multiarray scalarapi.c 377 error expected expression before xe2x80x98)xe2x80x99 token numpy core src multiarray scalarapi.c 694 error xe2x80x98PyStringObjectxe2x80x99 has no member named xe2x80x98ob shashxe2x80x99 numpy core src multiarray scalarapi.c 696 error xe2x80x98PyStringObjectxe2x80x99 has no member named xe2x80x98ob sstatexe2x80x99 numpy core src multiarray scalarapi.c 696 error xe2x80x98SSTATE NOT INTERNEDxe2x80x99 undeclared first use in this function) numpy core src multiarray scalarapi.c 710 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray scalarapi.c 716 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray scalarapi.c 717 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray scalarapi.c 718 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray scalarapi.c 719 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98lengthxe2x80x99 numpy core src multiarray scalarapi.c 720 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98hashxe2x80x99 numpy core src multiarray scalarapi.c 721 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98defencxe2x80x99 numpy core src multiarray scalarapi.c 748 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray datetime.c 3 22 error datetime.h No such file or directory numpy core src multiarray arraytypes.c.src 190 error xe2x80x98Py complexxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 190 error expected xe2x80x98;xe2x80x99 before xe2x80x98oopxe2x80x99 numpy core src multiarray arraytypes.c.src 204 error xe2x80x98oopxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 190 error xe2x80x98Py complexxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 190 error expected xe2x80x98;xe2x80x99 before xe2x80x98oopxe2x80x99 numpy core src multiarray arraytypes.c.src 204 error xe2x80x98oopxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 190 error xe2x80x98Py complexxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 190 error expected xe2x80x98;xe2x80x99 before xe2x80x98oopxe2x80x99 numpy core src multiarray arraytypes.c.src 204 error xe2x80x98oopxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 307 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray arraytypes.c.src 792 error xe2x80x98PyDateTime IMPORTxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 813 error xe2x80x98PyDateTime IMPORTxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 910 error xe2x80x98PyDateTime IMPORTxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 947 error xe2x80x98PyDateTime IMPORTxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 955 error xe2x80x98PyDateTime Deltaxe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 955 error expected expression before xe2x80x98)xe2x80x99 token numpy core src multiarray arraytypes.c.src 956 error expected expression before xe2x80x98)xe2x80x99 token numpy core src multiarray arraytypes.c.src 957 error expected expression before xe2x80x98)xe2x80x99 token numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray arraytypes.c.src 1494 error request for member xe2x80x98tp newxe2x80x99 in something not a structure or union numpy core src multiarray descriptor.c 212 error expected xe2x80x98 xe2x80x99 before xe2x80x98PyPyDict Checkxe2x80x99 numpy core src multiarray descriptor.c 1233 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray ctors.c 1586 error expected xe2x80x98 xe2x80x99 before xe2x80x98PyCObject Checkxe2x80x99 numpy core src multiarray iterators.c 202 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98stepxe2x80x99 numpy core src multiarray iterators.c 206 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98stepxe2x80x99 numpy core src multiarray iterators.c 217 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98startxe2x80x99 numpy core src multiarray iterators.c 221 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98startxe2x80x99 numpy core src multiarray iterators.c 235 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98stopxe2x80x99 numpy core src multiarray iterators.c 239 error xe2x80x98PySliceObjectxe2x80x99 has no member named xe2x80x98stopxe2x80x99 numpy core src multiarray ucsnarrow.c 94 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 95 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 100 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 101 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 102 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98lengthxe2x80x99 numpy core src multiarray ucsnarrow.c 103 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98hashxe2x80x99 numpy core src multiarray ucsnarrow.c 104 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98defencxe2x80x99 numpy core src multiarray ucsnarrow.c 116 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 117 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 117 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 118 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 119 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 123 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 numpy core src multiarray ucsnarrow.c 124 error xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98lengthxe2x80x99 numpy core src multiarray multiarraymodule.c 2255 error xe2x80x98PyOS sighandler txe2x80x99 undeclared first use in this function) numpy core src multiarray multiarraymodule.c 2255 error expected xe2x80x98;xe2x80x99 before xe2x80x98 npy sig savexe2x80x99 numpy core src multiarray multiarraymodule.c 2256 error xe2x80x98 npy sig savexe2x80x99 undeclared first use in this function) numpy core src multiarray multiarraymodule.c 2674 error xe2x80x98PyOS sighandler txe2x80x99 undeclared first use in this function) numpy core src multiarray multiarraymodule.c 2674 error expected xe2x80x98;xe2x80x99 before xe2x80x98 npy sig savexe2x80x99 numpy core src multiarray multiarraymodule.c 2674 error xe2x80x98 npy sig savexe2x80x99 undeclared first use in this function) numpy core src multiarray multiarraymodule.c 2812 error xe2x80x98PyComplex Typexe2x80x99 undeclared first use in this function) numpy core src multiarray multiarraymodule.c 2888 error request for member xe2x80x98tp hashxe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2893 error request for member xe2x80x98tp hashxe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2920 error request for member xe2x80x98tp hashxe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2927 error request for member xe2x80x98tp richcomparexe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2927 error request for member xe2x80x98tp comparexe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2927 error request for member xe2x80x98tp hashxe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2928 error request for member xe2x80x98tp richcomparexe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2928 error request for member xe2x80x98tp comparexe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 2928 error request for member xe2x80x98tp hashxe2x80x99 in something not a structure or union numpy core src multiarray multiarraymodule.c 3050 error xe2x80x98PyType GenericNewxe2x80x99 undeclared first use in this function)","bug"
|
|
"6013","2069","Assigning Transpose of slice of a 3d Array Always Gives a Symmetric Array Trac 1472).
|
|
Original ticket http projects.scipy.org numpy ticket 1472 on 2010 05 03 by trac user TomW, assigned to unknown.
|
|
A = numpy.random.rand 1,3,3) A[0] = A[0].T The array, A[0] is symmetric.
|
|
The lower triangular part of the original A[0] is copied to the upper triangular part.","bug"
|
|
"6013","2070","genfromtxt issue with EOL and or unicode Trac 1473).
|
|
Original ticket http projects.scipy.org numpy ticket 1473 on 2010 05 04 by vincentdavis, assigned to unknown.
|
|
Basic problem, I file saved on a Mac using excel as a csv cannot me opened with genfromtxt.
|
|
This will work, but if not expected to be necessary.
|
|
f = file x.csv , U ) genfromtxt f, ...) Stxc3xa9fan van der Walt has fixed the same issue in loadfromtxt http projects.scipy.org numpy changeset 8375 Long Story I ran into this issue and it was discussed on the pystatsmodels mailing list.
|
|
Here is the setup Running on a Mac 10.6 Using Office 2008 Saving an spreadsheet using excel save as a csv file.
|
|
Try to import using genfromtxt fails, report a EOL error I thought this was because the EOL was wrong, It seems the file has \r as the line ending this may be wrong) anyway I changed it to \n and it works fine.
|
|
I am told on the pystatsmodels mailing list) that this is actually because the file is in unicode and that genfromtxt does not read the EOL correctly.
|
|
To me it is a bug because one might expect a user to what to save a file from excel and read it using genfromtxt.
|
|
And for useres with little experience the problem is not obvious.
|
|
I guess this is not a problem with py3?
|
|
ORIGINAL ATTEMPT datatype = [ date , |S9 ), gpd , i8 ), temp , i8 precip , f16 )] data = np.genfromtxt waterdata.csv , delimiter= , , skip header=1, dtype=datatype) Traceback most recent call last) File Applications WingIDE.app Contents MacOS src debug tserver sandbox.py , line 1, in <module> Used internally for debug sandbox under external interpreter File Library Frameworks EPD64.framework Versions 6.1 lib python2.6 site packages numpy lib io.py , line 1048, in genfromtxt raise IOError End of file reached before encountering data. )
|
|
IOError End of file reached before encountering data.
|
|
THIS DOES NOT WORK >>> s = file data with CR.csv , r ) >>> data = np.genfromtxt s, delimiter= , , skip header=1, dtype=None) Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks EPD64.framework Versions 6.1 lib python2.6 site packages numpy lib io.py , line 1048, in genfromtxt raise IOError End of file reached before encountering data. )
|
|
IOError End of file reached before encountering data.
|
|
>>> data = np.genfromtxt s, delimiter= , , , dtype=None) File <stdin> , line 1 data = np.genfromtxt s, delimiter= , , , dtype=None) THIS DOES WORK >>> s = file data with CR.csv , U ) >>> data = np.genfromtxt s, delimiter= , , skip header=1, dtype=None) >>> data array [ 1 1 00 , 8021472, 52, 0.02), 1 2 00 , 9496016, 46, 0.059999999999999998), 1 3 00 , 8478792, 29, 0.0), ..., 12 29 02 , 10790000, 61, 0.0), 12 30 02 , 9501000, 44, 0.0), 12 31 02 , 9288000, 53, 0.0)], dtype=[ f0 , |S8 ), f1 , <i8 ), f2 , <i8 ), f3 , <f8 )])","bug"
|
|
"6013","2071","numpy.take has a bug ubuntu 10.4) Trac 1474).
|
|
Original ticket http projects.scipy.org numpy ticket 1474 on 2010 05 05 by trac user hillier dani, assigned to unknown.
|
|
import numpy a = numpy.ones 256,256,3)) b = numpy.take a,[1,0,2],axis=2) Traceback most recent call last) File <stdin> , line 1, in <module> File usr local lib python2.6 dist packages numpy 1.4.1 py2.6 linux x86 64.egg numpy core fromnumeric.py , line 103, in take return take indices, axis, out, mode) IndexError index out of range for array Example fails with above error numpy 1.4.1 python 2.6.5 ubuntu 10.4 Example works Mac OSX 10.6, same python numpy version.","bug"
|
|
"6013","2072","Cannot access date time internals without ctypes installed Trac 1475).
|
|
Original ticket http projects.scipy.org numpy ticket 1475 on 2010 05 05 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== ERROR test type check.TestDateTimeData.test basic Traceback most recent call last) File usr stsci pyssgdev 2.5.4 nose case.py , line 183, in runTest self.test self.arg) File usr stsci pyssgdev 2.5.4 numpy lib tests test type check.py , line 382, in test basic assert equal datetime data a.dtype), us , 1, 1, 1)) File usr stsci pyssgdev 2.5.4 numpy lib type check.py , line 610, in datetime data raise RuntimeError, Cannot access date time internals without ctypes installed RuntimeError Cannot access date time internals without ctypes installed Happens on Mac OSX Leopard, 32 bit python 2.5.4 Solaris 8, 32 bit Python 2.5.4 Not on Red Hat Enterprise 4 Linux, 32 or 64 bit python 2.5 Red Hat Enterprise 5 linux, 32 bit python 2.5 When you compile python, it automatically decides whether to include ctypes.
|
|
In my experience, it likes Linux but not Mac or Solaris.
|
|
It is not clear to me if this means that the test should be changed e.g.
|
|
to skip if ctypes is not available) or if ctypes should become a hard and fast requirement to run numpy.","bug"
|
|
"6013","2073","selector.choose more keyword list entries than argument specifiers Trac 1476).
|
|
Original ticket http projects.scipy.org numpy ticket 1476 on 2010 05 05 by stsci sienkiew, assigned to unknown.
|
|
numpy.test ) fails in many places with this error ====================================================================== ERROR test all test multiarray.TestArgmax) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test multiarray.py , line 681, in test all assert all amax == aargmax.choose a.transpose i, axes))) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR test basic test multiarray.TestChoose) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test multiarray.py , line 1190, in test basic A = np.choose self.ind, self.x, self.y)) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 295, in choose return wrapit a, choose , choices, out=out, mode=mode) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 37, in wrapit result = getattr asarray obj),method) args, kwds) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR test broadcast1 test multiarray.TestChoose) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test multiarray.py , line 1194, in test broadcast1 A = np.choose self.ind, self.x2, self.y2)) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 295, in choose return wrapit a, choose , choices, out=out, mode=mode) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 37, in wrapit result = getattr asarray obj),method) args, kwds) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR test broadcast2 test multiarray.TestChoose) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test multiarray.py , line 1198, in test broadcast2 A = np.choose self.ind, self.x, self.y2)) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 295, in choose return wrapit a, choose , choices, out=out, mode=mode) File usr stsci pyssgdev 2.5.4 numpy core fromnumeric.py , line 37, in wrapit result = getattr asarray obj),method) args, kwds) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR Test native double input with array min max.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 436, in test array double act = self.clip a, m, M) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 378, in clip return selector.choose a, m, M), out=out) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR Ensure that the clip ) function takes an out= argument.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 779, in test clip func takes out self.clip a, m, M, ac) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 378, in clip return selector.choose a, m, M), out=out) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR Test native double input with array min max Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 758, in test clip inplace array self.clip a, m, M, ac) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 378, in clip return selector.choose a, m, M), out=out) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR Test native double input with scalar min max Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 768, in test clip inplace simple self.clip a, m, M, ac) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 378, in clip return selector.choose a, m, M), out=out) RuntimeError more keyword list entries than argument specifiers ====================================================================== ERROR Test clip for non contiguous native input and native scalar min max.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 484, in test clip non contig act = self.clip a, 1.6, 1.7) File usr stsci pyssgdev 2.5.4 numpy core tests test numeric.py , line 378, in clip return selector.choose a, m, M), out=out) RuntimeError more keyword list entries than argument specifiers There are more, but this is a representative sample of the problem.
|
|
Happens on Mac OSX Leopard, 32 bit python 2.5.4 Solaris 8, 32 bit python 2.5.4 Red Hat Enterprise 4 Linux, 32 bit python 2.5.4 Red Hat Enterprise 4 Linux, 64 bit python 2.5.4 Not on Red Hat Enterprise 5 Linux, 64 bit python 2.5.4 First observed May 3 in version 2.0.0.dev8379","bug"
|
|
"6013","2074","NumPy inverts a singular matrix Trac 1477).
|
|
Original ticket http projects.scipy.org numpy ticket 1477 on 2010 05 05 by trac user egonschiele, assigned to pv.
|
|
This might be just a question of precision.
|
|
In the following code, A2 is a singular matrix.
|
|
NumPy calculates it s inverse and prints out a non zero determinant even though the matrix A2 is clearly singular A = array [[.1,.01,.3],[.2,.99,.3],[.7,0,.4]]) I = identity 3) A2 = A I this should be singular print inv A2) prints out a singular matrix !!)
|
|
print det A2) prints 2.33146835171e 18 instead of 0","bug"
|
|
"6013","2075","NumPy Homepage does not make license obvious Trac 1478).
|
|
Original ticket http projects.scipy.org numpy ticket 1478 on 2010 05 06 by trac user Figs, assigned to pv.
|
|
At the NumPy homepage, http numpy.scipy.org there is no indication of where to find copyright or license information.
|
|
As a potential new user for the library, this is unhelpful.
|
|
Searching on Google suggests that the project is released under the new BSD license is this correct?).
|
|
If so, this should be prominently displayed as it would be a major selling point for professional use of the library.","bug"
|
|
"6013","2076","test universal newline has string bug in python 3.1 Trac 1479).
|
|
Original ticket http projects.scipy.org numpy ticket 1479 on 2010 05 07 by charris, assigned to unknown.
|
|
ERROR test universal newline test io.TestLoadTxt) Traceback most recent call last) File usr local lib python3.1 site packages numpy lib tests test io.py , line 412, in test universal newline data = np.loadtxt name) File usr local lib python3.1 site packages numpy lib npyio.py , line 632, in loadtxt first vals = split line first line) File usr local lib python3.1 site packages numpy lib npyio.py , line 610, in split line line = line.split comments)[0].strip ) TypeError Can t convert bytes object to str implicitly","bug"
|
|
"6013","2077","apply over axes does not take the mask into account Trac 1480).
|
|
Original ticket http projects.scipy.org numpy ticket 1480 on 2010 05 07 by huard, assigned to unknown.
|
|
The mask of an array is ignored by apply over axes and apply along axis.
|
|
>>> o = np.ma.array [[1,1],[1,1]], mask=[[0,0],[0,1]]) >>> o.sum ) 3 >>> np.apply over axes np.sum, o, [0,1]) array [[4]]) There is no mention of masked arrays handling in the docstring.","enhancement"
|
|
"6013","2078","Move non function API from ndarrayobject.h into arraytypes.h Trac 1481).
|
|
Original ticket http projects.scipy.org numpy ticket 1481 on 2010 05 08 by trac user abingham, assigned to unknown.
|
|
This change moves the parts of ndarrayobject.h that don t depend on the PY ARRAY UNIQUE SYMBOL and NO IMPORT into a new header, arraytypes.h.
|
|
The intent of this change is to allow users to include the types, enums, etc.
|
|
associated with arrays without needing to worry about the macro definitions.","enhancement"
|
|
"6013","2079","lapack interface handles non native byte order in an invalid way Trac 1482).
|
|
Original ticket http projects.scipy.org numpy ticket 1482 on 2010 05 11 by yarikoptic, assigned to unknown.
|
|
Original report is at http bugs.debian.org cgi bin bugreport.cgi?bug=581043 but now I ve tested with current master trunk 8404 and issue present as well, e.g > python c import numpy as N; print N. version ;print N.linalg.cholesky N.array [[ 5.7998084, 2.1825367 ], [ 2.1825367, 9.85910595]], dtype= >f8 )) 2.0.0.dev Traceback most recent call last) File <string> , line 1, in <module> File home yoh proj numpy build lib.linux x86 64 2.5 numpy linalg linalg.py , line 517, in cholesky Cholesky decomposition cannot be computed numpy.linalg.linalg.LinAlgError Matrix is not positive definite Cholesky decomposition cannot be computed > python c import numpy as N; print N. version ;print N.linalg.cholesky N.array [[ 5.7998084, 2.1825367 ], [ 2.1825367, 9.85910595]], dtype= <f8 )) 2.0.0.dev [[ 2.40827914 0. ]
|
|
[ 0.90626401 3.00629199]] myself I consider it of higher than normal priority since the issue is quite nastily veiled some functions, e.g.
|
|
svd would just produce bogus results) and I am surprised noone seems hit it by now.","bug"
|
|
"6013","2080","Delayed segfault in numpy Python 2.4 Trac 1483).
|
|
Original ticket http projects.scipy.org numpy ticket 1483 on 2010 05 11 by pv, assigned to unknown.
|
|
Any ideas about this, ie.
|
|
can someone with Python 2.4 reproduce it?
|
|
Worksforme on Python 2.4.6.
|
|
It s currently blocking the Scipy doc wiki from updating, so it would be nice to have it fixed.
|
|
Looks like a refcount bug somewhere in Numpy.
|
|
Bisection search via git points that the first bad revision is r8262 Python 2.4.3 1, Jan 21 2009, 01 11 33) [GCC 4.1.2 20071124 Red Hat 4.1.2 42)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy, scipy, gc >>> numpy.
|
|
version 2.0.0.dev8404 >>> scipy.
|
|
version 0.8.0.dev5623 >>> import scipy.integrate >>> gc.collect ) python Modules gcmodule.c 275 visit decref Assertion `gc >gc.gc refs !
|
|
= 0 failed.
|
|
Program received signal SIGABRT, Aborted.
|
|
0x0000003238430215 in raise ) from lib64 libc.so.6 gdb) bt 0 0x0000003238430215 in raise ) from lib64 libc.so.6 1 0x0000003238431cc0 in abort ) from lib64 libc.so.6 2 0x0000003238429696 in assert fail ) from lib64 libc.so.6 3 0x0000003c2e0b9c5c in ?? )
|
|
from usr lib64 libpython2.4.so.1.0 4 0x0000003c2e05a89d in ?? )
|
|
from usr lib64 libpython2.4.so.1.0 5 0x0000003c2e0ba157 in ?? )
|
|
from usr lib64 libpython2.4.so.1.0 6 0x0000003c2e0babf4 in ?? )
|
|
from usr lib64 libpython2.4.so.1.0 7 0x0000003c2e094747 in PyEval EvalFrame ) from usr lib64 libpython2.4.so.1.0 8 0x0000003c2e0958a5 in PyEval EvalCodeEx ) from usr lib64 libpython2.4.so.1.0 9 0x0000003c2e0958f2 in PyEval EvalCode ) from usr lib64 libpython2.4.so.1.0 10 0x0000003c2e0b1f29 in ?? )
|
|
from usr lib64 libpython2.4.so.1.0 11 0x0000003c2e0b37e7 in PyRun InteractiveOneFlags ) from usr lib64 libpython2.4.so.1.0 12 0x0000003c2e0b38de in PyRun InteractiveLoopFlags ) from usr lib64 libpython2.4.so.1.0 13 0x0000003c2e0b39ec in PyRun AnyFileExFlags ) from usr lib64 libpython2.4.so.1.0 14 0x0000003c2e0b980d in Py Main ) from usr lib64 libpython2.4.so.1.0 15 0x000000323841d974 in libc start main ) from lib64 libc.so.6 16 0x0000000000400629 in start ) gdb) q Yeah, debugging symbols missing ATM, but I guess the backtrace would be of little help anyway since the crash is nonlocal.)","bug"
|
|
"6013","2081","Missing parenthesis Trac 1484).
|
|
Original ticket http projects.scipy.org numpy ticket 1484 on 2010 05 17 by trac user Ademan, assigned to unknown.
|
|
In some places, if statements whose only condition is a CPython macro, omit the parenthesis.
|
|
These aren t guaranteed by the C API to be macros, and so they require parenthesis.
|
|
Attached is a patch addressing this.","enhancement"
|
|
"6013","2082","free ) invalid next size fast) error on Ubuntu 9.10 and 10.04 Trac 1485).
|
|
Original ticket http projects.scipy.org numpy ticket 1485 on 2010 05 18 by trac user FarMcKon, assigned to unknown.
|
|
I am getting free ) invalid next size fast) when i run the code below.
|
|
I am using Ubuntu 9.10 64 bit) on a MacBook Pro 5,1.
|
|
I am using pyton 2.5 and numpy version 1.4.1 The code I run to generate the error is from numpy import c = linspace pi,pi,5) c When I hit enter to display the value of C, I get the error glibc detected python free ) invalid next size fast) 0x000000000264c6a0 sometimes it crashes at this point, sometimes I have to suspend python and kill 9 it.","bug"
|
|
"6013","2083","Add option to use stable sorting algorithm in unique 1d) Trac 1486).
|
|
Original ticket http projects.scipy.org numpy ticket 1486 on 2010 05 18 by trac user bverheg, assigned to unknown.
|
|
The unique 1d function in arraysetops.py returns the index of the unique elements in an array when using the argument return index=True), but in case of equal elements it is undetermined which of the multiple indices will be returned.
|
|
I often need to find the first occurence of each of the unique elements in the array.
|
|
The most elegant way I have come up with to achieve this, is to create a minor variation of unique 1d, using a stable sorting algorithm mergesort ) instead of the quicksort default, at this place in the code if return inverse or return index perm = ar.argsort kind= mergesort ) aux = ar[perm] I suggest therefore to add an argument to the unique 1d function to allow the selection of the stable mergesort algorithm at the place shown above.
|
|
In my opinion, it does not make much sense to use this slower algorithm also in the other sort operations in unique 1d.","enhancement"
|
|
"6013","2084","vectorize ) result depends on execution order; is this normal?
|
|
Trac 1487).
|
|
Original ticket http projects.scipy.org numpy ticket 1487 on 2010 05 21 by lebigot, assigned to unknown.
|
|
The dtype of the array returned by a vectorize ) d function depends on the execution order Session 1 In [1] f vec = vectorize lambda x x) In [3] f vec arange 3)) Out[3] array [0, 1, 2]) In [4] f vec arange 0.1, 3)) floats Out[4] array [0, 1, 2]) integers Session 2 In [1] f vec = vectorize lambda x x) In [2] f vec arange 0.1, 3)) floats, same as above Out[2] array [ 0.1, 1.1, 2.1]) floats, and not integers as above Thus, vectorize ) has a serious side effect the dtype of an array returned by a vectorized function depends on which vectorized function was called first!
|
|
This creates bugs that are hard to detect, because such a behavior is highly unusual.
|
|
I had understood from the documentation of vectorize f) that the type of the first value returned by f instead determined the dtype of the returned array","bug"
|
|
"6013","2085","vectorize ) could the dtype of the output array be the smallest possible dtype ?
|
|
Trac 1488).
|
|
Original ticket http projects.scipy.org numpy ticket 1488 on 2010 05 21 by lebigot, assigned to unknown.
|
|
It would be nice if vectorize ) d functions returned arrays with the smallest possible dtype .
|
|
Maybe such a feature could be added as a flag in vectorize s arguments?
|
|
Thus, the result of a vectorized function would be more robust; currently, if the first call to the vectorized function returned an array of integers, all its subsequent calls return integers even when the user really expects floats.
|
|
This problem might be hard to detect, in a long and complicated matrix calculation.
|
|
This could be one solution to the ticket in http projects.scipy.org numpy ticket 1487","enhancement"
|
|
"6013","2086","numpydoc sphinxext) broken by latest sphinx source.
|
|
Trac 1489).
|
|
Original ticket http projects.scipy.org numpy ticket 1489 on 2010 05 22 by mforbes, assigned to pv.
|
|
The numpydoc sphinxext) sphinx extension as of svn version 8426 is broken when used with the latest sphinx tip 2f0924a705b2).
|
|
The first hg version of sphinx where this starts failing is 98f9c2d6f51f where domains are introduced.
|
|
It is not clear to me right now how to fix this.
|
|
Using sphinx quickstart and adding the numpydoc extension no other modifications to the defaults and including autodoc) gives the following make html sphinx build b html d build doctrees .
|
|
build html Making output directory... Running Sphinx v1.0pre Exception occurred File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages numpydoc 0.3.dev py2.6.egg numpydoc numpydoc.py , line 137, in get directive raise RuntimeError No directive named s found name) RuntimeError No directive named function found ... make [html] Error 1 Here is the traceback Sphinx version 1.0pre Docutils version 0.6 release Jinja2 version 2.2.1 Traceback most recent call last) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages Sphinx 1.0pre 2f0924a705b2dev 20100521 py2.6.egg sphinx cmdline.py , line 172, in main warningiserror, tags) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages Sphinx 1.0pre 2f0924a705b2dev 20100521 py2.6.egg sphinx application.py , line 114, in init self.setup extension extension) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages Sphinx 1.0pre 2f0924a705b2dev 20100521 py2.6.egg sphinx application.py , line 251, in setup extension mod.setup self) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages numpydoc 0.3.dev py2.6.egg numpydoc numpydoc.py , line 120, in setup app.add directive np name, wrap mangling directive name, objtype)) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages numpydoc 0.3.dev py2.6.egg numpydoc numpydoc.py , line 140, in wrap mangling directive base directive = get directive base directive name) File Library Frameworks Python.framework Versions 6.0.4 lib python2.6 site packages numpydoc 0.3.dev py2.6.egg numpydoc numpydoc.py , line 137, in get directive raise RuntimeError No directive named s found name) RuntimeError No directive named function found","bug"
|
|
"6013","2087","searchsorted does not work properly for non native endianness Trac 1490).
|
|
Original ticket http projects.scipy.org numpy ticket 1490 on 2010 05 26 by trac user briangerke, assigned to charris.
|
|
I am currently using numpy version 1.4.1rc2 .
|
|
I have encountered this in several other versions, however.
|
|
The searchsorted function sometimes gives incorrect results when it is passed two arrays that are both in non native endianness.
|
|
The following is copied from an email I sent to another numpy user, to make sure I wasn t going crazy.
|
|
He confirmed the bug.
|
|
> Consider this running interactively in ipython on my little endian Intel MacBook) > > In [5] searchsorted arange 900, dtype= i4 ), array [894,895], dtype= i4 )) > > Out[5] array [894, 895]) > > That s working as it should searchsorted just says to insert 894 and 895 > before elements 894 and 895 in a range array that runs from 0 to 899.
|
|
> > But now if I turn both input arrays into big endian, things get weird > > In [6] searchsorted arange 900, dtype= >i4 ), array [894,895], > dtype= >i4 )) > > Out[6] array [894, 900]) > > Now searchsorted says to insert 894 before the 894th element correct) but > that 895 is bigger than 899 and should thus be inserted at the end what?).
|
|
> > Similar nonsense occurs if I use big endian floats > > In [13] searchsorted arange 900, dtype= >f4 ), array [894,895], > dtype= >f4 )) > > Out[13] array [679, 899]) > > Now it says that 894<679 and 895=899.
|
|
Huh?
|
|
> > As far as I can tell, this only occurs when BOTH arrays are explicitly > big endian > > In [7] searchsorted arange 900, dtype= >i4 ), array [894,895], dtype= i4 )) > > Out[7] array [894, 895]) > > In [8] searchsorted arange 900, dtype= i4 ), array [894,895], dtype= >i4 )) > > Out[8] array [894, 895]) > > Apparently if EITHER input array has native endian ness, there is no > problem.
|
|
I would be surprised to discover that this is expected behavior.
|
|
Surely the endianness ought not to matter.","bug"
|
|
"6013","2088","Bug when comparing str to object Trac 1491).
|
|
Original ticket http projects.scipy.org numpy ticket 1491 on 2010 05 26 by kwgoodman, assigned to unknown.
|
|
Comparisons of dtpye=str on the left hand side with dtype=object on the right hand side fail >> a1 = np.array [ a , b ], dtype=object) >> a2 = np.array [ a , b ]) >> >> a1 == a2 array [ True, True], dtype=bool) Looks good >> a2 == a1 False Bug Expected behavior >> a1 = np.array [1, 2], dtype=object) >> a2 = np.array [1, 2]) >> >> a1 == a2 array [ True, True], dtype=bool) >> a2 == a1 array [ True, True], dtype=bool) And similarly for the other comparison operators.","bug"
|
|
"6013","2089","numpy.random.dirichlet does not check parameters and hangs when invalid Trac 1492).
|
|
Original ticket http projects.scipy.org numpy ticket 1492 on 2010 05 27 by trac user JohnReid, assigned to unknown.
|
|
Invalid negative parameters can cause numpy.random.dirichlet to hang and never return.
|
|
Note the last parameter is very small and negative.
|
|
a=numpy.array [ 5.64081750e 01, 2.38946208e 02, 8.51495969e 02, 8.75933900e 02, 8.38591170e 02, 3.42876279e 03, 2.98989982e 04, 4.41411094e 04, 3.98247752e 06, 2.10975195e 05, 1.37975897e 07, 1.75958131e 07, 1.33898931e 07, 2.82288418e 07, 3.14258899e 09, 8.48688121e 10, 1.84816307e 10, 3.78878308e 11, 1.41103782e 11, 1.01214380e 11, 6.75038727e 12, 6.63904489e 12, 5.17487421e 13, 7.07797216e 14, 7.97992577e 15, 2.43240084e 14, 6.83348954e 15, 2.15791088e 15, 3.67340489e 15, 1.26240005e 15, 1.93752302e 17, 4.83345747e 17, 1.65380214e 18, 1.39628957e 18, 1.42096661e 18, 4.89995717e 19, 3.53568962e 20, 5.80340121e 20, 1.79688051e 19, 1.88465566e 16]) numpy.random.dirichlet a) I m running numpy 1.3.0 on Ubuntu and the code above just spun at 100 CPU without returning.","bug"
|
|
"6013","2090","Masked array equality with None is buggy Trac 1493).
|
|
Original ticket http projects.scipy.org numpy ticket 1493 on 2010 05 27 by huard, assigned to pierregm.
|
|
Equality and inequality of masked arrays with None raise an error.
|
|
>>> a = np.ma.array [3,4,5], mask=[1,1,1]) >>> a == None AttributeError Traceback most recent call last) home davhua1 <ipython console> in <module> ) usr local lib python2.6 dist packages numpy ma core.pyc in ne self, other) 3610 omask = getattr other, mask , nomask) 3611 if omask is nomask > 3612 check = ndarray.
|
|
ne self.filled 0), other).view type self)) 3613 check.
|
|
mask = self.
|
|
mask 3614 else AttributeError bool object has no attribute view","bug"
|
|
"6013","2091","functions return NotImplemented instead of raising error Trac 1494).
|
|
Original ticket http projects.scipy.org numpy ticket 1494 on 2010 05 27 by josef pkt, assigned to unknown.
|
|
examples >>> np.sqrt 5 ) NotImplemented >>> a = np.sqrt 5 ) >>> type a) <type NotImplementedType > >>> np.log np.array [ 5 ])) NotImplemented >>> np.power np.array [ 5 ]),5) NotImplemented >>> np.multiply np.array [ 5 ]),5) NotImplemented","bug"
|
|
"6013","2092","better error message for np.poly np.zeros 0,0))) Trac 1495).
|
|
Original ticket http projects.scipy.org numpy ticket 1495 on 2010 05 28 by josef pkt, assigned to unknown.
|
|
np.poly np.zeros 0,0))) crashed with numpy 1.4.0 but has been fixed in trunk 2.0.0) the error message could be more explicit In [1] np.poly np.zeros 0,0))) ValueError Traceback most recent call last) home charris <ipython console> in <module> ) usr local lib python2.6 dist packages numpy lib polynomial.pyc in poly seq of zeros) 126 pass 127 else > 128 raise ValueError, input must be 1d or square 2d array.
|
|
129 130 if len seq of zeros) == 0 ValueError input must be 1d or square 2d array.","bug"
|
|
"6013","2093","Numpy build on AIX 5.3 ERROR Undefined symbol .npy carg Trac 1496).
|
|
Original ticket http projects.scipy.org numpy ticket 1496 on 2010 05 28 by trac user gdrapala, assigned to unknown.
|
|
Encountered subject error while compiling Numpy from trunk r8445.
|
|
Using trunk to work around issue 1999 compiling with gcc on AIX 5.3).
|
|
After dump g build temp.aix 5.3 2.6 libnpymath.a showed all npy carg symbols were missing, made the following change in numpy core src npymath npy math complex.c.src.
|
|
Changed begin repeat1 kind = cabs,carg KIND = CABS,carg to begin repeat1 kind = cabs,carg KIND = CABS,CARG and it compiled successfully.","bug"
|
|
"6013","2094","Make nose detailed errors optional Trac 1497).
|
|
Original ticket http projects.scipy.org numpy ticket 1497 on 2010 05 30 by jseabold, assigned to unknown.
|
|
Right now the use of the detailed errors flag for nose tests is not optional.
|
|
I have attached a patch to make it optional.
|
|
The default behavior of using detailed errors is not changed though maybe it should be?
|
|
), as it is useful for the build bots.
|
|
I find it to be extra noise while debugging tests though.
|
|
Consider a failure with detailed errors flag in scipy r6447 scipy.test ) ====================================================================== FAIL test mio.test mat4 3d <type exceptions.DeprecationWarning >, <functools.partial object at 0x479a418>, <StringIO.StringIO instance at 0x753c6c8>, { a array [[[ 0, 1, 2, 3], Traceback most recent call last) File usr local lib python2.6 dist packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File usr local lib python2.6 dist packages numpy testing utils.py , line 982, in assert raises return nose.tools.assert raises args, kwargs) AssertionError DeprecationWarning not raised if hasattr excClass, name ) DeprecationWarning = excClass.
|
|
unittest else DeprecationWarning = str excClass) >> raise self.failureException, s not raised DeprecationWarning The same one without the flag scipy.test extra argv=[]) ====================================================================== FAIL test mio.test mat4 3d <type exceptions.DeprecationWarning >, <functools.partial object at 0x479a418>, <StringIO.StringIO instance at 0x6ac1248>, { a array [[[ 0, 1, 2, 3], Traceback most recent call last) File usr local lib python2.6 dist packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File usr local lib python2.6 dist packages numpy testing utils.py , line 982, in assert raises return nose.tools.assert raises args, kwargs) AssertionError DeprecationWarning not raised","enhancement"
|
|
"6013","2095","Make nose detailed errors optional Trac 1498).
|
|
Original ticket http projects.scipy.org numpy ticket 1498 on 2010 05 30 by jseabold, assigned to unknown.
|
|
Right now the use of the detailed errors flag for assert introspection in nose tests is not optional.
|
|
I have attached a patch to make it optional.
|
|
The default behavior of using introspection is not changed though maybe it should be?
|
|
), as it is useful for the build bots.
|
|
I find it to be extra noise while debugging tests though.
|
|
Consider a failure with detailed errors flag in scipy r6447 scipy.test ) ====================================================================== FAIL test mio.test mat4 3d <type exceptions.DeprecationWarning >, <functools.partial object at 0x479a418>, <StringIO.StringIO instance at 0x753c6c8>, { a array [[[ 0, 1, 2, 3], Traceback most recent call last) File usr local lib python2.6 dist packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File usr local lib python2.6 dist packages numpy testing utils.py , line 982, in assert raises return nose.tools.assert raises args, kwargs) AssertionError DeprecationWarning not raised if hasattr excClass, name ) DeprecationWarning = excClass.
|
|
unittest else DeprecationWarning = str excClass) >> raise self.failureException, s not raised DeprecationWarning The same one without the flag scipy.test extra argv=[]) ====================================================================== FAIL test mio.test mat4 3d <type exceptions.DeprecationWarning >, <functools.partial object at 0x479a418>, <StringIO.StringIO instance at 0x6ac1248>, { a array [[[ 0, 1, 2, 3], Traceback most recent call last) File usr local lib python2.6 dist packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File usr local lib python2.6 dist packages numpy testing utils.py , line 982, in assert raises return nose.tools.assert raises args, kwargs) AssertionError DeprecationWarning not raised","enhancement"
|
|
"6013","2096","test fails test ldexp test umath.TestLdexp) Trac 1499).
|
|
Original ticket http projects.scipy.org numpy ticket 1499 on 2010 06 02 by stsci sienkiew, assigned to unknown.
|
|
NumPy version 2.0.0.dev8449 ====================================================================== ERROR test ldexp test umath.TestLdexp) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test umath.py , line 363, in test ldexp assert almost equal ncu.ldexp np.array 2., np.longdouble), np.array 3, np.int16)), 16.)
|
|
TypeError function not supported for these types, and can t coerce safely to supported types Fails on Python 2.5.4, Solaris 8, sparc 32 bits.
|
|
Passes on Red Hat Enterprise Linux 4, 32 bits, python 2.5.4 RHE 4, 64 bits, python 2.5.4 RHE 5, 64 bits, python 2.5.4","bug"
|
|
"6013","2097","isfinite, isinf macros raise invalid value fp error flag Trac 1500).
|
|
Original ticket http projects.scipy.org numpy ticket 1500 on 2010 06 03 by efiring, assigned to unknown.
|
|
In npy math.h, if NPY HAVE DECL ISFINITE is not defined, the macro that is defined in place of a system supplied macro works by generating a nan when given an inf, and then using isnan to detect that case.
|
|
This raises the invalid value fp error flag, so every time numpy.isfinite encounters an inf, a warning is printed by default.
|
|
The situation is the same with isinf.
|
|
Here is the relevant code from npy math.h ifndef NPY HAVE DECL ISFINITE define npy isfinite x) !npy isnan x) x)) else define npy isfinite x) isfinite x)) endif ifndef NPY HAVE DECL ISINF define npy isinf x) !npy isfinite x) !npy isnan x)) else define npy isinf x) isinf x)) endif The problem became evident because of a bug in setup.py the NPY HAVE DECL macros were not being generated.
|
|
This was fixed in svn 8455, so now the problem will appear only on those very rare?)
|
|
platforms that do not have their own isinf and isfinite.","bug"
|
|
"6013","2098","Addition to the docstring standard an optional section for noting deprecation Trac 1501).
|
|
Original ticket http projects.scipy.org numpy ticket 1501 on 2010 06 03 by trac user dgoldsmith, assigned to pv.
|
|
Presently, the docstring standard does not specify how to note that an object is to be deprecated; it has been proposed that this needs to be rectified.
|
|
Obviously, this should be an optional section in general, but required for objects once it is decided that they are to be deprecated.
|
|
Discussion on scipy dev agreed that this section should be at or near the top, but at the top or between the One line and Extended Summaries have both been proposed we will try to reach a consensus here.
|
|
Proposed format is to utilize Sphinx .. deprecated directive; someone please provide a concrete example of what this looks like for example, does this directive support multi line content, and if so, what does that look like).
|
|
Proposed content summaries of deprecation schedule in version number time, not real time) and justification for deprecation e.g., being replaced, duplicates extant functionality elsewhere); existing alternatives to obtain the same functionality.
|
|
Feel strongly that it should contain something else?
|
|
Add it below as a comment.)
|
|
IMO, we should try to decide on this and update the standard by June 15 at the latest.
|
|
Have I forgotten anything?","enhancement"
|
|
"6013","2099","isnan and isinf not defined when using msvc9 Trac 1502).
|
|
Original ticket http projects.scipy.org numpy ticket 1502 on 2010 06 04 by cgohlke, assigned to unknown.
|
|
Changeset r8455, BUG fix missing macros definition for IEEE macros , breaks the build process on Windows when using Visual Studio 2008.
|
|
The changes in r8455 look correct.
|
|
The problem is that for some reason NPY HAVE DECL ISNAN and NPY HAVE DECL ISINF are defined in Python s pyconfig.h header, but the macros isnan and isinf are not.
|
|
Visual Studio does not have these C99 macros functions either, so the numpy build fails with unresolved external symbol isnan and ...isinf .
|
|
A workaround might be to use the msvc functions isnan ) and finite ) in npy math.h ifndef NPY HAVE DECL ISNAN define npy isnan x) x) != x)) else ifdef MSC VER define npy isnan x) isnan x)) else define npy isnan x) isnan x)) endif endif ifndef NPY HAVE DECL ISFINITE ifdef MSC VER define npy isfinite x) finite x)) else define npy isfinite x) !npy isnan x) x)) endif else define npy isfinite x) isfinite x)) endif ifndef NPY HAVE DECL ISINF define npy isinf x) !npy isfinite x) !npy isnan x)) else ifdef MSC VER define npy isinf x) !
|
|
finite x) !
|
|
isnan x)) else define npy isinf x) isinf x)) endif endif","bug"
|
|
"6013","2100","random.vonmises docstring shows incorrect function signature Trac 1503).
|
|
Original ticket http projects.scipy.org numpy ticket 1503 on 2010 06 04 by WarrenWeckesser, assigned to unknown.
|
|
The first line in the docstring of random.vonmises is vonmises mu=0.0, kappa=1.0, size=None) but the actual function definition is def vonmises self, mu, kappa, size=None) The first two arguments are positional, not keywords.","bug"
|
|
"6013","2101","Numpy Can t Link Against 64 bit threaded MKL libraries Trac 1504).
|
|
Original ticket http projects.scipy.org numpy ticket 1504 on 2010 06 04 by trac user SevenThunders, assigned to unknown.
|
|
The latest MKL libraries > 10.2) have somewhat complicated link requirements.
|
|
They require grouping flags and now apparently generate runtime errors if one tries to use numpy e.g.
|
|
dot )) after linking against the dynamic libraries in linux.
|
|
I m getting the errors described here http software.intel.com en us forums showthread.php?t=74037 The claim is that the dynamic loading of the mkl libraries is not in the correct order and cause undefined symbols errors.
|
|
The proposed solution is to link using static libraries, but the current build scheme does not offer that kind of control.
|
|
I thought perhaps the scons fork might get me there, but I currently can t install that because it seems to require a development version of NumScons 0.11 ?)
|
|
for which there are no downloadable binaries packages.
|
|
Seeing as how mkl is the currently the best way to get good numpy performance on Intel machines, I would think this would be a fairly high priority item.
|
|
I know I m not the only one with this problem.","bug"
|
|
"6013","2102","load ) yields corrupted incorrect data Trac 1505).
|
|
Original ticket http projects.scipy.org numpy ticket 1505 on 2010 06 04 by trac user RRosario, assigned to unknown.
|
|
Using a huge sparse 395000 x 395000) CSC matrix.
|
|
Tried to save data, indptr, indices and shape separately.
|
|
When I try to read back the indices pickle, I get incorrect data >> np.save indices.pickle , mymatrix.indices) >>> indices = np.load indices.pickle.npy ) >>> indices array [394852, 394649, 394533, ..., 0, 0, 0], dtype=int32) >>> mymatrix.indices array [394852, 394649, 394533, ..., 1557, 1223, 285], dtype=int32) The two arrays are not equal to each other.","bug"
|
|
"6013","2103","issctype returns True for string Trac 1506).
|
|
Original ticket http projects.scipy.org numpy ticket 1506 on 2010 06 07 by trac user pch, assigned to pv.
|
|
Summary says is all >>> issctype dtype str )) True If it s a feature instead of a bug, the behaviour should be documented.","bug"
|
|
"6013","2104","Error when build documentation Trac 1507).
|
|
Original ticket http projects.scipy.org numpy ticket 1507 on 2010 06 08 by trac user howarn, assigned to unknown.
|
|
cd doc make html mkdir p build touch build generate stamp mkdir p build html build doctrees LANG=C sphinx build b html d build doctrees source build html Running Sphinx v1.0b2 2.0.dev 2.0.0.dev Exception occurred File usr src RPM BUILD python module numpy 2.0.0 doc sphinxext numpydoc.py , line 137, in get directive raise RuntimeError No directive named s found name) RuntimeError No directive named function found The full traceback has been saved in usr src tmp sphinx err m2iNvQ.log, if you want to report the issue to the developers.
|
|
Please also report this if it was a user error, so that a better error message can be provided next time.
|
|
Either send bugs to the mailing list at <http groups.google.com group sphinx dev >, or report them in the tracker at <http bitbucket.org birkenfeld sphinx issues >.
|
|
Thanks!
|
|
make [html] Error 1","bug"
|
|
"6013","2105","isscalar doesn t handle ndarray s scalars Trac 1508).
|
|
Original ticket http projects.scipy.org numpy ticket 1508 on 2010 06 08 by trac user pch, assigned to unknown.
|
|
The following behaviour looks incoherent >>> a = array 1) >>> print numpy.isscalar a), numpy.rank a) False 0 Expected behaviour isscalar a) should return True","bug"
|
|
"6013","2106","Link in docstring standard to IEEE citation format is broken Trac 1509).
|
|
Original ticket http projects.scipy.org numpy ticket 1509 on 2010 06 11 by trac user dgoldsmith, assigned to unknown.
|
|
http www.ieee.org pubs transactions auinfo03.pdf, the current link in the docstring Standard for the text citation format of the IEEE, is obsolete.
|
|
The same information is now at http standards.ieee.org guides style 2009 Style Manual.pdf The link in the Standard needs to be updated accordingly.","bug"
|
|
"6013","2107","Feature higher order singular value decomposition HOSVD) Trac 1510).
|
|
Original ticket http projects.scipy.org numpy ticket 1510 on 2010 06 11 by trac user jiahao, assigned to pv.
|
|
Hi there, I recently need to compute higher order singular value decompositions HOSVDs) on some tensors.
|
|
I coded up a quick and dirty implementation as described in the original paper and thought it might be a useful contribution to the numpy.linalg library.
|
|
The code is attached and released to you under public domain.
|
|
If you do choose to incorporate the code, please note that 1) the code is a direct implementation of the algorithm described in the original paper, and could probably be done more quickly and cleverly, 2) it has only been tested on the few examples in the original paper 3) there are some discrepancies up to overall signs in some columns of the left singular matrices.
|
|
cite Lieven de Lathauwer, Bart de Moor, Joos Vandewalle, A multilinear singular value decomposition , SIAM J. Matrix Anal.
|
|
Appl.
|
|
21 4), 2000, 1253 1278","enhancement"
|
|
"6013","2108","define layout for HAVE LDOUBLE IEEE QUAD LE Trac 1511).
|
|
Original ticket http projects.scipy.org numpy ticket 1511 on 2010 06 13 by r0bertz, assigned to unknown.
|
|
Please check the attachment.
|
|
There is one thing that I am not sure.
|
|
That is the LDBL NBIT.
|
|
I checked glibc s info, it says the hidden bit is always 1 for normalized number and not stored.
|
|
So I guess it should be 0.","enhancement"
|
|
"6013","2109","Need tests for random.randn ) and similiar Trac 1512).
|
|
Original ticket http projects.scipy.org numpy ticket 1512 on 2010 06 13 by vincentdavis, assigned to unknown.
|
|
A Scipy test failed do to a seeded randn ) not returning the expected results details below).
|
|
I seemed to be an issue specific to enthought dist 6.2 1.
|
|
This highlights the need for a test of randn ) and similar.
|
|
Correct results as tested on several machines From Josef, I assume windows >>>>> np.random.seed 0) >>>>> np.random.randn 3) >> array [ 1.76405235, 0.40015721, 0.97873798]) Results from enthought 6.2 1 very different) On my machine OSX py 2.6.5, numpy 1.4.0 scipy 8.0b >>>> np.random.seed 0) >>>> np.random.randn 3) > array [ 0.06897149, 1.32078057, 1.5997924 ]) Also correct ubuntu 10.0.4 py 2.5.6 current scipy and numpy just built from the git source.
|
|
>>> np.random.seed 0) >>> np.random.randn 3) array [ 1.76405235, 0.40015721, 0.97873798]) http permalink.gmane.org gmane.comp.python.scientific.devel 13726 Failed scipy test on On my machine OSX py 2.6.5, numpy 1.4.0 scipy 8.0b enthought 6.2 1 FAIL test stats.test kstest Traceback most recent call last) File Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages nose case.py , line 186, in runTest self.test self.arg) File Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages scipy stats tests test stats.py , line 1228, in test kstest assert almost equal D, 0.12464329735846891, 15) File Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages numpy testing utils.py , line 459, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL 0.093893737596468518 DESIRED 0.1246432973584689","enhancement"
|
|
"6013","2110","flushing memmap incomplete Trac 1513).
|
|
Original ticket http projects.scipy.org numpy ticket 1513 on 2010 06 17 by rgommers, assigned to unknown.
|
|
Reported by Geoffrey Irving on the mailing list If I create an mmap ed array, and then generate another array referencing only its base, destruction of the original mmap ed array closes the mmap.
|
|
The second array is then prone to segfaults.
|
|
I think the best fix is to put the responsibility for flushing the mmap onto the actual mmap object and therefore the base object) directly, so that the numpy memmap object has no cleanup responsibilities.
|
|
A patch to core memmap.py follows.
|
|
It would be nicer to put the mmap flush class declaration somewhere outside the function, but that conflicts with the local import mmap.","bug"
|
|
"6013","2111","unique and NaN entries Trac 1514).
|
|
Original ticket http projects.scipy.org numpy ticket 1514 on 2010 06 18 by trac user rspringuel, assigned to unknown.
|
|
When unique operates on an array with multiple NaN entries its return includes a NaN for each entry that was NaN in the original array.
|
|
Examples a = random.randint 5,size=100).astype float) >>> a[12] = nan add a single nan entry >>> unique a) array [ 0., 1., 2., 3., 4., NaN]) >>> a[20] = nan add a second >>> unique a) array [ 0., 1., 2., 3., 4., NaN, NaN]) >>> a[13] = nan >>> unique a) and a third array [ 0., 1., 2., 3., 4., NaN, NaN, NaN]) This is probably due to the fact that x == y evaluates to False if both x and y are NaN.
|
|
Unique needs to have or isnan x) and isnan y)) added to the conditional that checks for the presence of a value in the already identified values.
|
|
I don t know were unique lives in numpy and couldn t find it when I went looking, so I can t make the change myself or even be sure what the exact syntax of the conditional should be).
|
|
Also, the following function can be used to patch over the behavior.
|
|
def nanunique x) a = numpy.unique x) r = [] for i in a if i in r or numpy.isnan i) and numpy.any numpy.isnan r))) continue else r.append i) return numpy.array r)","bug"
|
|
"6013","2112","Python string exceptions no more allowed in Python >= 2.6 Trac 1515).
|
|
Original ticket http projects.scipy.org numpy ticket 1515 on 2010 06 19 by sandrotosi, assigned to pearu.
|
|
Hello, I m forwarding here the Debian bugs http bugs.debian.org 585309 .
|
|
We recently run an archive check for string exceptions, and numpy resulted to contain one numpy f2py capi maps.py except raise sign2map expected complex number ` r,i)\ but got ` s\ as initial value of s. init,`a`) This code won t work with python version >= 2.6, so it would be really nice to have this fixed soon even because the solution is very simple, wrap the string in Exception )).
|
|
Thanks in advance, Sandro","bug"
|
|
"6013","2113","Numpy float cast to int when multiplying a python list Trac 1516).
|
|
Original ticket http projects.scipy.org numpy ticket 1516 on 2010 06 20 by tonysyu, assigned to unknown.
|
|
When multiplying a numpy float and a python list, the float is cast as an int and then does the usual python thing creates duplicates of the original sequence).
|
|
For example, >>> np.float64 2.3) [1, 2] [1, 2, 1, 2] Instead, the multiplication should either work like an array scalar >>> np.array 2.3) [1, 2] array [ 2.3, 4.6]) or throw an exception like a python float >>> 2.3 [1, 2] Traceback most recent call last) File <stdin> , line 1, in <module> TypeError can t multiply sequence by non int of type float","bug"
|
|
"6013","2114","numpy.load does not release file handle for npz files Trac 1517).
|
|
Original ticket http projects.scipy.org numpy ticket 1517 on 2010 06 22 by trac user etecman, assigned to unknown.
|
|
When opening a npz file with numpy.load, the handle to the file is not released such that the file cannot be deleted during the runtime of the script even if all returned values are deleted.
|
|
Error message WindowsError [Error 32] The process cannot access the file because it is being used by another process deleteme.npz Example import os import numpy a = numpy.array [[1, 2, 3], [4, 5, 6]]) filenameNPZ = deleteme.npz numpy.savez filenameNPZ, a = a) npz = numpy.load filenameNPZ) del npz os.remove filenameNPZ)","bug"
|
|
"6013","2115","numpy.concat does not appear to work across broadcast axes Trac 1518).
|
|
Original ticket http projects.scipy.org numpy ticket 1518 on 2010 06 22 by trac user eob, assigned to unknown.
|
|
When I m trying to concatenate two tensors together, the concatenate operation does not allow me to use broadcast dimensions using newaxis) in one of them.","bug"
|
|
"6013","2116","linalg tests test build.py incorrectly reports both g77 and gfortran Trac 1519).
|
|
Original ticket http projects.scipy.org numpy ticket 1519 on 2010 06 23 by trac user gvubrugier, assigned to unknown.
|
|
If libgfortran appears twice in the output of ldd, the test suite will fail although libg2c does not appear.
|
|
For example ldd usr lib python2.6 site packages numpy linalg lapack lite.so linux gate.so.1 => 0xffffe000) liblapack.so.3.0 => usr lib sse2 liblapack.so.3.0 0xb7105000) libf77blas.so.1.1 => usr lib sse2 libf77blas.so.1.1 0xb70e9000) libcblas.so.1.1 => usr lib sse2 libcblas.so.1.1 0xb70ca000) libatlas.so.3.0 => usr lib sse2 libatlas.so.3.0 0xb6b6e000) libgfortran.so.3 => usr lib libgfortran.so.3 0xb6ab3000) libm.so.6 => lib i686 libm.so.6 0xb6a8d000) libgcc s.so.1 => lib libgcc s.so.1 0xb6a7f000) libc.so.6 => lib i686 libc.so.6 0xb6931000) libblas.so.1.1 => usr lib sse2 libblas.so.1.1 0xb63a2000) libgfortran.so.2 => usr lib libgfortran.so.2 0xb6304000) lib ld linux.so.2 0xb77ff000) Possible solution The founds variable in grep dependencies ) should be a set.","bug"
|
|
"6013","2117","Numpy 1.5 for Python 2.7 and 3.1 Trac 1520).
|
|
Original ticket http projects.scipy.org numpy ticket 1520 on 2010 06 23 by cgohlke, assigned to unknown.
|
|
In the hope to get a numpy 1.5.x branch started soon, the attached patch against numpy svn trunk r8464 does [[BR]] 1) remove the datetime functionality.
|
|
[[BR]] 2) restore ABI compatibility with numpy 1.4.
|
|
[[BR]] 3) enable numpy to build and run on Python 2.7 and 3.1.
|
|
[[BR]] The patch is based on the following changesets tickets [[BR]] http projects.scipy.org numpy changeset 8098 [[BR]] http projects.scipy.org numpy changeset 8113 [[BR]] http projects.scipy.org numpy changeset 8107 [[BR]] http projects.scipy.org numpy changeset 8108 [[BR]] http projects.scipy.org numpy changeset 8115 [[BR]] http projects.scipy.org numpy ticket 1502 I verified that the patched numpy 1) builds and tests without crash) on Python 2.5, 2.6, 2.7, and 3.1 for Windows, 32 and 64 bit.
|
|
[[BR]] 2) works with some popular Python extensions, which were built against numpy 1.4.1 and Python 2.6 for Windows matplotlib, numexpr, h5py, pygame, pytables, scipy.
|
|
[[BR]] 3) can be used to build binaries of some popular Python packages that work with numpy 1.4.1 on Python 2.6 for Windows matplotlib, scipy.
|
|
The development and test environment is Windows 7, Visual Studio, Platform SDK, Intel Fortran MKL based.","enhancement"
|
|
"6013","2118","numpy.distutils mixes CFLAGS and CXXFLAGS Trac 1521).
|
|
Original ticket http projects.scipy.org numpy ticket 1521 on 2010 06 25 by trac user killua.eu, assigned to cournape.
|
|
Please see [https bugs.gentoo.org show bug.cgi?id=236332] for details.
|
|
Thanks!","bug"
|
|
"6013","2119","segfault in any ) on large object array Trac 1522).
|
|
Original ticket http projects.scipy.org numpy ticket 1522 on 2010 06 27 by trac user glub, assigned to unknown.
|
|
I am consistently getting segfaults when I call any ) on a particular array of 42364 elements.
|
|
I will try to attach the array as a text file.
|
|
There doesn t appear to be anything interesting about the array except that it does not have any zeros and appears to have a lot of smallish, duplicate values.
|
|
I got the same segfault in 1.3.0, 1.4.1, and r8464 from svn Program received signal SIGSEGV, Segmentation fault.
|
|
0x00007ffff03eac2f in PyUFunc Reduce self=<value optimized out>, args=<value optimized out>, kwds=<value optimized out>, operation=<value optimized out>) at numpy core src umath ufunc object.c 2785 2785 Py XINCREF PyObject )loop >castbuf)); gdb) bt 0 0x00007ffff03eac2f in PyUFunc Reduce self=<value optimized out>, args=<value optimized out>, kwds=<value optimized out>, operation=<value optimized out>) at numpy core src umath ufunc object.c 2785 1 PyUFunc GenericReduction self=<value optimized out>, args=<value optimized out>, kwds=<value optimized out>, operation=<value optimized out>) at numpy core src umath ufunc object.c 3352 2 0x00000035d1043db3 in PyObject Call ) from usr lib64 libpython2.6.so.1.0 3 0x00007ffff065ecbe in PyArray GenericReduceFunction m1=<value optimized out>, op=<value optimized out>, axis=<value optimized out>, rtype=<value optimized out>, out=0x0) at numpy core src multiarray number.c 173 4 0x00007ffff0682aab in PyArray Any self=<value optimized out>, axis=0, out=0x0) at numpy core src multiarray calculation.c 697 5 0x00007ffff0682b5e in array any self=0xd6ae90, args=<value optimized out>, kwds=<value optimized out>) at numpy core src multiarray methods.c 1825 6 0x00000035d10ddae6 in PyEval EvalFrameEx ) from usr lib64 libpython2.6.so.1.0 7 0x00000035d10de312 in PyEval EvalFrameEx ) from usr lib64 libpython2.6.so.1.0 8 0x00000035d10df4e9 in PyEval EvalCodeEx ) from usr lib64 libpython2.6.so.1.0 9 0x00000035d10dd897 in PyEval EvalFrameEx ) from usr lib64 libpython2.6.so.1.0 10 0x00000035d10de312 in PyEval EvalFrameEx ) from usr lib64 libpython2.6.so.1.0 11 0x00000035d10de312 in PyEval EvalFrameEx ) from usr lib64 libpython2.6.so.1.0 12 0x00000035d10df4e9 in PyEval EvalCodeEx ) from usr lib64 libpython2.6.so.1.0 13 0x00000035d10df5b2 in PyEval EvalCode ) from usr lib64 libpython2.6.so.1.0 14 0x00000035d10fa52c in ?? )
|
|
from usr lib64 libpython2.6.so.1.0 15 0x00000035d10fa600 in PyRun FileExFlags ) from usr lib64 libpython2.6.so.1.0 16 0x00000035d10fb9dc in PyRun SimpleFileExFlags ) from usr lib64 libpython2.6.so.1.0 17 0x00000035d110807d in Py Main ) from usr lib64 libpython2.6.so.1.0 18 0x00000034b341ec5d in libc start main ) from lib64 libc.so.6 19 0x0000000000400649 in start )","bug"
|
|
"6013","2120","numpy.lib.function base.sinc doesn t properly handle non array sequence input Trac 1523).
|
|
Original ticket http projects.scipy.org numpy ticket 1523 on 2010 06 27 by trac user dgoldsmith, assigned to unknown.
|
|
>>> np.sinc np.array 0,0.5))) array [ 1. , 0.63661977]) >>> np.sinc 0,0.5)) array [ NaN, 0.63661977]) >>> np.sinc [0,0.5]) array [ NaN, 0.63661977]) Simply needs an x = np.asanyarray x) before the rest of the code.
|
|
Thanks RK)","bug"
|
|
"6013","2121","Citation example in HOWTO DOCUMENT.txt slightly inconsistent w IEEE Standard Trac 1524).
|
|
Original ticket http projects.scipy.org numpy ticket 1524 on 2010 06 27 by trac user dgoldsmith, assigned to unknown.
|
|
Here s the example in HOWTO DOCUMENT.txt O. McNoleg, The integration of GIS, remote sensing, expert systems and adaptive co kriging for environmental habitat modelling of the Highland Haggis using object oriented, fuzzy logic and neural network techniques, Computers Geosciences, vol.
|
|
22, pp.
|
|
585 588, 1996.
|
|
Here s the spec in http standards.ieee.org guides style 2009 Style Manual.pdf 19.3 Articles in periodicals Articles listed shall include the following information in the order shown a) Last name of author or authors and first name or initials, or name of organization [Our example is in the wrong order] b) Title of article in quotation marks c) Title of periodical in full and set in italics [ours isn t italicized] d) Volume, number, and, if available, part e) First and last pages of article f) Date of issue [Their] Example [B1] Boggs, S. A., and Fujimoto, N., xe2x80x9cTechniques and instrumentation for measurement of transients in gas insulated switchgear,xe2x80x9d IEEE Transactions on Electrical Installation, vol.
|
|
ET 19, no.
|
|
2, pp.
|
|
87xe2x80x9392, Apr.
|
|
1984.","bug"
|
|
"6013","2122","Numpy is ignoring CFLAGS, which makes a 64 bit build difficult Trac 1525).
|
|
Original ticket http projects.scipy.org numpy ticket 1525 on 2010 06 28 by trac user drkirkby, assigned to cournape.
|
|
I am trying to get Numpy to build as a 64 bit executable, as part of a 64 bit port of Sage to !OpenSolaris.
|
|
But despite adding the compiler option m64 which generates 64 bit objects) to CFLAGS, Numpy ignores this so GCC generates an error.
|
|
The same issue has been known in Sage for a long time affecting OS X systems where building 32 bit objects is the default, so one needed the m64 compiler flag to generate the 64 bit code.
|
|
wrong ELF class ELFCLASS64 which is typical of when a program is expecting to find a 32 bit object, but finds a 64 bit one.
|
|
numpy 1.3.0.p2 src setup.py numpy 1.3.0.p2 src MANIFEST.in numpy 1.3.0.p2 .hgtags Finished extraction Host system uname a SunOS hawk 5.11 snv 111b i86pc i386 i86pc CC Version gcc v Using built in specs.
|
|
Target i386 pc solaris2.11 Configured with .. gcc 4.3.4 configure prefix= usr local gcc 4.3.4 GNU assembler Sun linker with as= usr local binutils 2.20 bin as with ld= usr ccs bin ld with gmp= usr local with mpfr= usr local Thread model posix gcc version 4.3.4 GCC) Running from numpy source directory.
|
|
F2PY Version 2 blas opt info blas mkl info libraries mkl,vml,guide not found in export home drkirkby sage 4.3.1 local lib NOT AVAILABLE atlas blas threads info Setting PTATLAS=ATLAS libraries ptf77blas,ptcblas,atlas not found in export home drkirkby sage 4.3.1 local lib NOT AVAILABLE atlas blas info FOUND libraries = [ f77blas , cblas , atlas ] library dirs = [ export home drkirkby sage 4.3.1 local lib ] language = c include dirs = [ export home drkirkby sage 4.3.1 local include ] export home drkirkby sage 4.3.1 spkg build numpy 1.3.0.p2 src numpy distutils command config.py 361 DeprecationWarning Usage of get output is deprecated please do not use it anymore, and avoid configuration checks involving running executable on the target machine.
|
|
DeprecationWarning) customize Sage FCompiler 1 customize Sage FCompiler 1 customize Sage FCompiler 1 using config compiling configtest.c This file is generated from numpy distutils system info.py void ATL buildinfo void); int main void) { ATL buildinfo ); return 0; } C compiler gcc fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes fPIC compile options c gcc configtest.c gcc configtest.o L export home drkirkby sage 4.3.1 local lib lf77blas lcblas latlas o configtest ld fatal file export home drkirkby sage 4.3.1 local lib libf77blas.so wrong ELF class ELFCLASS64 ld fatal file export home drkirkby sage 4.3.1 local lib libcblas.so wrong ELF class ELFCLASS64 ld fatal file export home drkirkby sage 4.3.1 local lib libatlas.so wrong ELF class ELFCLASS64 ld fatal file processing errors.
|
|
No output written to configtest collect2 ld returned 1 exit status ld fatal file export home drkirkby sage 4.3.1 local lib libf77blas.so wrong ELF class ELFCLASS64 ld fatal file export home drkirkby sage 4.3.1 local lib libcblas.so wrong ELF class ELFCLASS64 ld fatal file export home drkirkby sage 4.3.1 local lib libatlas.so wrong ELF class ELFCLASS64 ld fatal file processing errors.
|
|
No output written to configtest collect2 ld returned 1 exit status failure.
|
|
removing configtest.c configtest.o Status 255 Output FOUND libraries = [ f77blas , cblas , atlas ] library dirs = [ export home drkirkby sage 4.3.1 local lib ] language = c define macros = [ NO ATLAS INFO , 2)] include dirs = [ export home drkirkby sage 4.3.1 local include ] lapack opt info lapack mkl info","bug"
|
|
"6013","2123","python numpy 1.4.1 configtest.c 6 undefined reference to `exp Trac 1526).
|
|
Original ticket http projects.scipy.org numpy ticket 1526 on 2010 06 29 by sandrotosi, assigned to unknown.
|
|
Hello, this bug has been filed in the Debian BTS as http bugs.debian.org 527707 and I m now forwarding it here.
|
|
The report above is about 1.2.1 but it still applies in 1.4.1 compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core sr c multiarray Inumpy core src umath Inumpy core include I usr include python2.6 c gcc configtest.c configtest.c 1 warning conflicting types for built in function xe2x80x98expxe2x80x99 gcc pthread configtest.o o configtest configtest.o In function `main home morph deb build area python numpy 1.4.1 configtest.c 6 undefined reference to `exp collect2 ld returned 1 exit status configtest.o In function `main home morph deb build area python numpy 1.4.1 configtest.c 6 undefined reference to `exp collect2 ld returned 1 exit status failure.
|
|
removing configtest.c configtest.o C compiler gcc pthread fno strict aliasing DNDEBUG g fwrapv O2 Wall Wstrict prototypes g O2 g Wall O2 fPIC a complete log can be found here http people.debian.org morph tmp python numpy 1.4.1 1 amd64.build.bz2 Thanks in advance for your help, Sandro","bug"
|
|
"6013","2124","numpy.savez ) doesn t compress that much Trac 1527).
|
|
Original ticket http projects.scipy.org numpy ticket 1527 on 2010 06 29 by sandrotosi, assigned to unknown.
|
|
Hello, I m here forwarding the bug filed in the Debian BTS at http bugs.debian.org 564774 .
|
|
The bug refers to 1.3.0 but I m able to replicate it with 1.4.1 too.
|
|
The key part is for the rest, please read the link above) >>> import numpy as np >>> np.savez test.npz , N.ones 10000, 1000))) >>> ll h .npz rw r r 1 michael michael 77M 2010 01 11 15 52 test.npz >>> !gzip test.npz >>> ll h test .gz rw r r 1 michael michael 114K 2010 01 11 15 52 test.npz.gz Thanks for your support, Sandro","bug"
|
|
"6013","2125","numpy.savez ) doesn t compress that much Trac 1528).
|
|
Original ticket http projects.scipy.org numpy ticket 1528 on 2010 06 29 by sandrotosi, assigned to unknown.
|
|
Hello, I m here forwarding the bug filed in the Debian BTS at http bugs.debian.org 564774 .
|
|
The bug refers to 1.3.0 but I m able to replicate it with 1.4.1 too.
|
|
The key part is for the rest, please read the link above) >>> import numpy as np >>> np.savez test.npz , N.ones 10000, 1000))) >>> ll h .npz rw r r 1 michael michael 77M 2010 01 11 15 52 test.npz >>> !gzip test.npz >>> ll h test .gz rw r r 1 michael michael 114K 2010 01 11 15 52 test.npz.gz Thanks for your support, Sandro","bug"
|
|
"6013","2126","numpy.savez ) doesn t compress that much Trac 1529).
|
|
Original ticket http projects.scipy.org numpy ticket 1529 on 2010 06 29 by sandrotosi, assigned to unknown.
|
|
Hello, I m here forwarding the bug filed in the Debian BTS at http bugs.debian.org 564774 .
|
|
The bug refers to 1.3.0 but I m able to replicate it with 1.4.1 too.
|
|
The key part is for the rest, please read the link above) >>> import numpy as np >>> np.savez test.npz , N.ones 10000, 1000))) >>> ll h .npz rw r r 1 michael michael 77M 2010 01 11 15 52 test.npz >>> !gzip test.npz >>> ll h test .gz rw r r 1 michael michael 114K 2010 01 11 15 52 test.npz.gz Thanks for your support, Sandro","bug"
|
|
"6013","2127","please convert warnings to DeprecationWarning Trac 1530).
|
|
Original ticket http projects.scipy.org numpy ticket 1530 on 2010 06 29 by sandrotosi, assigned to unknown.
|
|
Hello, I m here forwarding the bug filed in the Debian bts at http bugs.debian.org 519483 .
|
|
The ticket refers to 1.2.1 but the same still applies in 1.4.1 too, for example python numpy 1.4.1 grep A4 warnings.warn numpy lib function base.py warnings.warn The histogram semantics being used is now deprecated and will disappear in NumPy 2.0.
|
|
Please update your code to use the default semantics.
|
|
, DeprecationWarning) warnings.warn The new semantics of histogram is now the default and the `new` keyword will be removed in NumPy 2.0. , Warning) a = asarray a) and there are several other places where warnings.warn is used python numpy 1.4.1 grep warnings.warn numpy rc | grep v 0 numpy core memmap.py 2 numpy core numeric.py 1 numpy core setup common.py 1 numpy ctypeslib.py 1 numpy distutils command config.py 2 numpy distutils core.py 5 numpy distutils cpuinfo.py 2 numpy distutils exec command.py 1 numpy distutils fcompiler gnu.py 1 numpy distutils misc util.py 1 numpy distutils system info.py 12 numpy lib arraysetops.py 1 numpy lib function base.py 2 numpy lib io.py 3 numpy lib polynomial.py 1 numpy lib utils.py 2 numpy ma core.py 6 numpy ma extras.py 1 numpy ma mrecords.py 1 numpy numarray alter code2.py 1 numpy numarray functions.py 4 numpy oldnumeric alter code2.py 1 numpy oldnumeric ma.py 1 numpy polynomial chebyshev.py 1 numpy polynomial polynomial.py 1 numpy testing tests test utils.py 2 If the warnings.warn is used to signal a deprecation coming in the near future, then using DeprecationWarning is more precise it specifies exactly that) and also would allow users to silence them.
|
|
Thanks for your support, Sandro","bug"
|
|
"6013","2128","Documentation fix for routines.fft Trac 1531).
|
|
Original ticket http projects.scipy.org numpy ticket 1531 on 2010 06 30 by trac user valhallasw, assigned to unknown.
|
|
http docs.scipy.org doc numpy reference routines.fft.html reads If A = fft a, n), then A[0] contains the zero frequency term the mean of the signal), which is always purely real for real inputs.
|
|
This is, however, incorrect.
|
|
A[0] contains the zero frequency term multiplied by the number of data points .
|
|
The same error is in more docs, for example http docs.scipy.org doc numpy reference generated numpy.fft.rfft.html numpy.fft.rfft.","bug"
|
|
"6013","2129","f2py does not respect assumed size arrays of fortran 90 Trac 1532).
|
|
Original ticket http projects.scipy.org numpy ticket 1532 on 2010 07 01 by trac user ozn, assigned to pearu.
|
|
consider the following fortran 90 file cat dsum.f90 subroutine dsum b, a, n) integer, intent in) n real,intent in) a n) real,intent out) b b=0.0d0 do i=1,n b=b a i) end do end subroutine locpt x0,y0,x,y,xd,yd,n,l,m) IMPLICIT NONE INTEGER, INTENT IN) xd,yd,n REAL, INTENT IN) x0, y0, x ), y ) INTEGER, INTENT OUT) l, m print , x0 print , y0 print , x print , y print , n , n call dsum x0,x,n) print , sum , x0 END SUBROUTINE locpt I compiled the filed with f2py c m dsum dsum.f90 and it crashes with oz safira2 test python Python 2.5.2 r252 60911, Jan 24 2010, 14 53 14) [GCC 4.3.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import dsum >>> dsum.locpt 2.0,2.0, 2.0,1.0,3.0), 5.0,6.0,8.0),7,7,3) 2.0000000 2.0000000 Segmentation fault if I replace the line declaring the array with REAL, INTENT IN) x0, y0, x xd), y yd) no problems occur.
|
|
It would be nice to be able to pass assumed size arrays too.
|
|
on the other hand, f2py is smart enough to conclude the follwoing print dsum2.
|
|
doc This module dsum2 is auto generated with f2py version 2 5237).
|
|
Functions b = dsum a,n=len a)) l,m = locpt x0,y0,x,y,n,xd=len x),yd=len y)) so, it figures automatically, that xd and yd, should be optional.
|
|
I don t know whether I d classify it as a bug feature wish.
|
|
But I hope it would at least be considered.
|
|
Cheers, Oz N P.S) I m using numpy 1.3 on debian lenny.","enhancement"
|
|
"6013","2130","no error for poor spelling in pyf file Trac 1533).
|
|
Original ticket http projects.scipy.org numpy ticket 1533 on 2010 07 02 by samtygier, assigned to pearu.
|
|
it took me a while to figure out why the doubles i was passing to c where arriving as 0. it turned out that i has miss spelt double precision in the pyf file.
|
|
there was no warning given for this.
|
|
here is an example include <stdio.h> void di int a, int b, double c, double d){ printf c a= d\n , a); printf c b= d\n , b); printf c c= f\n , c); printf c d= f\n , d); } pyf python module di interface !void foo int a, int b, double c, double d){ subroutine di a,b,c,d) intent c) di intent c) integer intent in) a integer intent in) b double precission intent in) c double precision intent in) d end subroutine di end interface end python module di and the debug output when i run it >>> import di; di.di 1,2,3,4) debug capi Python C API function di.di a,b,c,d) debug capi int a= input,required,scalar debug capi a=1 debug capi float c= input,required,scalar debug capi c=3 debug capi int b= input,required,scalar debug capi b=2 debug capi double d= input,required,scalar debug capi d=4 debug capi Fortran subroutine `di a,b,c,d) c a=1 c b=2 c c=0.000000 c d=4.000000 debug capi a=1 debug capi b=2 debug capi c=3 debug capi d=4 debug capi Building return value.
|
|
debug capi Python C API function di.di successful.
|
|
debug capi Freeing memory.
|
|
This ought to give an error when calling f2py.
|
|
i found this with numpy 1.3, but a quick test with 1.4.1 seems to still have it.","enhancement"
|
|
"6013","2131","FAIL test print.test complex types <type numpy.complex64 >,) Trac 1534).
|
|
Original ticket http projects.scipy.org numpy ticket 1534 on 2010 07 03 by vincentdavis, assigned to unknown.
|
|
Results as run on Mac os x 10.6.4, python 2.7, numpy 2.0.0Dev, built 64bit ====================================================================== FAIL test print.test complex types <type numpy.complex64 >,) Check formatting of complex types.
|
|
Traceback most recent call last) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 186, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core tests test print.py , line 61, in check complex type err msg= Failed str formatting for type s tp) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed str formatting for type <type numpy.complex64 > ACTUAL 1j DESIRED 0 1j) >> raise AssertionError \nItems are not equal Failed str formatting for type <type numpy.complex64 >\n ACTUAL 1j \n DESIRED 0 1j) ) ====================================================================== FAIL test print.test complex types <type numpy.complex128 >,) Check formatting of complex types.
|
|
Traceback most recent call last) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 186, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core tests test print.py , line 61, in check complex type err msg= Failed str formatting for type s tp) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed str formatting for type <type numpy.complex128 > ACTUAL 1j DESIRED 0 1j) >> raise AssertionError \nItems are not equal Failed str formatting for type <type numpy.complex128 >\n ACTUAL 1j \n DESIRED 0 1j) ) ====================================================================== FAIL test print.test complex types <type numpy.complex256 >,) Check formatting of complex types.
|
|
Traceback most recent call last) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 186, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core tests test print.py , line 61, in check complex type err msg= Failed str formatting for type s tp) File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal Failed str formatting for type <type numpy.complex256 > ACTUAL 1j DESIRED 0 1j) >> raise AssertionError \nItems are not equal Failed str formatting for type <type numpy.complex256 >\n ACTUAL 1j \n DESIRED 0 1j) ) Ran 2466 tests in 11.029s FAILED KNOWNFAIL=4, SKIP=1, failures=3) <nose.result.TextTestResult run=2466 errors=0 failures=3>","bug"
|
|
"6013","2132","numpy.random.permutations fails for non integer arguments Trac 1535).
|
|
Original ticket http projects.scipy.org numpy ticket 1535 on 2010 07 05 by robince, assigned to unknown.
|
|
When calling numpy.random.permutation with non integer arguments it fails with a cryptic error message In [49] np.random.permutation 12) Out[49] array [10, 9, 4, 7, 3, 8, 0, 6, 5, 1, 11, 2]) In [50] np.random.permutation 12.0) TypeError Traceback most recent call last) D \robin\pyentropy 0.4.0\<ipython console> in <module> ) C \Python26\lib\site packages\numpy\random\mtrand.pyd in mtrand.RandomState.permutation build\scons\numpy\random\mtrand \mtrand.c 18669) ) C \Python26\lib\site packages\numpy\random\mtrand.pyd in mtrand.RandomState.shuffle build\scons\numpy\random\mtrand\mtr and.c 18261) ) TypeError len ) of unsized object > d \robin\pyentropy 0.4.0\mtrand.pyx 4174)mtrand.RandomState.shuffle build\scons\numpy\random\mtrand\mtrand.c 18261) ) This also happens with long arguments, and so np.random.permutation X.shape[0]) where X is an array fails on 64 bit windows where shape is a tuple of longs).
|
|
It would be great if it could cast to integer or at least raise a proper error for non integer types.","bug"
|
|
"6013","2133","empty function problem Trac 1536).
|
|
Original ticket http projects.scipy.org numpy ticket 1536 on 2010 07 06 by trac user sdnxpt, assigned to unknown.
|
|
Hi, I use Python 2.6.5 installed on Ubuntu 10.04, and the following part of code v = np.ones self.n) generates this error glibc detected usr bin python2.6 free ) invalid next size fast) 0x000000000355e1b0 ======= Backtrace ========= lib libc.so.6 0x775b6)[0x7fec5335a5b6] lib libc.so.6 cfree 0x73)[0x7fec53360e53] usr local lib python2.6 dist packages numpy 1.4.1 py2.6 linux x86 64.egg numpy core multiarray.so 0x4183c)[0x7fec5005a83c] usr bin python2.6[0x46a50d] usr local lib python2.6 dist packages numpy 1.4.1 py2.6 linux x86 64.egg numpy core multiarray.so 0x4fbd8)[0x7fec50068bd8] usr local lib python2.6 dist packages numpy 1.4.1 py2.6 linux x86 64.egg numpy core multiarray.so 0x5aded)[0x7fec50073ded] usr bin python2.6 PyEval EvalFrameEx 0x516e)[0x4a7c5e] usr bin python2.6 PyEval EvalFrameEx 0x5a60)[0x4a8550] usr bin python2.6 PyEval EvalCodeEx 0x911)[0x4a9671] usr bin python2.6[0x537620] usr bin python2.6 PyObject Call 0x47)[0x41f0c7] usr bin python2.6[0x427dff] usr bin python2.6 PyObject Call 0x47)[0x41f0c7] usr bin python2.6[0x477bff] usr bin python2.6[0x46f47f] usr bin python2.6 PyObject Call 0x47)[0x41f0c7] usr bin python2.6 PyEval EvalFrameEx 0x4888)[0x4a7378] usr bin python2.6 PyEval EvalCodeEx 0x911)[0x4a9671] usr bin python2.6 PyEval EvalFrameEx 0x4d19)[0x4a7809] usr bin python2.6 PyEval EvalCodeEx 0x911)[0x4a9671] usr bin python2.6 PyEval EvalFrameEx 0x4d19)[0x4a7809] usr bin python2.6 PyEval EvalCodeEx 0x911)[0x4a9671] usr bin python2.6[0x537620] usr bin python2.6 PyObject Call 0x47)[0x41f0c7] usr bin python2.6 PyEval CallObjectWithKeywords 0x43)[0x4a1b03] usr local lib python2.6 dist packages numpy 1.4.1 py2.6 linux x86 64.egg numpy core multiarray.so 0x2ee94)[0x7fec50047e94] usr bin python2.6 PyObject Str 0x61)[0x454a81] usr bin python2.6 PyString Format 0x85c)[0x4670ac] usr bin python2.6[0x41f75f] usr bin python2.6 PyNumber Remainder 0x1e)[0x42251e] usr bin python2.6 PyEval EvalFrameEx 0x2624)[0x4a5114] usr bin python2.6 PyEval EvalFrameEx 0x5a60)[0x4a8550] On Windows i don t have this problem.","bug"
|
|
"6013","2134","import numpy fails using python 2.6 on Windows Trac 1537).
|
|
Original ticket http projects.scipy.org numpy ticket 1537 on 2010 07 07 by trac user hamstervision, assigned to rgommers.
|
|
I installed NumPy 1.4.1 using the following installer numpy 1.4.1 win32 superpack python2.6.exe I added the following line to a functional script import numpy and I get the following Traceback import numpy File C \Python26\lib\site packages\numpy\ init .py , line 132, in <module> import add newdocs File C \Python26\lib\site packages\numpy\add newdocs.py , line 9, in <module> from lib import add newdoc File C \Python26\lib\site packages\numpy\lib\ init .py , line 4, in <module> from type check import File C \Python26\lib\site packages\numpy\lib\type check.py , line 8, in <module> import numpy.core.numeric as nx File C \Python26\lib\site packages\numpy\core\ init .py , line 5, in <module> import multiarray ImportError DLL load failed 1 is not a valid Win32 application.
|
|
I m running Windows 7 Professional 64 bit.","bug"
|
|
"6013","2135","finfo.eps returns scalar, finfo.epsneg returns array Trac 1538).
|
|
Original ticket http projects.scipy.org numpy ticket 1538 on 2010 07 07 by trac user dgoldsmith, assigned to unknown.
|
|
Python 2.6.5 r265 79096, Mar 19 2010, 18 02 59) [MSC v.1500 64 bit AMD64)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.version.version 1.4.1 >>> np.finfo float ).eps 2.2204460492503131e 16 >>> np.finfo float ).epsneg array 1.1102230246251565e 16)","bug"
|
|
"6013","2136","MSVC specific TypeError when using double, longdouble in numpy.dot Trac 1539).
|
|
Original ticket http projects.scipy.org numpy ticket 1539 on 2010 07 08 by cgohlke, assigned to unknown.
|
|
This issue has been discussed at [http mail.scipy.org pipermail numpy discussion 2010 July 051352.html] and is the cause of the scipy.sparse TypeError test failures reported at [http mail.scipy.org pipermail scipy user 2010 July 025961.html].
|
|
Using Python 2.6.5 32 bit on Windows and numpy 1.4.1 built with Visual Studio 2008 msvc9), I get the following >>> a = numpy.array [[1.0]]) >>> numpy.dot a.astype double ), a.astype longdouble )) array [[ 1.]])
|
|
>>> a = numpy.array [[1]]) >>> numpy.dot a.astype double ), a.astype longdouble )) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError array cannot be safely cast to required type >>> a = numpy.array [[1]]) >>> numpy.dot a.astype single ), a.astype longdouble )) array [[1.0]], dtype=float64) >>> a = numpy.array [[1]]) >>> numpy.dot a.astype longdouble ), a.astype double )) array [[1.0]], dtype=float64) >>> numpy.array [1]).astype longdouble ).dtype.num 13 >>> numpy.array [1.0]).astype longdouble ).dtype.num 12 The TypeError is unexpected.
|
|
The same results are obtained when using a msvc7 build of numpy 1.4.1 on Python 2.5, or a msvc9 64 bit build of numpy 2.0.dev on Python 3.1.
|
|
Using numpy 1.4.1 built with mingw does not raise a TypeError >>> a = numpy.array [[1]]) >>> numpy.dot a.astype double ), a.astype longdouble )) array [[1.0]], dtype=float96) Note that for MSVC compilers sizeof longdouble) = sizeof double) = 8","bug"
|
|
"6013","2137","numpy.savez has race condition Trac 1540).
|
|
Original ticket http projects.scipy.org numpy ticket 1540 on 2010 07 09 by trac user Koen, assigned to unknown.
|
|
When saving a variable named test , numpy.savez will always use a temporary file called tmp test.npy.
|
|
When running multiple programs saving the same test variable at the same time, this leads to race conditions.
|
|
I have modified the code below) to use a proper temporary filename, and reuse that same temporary file for all variables within the same program.
|
|
def savez file, args, kwds) Save several arrays into a single, compressed file with extension .npz If keyword arguments are given, the names for variables assigned to the keywords are the keyword names not the variable names in the caller).
|
|
If arguments are passed in with no keywords, the corresponding variable names are arr 0, arr 1, etc.
|
|
Parameters file Either the filename string) or an open file file like object) If file is a string, it names the output file.
|
|
.npz will be appended if it is not already there.
|
|
args Arguments Any function arguments other than the file name are variables to save.
|
|
Since it is not possible for Python to know their names outside the savez function, they will be saved with names arr 0 , arr 1 , and so on.
|
|
These arguments can be any expression.
|
|
kwds Keyword arguments All keyword=value pairs cause the value to be saved with the name of the keyword.
|
|
See Also save Save a single array to a binary file in NumPy format savetxt Save an array to a file as plain text Notes The .npz file format is a zipped archive of files named after the variables they contain.
|
|
Each file contains one variable in .npy format.
|
|
Import is postponed to here since zipfile depends on gzip, an optional component of the so called standard library.
|
|
import zipfile if isinstance file, basestring) if not file.endswith .npz ) file = file .npz namedict = kwds for i, val in enumerate args) key = arr d i if key in namedict.keys ) raise ValueError, Cannot use un named variables and keyword s key namedict[key] = val zip = zipfile.ZipFile file, mode= w ) Place to write temporary .npy files before storing them in the zip import tempfile fid, tmpname) = tempfile.mkstemp .npy , numpy ) os.close fid) for key, val in namedict.iteritems ) fname = key .npy fid = open tmpname, wb ) format.write array fid, np.asanyarray val)) fid.close ) zip.write tmpname, arcname=fname) zip.close ) os.remove tmpname)","bug"
|
|
"6013","2138","random.uniform gives inf when using finfo float ).min, finfo float ).max as intervall Trac 1541).
|
|
Original ticket http projects.scipy.org numpy ticket 1541 on 2010 07 10 by trac user bowie 22, assigned to unknown.
|
|
Using the following statement gives also inf as output.
|
|
print np.random.uniform np.finfo float ).min, np.finfo float ).max, size= 2,2))) [[ Inf Inf] [ Inf Inf]] some experiments have shown the follwing behavior print np.random.uniform 1.7976931348623157e 292, np.finfo float ).max, size= 2,2))) [[ Inf Inf] [ Inf Inf]] print np.random.uniform 1.7976931348623157e 291, np.finfo float ).max, size= 2,2))) [[ 3.02733730e 307 5.55521694e 307] [ 1.64373640e 308 1.54380200e 308]] Where print np.finfo float ).max) 1.79769313486e 308 and print np.finfo float ).min) 1.79769313486e 308","bug"
|
|
"6013","2139","Memory leak casting view of object type ndarray to bool Trac 1542).
|
|
Original ticket http projects.scipy.org numpy ticket 1542 on 2010 07 12 by wesm, assigned to unknown.
|
|
Seen in NumPy 1.3.0 and NumPy 1.4.1 on Windows XP.
|
|
from datetime import datetime import numpy as np import sys def foo verbose=True) arr = np.array [datetime.today ) for in xrange 1000)]) arr = arr.reshape 500, 2)) sl = arr[ , 0] if verbose print Rec ct of index 0 d sys.getrefcount sl[0]) for in xrange 10) foo = sl.astype bool) if verbose print Rec ct of index 0 d sys.getrefcount sl[0]) if name == main foo ) leaks memory for i in xrange 10000) if not i 1000 print i foo verbose=False)","bug"
|
|
"6013","2140","assert almost equal has bad behavior decimals=) Trac 1543).
|
|
Original ticket http projects.scipy.org numpy ticket 1543 on 2010 07 14 by pv, assigned to unknown.
|
|
The `decimals=` absolute tolerance only approach used in `assert almost equal` seems wrong in comparing floating point numbers.
|
|
Not to mention that specifying tolerance by the number of decimals is a bit clunky.)
|
|
People may look at the name of the function, and be surprised by things like this >>> x = 1e99 >>> y = np.nextafter x, 1) x and y are neighbouring fp numbers!
|
|
>>> np.testing.assert almost equal x, y) Traceback most recent call last) ...
|
|
Items are not equal ... >>> x = 1e 19 >>> y = 1e 99 >>> np.testing.assert almost equal x, y) passes The best choice, IMHO, would be to deprecate the `assert almost equal` function we already have an ULP comparison function copy the `assert tol equal` from `scipy.special` to numpy","bug"
|
|
"6013","2141","cannot set dtype on record array with 04 datetime records Trac 1544).
|
|
Original ticket http projects.scipy.org numpy ticket 1544 on 2010 07 14 by trac user jdh2358, assigned to unknown.
|
|
I use record arrays extensively with python datetimes, which works if you pass in a list of lists of data with the names.
|
|
numpy can accurately infer the dtypes and create a usable record array.
|
|
Eg, import datetime import numpy as np rows = [ [datetime.date 2001,1,1), 12, 23.
|
|
], [datetime.date 2002,1,1), 10, 13.
|
|
], [datetime.date 2003,1,1), 2, 1.
|
|
], ] r1 = np.rec.fromrecords rows, names= a,b,c ) print r1.dtype prints out [ a , |O4 ), b , <i4 ), c , <f8 )] but if I want to speed things up by providing the dtype, numpy raises ValueError dtype = [ a , |O4 ), b , <i4 ), c , <f8 )] r2 = np.rec.fromrecords rows, dtype=dtype) home titan johnh test.py 12 dtype = [ a , |O4 ), b , <i4 ), c , <f8 )] 13 > 14 r2 = np.rec.fromrecords rows, dtype=dtype) 15 16 home titan johnh dev lib python2.4 site packages numpy core records.pyc in fromrecords recList, dtype, shape, formats, names, titles, aligned, byteorder) 610 611 try > 612 retval = sb.array recList, dtype=descr) 613 except TypeError list of lists instead of list of tuples 614 if shape is None or shape == 0) ValueError Setting void array with object members using buffer.
|
|
WARNING Failure executing file <test.py> Running from svn HEAD In [2] numpy.
|
|
version Out[2] 2.0.0.dev8480 Here is a complete script import datetime import numpy as np rows = [ [datetime.date 2001,1,1), 12, 23.
|
|
], [datetime.date 2002,1,1), 10, 13.
|
|
], [datetime.date 2003,1,1), 2, 1.
|
|
], ] r1 = np.rec.fromrecords rows, names= a,b,c ) print r1.dtype dtype = [ a , |O4 ), b , <i4 ), c , <f8 )] r2 = np.rec.fromrecords rows, dtype=dtype)","bug"
|
|
"6013","2142","Add a norm keyword arg to fft functions Trac 1545).
|
|
Original ticket http projects.scipy.org numpy ticket 1545 on 2010 07 15 by trac user dgoldsmith, assigned to unknown.
|
|
Presently, the fft functions return the unscaled transform and the ifft functions return the 1 n) scaled inverse transform, meaning that neither result is unitary.
|
|
In contrast, the discrete cosine transform DCT) functions in scipy.fftpack include a keyword argument norm) which, when set to ortho, normalize both the forward and inverse transforms both functions have the keyword argument) by 1 sqrt n), thus returning unitary results.
|
|
There appears to be no reason why the fft functions couldn t also support this.","enhancement"
|
|
"6013","2143","function prod fails over arrays without warm Trac 1546).
|
|
Original ticket http projects.scipy.org numpy ticket 1546 on 2010 07 15 by trac user mmarquez, assigned to unknown.
|
|
I m a begginer using python.
|
|
As a very simply exercise I was defining the fact function, but I had a very bad surprise.
|
|
The output of import numpy def fact n) return numpy.prod range 2,n 1)) for i in range 2,50,5) print Fact of i d is d i,fact i)) is Fact of i 2 is 2 Fact of i 7 is 5040 Fact of i 12 is 479001600 Fact of i 17 is 288522240 Fact of i 22 is 522715136 Fact of i 27 is 1484783616 Fact of i 32 is 2147483648 Fact of i 37 is 0 Fact of i 42 is 0 Fact of i 47 is 0 which is obviously wrong .
|
|
But even more, this happen without any warning or error messages .
|
|
This is wrong even if the computations are made using modular arithmetic as claimed by the help.","bug"
|
|
"6013","2144","np.isinf broken on solaris Trac 1547).
|
|
Original ticket http projects.scipy.org numpy ticket 1547 on 2010 07 15 by trac user jdh2358, assigned to unknown.
|
|
I am seeing a problem on Solaris since I upgraded to svn HEAD.
|
|
np.isinf does not handle np.inf.
|
|
See ipython session below.
|
|
I am not seeing this problem w HEAD on an ubuntu linux box I tested on In [1] import numpy as np In [2] np.
|
|
version Out[2] 2.0.0.dev8480 In [3] x = np.inf np.inf np.info np.infty In [3] x = np.inf In [4] np.isinf x) Warning invalid value encountered in isinf Out[4] True In [5] np.seter np.seterr np.seterrcall np.seterrobj In [5] np.seterr all= raise ) Out[5] { over print , divide print , invalid print , under ignore } In [6] np.isinf x) FloatingPointError Traceback most recent call last) home titan johnh <ipython console> FloatingPointError invalid value encountered in isinf In [7] !uname a SunOS udesktop191 5.10 Generic 139556 08 i86pc i386 i86pc In [43] !gcc version gcc GCC) 3.4.3 csl sol210 3 4 branch sol rpath)","bug"
|
|
"6013","2145","matrix unbounded memory leak using = Trac 1548).
|
|
Original ticket http projects.scipy.org numpy ticket 1548 on 2010 07 16 by trac user cardinal, assigned to unknown.
|
|
The following code and many variants) leaks memory unboundedly from numpy import niter = 2000000 X = matrix zeros 4), float).reshape 2,2) for iter in range niter) X = X Replacing X = X with X = X X solves the problem.
|
|
Things like X = a where a is a scalar cause the same problem and have the same fix.
|
|
This occurs in Python 2.5.1 r251 54863, Jan 26 2008, 16 54 44) with numpy 1.0.4 on a Linux box and Python 2.6.1 r261 67515, Dec 17 2009, 00 59 15) on Mac OS X laptop with numpy 1.2.1.","bug"
|
|
"6013","2146","base repr 243, 3) returns 30000 Trac 1549).
|
|
Original ticket http projects.scipy.org numpy ticket 1549 on 2010 07 17 by trac user kmaglione, assigned to unknown.
|
|
Calling `base repr 343, 3)` returns 30000 rather than the expected 100000 .
|
|
NumPy 1.4.1 python V Python 2.6.5 uname a Linux jg 2.6.34 ARCH 1 SMP PREEMPT Sat Jun 19 00 07 49 CEST 2010 x86 64 AMD Athlon tm) 64 X2 Dual Core Processor TK 55 AuthenticAMD GNU Linux","bug"
|
|
"6013","2147","Memmap test failure Trac 1550).
|
|
Original ticket http projects.scipy.org numpy ticket 1550 on 2010 07 20 by cgohlke, assigned to unknown.
|
|
I get the following test failure on all msvc9 builds of numpy 1.5.0.dev for Windows.
|
|
A patch is attached.
|
|
====================================================================== ERROR test filename test memmap.TestMemmap) Traceback most recent call last) File X \Python26 x64\lib\site packages\numpy\core\tests\test memmap.py , line 60, in test filename os.unlink tmpname) WindowsError [Error 32] The process cannot access the file because it is being used by another process","bug"
|
|
"6013","2148","Deprecate old Numeric type codes Trac 1551).
|
|
Original ticket http projects.scipy.org numpy ticket 1551 on 2010 07 20 by pv, assigned to unknown.
|
|
The old Numeric style type codes Float8 , Complex32 , ...) should probably be deprecated; at least for complex numbers they are confusing >>> np.array [0], dtype= Complex32 ) array [ 0.
|
|
0.j], dtype=complex64)","bug"
|
|
"6013","2149","fixed length strings automatically truncate right side \x00 bytes Trac 1552).
|
|
Original ticket http projects.scipy.org numpy ticket 1552 on 2010 07 23 by trac user chrismurf, assigned to unknown.
|
|
Trivial Example !rst .. code block python map ord, N.array [ \x00 ])[0]) returns [] map ord, N.array [ \x01 ])[0]) returns [1] This seems highly inconsistent.
|
|
All other strings in python store size seperately, rather than truncating following null bytes.
|
|
I can see where for text strings this may be desirably behavior, but for binary strings my use case) it is far from it.
|
|
As is, I need to wrap every returned value with a call to ljust x, \x00 ) or equivalent.
|
|
It would be nice to have an actual fixed length string dtype, or have a way to disable this behavior.","bug"
|
|
"6013","2150","clip out=...) is incorrect for non contiguous input Trac 1553).
|
|
Original ticket http projects.scipy.org numpy ticket 1553 on 2010 07 24 by trac user yury, assigned to unknown.
|
|
clip always copies data to out as if the input were contiguous import numpy as np x = np.random.random 2,3)) y = np.zeros 3,2)) assert x.T.clip np.infty, np.infty, y) == x.T).all ) This bug may be related http projects.scipy.org numpy ticket 1352","bug"
|
|
"6013","2151","polynomial.polyutils.mapdomain chokes on multi d first argument Trac 1554).
|
|
Original ticket http projects.scipy.org numpy ticket 1554 on 2010 07 25 by trac user dgoldsmith, assigned to charris.
|
|
polynomial.polyutils.mapdomain x, old, new) raises ValueError Coefficient array is not 1 d if x.ndim > 1.","bug"
|
|
"6013","2152","fromstring on Unicode objects behavior change on Py3 Trac 1555).
|
|
Original ticket http projects.scipy.org numpy ticket 1555 on 2010 07 25 by pv, assigned to unknown.
|
|
Python 2 >>> np.fromstring \xe4 .decode latin1 ), dtype=np.uint8) Traceback most recent call last) File <stdin> , line 1, in <module> UnicodeEncodeError ascii codec can t encode character u \xe4 in position 0 ordinal not in range 128) Python 3 >>> np.fromstring b \xe4 .decode latin1 ), dtype=np.uint8) array [195, 164], dtype=uint8) The origin is a behavior change in `PyArg ParseTuple ` `s ` format Python 2 Unicode objects pass back a pointer to the default encoded string version of the object if such a conversion is possible.
|
|
http docs.python.org c api arg.html Python 3 Unicode objects are converted to C strings using utf 8 encoding.
|
|
http docs.python.org py3k c api arg.html","bug"
|
|
"6013","2153","numpy.linalg.det does not handle scalars or 1d array scalars Trac 1556).
|
|
Original ticket http projects.scipy.org numpy ticket 1556 on 2010 07 26 by jseabold, assigned to pv.
|
|
Current behavior In [20] np.linalg.det [[1]]) Out[20] 1.0 but In [21] np.linalg.det [1]) <snip> LinAlgError 1 dimensional array given.
|
|
Array must be two dimensional After applying the attached diff In [3] np.linalg.det [[1]]) Out[3] 1.0 In [4] np.linalg.det [1]) Out[4] 1.0 In [5] np.linalg.det 1) Out[5] 1.0 It seems to me that checking if an array is 2d then checking if it s square is redundant.","enhancement"
|
|
"6013","2154","Bug with dtype.type for structured arrays Trac 1557).
|
|
Original ticket http projects.scipy.org numpy ticket 1557 on 2010 07 27 by astrofrog, assigned to unknown.
|
|
If I create a structured array with vector columns >>> array = np.array zip [[1,2],[1,2],[1,3]]),dtype=[ a ,float,2)]) then examine the type of the column, I get >>> array.dtype[0] dtype float64 , 2,))) Then, if I try and view the numerical type, I see >>> array.dtype[0].type <type numpy.void > I have to basically do >>> array.dtype[0].subdtype[0] dtype float64 ) to get what I need.
|
|
I seem to remember that this used not to be the case, and that even for vector columns, one could access array.dtype[0].type to get the numerical type.","bug"
|
|
"6013","2155","random.uniform ) default value for low Trac 1558).
|
|
Original ticket http projects.scipy.org numpy ticket 1558 on 2010 07 27 by trac user heitzig j, assigned to unknown.
|
|
In Numpy 1.3.0, the default value for the parameter low in random.uniform is documented as 0 but seems to be 1 actually.","bug"
|
|
"6013","2156","atleast ?d ) breaks masked arrays Trac 1559).
|
|
Original ticket http projects.scipy.org numpy ticket 1559 on 2010 07 27 by WeatherGod, assigned to pierregm.
|
|
calling a function like .atleast 2d ) to change the number of dimensions an array has can break the original masked array object.
|
|
See the following example using a 1d masked array >>> import numpy >>> a = numpy.ma.masked array [0.0, 1.2, 3.5], mask=[False, True, False]) >>> b = numpy.atleast 2d a) >>> b masked array data = [[0.0 3.5]], mask = [[False True False]], fill value = 1e 20) >>> a Traceback most recent call last) File <stdin> , line 1, in <module> File home bvr Programs numpy numpy ma core.py , line 3570, in repr data=str self), mask=str self.
|
|
mask), File home bvr Programs numpy numpy ma core.py , line 3554, in str res[m] = f ValueError boolean index array should have 1 dimension >>> The problem does not occur if there is no change to the number of dimensions.","bug"
|
|
"6013","2157","atleast 3d ) strips subclass info Trac 1560).
|
|
Original ticket http projects.scipy.org numpy ticket 1560 on 2010 07 27 by WeatherGod, assigned to unknown.
|
|
The .atleast 3d ) function can not be used with arrays that are subclassed from ndarray ).
|
|
Because the function uses .asarray ) instead of .asanyarray ), subclass info such as masks for masked arrays are stripped from the results.
|
|
See the following example >>> a = numpy.ma.masked array [0.0, 1.2, 3.5], mask=[False, True, False]) >>> b = numpy.atleast 3d a) >>> b array [[[ 0.
|
|
], [ 1.2], [ 3.5]]]) >>> a masked array data = [0.0 3.5], mask = [False True False], fill value = 1e 20) >>> Note that the .atleast 1d ) and .atleast 2d ) functions are safe for subclassed arrays.
|
|
I believe that simply replacing .asarray ) with .asanyarray ) should fix the issue.","bug"
|
|
"6013","2158","Wrong type for PyArray SimpleNew and alike Trac 1561).
|
|
Original ticket http projects.scipy.org numpy ticket 1561 on 2010 07 27 by trac user PhiSch, assigned to unknown.
|
|
All the function that create new PyArray objects in the C API which have an dims argument use npy intp as type.
|
|
However the accepted structure is a structure in the form npy int or npy intp, e.g.
|
|
a pointer to npy int values which describe the dimensions instead of a pointer to pointers to npy int which describe the dimensions.
|
|
Best regards Philipp","bug"
|
|
"6013","2159","potentially unexpected results with loadtxt ) Trac 1562).
|
|
Original ticket http projects.scipy.org numpy ticket 1562 on 2010 07 27 by WeatherGod, assigned to unknown.
|
|
Consider the case where a user programs a script to load multi row, multi column data from a text file.
|
|
The user will then subsequently program the rest of code assuming that he has a 2 d array or a 1 d record array).
|
|
Then, when the user loads a 1 line text file, loadtxt ) will return a 1 D array or a scalar for a record array) and would cause unexpected results with the rest of the code.
|
|
This is due to loadtxt ) calling squeeze ) on the data before returning it.
|
|
While this action is nice as it magically anticipates the users intentions, it has some edge cases.
|
|
Note, this would happen with a person working with multi line, single column data, or even single line, multi column data and encounters a file with a single value in it.
|
|
My suggestion would be to have a keyword argument ndmin that would default to 0 for backwards compatibility) that would allow the users to specify what they always expect from loadtxt ).
|
|
By having this keyword and documenting it well, it should raise awareness among programmers of these sort of edge cases.
|
|
How exactly to use the ndmin keyword is an interesting question.
|
|
Could, perhaps, it get passed to numpy.squeeze, indicating to leave so many dimensions alone?","enhancement"
|
|
"6013","2160","setitem in subclass of ndarray not working for 1D arrays with tuples Trac 1563).
|
|
Original ticket http projects.scipy.org numpy ticket 1563 on 2010 07 28 by trac user ds, assigned to unknown.
|
|
For a subclass of ndarray the setitem method is not working with tuples, but only if the array is 1D.
|
|
from numpy import ndarray class C ndarray) def new cls,i) return N.ones i,)).view cls) a = C 5) a[ 0,)] = 2. raises RuntimeError Getitem not returning array.
|
|
this would work if a is a normal ndarray note a[0] = 2. is working b = a[ 0,)] is working and if the subclass is an ndarray of higher dimension then also e.g.
|
|
for 2D array subclass) a[ 0,1)] = 2. is working","bug"
|
|
"6013","2161","numpy.arange returns array with wrong length Trac 1564).
|
|
Original ticket http projects.scipy.org numpy ticket 1564 on 2010 07 28 by trac user crassus, assigned to unknown.
|
|
I m running the scipy package within a VM for sage http sagemath.org numpy.version.version 1.3.0 CPU Intel i7 Under certain circumstances numpy.arange returns an array that contains the stop parameter.
|
|
But according to the docs stop number End of interval.
|
|
The interval does not include this value Here is the example Works fine ydat= numpy.zeros 970) step= float 1.0 400000) xdat= numpy.arange 0, len ydat) step, step) print len ydat) print len xdat) print xdat[ 1]==len ydat) step Bug ydat= numpy.zeros 971) step= float 1.0 400000) xdat= numpy.arange 0, len ydat) step, step) print len ydat) print len xdat) print xdat[ 1]==len ydat) step","bug"
|
|
"6013","2162","loadtxt fails to load large unsigned int64 integers.
|
|
Trac 1565).
|
|
Original ticket http projects.scipy.org numpy ticket 1565 on 2010 07 28 by fengy research, assigned to unknown.
|
|
Prepare the following file file tmp test 9223372043271415339 9223372043271415853 9223372043271415612 9223372043271416107 9223372043271415594 9223372043271415836 9223372043761290139 9223372044088967272 9223372044088967273 9223372043925949039 end of file And run the following code In [16] print loadtxt tmp test , dtype= uint64 ) > print loadtxt tmp test , dtype= uint64 )) [9223372043271415808 9223372043271415808 9223372043271415808 9223372043271415808 9223372043271415808 9223372043271415808 9223372043761289216 9223372044088967168 9223372044088967168 9223372043925948416] On the other hand, with fromfile ), In [2] print fromfile tmp test , dtype= uint64 , sep= ) > print fromfile tmp test , dtype= uint64 , sep= )) [9223372043271415339 9223372043271415853 9223372043271415612 9223372043271416107 9223372043271415594 9223372043271415836 9223372043761290139 9223372044088967272 9223372044088967273 9223372043925949039] Clearly the first few numbers are wrongly converted by loadtxt The problem was tracked to line 453 in numpy lib io.py, getconv.
|
|
The conversion for np.integer is int float x)), which is inexact for large integers.
|
|
I don t know if a priority of normal is appropriate, as this bug will produce hidden errors in programs that use numpy.","bug"
|
|
"6013","2163","fmt= u in savetxt behaves the same as d .
|
|
Trac 1566).
|
|
Original ticket http projects.scipy.org numpy ticket 1566 on 2010 07 28 by fengy research, assigned to unknown.
|
|
According to the document of savetxt ``u`` unsigned decimal integer ``x,X`` unsigned hexadecimal integer However, the behavior is not as documented in [38] print IDs[0 10] > print IDs[0 10]) [9223372043271415339 9223372043271415853 9223372043271415612 9223372043271416107 9223372043271415594 9223372043271415836 9223372043761290139 9223372044088967272 9223372044088967273 9223372043925949039] In [26] savetxt tmp IDs , IDs, u ) In [27] !head tmp IDs 9223372030438136277 9223372030438135763 9223372030438136004 9223372030438135509 9223372030438136022 9223372030438135780 9223372029948261477 9223372029620584344 9223372029620584343 9223372029783602577 The problem is on lib io.py 1542, fh.write format tuple row) \n ) Apparently python doesn t support u in the operator.
|
|
In [44] u , IDs[0] Out[44] u , 9223372043271415339) In [45] print u IDs[0] > print u IDs[0]) 9223372030438136277 Thus I don t see any easy fix for the problem.
|
|
What is ironic is that IDs[0] is however properly formatted by print .","bug"
|
|
"6013","2164","meshgrid reverses dimensions of output?
|
|
Trac 1567).
|
|
Original ticket http projects.scipy.org numpy ticket 1567 on 2010 07 29 by trac user auser2, assigned to unknown.
|
|
In the following code, wouldn t it make more sense if shape Xmesh)== NX,NY) rather than the current way around?
|
|
The order of the arguements to meshgrid is X,Y), so why shouldn t that be the order of the output s scripting?
|
|
Code below from numpy import meshgrid,arange,shape; NX=5; NY=10; X=arange NX); Y=arange NY); Xmesh, Ymesh = meshgrid X,Y); assert shape Xmesh) == NY,NX));","bug"
|
|
"6013","2165","Changeset 8541 breaks numscons build Trac 1568).
|
|
Original ticket http projects.scipy.org numpy ticket 1568 on 2010 07 31 by cgohlke, assigned to unknown.
|
|
On platforms without endian.h , changesets 8541 3 break the build process using numscons.
|
|
The file numpyconfig.h ends up containing the following line with illegal C syntax DEFINE NPY HAVE ENDIAN H A patch for numpy 1.5.x is attached.
|
|
Probably also applies to trunk and 1.4.x.","bug"
|
|
"6013","2166","Migrate sparse matrices to Numpy Trac 1569).
|
|
Original ticket http projects.scipy.org numpy ticket 1569 on 2010 08 01 by trac user NabilS, assigned to unknown.
|
|
I propose that sparse matrices be migrated to numpy from scipy) for these reasons They are essentially another type of array, thus very basic.
|
|
It would allow people to use them for instance in information retrieval scenarios) without all the irrelevant features of scipy making distribution simpler).","enhancement"
|
|
"6013","2167","IndexError index Ld) out of range 0<=index< Ld) in dimension d Trac 1570).
|
|
Original ticket http projects.scipy.org numpy ticket 1570 on 2010 08 02 by cgohlke, assigned to unknown.
|
|
Using numpy 1.4.1 or 1.5b1, built with msvc9 for win amd64 >>> a = np.arange 10) >>> a.shape = 2, 5 >>> a[1, 10] Traceback most recent call last) File <input> , line 1, in <module> IndexError index Ld) out of range 0<=index< Ld) in dimension d The error message should be IndexError index 10) out of range 0<=index<5) in dimension 1 The attached patch uses lld instead of Ld as an int64 format specifier.
|
|
This currently works only on Python 2.7 since lld is only processed starting with Python 2.7 see http bugs.python.org issue7228).
|
|
No additional tests fail with this patch applied.","bug"
|
|
"6013","2168","Python3.2a1 errors Trac 1571).
|
|
Original ticket http projects.scipy.org numpy ticket 1571 on 2010 08 03 by bsouthey, assigned to unknown.
|
|
First, string.maketrans function that was depreciated in Python 3.1 was removed in Python3.2.
|
|
http docs.python.org dev py3k whatsnew 3.2.html Consequently a minor change is needed in line 116 of numpy core code generators generate umath.py After doing that, the build on Linux 64 bit, python3.2 crashes due to undefined symbol PyCObject AsVoidPtr .
|
|
python3.2 c import numpy; numpy.test ) Traceback most recent call last) File <string> , line 1, in <module> File usr local lib python3.2 site packages numpy init .py , line 136, in <module> from .
|
|
import add newdocs File usr local lib python3.2 site packages numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File usr local lib python3.2 site packages numpy lib init .py , line 4, in <module> from .type check import File usr local lib python3.2 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File usr local lib python3.2 site packages numpy core init .py , line 5, in <module> from .
|
|
import multiarray ImportError usr local lib python3.2 site packages numpy core multiarray.so undefined symbol PyCObject AsVoidPtr From the build logs, this is appears to be this section In file included from usr local include python3.2 Python.h 8, from numpy core src multiarray common.c 2, from numpy core src multiarray multiarraymodule onefile.c 8 usr local include python3.2 pyconfig.h 137 1 warning this is the location of the previous definition In file included from numpy core src multiarray multiarraymodule onefile.c 14 numpy core src multiarray arraytypes.c.src In function xe2x80x98VOID getitemxe2x80x99 numpy core src multiarray arraytypes.c.src 633 warning assignment from incompatible pointer type In file included from numpy core src multiarray multiarraymodule onefile.c 21 numpy core src multiarray ctors.c In function xe2x80x98 array from buffer 3118xe2x80x99 numpy core src multiarray ctors.c 1656 warning assignment from incompatible pointer type In file included from numpy core src multiarray multiarraymodule onefile.c 25 numpy core src multiarray getset.c In function xe2x80x98array data getxe2x80x99 numpy core src multiarray getset.c 283 warning passing argument 1 of xe2x80x98PyMemoryView FromObjectxe2x80x99 from incompatible pointer type usr local include python3.2 memoryobject.h 54 note expected xe2x80x98struct PyObject xe2x80x99 but argument is of type xe2x80x98struct PyArrayObject xe2x80x99 In file included from numpy core src multiarray multiarraymodule onefile.c 28 numpy core src multiarray convert datatype.c In function xe2x80x98PyArray GetCastFuncxe2x80x99 numpy core src multiarray convert datatype.c 106 warning implicit declaration of function xe2x80x98PyCObject AsVoidPtrxe2x80x99 numpy core src multiarray convert datatype.c 106 warning assignment makes pointer from integer without a cast","bug"
|
|
"6013","2169","FloatingPointError invalid value encountered in isfinite Trac 1572).
|
|
Original ticket http projects.scipy.org numpy ticket 1572 on 2010 08 04 by trac user nwatanabe, assigned to pierregm.
|
|
numpy.ma.log10 [0.0]) raises an unexpected FloatingPointError in NumPy v1.4.1 >>> j = numpy.seterr invalid= ignore ) >>> numpy.log10 0.0) inf >>> numpy.ma.log10 [0.0]) masked array data = [ ], mask = [ True], fill value = 1e 20) >>> j = numpy.seterr invalid= raise ) >>> numpy.log10 0.0) inf >>> numpy.ma.log10 [0.0]) Traceback most recent call last) File <stdin> , line 1, in <module> File share apps python 2.6.5 lib python2.6 site packages numpy ma core.py , line 856, in call m = umath.isfinite result) FloatingPointError invalid value encountered in isfinite","bug"
|
|
"6013","2170","savetxt complex output Trac 1573).
|
|
Original ticket http projects.scipy.org numpy ticket 1573 on 2010 08 06 by nbecker, assigned to unknown.
|
|
savetxt does not work properly for complex.
|
|
Seems to only save real part.
|
|
I believe savetxt should save complex arrays with elements in the same format python using for complex numbers for e in u print e This gives r0 i0j) r1 i1j) ...","bug"
|
|
"6013","2171","1.5.0b1 umath complex failures on Windows Trac 1574).
|
|
Original ticket http projects.scipy.org numpy ticket 1574 on 2010 08 08 by rgommers, assigned to unknown.
|
|
Reported on Windows with Python 2.7 by two people with the 1.5.0b1 binary.
|
|
Python 2.7 r27 82525, Jul 4 2010, 09 01 59) [MSC v.1500 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.test ) Running unit tests for numpy NumPy version 1.5.0b1 NumPy is installed in C \Python27\lib\site packages\numpy Python version 2.7 r27 82525, Jul 4 2010, 09 01 59) [MSC v.1500 32 bit Intel)] nose version 0.11.0 .......................................................................................... ....................E..................................................................... .......................................................................................... .......................................................................................... .......................................................................................... .....................K.................................................................... .......................................................................................... ..................................................................K.......................
|
|
...................................K..K.............Warning invalid value encountered in absolute .Warning invalid value encountered in absolute ................K........F....FFFF.....K.F.........FF..FFF.....F.......................... ......................................................S................................... .......................................................................................... .......................................................................................... .......................................................................................... .......................................................................................... .......................................K.........K........................................ .......................................................................................... .......................................................................................... .......................................................................................... .......................................................................................... ..........................................S............................................... .......................................................................................... .......................................................................................... .......................................................................................... .......................................................................................... .......................................................................................... .......................................................................................... ................................................................ ====================================================================== ERROR test filename test memmap.TestMemmap) Traceback most recent call last) File C \Python27\lib\site packages\numpy\core\tests\test memmap.py , line 60, in test f ilename os.unlink tmpname) WindowsError [Error 32] The process cannot access the file because it is being used by an other process c \\users\\alanis 1\\appdata\\local\\temp\\mmapoo5c2b ====================================================================== FAIL test umath complex.TestCexp.test special values <ufunc exp >, inf, 0, inf, 0) test umath complex.TestCexp.test special values <ufunc exp >, inf, 0, inf, 0) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ inf nanj]) y array inf 0j)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ inf nanj])\n y array inf 0j)) ) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) test umath complex.TestCexp.test special values None,) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 60, in check ninf inf raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, inf) is nan, nan), expected 0, 0) >> raise AssertionError cexp inf, inf) is f, f), expected 0, 0) nan nan j).
|
|
real, nan nan j).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) test umath complex.TestCexp.test special values None,) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 73, in check inf inf raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, inf) is nan, nan), expected inf, nan) >> raise AssertionError cexp inf, inf) is f, f), expected inf, nan) nan nan j ).real, nan nan j).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) test umath complex.TestCexp.test special values None,) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 86, in check ninf nan raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, nan) is nan, nan), expected 0, 0) >> raise AssertionError cexp inf, nan) is f, f), expected 0, 0) nan nan j) .real, nan nan j).imag)) ====================================================================== FAIL test umath complex.TestCexp.test special values None,) test umath complex.TestCexp.test special values None,) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 99, in check inf nan raise AssertionError msgform z.real, z.imag)) AssertionError cexp inf, nan) is nan, nan), expected inf, nan) >> raise AssertionError cexp inf, nan) is f, f), expected inf, nan) nan nan j).real, nan nan j).imag)) ====================================================================== FAIL test special values test umath complex.TestClog) Traceback most recent call last) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 144, i n test special values self.assertRaises FloatingPointError, np.log, x) AssertionError FloatingPointError not raised ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ inf nanj]) y array inf infj)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ inf nanj])\n y array inf infj)) ) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ nan infj]) y array inf infj)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ nan infj])\n y array inf infj)) ) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf) test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ inf nanj]) y array inf infj)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ inf nanj])\n y array inf infj)) ) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf ) test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ nan infj]) y array inf infj)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ nan infj])\n y array inf infj)) ) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, nan, inf, inf, inf) test umath complex.TestCsqrt.test special values <ufunc sqrt >, nan, inf, inf, inf) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 539, i n check complex value assert equal f z1), z2) File C \Python27\lib\site packages\numpy\testing\utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File C \Python27\lib\site packages\numpy\testing\utils.py , line 686, in assert array e qual verbose=verbose, header= Arrays are not equal ) File C \Python27\lib\site packages\numpy\testing\utils.py , line 596, in assert array c ompare raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch [ True], False mismatch) x array [ nan nanj]) y array inf infj)) >> raise AssertionError \nArrays are not equal\n\n x and y nan location mismatch [ True] , False mismatch)\n x array [ nan nanj])\n y array inf infj)) ) ====================================================================== FAIL test umath complex.TestCsqrt.test special values None,) test umath complex.TestCsqrt.test special values None,) Traceback most recent call last) File C \Python27\lib\site packages\nose\case.py , line 183, in runTest self.test self.arg) File C \Python27\lib\site packages\numpy\core\tests\test umath complex.py , line 340, i n check ninf nan raise AssertionError msgform z.real, z.imag)) AssertionError csqrt inf, nan) is nan, nan), expected nan, inf) >> raise AssertionError csqrt inf, nan) is f, f), expected nan, inf) nan nan j).real, nan nan j).imag)) Ran 2456 tests in 19.070s FAILED KNOWNFAIL=8, SKIP=2, errors=1, failures=12) <nose.result.TextTestResult run=2456 errors=1 failures=12>","bug"
|
|
"6013","2172","format gives ValueError when passed float32 Trac 1575).
|
|
Original ticket http projects.scipy.org numpy ticket 1575 on 2010 08 09 by trac user crm, assigned to unknown.
|
|
Passing a float32 to the inbuilt format function returns a value error >>> format np.float32 1.23), .1f ) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError Unknown format code f for object of type str float64 works as expected >>> format np.float64 1.23), .1f ) 1.2 python version 2.6.5[[BR]] numpy version 1.4.1[[BR]] OS WindowsXP","bug"
|
|
"6013","2173","numpy.ma.flatnotmasked contiguous error in slice.stop values Trac 1576).
|
|
Original ticket http projects.scipy.org numpy ticket 1576 on 2010 08 09 by trac user gdmcbain, assigned to pierregm.
|
|
The function numpy.ma.flatnotmasked contiguous returns slices which miss the last element; for example, based on the example at [http docs.scipy.org doc numpy reference generated numpy.ma.flatnotmasked contiguous.html] import numpy as np a = np.arange 10) am = np.ma.array a, mask = a < 3) | a > 8) | a == 5)) segments = np.ma.flatnotmasked contiguous am) print the masked array , am print flatnotmasked contiguous , segments print am[segments[0]], am[segments[1]] prints the masked array [ 3 4 6 7 8 ] flatnotmasked contiguous [slice 3, 4, None), slice 6, 8, None)] [3] [6 7] which isn t really what we re after; I would have expected [3 4] and [6 7 8].
|
|
Compare the similarly named matplotlib.mlab.contiguous regions which does have the expected behaviour although its output format is different) import matplotlib.mlab segments1 = matplotlib.mlab.contiguous regions am) print contiguous regions , segments1 print am[slice segments1[0])], am[slice segments1[1])] which prints contiguous regions [ 3, 5), 6, 9)] [3 4] [6 7 8]","bug"
|
|
"6013","2174","numpy.polyfit ignores masks Trac 1577).
|
|
Original ticket http projects.scipy.org numpy ticket 1577 on 2010 08 10 by trac user gdmcbain, assigned to pierregm.
|
|
A masked array of abscissae gives the same results as if unmasked; compressing it gives a different result, which is perhaps the expected one; e.g., based on the example at [http docs.scipy.org doc numpy reference generated numpy.ma.polyfit.html?highlight=polyfit numpy.ma.polyfit], the script import numpy as np import numpy.ma as ma x = np.array [0.0, 1.0, 2.0, 3.0, 4.0, 5.0]) y = np.array [0.0, 0.8, 0.9, 0.1, 0.8, 1.0]) print unmasked , np.polyfit x, y, 3) mask = np.zeros x.shape, dtype = np.bool) mask[2] = True mx = ma.array x, mask = mask) my = ma.array y, mask = mask) print masked , np.polyfit mx, my, 3) ignores mask!
|
|
print compressed , np.polyfit mx.compressed ), my.compressed ), 3) produces unmasked [ 0.08703704 0.81349206 1.69312169 0.03968254] masked [ 0.08703704 0.81349206 1.69312169 0.03968254] compressed [ 0.07941176 0.74159664 1.5210084 0.01680672]","bug"
|
|
"6013","2175","test from unicode crashes on Python 2.7 and Numpy 2.0.0 devel Trac 1578).
|
|
Original ticket http projects.scipy.org numpy ticket 1578 on 2010 08 10 by trac user mbudisic, assigned to unknown.
|
|
I am trying to build my entire Python stack from source on x86 64 machine under CentOS, GCC 4.5.0, with custom compiled ATLAS.
|
|
Numpy builds but its test crash with SIGABRT signal, when numpy.test ) is run.
|
|
This is output into Python shell before exit.
|
|
test from unicode test defchararray.TestBasic) ... Debug memory block at address p=0x17941d0 API m 4 bytes originally requested The 7 pad bytes at p 7 are FORBIDDENBYTE, as expected.
|
|
The 8 pad bytes at tail=0x17941d4 are FORBIDDENBYTE, as expected.
|
|
The block was made by call 814353 to debug malloc realloc.
|
|
Data at p ff fe 00 00 Fatal Python error bad ID Allocated using API m , verified using API o I m attaching full gdb backtrace and pystack outputs.","bug"
|
|
"6013","2176","incorrect check docstring for tri[ul] indices from Trac 1579).
|
|
Original ticket http projects.scipy.org numpy ticket 1579 on 2010 08 10 by yarikoptic, assigned to unknown.
|
|
incorrect check if not arr.ndim==2 and arr.shape[0] == arr.shape[1] should be if not arr.ndim==2 and arr.shape[0] == arr.shape[1]) incorrect docstring signature is {{{arr,k=0}} while docstring describes n and k as in tri?
|
|
indices functions)","bug"
|
|
"6013","2177","shape lost inside record arrays Trac 1580).
|
|
Original ticket http projects.scipy.org numpy ticket 1580 on 2010 08 12 by trac user batripler, assigned to unknown.
|
|
The following code replicates the problem import numpy def foo dim) sdt=numpy.dtype [ v , float, 1,)))]) print numpy.empty dim,sdt)[ v ].shape print numpy.empty dim,numpy.dtype [ d1 ,sdt), d2 ,sdt)]))[ d1 ][ v ].shape def test ) foo 2,3)) foo 0,3)) Output 2, 3, 1) 2, 3, 1) 0, 3, 1) 1, 0, 3) The 1st line matches the 2nd line of output, since the array is of non zero size.
|
|
The same should be true for the 3rd and 4th lines, when the array happens to be zero sized i.e.
|
|
zero along at least one dimension).
|
|
Indexing zero sized record arrays with subarray fields doesn t return arrays of the correct shape.","bug"
|
|
"6013","2178","arr.copy order={ C , F }) raises TypeError Trac 1581).
|
|
Original ticket http projects.scipy.org numpy ticket 1581 on 2010 08 13 by trac user kwmsmith, assigned to unknown.
|
|
I d really like `arr.copy order= F )` to work is it supposed to as its docstring says, or is it supposed to raise a `TypeError` as it does now?
|
|
This is on numpy 1.4 >>> import numpy as np >>> a = np.arange 10).reshape 5,2) >>> a array [[0, 1], [2, 3], [4, 5], [6, 7], [8, 9]]) >>> print a.copy.
|
|
doc a.copy order= C ) Return a copy of the array.
|
|
Parameters order { C , F , A }, optional By default, the result is stored in C contiguous row major) order in memory.
|
|
If `order` is `F`, the result has Fortran column major) order.
|
|
If order is A Any ), then the result has the same order as the input.
|
|
Examples >>> x = np.array [[1,2,3],[4,5,6]], order= F ) >>> y = x.copy ) >>> x.fill 0) >>> x array [[0, 0, 0], [0, 0, 0]]) >>> y array [[1, 2, 3], [4, 5, 6]]) >>> y.flags[ C CONTIGUOUS ] True >>> a.copy order= C ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError copy ) takes no keyword arguments >>> a.copy order= F ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError copy ) takes no keyword arguments >>> From Pauli Virtanen The problem is probably that the function is using `PyArg ParseTuple` instead of `PyArg ParseTupleAndKeywords`.","bug"
|
|
"6013","2179","numpy bug in AIX problem with empty arrays Trac 1582).
|
|
Original ticket http projects.scipy.org numpy ticket 1582 on 2010 08 13 by trac user davide, assigned to unknown.
|
|
This error is correctly detected by a test case, and it boils down to the following >>> import cPickle, numpy >>> cPickle.loads cPickle.dumps numpy.array [], dtype=object))) Traceback most recent call last) File <stdin> , line 1, in <module> MemoryError The error is due to the different behavior malloc 0) has on AIX compared to linux see here for an example of what I mean http whatilearned2day.wordpress.com 2006 07 13 zero sized allocation using malloc on aix ) I am attaching two patches, that both fix this issue in a different way, but I don t know much about numpy internals, so you might develop a third, better one.","bug"
|
|
"6013","2180","fromfile tofile py3k errors on Windows Trac 1583).
|
|
Original ticket http projects.scipy.org numpy ticket 1583 on 2010 08 14 by rgommers, assigned to unknown.
|
|
In total 38 errors see attached), all like this ====================================================================== ERROR test multiarray.TestIO.test ascii Traceback most recent call last) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\nose 3.0.0.dev py3.1.egg\nose\case.py , line 178, in runTest self.test self.arg) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\core\tests\test multiarray.py , line 1017, in test ascii self.
|
|
check from 1 , 2 , 3 , 4 , [1.,2.,3.,4.
|
|
], sep= , ) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\core\tests\test multiarray.py , line 980, in check from y = np.fromfile self.filename, kw) IOError first argument must be an open file ====================================================================== ERROR test multiarray.TestIO.test ascii Traceback most recent call last) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\nose 3.0.0.dev py3.1.egg\nose\case.py , line 370, in tearDown try run self.inst, teardown , tearDown )) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\nose 3.0.0.dev py3.1.egg\nose\util.py , line 469, in try run return func ) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\core\tests\test multiarray.py , line 918, in tearDown os.unlink self.filename) WindowsError [Error 32] Sharing violation z \\var\\folders\\uu\\uuxfo1nlfae4yyypscz xe ti\\ tmp \\tmpdtktwh Looks like 1943 is related.","bug"
|
|
"6013","2181","Contradictory info in MaskedArray documentation Trac 1584).
|
|
Original ticket http projects.scipy.org numpy ticket 1584 on 2010 08 17 by trac user obernardo, assigned to pierregm.
|
|
>>> help np.ma.MaskedArray) returns a page informing that in MaskedArray constructor, copy = True and then says that the default value is False Construction | | x = MaskedArray data, mask=nomask, dtype=None, copy=True, | fill value=None, keep mask=True, hard mask=False, | shrink=True) | | Parameters | ... | copy bool, optional | Whether to copy the input data True), or to use a reference instead.
|
|
| Default is False.","bug"
|
|
"6013","2182","numpy.ma.core.
|
|
print templates missing long key Trac 1586).
|
|
Original ticket http projects.scipy.org numpy ticket 1586 on 2010 08 18 by trac user alue, assigned to pierregm.
|
|
Note I m not sure what s causing this problem.
|
|
After downloading from the repository, building, and installing, `numpy.ma.core.
|
|
print templates` has no long key, which is referenced in `numpy.ma.core.MaskedArray.
|
|
repr )`.
|
|
To reproduce uname a Linux HOSTNAME 2.6.32 24 generic 39 Ubuntu SMP Wed Jul 28 06 07 29 UTC 2010 i686 GNU Linux lsb release a No LSB modules are available.
|
|
Distributor ID Ubuntu Description Ubuntu 10.04.1 LTS Release 10.04 Codename lucid svn co http svn.scipy.org svn numpy trunk numpy > dev null cd numpy python3 setup.py build Although numpy ma core.py does not contain this problem, build py3k numpy ma core.py does.
|
|
One of the consequences is that calls to `ma.array )` fail.","bug"
|
|
"6013","2183","Silence test warnings in numpy 1.5 Trac 1587).
|
|
Original ticket http projects.scipy.org numpy ticket 1587 on 2010 08 19 by cgohlke, assigned to unknown.
|
|
This issue was discussed at http mail.scipy.org pipermail numpy discussion 2010 August 052283.html NumPy 1.5.0 beta 2 built with msvc9 mkl for Python 2.6, 2.7 and 3.1 32 and 64 bit) reports many > 200) warnings.
|
|
The warnings are of the type Warning invalid value encountered in for the functions reduce, fmax, fmin, logaddexp, maximum, greater, less equal, greater equal, absolute, and others.
|
|
When using MSVC compilers, ordered comparisons involving NaN raise an exception [1], i.e.
|
|
set the invalid x87 status bit.
|
|
The attached patch silences those warnings during the numpy tests.
|
|
It basically wraps the warning emitting code into err = np.seterr invalid= raise ) try ... finally np.seterr err) The patch also silences the divide by zero encountered in log warning in the masked array module tests.
|
|
This warning is also observed on other platforms.
|
|
[1] http msdn.microsoft.com en us library e7s85ffb 28v=VS.90 29.aspx","enhancement"
|
|
"6013","2184","numpy.ndarray.clip segfaults on big endian dtype Trac 1588).
|
|
Original ticket http projects.scipy.org numpy ticket 1588 on 2010 08 20 by trac user bdkern, assigned to unknown.
|
|
import numpy numpy.zeros 1,), dtype= >i2 ).clip min=0) segfaults on three systems tried Ubuntu Python 2.5.2 Numpy 1.1.1, Red Hat py 2.6.4 np 1.3.0, WinXP py 2.5.4 np 1.3.0.
|
|
This same setup also segfaults for dtypes >f4 , >f8 , etc.","bug"
|
|
"6013","2185","loadtxt reads floats in as float32 instead of float64 under certain cicumstances Trac 1589).
|
|
Original ticket http projects.scipy.org numpy ticket 1589 on 2010 08 21 by astrofrog, assigned to unknown.
|
|
I have a file called data.txt with the following contents cat data.txt 9.9999999961922814E 01 9.9999999996192290E 01 9.9999999999619227E 01 9.9999999999961919E 01 9.9999999999996192E 01 9.9999999999999611E 01 1.0000000000000000E 00 If I try and read this in using loadtxt, which should read numbers in using 64 bit) float by default, I get Python 2.6.1 r261 67515, Feb 11 2010, 00 51 29) [GCC 4.2.1 Apple Inc. build 5646)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 2.0.0.dev8657 >>> np.loadtxt data.txt ) array [ 1., 1., 1., 1., 1., 1., 1.])
|
|
If I now create a file called data2.txt with only the first line cat data2.txt 9.9999999961922814E 01 loadtxt works correctly Python 2.6.1 r261 67515, Feb 11 2010, 00 51 29) [GCC 4.2.1 Apple Inc. build 5646)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 2.0.0.dev8657 >>> np.loadtxt data2.txt ) array 0.99999999961922814)","bug"
|
|
"6013","2186","complex array to scalar conversion fails Trac 1590).
|
|
Original ticket http projects.scipy.org numpy ticket 1590 on 2010 08 22 by trac user drizzd, assigned to unknown.
|
|
While float array [1])) returns 1.0 as expected, complex array [1j])) throws TypeError can t convert complex to float I expected it to succeed and return 1.0j.
|
|
I suspect this is the same error triggered by the following code.
|
|
a = zeros 2, complex) b = ones 1) 1j a[0] = b","bug"
|
|
"6013","2187","flatten dtype does not handle titles Trac 1591).
|
|
Original ticket http projects.scipy.org numpy ticket 1591 on 2010 08 23 by trac user nasturtium86, assigned to unknown.
|
|
numpy.lib.io.flatten dtype ) raises ValueError too many values to unpack when you run it on a dtype that has titles.
|
|
For example, this code works import numpy dtype1 = numpy.dtype [ Current , f8 ), Voltage , f8 ), Power , f8 )]) print numpy.lib.io.flatten dtype dtype1) But this code doesn t import numpy dtype2 = numpy.dtype [ Amps , Current ), f8 ), Volts , Voltage ), f8 ), Watts , Power ), f8 )]) print numpy.lib.io.flatten dtype dtype2) The following monkey patch fixes the problem def flatten dtype ndtype) Unpack a structured data type.
|
|
names = ndtype.names if names is None return [ndtype] else types = [] for field in names typ fields = ndtype.fields[field] flat dt = flatten dtype typ fields[0]) types.extend flat dt) return types numpy.lib.io.flatten dtype=flatten dtype With this monkey patch in place, both dtype1 and dtype2 are handled correctly.
|
|
There is a thread open about this issue at [http www.mail archive.com numpy discussion scipy.org msg27441.html]","bug"
|
|
"6013","2188","Formatting with arrayprint goes wrong Trac 1592).
|
|
Original ticket http projects.scipy.org numpy ticket 1592 on 2010 08 24 by trac user gaubi, assigned to unknown.
|
|
If there is an array with no number smaller than 1.e 4 and precision is set to smaller than 4, values are formatted wrong and printed 0. only gaubitzer asterix test python Python 2.6.6rc1 r266rc1 83691, Aug 5 2010, 17 07 04) [GCC 4.4.5 20100728 prerelease)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> a = numpy.array [1.e 3,1.e 4]) >>> b = numpy.array [1.e 4,1.e 5]) >>> print a,b [ 0.001 0.0001] [ 1.00000000e 04 1.00000000e 05] >>> numpy.set printoptions precision=3) >>> print a,b [ 0.001 0. ]
|
|
[ 1.000e 04 1.000e 05]","bug"
|
|
"6013","2189","genfromtxt skip footer has problems with comments Trac 1593).
|
|
Original ticket http projects.scipy.org numpy ticket 1593 on 2010 08 26 by trac user miguel, assigned to unknown.
|
|
It seems that genfromtxt s skip footer parameter in numpy 1.4.1 ignores lines with comments or data with different layout from previous rows.
|
|
>>>print np.genfromtxt StringIO 4 1 2\n 3 12\n ), skip header=2, skip footer=3) []","bug"
|
|
"6013","2190","numpy.dtype float64 ) == None returns True Trac 1594).
|
|
Original ticket http projects.scipy.org numpy ticket 1594 on 2010 08 26 by trac user Zbyszek Szmek, assigned to unknown.
|
|
Automatic dtype conversion is performed when dtype object is compared with something of different class.
|
|
A special case is that None is converted to the default dtype.
|
|
This result in surprising behavior when comparing with None >>> numpy.dtype float64 ) == None True >>> numpy.dtype float32 ) == None False This behaviour is present in numpy at least 1.3.0 1.5.0.","bug"
|
|
"6013","2191","Add minlength keyword to bincount for minimum size of output array.
|
|
Trac 1595).
|
|
Original ticket http projects.scipy.org numpy ticket 1595 on 2010 08 27 by dwf, assigned to unknown.
|
|
The patch adds a minlength keyword argument to bincount, allowing users to specify a minimum size for the output array, e.g.
|
|
>>> import numpy as np >>> np.bincount np.random.random integers 0,1,size=40),minlength=3) array [23, 17, 0]) The patch includes docs for the new behaviour as well as several tests.","enhancement"
|
|
"6013","2192","error building numpy 1.5.0 docs Trac 1596).
|
|
Original ticket http projects.scipy.org numpy ticket 1596 on 2010 08 31 by lebedov, assigned to pv.
|
|
When I attempted to build the numpy 1.5.0 html docs, I observed the following error reading sources... [ 3 ] reference generated numpy.atleast 1d reading sources... [ 3 ] reference generated numpy.atleast 2d reading sources... [ 3 ] reference generated numpy.atleast 3d reading sources... [ 3 ] reference generated numpy.average reading sources... [ 3 ] reference generated numpy.bartlett Exception occurred File usr lib python2.6 site packages docutils statemachine.py , line 1193, in radd raise TypeError adding ViewList to a non ViewList ) TypeError adding ViewList to a non ViewList The full traceback has been saved in tmp sphinx err 90cDJW.log, if you want to report the issue t o the developers.
|
|
Please also report this if it was a user error, so that a better error message can be provided nex t time.
|
|
Either send bugs to the mailing list at <http groups.google.com group sphinx dev >, or report them in the tracker at <http bitbucket.org birkenfeld sphinx issues >.
|
|
Thanks!
|
|
I m running the build on 64 bit Linux with python 2.6.5, sphinx 1.0.3, docutils 0.7, jinja2 2.5.1, and pygments 1.3.1.","bug"
|
|
"6013","2193","loadtxt gets stuck on empty files Trac 1597).
|
|
Original ticket http projects.scipy.org numpy ticket 1597 on 2010 09 01 by trac user ctmedra, assigned to unknown.
|
|
I have found that loadtxt gets stuck on empty files.
|
|
My numpy.
|
|
version is 1.2.1 For instance In [1] import numpy In [2] fp=open foo.dat , w ) In [3] fp.close ) In [4] numpy.loadtxt foo.dat ) then I wait with no answer.","bug"
|
|
"6013","2194","Segfault when changing the filed names of an structured array Trac 1598).
|
|
Original ticket http projects.scipy.org numpy ticket 1598 on 2010 09 01 by FrancescAlted, assigned to unknown.
|
|
The next reproduces the segfault In [1] import numpy as np In [2] np.
|
|
version Out[2] 1.5.0 In [3] ra = np.fromiter i 3, i 2) for i in xrange 10)), dtype= i8,f8 ) In [4] ra.dtype.names = f1 , f2 ) In [5] ra Segmentation fault I don t know if the `names` filed of a dtype can be changed like this, but numpy let me do that, so...","bug"
|
|
"6013","2195","as strided arrays are not reshape able Trac 1599).
|
|
Original ticket http projects.scipy.org numpy ticket 1599 on 2010 09 01 by trac user dalleyg, assigned to unknown.
|
|
Arrays created by as strided are not reshape able.
|
|
If one attempts to reshape, the operation will appear to succeed, but corrupted data will result.
|
|
Here s a test example.
|
|
import numpy; print numpy.
|
|
version ; xx = numpy.arange 3, dtype=int)[None, None, ]; xx2 = numpy.lib.stride tricks.as strided xx, shape= 2,1,3), strides=xx.strides); xx3=xx2.reshape 2,3)); print xx3 It prints things like [[ 0 1 2] [5602136 33 0]] Note that the second row should look like the first, but it doesn t. This issue exists in numpy 1.3.0 and numpy 1.4.1 at least).
|
|
I do not know if it persists in numpy 1.5.0.
|
|
IMHO, reshape should be generalized to handle restrided arrays at least ones that are restrided to perform a cheap tile operation), copy the array if the strides are too nonstandard, and or raise an exception if the strides are too nonstandard.","bug"
|
|
"6013","2196","PyArray DescrConverter inconsistency with empty arrays in records Trac 1600).
|
|
Original ticket http projects.scipy.org numpy ticket 1600 on 2010 09 02 by trac user steveb, assigned to unknown.
|
|
When a record type is defined using a dictionary, dtype doesn t allow zero length arrays.
|
|
np.dtype [ val1 , np.float32, 0,))]) OK np.dtype { names [ val1 ], formats [ np.float32, 0,))]}) Throws TypeError It isn t clear to me why they shouldn t be allowed in this case.","bug"
|
|
"6013","2197","distutils ccompiler.py 68 UnicodeEncodeError on Python 3.2 Trac 1601).
|
|
Original ticket http projects.scipy.org numpy ticket 1601 on 2010 09 02 by trac user srid, assigned to cournape.
|
|
Python 3.2 py3k branch | Linux 64 bit FC4) | GCC 4.0.x | numpy 1.5.0 | ATLAS Running from numpy source directory.Traceback most recent call last) File setup.py , line 211, in <module> setup package ) File setup.py , line 204, in setup package configuration=configuration ) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils core.py , line 186, in setup return old setup new attr) File opt ActivePython 3.2 lib python3.2 distutils core.py , line 149, in setup dist.run commands ) File opt ActivePython 3.2 lib python3.2 distutils dist.py , line 919, in run commands self.run command cmd) File opt ActivePython 3.2 lib python3.2 distutils dist.py , line 938, in run command cmd obj.run ) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command build.py , line 37, in run old build.run self) File opt ActivePython 3.2 lib python3.2 distutils command build.py , line 128, in run self.run command cmd name) File opt ActivePython 3.2 lib python3.2 distutils cmd.py , line 315, in run command self.distribution.run command command) File opt ActivePython 3.2 lib python3.2 distutils dist.py , line 938, in run command cmd obj.run ) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command build src.py , line 152, in run self.build sources ) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command build src.py , line 169, in build sources self.build extension sources ext) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command build src.py , line 328, in build extension sources sources = self.generate sources sources, ext) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command build src.py , line 385, in generate sources source = func extension, build dir) File numpy core setup.py , line 399, in generate config h moredefs, ignored = cocache.check types config cmd, ext, build dir) File numpy core setup.py , line 41, in check types out = check types a, kw) File numpy core setup.py , line 294, in check types res = config cmd.check type size complex def, expected=2 expected[type]) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command config.py , line 236, in check type size headers, include dirs, c ) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command config.py , line 99, in compile body, headers, include dirs, lang)) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils command config.py , line 89, in wrap method ret = mth self,) args)) File opt ActivePython 3.2 lib python3.2 distutils command config.py , line 135, in compile self.compiler.compile [src], include dirs=include dirs) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils ccompiler.py , line 33, in <lambda> m = lambda self, args, kw func self, args, kw) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils ccompiler.py , line 206, in CCompiler compile self.
|
|
compile obj, src, ext, cc args, extra postargs, pp opts) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils ccompiler.py , line 33, in <lambda> m = lambda self, args, kw func self, args, kw) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils unixccompiler.py , line 23, in UnixCCompiler compile extra postargs, display = display) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils ccompiler.py , line 33, in <lambda> m = lambda self, args, kw func self, args, kw) File tmp tmpJXOWDV pypm numpy 1.5.0 numpy 1.5.0 build py3k numpy distutils ccompiler.py , line 68, in CCompiler spawn print o) UnicodeEncodeError ascii codec can t encode characters in position 50 52 ordinal not in range 128) BTW, there is no 1.5.0 version in Version list)","bug"
|
|
"6013","2198","Py3k syntax error in setupscons.py Trac 1602).
|
|
Original ticket http projects.scipy.org numpy ticket 1602 on 2010 09 02 by trac user srid, assigned to cournape.
|
|
numpy 1.5.0 and Python 3.1.2 python3 setupscons.py File setupscons.py , line 61 except ImportError, e SyntaxError invalid syntax Also import builtin need to be changed to import builtins on Py3k.","bug"
|
|
"6013","2199","in1d speed up Trac 1603).
|
|
Original ticket http projects.scipy.org numpy ticket 1603 on 2010 09 02 by trac user nhmc, assigned to unknown.
|
|
Robert Kern mentioned that in1d ar1, ar2) can be slow for the case when len ar2) is very small http article.gmane.org gmane.comp.python.numeric.general 40077 I ve attached a script that compares timings for the existing version of in1d and the kern in function described in the thread above; I use this to work out which algorithm is fastest for given lengths of ar1 and ar2 see also the attached plot).
|
|
Also attached is a patch that changes in1d to use the kern in algorithm when it results in a speed up.
|
|
I think the speedup, which can be > 10x for very large ar1 and very small ar2, is worth the minor increase in code complexity.","enhancement"
|
|
"6013","2200","distutils command config.py binary search needs floordiv to work under Python3 Trac 1604).
|
|
Original ticket http projects.scipy.org numpy ticket 1604 on 2010 09 04 by trac user cb123, assigned to unknown.
|
|
There is a bug in the build under Python3.
|
|
The symptom is an infinite loop of recompiling a configtest.c file.
|
|
numpy distutils command config.py uses hi lo) 2 in a binary search.
|
|
In py2 is integer floor div but in py3 is float converting div .
|
|
The fix is just using which has been available since python2.2 or >> 1 instead of 2 ).
|
|
As coded, binary search requires floordiv to converge.
|
|
Beyond the non convergence issue, the tests themselves fail erroneously.
|
|
Using a float divide creates a float.
|
|
Upon conversion to a string via size)s , that creates constants in the generated code like <= 16.0 .
|
|
Though the value of the expression is integral, as required, these being .0 instead of round numbers cause the left operand of the inequality to be promoted to double and <= to be evaluated in compile time floating point arithmetic.
|
|
Some compilers, e.g.
|
|
gcc 4.5.1, do not support FP math for compile time constants like an array length even if it evalutes to an integer.
|
|
Other compilers, e.g.
|
|
clang 1.1, g 4.5.1, work fine with it.
|
|
I have not looked into what relevant standards say, but using makes forward python behavior imitate the python 2.7 and early behavior and seems a reasonable fix.
|
|
Ok.
|
|
This is possibly the longest explanation I have ever had for what amounts to a one character patch. )","bug"
|
|
"6013","2201","Bug in numpy version 1.5.x RuntimeWarning tp compare didn t return 1 or 2 for exception Trac 1605).
|
|
Original ticket http projects.scipy.org numpy ticket 1605 on 2010 09 04 by trac user maldun, assigned to unknown.
|
|
We tried to upgrade NumPy in Sage.
|
|
We managed to get 1.4.1 working.
|
|
see http trac.sagemath.org sage trac ticket 9808 for more info) numpy 1.5.0 should work also, but we have problems with some of the doctests.
|
|
Example sage t valgrind devel sage sage rings polynomial polynomial element.pyx Total time for all tests 716.4 seconds maldun zauberbuch sage sage 4.5.2 sage t valgrind devel sage sage rings polynomial real roots.pyx > ERROR File .
|
|
devel sage sage rings polynomial real roots.pyx is missing The following tests failed .
|
|
devel sage sage rings polynomial real roots.pyx File not found Total time for all tests 0.0 seconds maldun zauberbuch sage sage 4.5.2 sage t valgrind devel sage sage rings polynomial real roots.pyx sage t valgrind devel sage sage rings polynomial real roots.pyx File home maldun sage sage 4.5.2 devel sage sage rings polynomial real roots.pyx , line 1819, in main .example 76 Failed example oc.find roots ) line 3064 sage >>> oc.find roots ) Expected nothing Got doctest 1 RuntimeWarning tp compare didn t return 1 or 2 for exception File home maldun sage sage 4.5.2 devel sage sage rings polynomial real roots.pyx , line 1840, in main .example 77 Failed example oc.find roots ) line 3085 sage >>> oc.find roots ) Expected nothing Got doctest 1 RuntimeWarning tp compare didn t return 1 or 2 for exception File home maldun sage sage 4.5.2 devel sage sage rings polynomial real roots.pyx , line 1934, in main .example 80 Failed example oc.find roots ) line 3157 sage >>> oc.find roots ) Expected nothing Got doctest 1 RuntimeWarning tp compare didn t return 1 or 2 for exception File home maldun sage sage 4.5.2 devel sage sage rings polynomial real roots.pyx , line 2320, in main .example 98 Failed example real roots x Integer 5) x Integer 2) Integer 9999)) Integer 2) Integer 1)) line 3870 sage >>> real roots x 5 x 2 9999) 2 1) Expected [ 29274496381311 9007199254740992, 419601125186091 2251799813685248), 1), 2126658450145849453951061654415153249597 21267647932558653966460912964485513216, 4253316902721330018853696359533061621799 42535295865117307932921825928971026432), 1), 1063329226287740282451317352558954186101 10633823966279326983230456482242756608, 531664614358685696701445201630854654353 5316911983139663491615228241121378304), 1)] Got doctest 1 RuntimeWarning tp compare didn t return 1 or 2 for exception [ 29274496381311 9007199254740992, 419601125186091 2251799813685248), 1), 2126658450145849453951061654415153249597 21267647932558653966460912964485513216, 4253316902721330018853696359533061621799 42535295865117307932921825928971026432), 1), 1063329226287740282451317352558954186101 10633823966279326983230456482242756608, 531664614358685696701445201630854654353 5316911983139663491615228241121378304), 1)] 4 items had failures 1 of 10 in main .example 76 1 of 9 in main .example 77 1 of 10 in main .example 80 1 of 44 in main .example 98 Test Failed 4 failures.
|
|
[227.6 s] The following tests failed sage t valgrind devel sage sage rings polynomial real roots.pyx Total time for all tests 227.6 seconds without valgrind we get memory leak It seems that it is related to the following bug http groups.google.com group cython users browse thread thread 624c696293b7fe44 If this problem can be solved we could move to numpy 1.5.0 in sage.","bug"
|
|
"6013","2202","broadcasting masked arrays in exponents Trac 1606).
|
|
Original ticket http projects.scipy.org numpy ticket 1606 on 2010 09 07 by WeatherGod, assigned to pierregm.
|
|
Here is a fun one... import numpy as np a 2d = np.random.random 3, 5)) b 1d = np.random.random 5) b 2d = np.vstack b 1d, b 1d, b 1d)) a ma 2d = np.ma.masked array a 2d, mask= numpy.random.random 3, 5)) < 0.25)) b ma 1d = np.ma.masked array b 1d, mask= numpy.random.random 5) < 0.25)) b ma 2d = np.ma.masked array b 2d, mask= numpy.random.random 3, 5)) < 0.25)) a b without broadcasting works) print a 2d b 2d a b with broadcasting works) print a 2d b 1d a ma b ma without broadcasting works) print a ma 2d b ma 2d a ma b ma with broadcasting works) print a ma 2d b ma 1d a b ma without broadcasting works) print a 2d b ma 2d a b ma with broadcasting FAILS) print a 2d b ma 1d Traceback most recent call last) File <stdin> , line 1, in <module> File home bvr Programs numpy numpy ma core.py , line 3697, in rpow return power other, self) File home bvr Programs numpy numpy ma core.py , line 6043, in power m |= invalid ValueError invalid return array shape Now, test broadcasting of the other side of the operator c = np.random.random 5) c ma = np.ma.masked array c, mask= np.random.random 10) < 0.25)) c b ma with broadcasting works) print c b ma 2d c ma b with broadcasting FAILS) print c ma b 2d Traceback most recent call last) File <stdin> , line 1, in <module> File home bvr Programs numpy numpy ma core.py , line 3693, in pow return power self, other) File home bvr Programs numpy numpy ma core.py , line 6043, in power m |= invalid ValueError invalid return array shape So, this fails if we broadcast the masked array portion of a power expression and the other value is a ndarray.
|
|
Note, the same also occurs if we broadcast a masked array in order to raise it to powers indicated by an ndarray.
|
|
This bug was originally reported in http www.mail archive.com numpy discussion scipy.org msg27679.html","bug"
|
|
"6013","2203","numpy.genfromtxt ) with Python 3.1.2 fails Trac 1607).
|
|
Original ticket http projects.scipy.org numpy ticket 1607 on 2010 09 09 by trac user magiclamp, assigned to unknown.
|
|
Hello genfromtxt ) fails with the following TypeError.
|
|
Thanks for the nice package!
|
|
[search dist][10273] > cat z.dat 0.01604 0.02668 0.03254 0.03254 0.15647 0.20313 0.02214 0.02214 [search dist][10274] > python Python 3.1.2 r312 79147, Sep 2 2010, 17 16 43) [GCC 4.5.0] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.version.version 1.5.0 >>> numpy.genfromtxt z.dat ) Traceback most recent call last) File <stdin> , line 1, in <module> File usr intel pkgs python 3.1.2 lib python3.1 site packages numpy lib npyio.py , line 1126, in genfromtxt first values = split line first line) File usr intel pkgs python 3.1.2 lib python3.1 site packages numpy lib iotools.py , line 206, in delimited splitter line = line.split self.comments)[0].strip asbytes \r\n )) TypeError Can t convert bytes object to str implicitly","bug"
|
|
"6013","2204","numpy.sort on x.flat modifies x Trac 1608).
|
|
Original ticket http projects.scipy.org numpy ticket 1608 on 2010 09 09 by trac user carlscheffler, assigned to unknown.
|
|
The code example below should be self explanatory.)
|
|
The sort function has the side effect of changing removing the view information of its argument when called with argument.flat.
|
|
This side effect does not appear when using sort ravel argument)) rather than sort argument.flat).
|
|
Tested on Python version 2.6.5, NumPy version 1.3.0; and Python version 2.6.2, NumPy version 1.2.1. import numpy x = numpy.array [[1,2],[3,4]]) x = x.transpose ) Create a transposed view of the original array print x [[1, 3], [2, 4]] OK y = numpy.sort numpy.ravel x)) print x [[1, 3], [2, 4]] OK y = numpy.sort x.flat) print x [[1, 2], [3, 4]] ERROR","bug"
|
|
"6013","2205","dotblas.dot ) not used on Python 3.x Trac 1609).
|
|
Original ticket http projects.scipy.org numpy ticket 1609 on 2010 09 11 by trac user hagen, assigned to unknown.
|
|
Under Python 3.x, the dotblas version of numpy.dot ) never gets used.
|
|
This is because the import line from dotblas import dot, ... in core numeric.py doesn t get fixed properly by 2to3, apparently because dotblas.so isn t there when 2to3 is run.","bug"
|
|
"6013","2206","fromfile tofile mess up file handle position on Python 3 Trac 1610).
|
|
Original ticket http projects.scipy.org numpy ticket 1610 on 2010 09 11 by pv, assigned to unknown.
|
|
>>> import numpy as np >>> f = open foo.dat , w b ) >>> f.write np.arange 255, dtype= u1 ).tostring )) 255 >>> f.seek 0) 0 >>> np.fromfile f, dtype= u1 , count=4) array [0, 1, 2, 3], dtype=uint8) >>> f.tell ) 255","bug"
|
|
"6013","2207","Different behaviour in ndarray.astype ) for scalars and arrays Trac 1611).
|
|
Original ticket http projects.scipy.org numpy ticket 1611 on 2010 09 14 by trac user lorenz, assigned to unknown.
|
|
There is an issue when using astype ) on a scalar array, i.e.
|
|
with shape == ), when doing an endiannes change possible in other cases too?)
|
|
Example >>> import numpy as np >>> >>> np.array 42).astype >i4 ).dtype dtype >i4 ) >>> np.array 42, dtype= >i4 ).astype >i4 ).dtype dtype int32 ) >>> np.array [42]).astype >i4 ).dtype dtype >i4 ) >>> np.array [42], dtype= >i4 ).astype >i4 ).dtype dtype >i4 ) Above one should get dtype >i4 ) in both cases.
|
|
>>> np.array 42, dtype= >f4 ).astype >f4 ).dtype dtype float32 ) >>> np.array [42], dtype= >f4 ).astype >f4 ).dtype dtype >f4 ) And here dtype >f4 ) in both cases.
|
|
>>> np.array 42, dtype= >f8 ).astype >f8 ).dtype dtype float64 ) >>> np.array [42], dtype= >f8 ).astype >f8 ).dtype dtype >f8 ) And here dtype >f8 ) in both cases.
|
|
My numpy version >>> np.version.version 1.4.0.dev7417","bug"
|
|
"6013","2208","Alignment don t work in subarrays Trac 1612).
|
|
Original ticket http projects.scipy.org numpy ticket 1612 on 2010 09 14 by trac user Ihor.Melnyk, assigned to unknown.
|
|
Specifying alignment in dtype specification does not work alignment should be 4 in both cases) >>> t = np.dtype 1i4 , align=True) >>> t.alignment 4 >>> t = np.dtype 2i4 , align=True) >>> t.alignment 1 This is critical to ensure that NPY binary data matches exactly complex structures defined in C. Proposed solution Index descriptor.c =================================================================== descriptor.c revision 8715) descriptor.c working copy) 254,6 254,7 newdescr >subarray = pya malloc sizeof PyArray ArrayDescr)); newdescr >subarray >base = type; newdescr >flags = type >flags; newdescr >alignment = type >alignment; Py INCREF val); newdescr >subarray >shape = val; Py XDECREF newdescr >fields);","bug"
|
|
"6013","2209","Documentation change Datetime requires ctypes with Python2.4 Trac 1613).
|
|
Original ticket http projects.scipy.org numpy ticket 1613 on 2010 09 15 by bsouthey, assigned to pv.
|
|
datetime requires ctypes otherwise it causes `RunTime` error.
|
|
However, ctypes has only been included in Python since version 2.5.
|
|
A standalone version of ctypes for Python2.4 is available at [http starship.python.net crew theller ctypes http starship.python.net crew theller ctypes ] Consequently the documentation needs to be updated to refer to this dependency.","bug"
|
|
"6013","2210","f2py seems to have problem processing extended lines?
|
|
Trac 1614).
|
|
Original ticket http projects.scipy.org numpy ticket 1614 on 2010 09 15 by trac user gbrault, assigned to pearu.
|
|
I have this example fortran file from the slicot library, where the sign is used to make line extension.
|
|
I get an error with f2py sage ubuntu sage slicot src py f2py c m slycot MD03BD.f llapack running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src building extension slycot sources f2py options [] f2py > tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c creating tmp tmpdSKvbW creating tmp tmpdSKvbW src.linux i686 2.6 Reading fortran codes... Reading file MD03BD.f format fix,strict) Post processing... Block slycot Block md03bd Block fcn Block qrfact Block lmparm Post processing stage 2)... Building modules... Constructing call back function cb fcn in md03bd user routines getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of def fcn iflag,m,n,ipar,lipar,dpar1,ldpar1,dpar2,ldpar2,x,nfevl,dwork,dwork,ldjsav,dwork,ldwork,infol) return Constructing call back function cb lmparm in md03bd user routines getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of def lmparm cond,n,ipar,lipar,e dwork jac e,ldj,iwork,diag,e dwork e e,delta,par,e iwork n 1 e,e dwork iw1 e,e dwork iw2 e,toldef,e dwork iw3 e,e ldwork iw3 1 e,infol) return Constructing call back function cb qrfact in md03bd user routines getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of def qrfact n,ipar,lipar,fnorm,e dwork jac e,ldj,e dwork e e,e dwork jw1 e,gnorm,iwork,e dwork jw2 e,e ldwork jw2 1 e,infol) return Building module slycot ... Constructing wrapper function md03bd ... getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of getarrdims warning assumed shape array, using 0 instead of md03bd xinit,scale,cond,fcn,qrfact,lmparm,m,n,itmax,factor,nprint,ipar,lipar,dpar1,ldpar1,dpar2,ldpar2,x,diag,nfev,njev,ftol,xtol,gtol,tol,iwork,dwork,ldwork,iwarn,info,[fcn extra args,qrfact extra args,lmparm extra args]) Wrote C API module slycot to file tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c adding tmp tmpdSKvbW src.linux i686 2.6 fortranobject.c to sources.
|
|
adding tmp tmpdSKvbW src.linux i686 2.6 to include dirs.
|
|
copying usr lib python2.6 dist packages numpy f2py src fortranobject.c > tmp tmpdSKvbW src.linux i686 2.6 copying usr lib python2.6 dist packages numpy f2py src fortranobject.h > tmp tmpdSKvbW src.linux i686 2.6 running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 customize IntelFCompiler Could not locate executable ifort Could not locate executable ifc customize LaheyFCompiler Could not locate executable lf95 customize PGroupFCompiler Could not locate executable pgf90 Could not locate executable pgf77 customize AbsoftFCompiler Could not locate executable f90 customize NAGFCompiler Found executable usr bin f95 customize VastFCompiler customize GnuFCompiler customize CompaqFCompiler Could not locate executable fort customize IntelItaniumFCompiler Could not locate executable efort Could not locate executable efc customize IntelEM64TFCompiler customize Gnu95FCompiler Found executable usr bin gfortran customize Gnu95FCompiler customize Gnu95FCompiler using build ext building slycot extension compiling C sources C compiler gcc pthread fno strict aliasing DNDEBUG g fwrapv O2 Wall Wstrict prototypes fPIC creating tmp tmpdSKvbW tmp creating tmp tmpdSKvbW tmp tmpdSKvbW creating tmp tmpdSKvbW tmp tmpdSKvbW src.linux i686 2.6 compile options I tmp tmpdSKvbW src.linux i686 2.6 I usr lib python2.6 dist packages numpy core include I usr include python2.6 c gcc tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 error redefinition of parameter xe2x80x98dworkxe2x80x99 tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 note previous definition of xe2x80x98dworkxe2x80x99 was here tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 error redefinition of parameter xe2x80x98dworkxe2x80x99 tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 note previous definition of xe2x80x98dworkxe2x80x99 was here tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 error redefinition of parameter xe2x80x98dworkxe2x80x99 tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 note previous definition of xe2x80x98dworkxe2x80x99 was here tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 error redefinition of parameter xe2x80x98dworkxe2x80x99 tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c 379 note previous definition of xe2x80x98dworkxe2x80x99 was here error Command gcc pthread fno strict aliasing DNDEBUG g fwrapv O2 Wall Wstrict prototypes fPIC I tmp tmpdSKvbW src.linux i686 2.6 I usr lib python2.6 dist packages numpy core include I usr include python2.6 c tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.c o tmp tmpdSKvbW tmp tmpdSKvbW src.linux i686 2.6 slycotmodule.o failed with exit status 1 The only difference I see is that this module use for line continuation.
|
|
Do you think this is the issue?
|
|
I have uploaded the fortran file for you to try BR","bug"
|
|
"6013","2211","genfromtxt reading quoted csv files enhancement Trac 1615).
|
|
Original ticket http projects.scipy.org numpy ticket 1615 on 2010 09 15 by trac user alefnula, assigned to unknown.
|
|
genfromtxt cannot handle the csv files that use quoting.
|
|
For example This is my text, that has a comma inside , Other value , 3 Another text, with coma , More text, with comma ,5 This is a csv text where the delimiter is , , but the values also contain , .
|
|
Here is the pach that enables the user to specify the quoter the quoting character).
|
|
The default behaviour is the same as the old behaviour of genfromtxt function, but if the quoter is set, quoting is taken into account.
|
|
Patch is in the attachment.","enhancement"
|
|
"6013","2212","numpy.loadtxt uses both file like and iterator interfaces instead of just one Trac 1616).
|
|
Original ticket http projects.scipy.org numpy ticket 1616 on 2010 09 19 by zachrahan, assigned to unknown.
|
|
numpy.loadtxt takes as input a filename or file like object; for many useful tasks like stopping reading lines after some specific delimiter) it would be easiest to just wrap a real file in some shim object that behaves as desired, rather than adding more options to loadtxt itself.
|
|
However, loadtxt requires that file like objects passed in expose both the iterator interface and a readline attribute.
|
|
This makes the task unnecessarily hard, and makes it impossible to just pass in a generator.
|
|
The following patch removes the use of readline without altering the behavior of the code at all.
|
|
Index numpy lib npyio.py =================================================================== numpy lib npyio.py revision 8716) numpy lib npyio.py working copy) 597,10 597,11 fh = bz2.BZ2File fname) else fh = open fname, U ) elif hasattr fname, readline ) fh = fname else raise ValueError fname must be a string or file handle ) try fh = iter fname) except raise ValueError fname must be a string or file handle ) X = [] def flatten dtype dt) 633,14 634,18 Skip the first `skiprows` lines for i in xrange skiprows) fh.readline ) try fh.next ) except StopIteration raise IOError End of file reached before encountering data. )
|
|
Read until we find a line with some values, and use it to estimate the number of columns, N. first vals = None while not first vals first line = fh.readline ) if not first line EOF reached try first line = fh.next ) except StopIteration raise IOError End of file reached before encountering data. )
|
|
first vals = split line first line) N = len usecols or first vals)","enhancement"
|
|
"6013","2213","Casting numpy.complex64 to Python) complex chops off imaginary part Trac 1617).
|
|
Original ticket http projects.scipy.org numpy ticket 1617 on 2010 09 20 by inducer, assigned to unknown.
|
|
Example session >>> import numpy >>> numpy.complex64 1 2j) 1 2j) >>> complex numpy.complex64 1 2j)) 1 0j)","bug"
|
|
"6013","2214",".real on object array misbehaves Trac 1618).
|
|
Original ticket http projects.scipy.org numpy ticket 1618 on 2010 09 21 by pv, assigned to unknown.
|
|
It probably shouldn t do this >>> import numpy as np >>> x = np.array 1 2j, object) >>> x.real array 1 2j), dtype=object) The implementation of .real for object arrays seems suspicious.","bug"
|
|
"6013","2215","Loading aligned dtype error Trac 1619).
|
|
Original ticket http projects.scipy.org numpy ticket 1619 on 2010 09 24 by trac user Ihor.Melnyk, assigned to unknown.
|
|
import numpy as np t = np.dtype i1, i4, i1 , align=True) d = np.zeros 1, t) np.save test.npy , d) data = np.load test.npy ) Traceback most recent call last) File D \Projects\Cuda\Cuda Git\pathwise\liinc\model\feeds\numpy bug.py , line 8, in <module> data = np.load test.npy ) File D \Projects\Cuda\Cuda Git\pathwise\pathwise\vendors\lib64\x64\python\numpy\lib\npyio.py , line 314, in load return format.read array fid) File D \Projects\Cuda\Cuda Git\pathwise\pathwise\vendors\lib64\x64\python\numpy\lib\format.py , line 440, in read array shape, fortran order, dtype = read array header 1 0 fp) File D \Projects\Cuda\Cuda Git\pathwise\pathwise\vendors\lib64\x64\python\numpy\lib\format.py , line 358, in read array header 1 0 dtype = numpy.dtype d[ descr ]) ValueError two fields with the same name","bug"
|
|
"6013","2216","axis parameter support for matrix inverse function numpy.linalg.inv) Trac 1620).
|
|
Original ticket http projects.scipy.org numpy ticket 1620 on 2010 09 25 by trac user rogerbrent, assigned to pv.
|
|
This would make it possible to compute multiple matrix inverses at once for arrays with shapes like 10, 2, 2), without having to resort to a loop.
|
|
I don t believe there is currently anything in NumPy that lets you do this.","enhancement"
|
|
"6013","2217","numpy 1.5.0 win32 superpack python2.7 Installer fails on 64bit windows Trac 1621).
|
|
Original ticket http projects.scipy.org numpy ticket 1621 on 2010 09 26 by trac user alxx, assigned to rgommers.
|
|
When installing numpy 1.5.0 win32 superpack python2.7 on windows7 64 bit, the installer can not find python2.7 in the registry.
|
|
Its a know problem with the distutils installer.
|
|
see http bugs.python.org issue6792 and http stackoverflow.com questions 3652625 installing setuptools on 64 bit windows Need to have a separate 64 bit windows installer or look for the additional registry entries to find the install location or have a fall through that allows a user to still install even if it doesn t find python 2.7.
|
|
PyQt v4.7.7 uses the fall through option.
|
|
wxpython has a separate installer Someone provides unofficial numpy and scipy) binaries here http www.lfd.uci.edu gohlke pythonlibs but I d much prefer using an official package if possible)","enhancement"
|
|
"6013","2218","Bus error when attempting to use itemset ) on a read only memory mapped array Trac 1622).
|
|
Original ticket http projects.scipy.org numpy ticket 1622 on 2010 09 26 by WarrenWeckesser, assigned to unknown.
|
|
From http projects.scipy.org scipy ticket 1202 In [1] import numpy as np In [2] np.
|
|
version Out[2] 1.5.0.dev8712 In [3] !cat tmp.dat DBC In [4] m = np.memmap tmp.dat , np.uint8, r ) In [5] m[0 1].itemset 65) Bus error The bus error does not occur if the file is opened with mode r .","bug"
|
|
"6013","2219","Masked array equivalent append function Trac 1623).
|
|
Original ticket http projects.scipy.org numpy ticket 1623 on 2010 09 29 by trac user gsever, assigned to pierregm.
|
|
Current np.append implementation can t handle masked array appending.
|
|
I[1] a = np.ma.masked equal [1,2,3], value=2) I[2] b = np.ma.masked equal [4,3,2], value=2) I[3] np.append a,b) O[3] masked array data = [1 2 3 4 3 2], mask = False, fill value = 999999) It would be nice to have np.ma.append ) to operate on masked arrays properly.
|
|
The desired output from that given piece would be np.ma.append a,b) masked array data = [1 3 4 3 ], mask = [False, True, False, False, False, True], fill value = 999999)","enhancement"
|
|
"6013","2220","dot gives wrong result for large array with float32 objects Trac 1624).
|
|
Original ticket http projects.scipy.org numpy ticket 1624 on 2010 10 01 by trac user c huettel, assigned to unknown.
|
|
with Python 2.6.5 r265 79063, Apr 16 2010, 13 09 56) [GCC 4.4.3] on linux2 do import numpy as np a = map lambda i np.float32 i ), range 100000 ) ) x = np.array a ) r0 = np.float32 0 ) for b in a r0 = b b ra = reduce lambda s,i s x[i] x[i], range len x)), np.float32 0) ) rx = np.dot x,x) print 1.0f r0), 1.0f ra), 1.0f rx), ra rx this gives 333328452419584 333328452419584 333328318201856 1.34218e 08 rx, the dot product from numpy, gives a different result as the much slower ) r0 and ra computations.
|
|
Same results are gotten when 1) the range is only up to 10000 2) float instead of float32 is used What is the reason?","bug"
|
|
"6013","2221","Numpy is ignoring the fact isfinite is undefined Trac 1625).
|
|
Original ticket http projects.scipy.org numpy ticket 1625 on 2010 10 05 by trac user drkirkby, assigned to unknown.
|
|
On Solaris or !OpenSolaris, when Python is built, it is not compiled as a C99 application.
|
|
That s a choice the Python developers took).
|
|
Therefore, `isfinite`, which was not defined in the C standard until C99, does not exist.
|
|
Hence when Python builds, the config.log shows ac cv have decl isfinite=no The latest Numpy, and this includes the latest snapshot taken only an hour or two) ago builds OK on Solaris.
|
|
However, when one tries to import this into Python with import numpy it generates an error ImportError ld.so.1 python fatal relocation error file home palmieri fulvia numpy sage 4.6.alpha2 local lib python2.6 site packages numpy core multiarray.so symbol isfinite referenced symbol not found It would appear that Numpy is trying to use `isfinite`, despite the fact that Python is built without adding specific C99 support.
|
|
Linux systems, which I assume is what most Numpy developers use, defines `isfinite` when one includes `math.h` in a file.
|
|
But this should not happen unless one builds C99 code, as `isfinite` was not defined until the C99 standard.
|
|
That s a problem with Linux it tends not to conform to standards too well.
|
|
Solaris in comparison will not define `isfinite` unless the code is C99.
|
|
`isfinite` is defined in a header file ` usr include iso math c99.h` but not in ` usr include math.h`.
|
|
However, if the compiler is building C99 code, then the contents of ` usr include iso math c99.h` will be included, so `isfinite` will be defined.
|
|
I believe this is a bug in Numpy, as it should respect the fact the Python build has determined `isfinite` is undefined so sets `ac cv have decl isfinite=no`) and so therefore Numpy should not try to use `isfinite`.
|
|
This came up as a bug when we tried to update Numpy in Sage.
|
|
It is documented on http trac.sagemath.org sage trac ticket 9808 but I believe you will find the information there confusing, as it covers many issues associated with upgrading Numpy and Scipy in Sage.
|
|
The basic problem is Numpy is using `isfinite` even though Python has worked out that it is undefined.
|
|
Dave","bug"
|
|
"6013","2222","numpy compilation fails on Solaris 10 with SunStudio 12.1 compiler Trac 1626).
|
|
Original ticket http projects.scipy.org numpy ticket 1626 on 2010 10 06 by trac user plubans580, assigned to unknown.
|
|
numpy.core.umath fails to build on Solaris 10 with Sun Studio 12.1 when compiling numpy core src umath umathmodule onefile.c.
|
|
The build gives the following error ... building numpy.core.umath extension compiling C sources C compiler u peterlu solaris sunstudio12.1 bin cc DNDEBUG O I u peterlu solaris python deploy include Kpic creating build temp.solaris 2.10 i86pc 2.5 numpy core src umath compile options Ibuild src.solaris 2.10 i86pc 2.5 numpy core src umath \ Inumpy core include Ibuild src.solaris 2.10 i86pc 2.5 numpy core include numpy \ Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath \ Inumpy core src multiarray Inumpy core src umath Inumpy core include \ I u peterlu solaris python deploy include python2.5 \ Ibuild src.solaris 2.10 i86pc 2.5 numpy core src multiarray \ Ibuild src.solaris 2.10 i86pc 2.5 numpy core src umath c cc numpy core src umath umathmodule onefile.c numpy core include numpy ufuncobject.h , line 384 warning statement not reached numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath ufunc object.c , line 3823 warning operands have incompatible pointer types op == cc acomp failed for numpy core src umath umathmodule onefile.c numpy core include numpy ufuncobject.h , line 384 warning statement not reached numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1386 operand must have real floating type op isnan numpy core src umath loops.c.src , line 1387 operand must have real floating type op isnan numpy core src umath ufunc object.c , line 3823 warning operands have incompatible pointer types op == cc acomp failed for numpy core src umath umathmodule onefile.c error Command u peterlu solaris sunstudio12.1 bin cc DNDEBUG O \ I u peterlu solaris python deploy include Kpic Ibuild src.solaris 2.10 i86pc 2.5 numpy core src umath \ Inumpy core include Ibuild src.solaris 2.10 i86pc 2.5 numpy core include numpy \ Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath \ Inumpy core src multiarray Inumpy core src umath Inumpy core include \ I u peterlu solaris python deploy include python2.5 Ibuild src.solaris 2.10 i86pc 2.5 numpy core src multiarray \ Ibuild src.solaris 2.10 i86pc 2.5 numpy core src umath \ c numpy core src umath umathmodule onefile.c \ o build temp.solaris 2.10 i86pc 2.5 numpy core src umath umathmodule onefile.o failed with exit status 1 > It looks like the sun compiler is very strict about what arguments its isnan built in will accept.
|
|
Replacing the literal 0 s with 0.0 s allows the build to complete numpy 1.5.0 numpy core src umath loops.c.src 2010 08 25 08 32 15.000000000 0400 numpy 1.5.0.new numpy core src umath loops.c.src 2010 10 06 01 29 46.735838000 0400 1376,22 1376,22 } } NPY NO EXPORT void C TYPE sign char args, intp dimensions, intp steps, void NPY UNUSED func)) { fixme sign of nan is currently 0 UNARY LOOP { const type in1r = type )ip1)[0]; const type in1i = type )ip1)[1]; type )op1)[0] = CGT in1r, in1i, 0, 0) ?
|
|
1 CLT in1r, in1i, 0, 0) ?
|
|
1 type )op1)[0] = CGT in1r, in1i, 0.0, 0.0) ?
|
|
1 CLT in1r, in1i, 0.0, 0.0) ?
|
|
1 CEQ in1r, in1i, 0, 0) ?
|
|
0 NPY NAN C )); type )op1)[1] = 0; } } begin repeat1 kind = maximum, minimum OP = CGE, CLE NPY NO EXPORT void","bug"
|
|
"6013","2223","Docs for genfromtxt does not properly explain unmask Trac 1627).
|
|
Original ticket http projects.scipy.org numpy ticket 1627 on 2010 10 07 by fonnesbeck, assigned to pv.
|
|
The description of unmask should include the limitation that all columns must be of the same type.
|
|
Otherwise, transposition cannot occur, and the resulting output appears to be a bug.","bug"
|
|
"6013","2224","density keyword for histogram Trac 1628).
|
|
Original ticket http projects.scipy.org numpy ticket 1628 on 2010 10 08 by rgommers, assigned to unknown.
|
|
In this thread http thread.gmane.org gmane.comp.python.numeric.general 39746 focus=40089 the preferred solution to solve a long standing bug with `normed=True` in histogram seemed to be to add a new density keyword and deprecate the normed keyword.
|
|
The fix now in trunk is different, see r8672, r8674, r8678.
|
|
Those changesets are not in 1.5.x.
|
|
This should be resolved for the 2.0.0 release.","bug"
|
|
"6013","2225","np.prod returns 0 in case of integer overflow Trac 1629).
|
|
Original ticket http projects.scipy.org numpy ticket 1629 on 2010 10 10 by trac user kilian, assigned to unknown.
|
|
I am using np.prod to calculate the number of elements of a sparse matrix np.prod C.shape)) and I noticed the following behavior In case the result is greater than 2 31, zero is returned.
|
|
Instead, the result should be converted to int long int or at least an exception should be raised).
|
|
>>> import numpy as np >>> np.version.version 1.4.0 >>> np.prod [2 16,2 16]) 0","bug"
|
|
"6013","2226","Documentation of known failure testing decorator is incomplete Trac 1630).
|
|
Original ticket http projects.scipy.org numpy ticket 1630 on 2010 10 10 by trac user jdmarch, assigned to pv.
|
|
[http projects.scipy.org numpy wiki TestingGuidelines known failures skipping tests] describes decorators numpy.testing.dec.skipif and knownfailureif.
|
|
It implies that these two decorators are equally available.
|
|
However skipif can be used in tests run from the command line with nosetests, and knownfailureif cannot.
|
|
Ralf Gommers explained Knownfailure is a numpy specific plugin, which needs to be loaded for tests decorated with dec.knownfailureif to be marked as K instead of E. This plugin loading is done by the `NoseTester` class defined in numpy testing nosetester.py) which is used when tests are run through numpy.test ).","bug"
|
|
"6013","2227","Solaris10 SPARC gcc 4.5.1 32bit SIGBUS in test mrecords.py Trac 1631).
|
|
Original ticket http projects.scipy.org numpy ticket 1631 on 2010 10 10 by nstange, assigned to charris.
|
|
Hi everybody, for numpy 1.5.0, the testsuite crashes here with an SIGBUS here Test filled w flexible dtype ...
|
|
Bus Error core dumped) numpy 1.5.0 numpy ma tests test core.py) and here Test that exotic formats are processed properly ...
|
|
Bus Error core dumped) numpy 1.5.0 numpy ma tests test mrecords.py).
|
|
The problem is that SPARC processors require alignment, that is 8 byte values have to be aligned at 8 byte memory boundraries.
|
|
The good point is numpy knows that the array is misaligned verified using gdb) and thus, the fix is easy see attached diff 02 put mask only on behaved arrays.diff) The check for ISCONTINOUS isn t enough in PyArray PutMask item selection.c), also check for ISBEHAVED = ISALIGNED ISWRITABLE).
|
|
But now, another problem arises a SIGSEGV here Tests fields retrieval numpy 1.5.0 numpy ma tests test mrecords.py 77).
|
|
The problem is in copy same shape numpy 1.5.0 numpy core src multiarrayctors.c 732) dest >dimensions == NULL.
|
|
Please also note that maxaxis == 1 at that point.
|
|
Check the attached diff 04 copy from same shape zerodim fix.diff for details.
|
|
There s one point in the diff I m unsure about The right position of PyArray INCREF src); PyArray XDECREF dest); I tried to resemble the original logical position, but since I have no clue about Python s reference counting, please have a look.
|
|
While debugging the last SIGSEGV issue, I stumbled over another mistake at least I think so) In PyArray IterAllButAxis numpy 1.5.0 numpy core src multiarray iterators.c), the minaxis won t be set if the first nonzero stride is the smallest one.
|
|
See attached diff 03 fix iterallbutaxis minstride search.diff) for a fix.
|
|
The reason why I poste those three issues into one report is that the testsuite still doesn t succeed and thus, I m unsure if I ve broken sth.
|
|
with my diffs.
|
|
What I get now is ====================================================================== FAIL Ticket 1897 second test Traceback most recent call last) File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy core tests test regression.py , line 1255, in test structured a rrays with objects2 assert sys.getrefcount strb) == numb AssertionError 7 = <module sys built in)>.getrefcount aaaa ) array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]) = <module numpy from pf m m2 22086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 site packages numpy init .pyc >.array [[ 0, aaaa ), 1, bbbb )]], i8,O ) >> array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )])[array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]).nonzero )] = array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]).ravel )[ 1] assert <module sys built in)>.getrefcount bbbb ) == 7 assert <module sys built in)>.getrefcount aaaa ) == 7 2 ====================================================================== FAIL Test filled w mvoid Traceback most recent call last) File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy ma tests test core.py , line 518, in test filled w mvoid assert equal tuple test), 1, default fill value 1.)))
|
|
File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy ma testutils.py , line 94, in assert equal return assert equal on sequences actual, desired, err msg= ) File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy ma testutils.py , line 66, in assert equal on sequences assert equal actual[k], desired[k], item= r\n s k,err msg)) File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy ma testutils.py , line 98, in assert equal raise AssertionError msg) AssertionError Items are not equal item=1 ACTUAL 2.0 DESIRED 1e 20 >> raise AssertionError \nItems are not equal \nitem=1\n\n ACTUAL 2.0\n DESIR ED 1e 20 ) At least not segfaults bus errors anymore ;).
|
|
Btw.
|
|
I don t know what mvoid is, but have a look at the following myuid myhost xas solaris10 python2 python 2.7 ve0 gcc bin python Python 2.7 r27 82500, Oct 9 2010, 17 26 38) [GCC 4.5.1] on sunos5 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> import numpy.ma as ma >>> import numpy.ma.core >>> from numpy.ma.core import >>> x = ma.array [ 1,2.
|
|
)], mask=[ 0,1)], dtype=[ a , int), b , float)]) >>> print x.filled ) [ 1, 1e 20)] >>> x = mvoid 1,2.
|
|
), mask=[ 0,1)], dtype=[ a , int), b , float)]) >>> print x.filled ) 1, 2.0) >>> Since the non working mvoid mask could have an impact on results, I cannot release that build to our the scientists at our site.
|
|
Do you have any idea what the issue could be?
|
|
Do you know what a mvoid is?
|
|
I can t find any documentation about it.
|
|
I m seriously thinking about just removing that class from my numpy build If I knew that there weren t any dependants outside of numpy)... Wishes Nicolai","bug"
|
|
"6013","2228","test structured arrays with objects2 fails with too high refcounts due to gcc 4.5.x bug Trac 1632).
|
|
Original ticket http projects.scipy.org numpy ticket 1632 on 2010 10 11 by nstange, assigned to unknown.
|
|
Hi everybody, this isn t really a numpy bug but numpy is affected by a gcc bug that is UNCONFIRMED right now) I ve found while debugging the following numpy testsuite failer ====================================================================== FAIL Ticket 1299 second test Traceback most recent call last) File pf m m222086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 sit e packages numpy core tests test regression.py , line 1255, in test structured a rrays with objects2 assert sys.getrefcount strb) == numb AssertionError 7 = <module sys built in)>.getrefcount aaaa ) array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]) = <module numpy from pf m m2 22086 xas solaris10 python2 python 2.7 ve0 gcc lib python2.7 site packages numpy init .pyc >.array [[ 0, aaaa ), 1, bbbb )]], i8,O ) >> array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )])[array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]).nonzero )] = array [[ 0L, aaaa ), 0L, aaaa )]], dtype=[ f0 , >i8 ), f1 , |O4 )]).ravel )[ 1] assert <module sys built in)>.getrefcount bbbb ) == 7 assert <module sys built in)>.getrefcount aaaa ) == 7 2 You ll find the gcc bug here http gcc.gnu.org bugzilla show bug.cgi?id=45967 Maybe we just wait for what the gcc people say about this issue; in the meanwhile you can use the attached patch against numpy 1.5.0 that fixes that issue with a workaround.
|
|
The workaround isn t very clear code and not commented ;) and thus, it either shouldn t be applied to the main branch at all or at least it should be removed again as soon as the issue has been fixed in gcc again if it is really a gcc bug).
|
|
Wishes Nicolai","bug"
|
|
"6013","2229","Changeset 8438 breaks ldexp on 64bit bigendian systems Trac 1633).
|
|
Original ticket http projects.scipy.org numpy ticket 1633 on 2010 10 11 by nstange, assigned to unknown.
|
|
Testsuite fails with ====================================================================== FAIL test ldexp test umath.TestLdexp) Traceback most recent call last) File pf m m222086 xas solaris10 64 python2 python 2.7 ve0 gcc lib python2.7 site packages numpy core tests test umath.py , line 391, in test ldexp assert almost equal ncu.ldexp 2., 3), 16.)
|
|
File pf m m222086 xas solaris10 64 python2 python 2.7 ve0 gcc lib python2.7 site packages numpy testing utils.py , line 463, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL 2.0 DESIRED 16.0 >> raise AssertionError \nArrays are not almost equal\n ACTUAL 2.0\n DESIRED 16.0 ) The issue has been introduced here http projects.scipy.org numpy changeset 8438 The reason is in DOUBLE ldexp loops.c.src 1122) const int in2 = int )ip2; Since ip2 is being cast from long ) to int ), and thus in2 exponent) will only receive the first 4 bytes, that is the 4 most significant bytes on bigendian, that is zero on 64bit.
|
|
System s ldexp s exp argument is always an int, and thus in my opinion, the correct behaviour is to reject any exp argument we cannot convert to an system s int, that is an np.intc, and adjusting the docs and the testsuite rather than silently giving invalid results.
|
|
The attached diff fixes the ldexp signatures to match the system s ldexp again broken by changeset 8438) adjusts the testsuite to only put in system s ints","bug"
|
|
"6013","2230","numpy.fromfile does not accept StringIO object Trac 1634).
|
|
Original ticket http projects.scipy.org numpy ticket 1634 on 2010 10 11 by trac user foobaron, assigned to unknown.
|
|
aArray = numpy.fromfile signalfile, numpy.dtype u1 ), number of bytes) signalfile = <StringIO.StringIO instance at 0xa0008ec>1 > IOError first argument must be an open file used python version 2.6.5","enhancement"
|
|
"6013","2231","Wrong isnan usage Trac 1635).
|
|
Original ticket http projects.scipy.org numpy ticket 1635 on 2010 10 12 by nstange, assigned to unknown.
|
|
Just for completeness Don t give system s isnan an int!
|
|
gcc 4.5.0 will reject this.","bug"
|
|
"6013","2232","genfromtxt bug with usecols and names=True Trac 1636).
|
|
Original ticket http projects.scipy.org numpy ticket 1636 on 2010 10 12 by jseabold, assigned to unknown.
|
|
from stringIO import stringIO import numpy as np np.
|
|
version 1.5.0 s = StringIO col0,col1,col2,col3,col4\n0,1,2,3,4\n0,1,2,3,4\n0,1,2,3,4 ) arr = np.genfromtxt s, names=True, usecols= 0,2,4), delimiter= , , dtype=[ int , int , int ]) arr.dtype.names col0 , col1 , col2 ) Interestingly in trying to reproduce, I mistyped the dtype and got the expected names but no error on my bad dtype).
|
|
arr = np.genfromtxt s, names=True, usecols= 0,2,4), delimiter= , , dtype=[ int , int , int , int , int ]) arr.dtype.names col0 , col2 , col4 ) It also works as expected with dtype=None.","bug"
|
|
"6013","2233","Dtypes should not use = in hashing Trac 1637).
|
|
Original ticket http projects.scipy.org numpy ticket 1637 on 2010 10 13 by pv, assigned to unknown.
|
|
Dtypes with byteorder== < and byteorder== = should hash equal on big endian systems).
|
|
This does not seem the case now http projects.scipy.org scipy ticket 1300 Normally, it s not possible to create dtype objects with the native variant of < and > as byte order, as that is converted to = in PyArray DescrConverter.
|
|
However, they might appear from pickling, 3rd party extensions etc., so doing the right thing would be useful.","bug"
|
|
"6013","2234","Provide .H for arrays Trac 1638).
|
|
Original ticket http projects.scipy.org numpy ticket 1638 on 2010 10 14 by trac user nicki, assigned to unknown.
|
|
Hi, I noticed that one difference between numpy arrays and matrices is that .H transpose conjugation) is only implemented for matrices.
|
|
.T , however, being structurally completely equivalent, is implemented for both.
|
|
A use case for .H would be mass dot products for multivectors.
|
|
In any case, this is an inconsistency in numpy.
|
|
Cheers, Nico","enhancement"
|
|
"6013","2235","FAIL test complex scalar warning test regression.TestRegression) Trac 1639).
|
|
Original ticket http projects.scipy.org numpy ticket 1639 on 2010 10 14 by stsci sienkiew, assigned to pv.
|
|
====================================================================== FAIL test complex scalar warning test regression.TestRegression) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 numpy core tests test regression.py , line 1407, in test complex scalar warning assert warns np.ComplexWarning, float, x) File usr stsci pyssgdev 2.5.4 numpy testing utils.py , line 1432, in assert warns func.
|
|
name ) AssertionError No warning raised when calling float Python 2.5.4 Red Hat Enterprise 5, 64 bit Mac Leopard Solaris Red Hat Enterprise 4, 32 bit Red Hat Enterprise 4, 64 bit This was recently observed in origin?http github.com numpy numpy.git fetch) origin?http github.com numpy numpy.git push) 1ee71d93fd069a51f45b4fa91e9e91d083a9334e but has been happening at least since Sept 30.","bug"
|
|
"6013","2236","structured array incorrectly gives a broadcast error Trac 1640).
|
|
Original ticket http projects.scipy.org numpy ticket 1640 on 2010 10 15 by mwiebe, assigned to unknown.
|
|
Here s a simple failure that should work >>> p = np.rec.recarray 2,),dtype=[ f0 , f4 ,3), f1 , f4 ,1)]) >>> p==p Traceback most recent call last) File <stdin> , line 1, in <module> ValueError shape mismatch objects cannot be broadcast to a single shape A slightly different dtype, and it works >>> p = np.rec.recarray 2,),dtype=[ f0 , f4 ,2), f1 , f4 ,1)]) >>> p==p rec.array [[ True, True], [ True, True]], dtype=bool)","bug"
|
|
"6013","2237","Strange behavior of ldexp on Python2.5 Wine Trac 1641).
|
|
Original ticket http projects.scipy.org numpy ticket 1641 on 2010 10 19 by pv, assigned to unknown.
|
|
On Python 2.5 MinGW Wine and probably on a real Windows, too) >>> import numpy as np >>> np.iinfo l ).max 2147483647 >>> np.ldexp 2, 2147483647) 0.0 >>> np.ldexp 2, 2147483646) 0.0 >>> np.ldexp 2, 2147483645) Warning overflow encountered in ldexp inf On Python 2.6 the above works correctly.
|
|
This occurs also on [d24db3430a] http github.com numpy numpy commit d24db3430a) and is therefore independent of 2229.","bug"
|
|
"6013","2238","FAIL test special values test umath complex.TestClog) on Solaris Sun Studio Trac 1642).
|
|
Original ticket http projects.scipy.org numpy ticket 1642 on 2010 10 19 by trac user opk, assigned to unknown.
|
|
I m trying to build 1.5.1rc1 on Solaris using Sun Studio.
|
|
I m getting messages saying undefined symbol for exp.
|
|
lm is needed for exp ).
|
|
Even the Linux man page for exp suggests linking with lm.
|
|
Error message is below.
|
|
Thank you compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I opt csw include python2.6 c cc configtest.c opt SUNWspro bin cc configtest.o o configtest Undefined first referenced symbol in file exp configtest.o ld fatal Symbol referencing errors.
|
|
No output written to configtest Undefined first referenced symbol in file exp configtest.o ld fatal Symbol referencing errors.
|
|
No output written to configtest failure.
|
|
removing configtest.c configtest.o","bug"
|
|
"6013","2239","[Py3] numpy arrays not recognized by struct module Trac 1643).
|
|
Original ticket http projects.scipy.org numpy ticket 1643 on 2010 10 20 by inducer, assigned to unknown.
|
|
This error is new in Py3 Numpy 1.5 and looks spurious >>> import numpy, struct >>> struct.pack I , numpy.uint32 0)) Traceback most recent call last) File <stdin> , line 1, in <module> struct.error required argument is not an integer Version Info Python 3.1.2 release31 maint, Jul 8 2010, 09 18 08) >>> numpy.
|
|
version 1.5.1rc1","bug"
|
|
"6013","2240","numpy.test ) fails test umath complex.TestClog on Solaris 5.10 Trac 1644).
|
|
Original ticket http projects.scipy.org numpy ticket 1644 on 2010 10 20 by trac user lytenyn, assigned to unknown.
|
|
Running numpy.test ) fails one test on Sun Solaris 5.10.
|
|
This is IMHO not a critical test.
|
|
Do you agree with that?
|
|
If so, could you please mark this test as a known failure, in order to not scare other users?
|
|
Thanks in advance!
|
|
The following test is failing FAIL test special values test umath complex.TestClog) Traceback most recent call last) File opt csw lib python site packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File opt csw lib python site packages numpy core tests test umath complex.py , line 173, in test special values assert almost equal np.log x), y) File opt csw lib python site packages numpy testing utils.py , line 443, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL array [ inf 0.j]) DESIRED inf 3.1415926535897931j) >> raise AssertionError \nArrays are not almost equal\n ACTUAL array [ inf 0.j])\n DESIRED inf 3.1415926535897931j) ) The full output of numpy.test ) is >>> numpy.test ) Running unit tests for numpy NumPy version 1.5.0 NumPy is installed in opt csw lib python site packages numpy Python version 2.6.6 r266 84292, Oct 12 2010, 15 25 47) [C] nose version 0.11.2 .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... ................................................................K... .................................................................... .................................................................... ....................................................................
|
|
...................K.........................................Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power ...................................................K................ ......K.F........................................................... ...........................................S........................ .................................................................... ........................................................................................................................................ .................................................................... .................................................................... .................................................................... ................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... ............................................................................................................................S........... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................... .................................................................................................... ====================================================================== FAIL test special values test umath complex.TestClog) Traceback most recent call last) File opt csw lib python site packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File opt csw lib python site packages numpy core tests test umath complex.py , line 173, in test special values assert almost equal np.log x), y) File opt csw lib python site packages numpy testing utils.py , line 443, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal ACTUAL array [ inf 0.j]) DESIRED inf 3.1415926535897931j) >> raise AssertionError \nArrays are not almost equal\n ACTUAL array [ inf 0.j])\n DESIRED inf 3.1415926535897931j) ) Ran 2608 tests in 60.639s FAILED KNOWNFAIL=4, SKIP=2, failures=1) <nose.result.TextTestResult run=2608 errors=0 failures=1>","bug"
|
|
"6013","2241","Setting a value to an indexed array with composite dtype using where or a boolean mask silently fails Trac 1645).
|
|
Original ticket http projects.scipy.org numpy ticket 1645 on 2010 10 20 by trac user pch, assigned to unknown.
|
|
The following ways to set an array with composite dtype or fwiw a recarray) fail.
|
|
Only using a slice works correctly >>> a=zeros 10,), dtype=[ f ,int)]) >>> a array [ 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,)], dtype=[ f , <i8 )]) >>> i=array range 10)) >>> a[i<=5][ f ] = 1 >>> a array [ 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,)], dtype=[ f , <i8 )]) >>> a[where i<=5)][ f ] = 1 >>> a array [ 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,), 0,)], dtype=[ f , <i8 )]) >>> a[0 5 2][ f ] = 1 >>> a array [ 1,), 0,), 1,), 0,), 1,), 0,), 0,), 0,), 0,), 0,)], dtype=[ f , <i8 )])","bug"
|
|
"6013","2242","numpy crash inside anki and mathplotlib) Trac 1646).
|
|
Original ticket http projects.scipy.org numpy ticket 1646 on 2010 10 24 by trac user batusek, assigned to unknown.
|
|
For a long time I have been trying to solve the issue of crashing anki spaced repetition learning program) in the numpy library.
|
|
The issue is tracked here http code.google.com p anki issues detail?id=1570), I will try to summarize it.
|
|
Anki for Windows crashes with the message like the offset changes in time).
|
|
AppName anki.exe AppVer 0.1.0.0 ModName numpy.core.
|
|
dotblas.pyd ModVer 0.0.0.0 Offset 00010ffd Now some facts numpy is called from mathplotlib anki crashes only in some environments e.g on my desktop PC with Windows XP CZ SP3), while it perfectly works on the others e.g.
|
|
my laptop with Windows XP CZ SP3) the crash doesn t occur, when development version of numpy is used I mean .py instead of the .pyd library) several numpy versions have been tried and all cause the same behavior The error probably occurs somewhere around the following lines FigureCanvas.setSizePolicy self, QSizePolicy.Expanding, QSizePolicy.Fixed) FigureCanvas.updateGeometry self) I am able to do some traces, if you provide my with a debug version of the library.
|
|
Or, I would at least need any ideas how to reproduce the bug in the development environment.
|
|
Or I would need some instructions how to build my own version of the numpy.core.
|
|
dotblas.pyd library so that I can add there some debug traces.
|
|
I am able to cooperate to find the root cause of this defect.","bug"
|
|
"6013","2243","int32 ) different behaviour under 1.3.0 and 1.4.1 for single element arrays Trac 1647).
|
|
Original ticket http projects.scipy.org numpy ticket 1647 on 2010 10 24 by trac user phaustus, assigned to unknown.
|
|
Hi, Under numpy 1.4.1, the output of > int32 array [0.]))
|
|
is simply 0 , whereas under numpy 1.3.0, the output is array [0], dtype=int32) .
|
|
While the former is a number, the latter is an array.
|
|
Interestingly, the output for > int32 array [0.,1.]))
|
|
and for > int32 array [])) is the same in each version i.e., both return arrays).
|
|
I apologize if this has already been addressed.
|
|
I did a search and found nothing.
|
|
Cheers, Phaustus","bug"
|
|
"6013","2244","Masked round ) doesn t work on zero dimensional vectors Trac 1648).
|
|
Original ticket http projects.scipy.org numpy ticket 1648 on 2010 10 25 by trac user andylei, assigned to unknown.
|
|
>>> import numpy as np >>> x = np.ma.array np.array 1.1)) >>> np.round x) AttributeError >>> np.ma.round x) AttributeError","bug"
|
|
"6013","2245","Strange array behavior in numpy c extension Trac 1649).
|
|
Original ticket http projects.scipy.org numpy ticket 1649 on 2010 10 25 by trac user c00kiemonster, assigned to rgommers.
|
|
I don t know whether I have made some very trivial mistake here, but I think there is something fishy going on in this particular example of mine.
|
|
I have attached four files C arraytest.h setup.py tests.py weird.c I am basically making some numpy c extensions, as described on this page http www.scipy.org Cookbook C Extensions NumPy arrays).
|
|
In this very simple example of mine I have included a c extension that returns the same 1d array that it has been given as a parameter ie a new array, but identical in size and data).
|
|
Everything works as expected, at least until i feed it with a 1d array that has come from slicing a column from a 2d array.
|
|
In that case it doesn t return the expected values, instead it seems to be taking the values from the original 2d array.
|
|
I think this has something to do with the way the pointer is iterating through the data of the 2d array and not the sliced out 1d array.
|
|
Feel free to compile the c extension weird.c) and run the tests.py script to replicate the behavior.
|
|
I don t really know if this is a bug, but it took me a very long time to weed this particular behavior out the script in which I found the issue is very much larger) so I thought I d let you guys know in case there is some bug hidden in there.","bug"
|
|
"6013","2246","Notify re distributors of binaries on SSE2 3 issues Trac 1650).
|
|
Original ticket http projects.scipy.org numpy ticket 1650 on 2010 10 27 by pv, assigned to rgommers.
|
|
The fact that the superpack Windows binary is SSE3 optimized can be confusing for binary redistributors e.g.
|
|
users of `py2exe` or similar tools).
|
|
See 2242 and http code.google.com p anki issues detail?id=1570 Somewhere preferably in the installer?)
|
|
we probably need to add a notice for binary re distributors that if they intend to share the installed Numpy libraries further, they should pick the no SSE version.
|
|
It might also be nice if Numpy could check what is was compiled with and fail on import with an error message...)","bug"
|
|
"6013","2247","Inconsistency with index ) for rank 1 arrays?
|
|
Trac 1651).
|
|
Original ticket http projects.scipy.org numpy ticket 1651 on 2010 10 28 by FrancescAlted, assigned to unknown.
|
|
I find this a bit misleading >>> a = np.arange 10) >>> a[np.array 0)] 0 >>> a[np.array [0])] array [0]) >>> a[[0]] array [0]) But, for regular python lists we have >>> l = a.tolist ) >>> l[np.array 0)] 0 >>> l[np.array [0])] 0 i.e.
|
|
indexing with a rank 0 array and a rank 1 array with one single element return the same result, which I find inconsistent with the expected behaviour for this case, i.e.
|
|
>>> l[[0]] TypeError Traceback most recent call last) tmp tables 2.2 <ipython console> in <module> ) TypeError list indices must be integers, not list The ultimate reason for all of this is >>> np.array 0).
|
|
index ) 0 >>> np.array [0]).
|
|
index ) 0 But I wonder why !NumPy needs the latter behaviour, instead of the more logical >>> np.array [0]).
|
|
index ) TypeError Traceback most recent call last) tmp tables 2.2 <ipython console> in <module> ) TypeError only rank 0 integer arrays can be converted to an index This inconsistency has indeed introduced a bug in my application and for solving this I d need something like def is idx index) Check if an object can work as an index or not.
|
|
if hasattr index, index ) Only works on Python 2.5 on if hasattr index, shape ) and index.shape == 1,)) return False try idx = index.
|
|
index ) return True except TypeError return False return False i.e.
|
|
for determining if an object can be an index or not, I need to explicitly check for a shape different from 1,), which is unnecessarily complicated.
|
|
So I find the current behaviour prone to introduce errors in apps and I m wondering why exactly `np.array [1])` should work as an index at all.
|
|
It would not be better if `np.array [1]).
|
|
index )` would raise a `TypeError`?
|
|
Thanks","bug"
|
|
"6013","2248","tofile ought to support anything with a write method Trac 1652).
|
|
Original ticket http projects.scipy.org numpy ticket 1652 on 2010 10 28 by gerritholl, assigned to unknown.
|
|
It would be good if .tofile ) supported anything with a write method.
|
|
I was debugging my code and wanted to see exactly what my array.tofile sys.stdout, ) would write.
|
|
However, as I was working in WingIDE, sys.stdout was actually a StringIO, and this code failed.
|
|
>>> type r[field]) <type numpy.float32 > >>> sys.stdout <StringIO.StringIO instance at 0x161a170> >>> r[field].tofile sys.stdout, , s ) Traceback most recent call last) File <string> , line 1, in <fragment> IOError first argument must be a string or open file See also http projects.scipy.org numpy ticket 949","enhancement"
|
|
"6013","2249","Enable future statements in sphinxext plot directive Trac 1653).
|
|
Original ticket http projects.scipy.org numpy ticket 1653 on 2010 10 29 by mforbes, assigned to pv.
|
|
Including statements like from future import division in the plot pre code fails because these are module specific.
|
|
One could fix this by changing line 464 of plot directive.py to something like futures = [l for l in setup.config.plot pre code if from future import in l] exec \n .join futures [code]) in ns This could probably be made slightly more robust.
|
|
There may be another way of doing this, but I cannot see how right now.
|
|
Michael.","enhancement"
|
|
"6013","2250","possible race condition with savez Trac 1654).
|
|
Original ticket http projects.scipy.org numpy ticket 1654 on 2010 10 29 by trac user unclepedro, assigned to unknown.
|
|
Hi, np.savez is not multi process safe, because it always uses gettempdir ) key .npy .
|
|
So, if you re running the same script over different data sets, you will roach your own data when one process removes the tmp arr 0.npy of another process.
|
|
As luck would have it, I m in that exact situation, so I have an example error Traceback most recent call last) File home pedro prospectus prelim bbvs bin skldthresh , line 99, in <module> main ) File home pedro prospectus prelim bbvs bin skldthresh , line 90, in main np.savez smask t 0.2f outfile,threshold), result) File usr lib python2.5 site packages numpy lib io.py , line 230, in savez zip.write filename, arcname=fname) File usr lib python2.5 zipfile.py , line 541, in write st = os.stat filename) OSError [Errno 2] No such file or directory tmp arr 0.npy Fortunately, replacing tempfile.gettempdir ) with tempfile.mkdtemp ) AND remembering to clean up at the end with os.rmdir direc) is all it takes to fix it as far as I can tell.
|
|
Attached is a patch.","bug"
|
|
"6013","2251","Poor precision using linalg.solve for certain matrices Trac 1655).
|
|
Original ticket http projects.scipy.org numpy ticket 1655 on 2010 10 30 by trac user zoof, assigned to pv.
|
|
There appears to be a bug in either the linear system solver or in the way than linalg computes LU decompositions.
|
|
On a number of different distributions, using somewhat different versions of numpy, I get much better precision by inverting the matrix linalg.inv) and multiplying than by either linalg.solve alone or linalg.cholesky in combination with linalg.solve.
|
|
An example has been posted to http pastebin.com KpXdTVh7 numpy 1.3.0 on Ubuntu Lucid numpy 1.5.0 on Arch numpy 1.5.1rc1 on Centos","bug"
|
|
"6013","2252","genfromtxt bug with delimiter is int type Trac 1656).
|
|
Original ticket http projects.scipy.org numpy ticket 1656 on 2010 10 31 by trac user sunqiang, assigned to pierregm.
|
|
from userguide http docs.scipy.org doc numpy user basics.io.genfromtxt.html the delimiter argument the second example) >>> import numpy as np >>> from StringIO import StringIO >>> data = 1 2 3\n 4 5 67\n890123 4 >>> np.genfromtxt StringIO data), delimiter=3) Traceback most recent call last) File <ipython console> , line 1, in <module> File C \Python27\lib\site packages\numpy\lib\npyio.py , line 1451, in genfrom txt raise ValueError errmsg) ValueError Some errors were detected !
|
|
Line 3 got 3 columns instead of 4) >>> np.
|
|
version 1.5.1rc1 quick and dirty fix insert line = line.rstrip ) at line 212 of C \Python27\Lib\site packages\numpy\lib\ iotools.py, aka, the first line of method fixedwidth splitter self, line) of class LineSplitter","bug"
|
|
"6013","2253","np.ma.masked invalid and precision Trac 1657).
|
|
Original ticket http projects.scipy.org numpy ticket 1657 on 2010 11 01 by trac user bevanj, assigned to pierregm.
|
|
It seems that the use of np.ma.masked invalid results in arrays printing precision of 12 regardless of the default setting or any subsequent setting.
|
|
Generating masked arrays does not seem to have this behaviour.
|
|
In [1] eg = np.random.random 1) In [2] eg mask = np.ma.masked array eg) In [3] eg mask invalid = np.ma.masked invalid eg) In [4] eg Out[4] array [ 0.78592569]) In [5] eg mask Out[5] masked array data = [ 0.78592569], mask = False, fill value = 1e 20) In [6] eg mask invalid Out[6] masked array data = [0.785925693305], mask = [False], fill value = 1e 20) In [7] np.set printoptions precision=3) In [8] eg Out[8] array [ 0.786]) In [9] eg mask Out[9] masked array data = [ 0.786], mask = False, fill value = 1e 20) In [10] eg mask invalid Out[10] masked array data = [0.785925693305], mask = [False], fill value = 1e 20)","bug"
|
|
"6013","2254","arrays with 0 length unicode strings does not unpickle well Trac 1658).
|
|
Original ticket http projects.scipy.org numpy ticket 1658 on 2010 11 02 by FrancescAlted, assigned to unknown.
|
|
This reproduces the problem >>> pickle.loads pickle.dumps np.array u a ))) 1 length works well array u a , dtype= <U1 ) >>> pickle.loads pickle.dumps np.array u ))) 0 length does not work TypeError Traceback most recent call last) tmp tables 2.2 <ipython console> in <module> ) usr local lib python2.6 pickle.pyc in loads str) 1372 def loads str) 1373 file = StringIO str) > 1374 return Unpickler file).load ) 1375 1376 Doctest usr local lib python2.6 pickle.pyc in load self) 856 while 1 857 key = read 1) > 858 dispatch[key] self) 859 except Stop, stopinst 860 return stopinst.value usr local lib python2.6 pickle.pyc in load reduce self) 1131 args = stack.pop ) 1132 func = stack[ 1] > 1133 value = func args) 1134 stack[ 1] = value 1135 dispatch[REDUCE] = load reduce TypeError data type not understood","bug"
|
|
"6013","2255","docstring for spacing has a mistake.
|
|
Trac 1659).
|
|
Original ticket http projects.scipy.org numpy ticket 1659 on 2010 11 02 by WarrenWeckesser, assigned to unknown.
|
|
The example in the docstring for the spacing function has a mistake; the line >>> np.spacing 1, 2) == np.finfo np.float64).eps should be >>> np.spacing 1) == np.finfo np.float64).eps","bug"
|
|
"6013","2256","2 32 byte tofile ) fromfile ) limit in 64 bit Windows Trac 1660).
|
|
Original ticket http projects.scipy.org numpy ticket 1660 on 2010 11 03 by trac user mspacek, assigned to charris.
|
|
I m using Christoph Gohlke s amd64 builds http www.lfd.uci.edu gohlke pythonlibs numpy) on 64 bit Windows 7, with the official Python 2.6.6 amd64 install.
|
|
I can t tofile ) or save ) or fromfile ) numpy arrays roughly > 2 32 bytes.
|
|
For save ) and tofile ) Python hangs, using 100 CPU on one core.
|
|
For fromfile ), it raises an IOError.
|
|
This doesn t happen in 64 bit Linux on the same machine.
|
|
load ) seems to work on any size of file.
|
|
Here are some examples, and some caveats if .tofile ) hangs, it hangs immediately np.zeros 2 32 2 16, np.int8).tofile test.npy ) hangs, file is 2 16 bytes np.zeros 2 32 2 2, np.int8).tofile test.npy ) hangs, file is 0 bytes np.zeros 2 32 2, np.int8).tofile test.npy ) hangs, file is 0 bytes np.zeros 2 32 1, np.int8).tofile test.npy ) hangs, file is 0 bytes np.zeros 2 32, np.int8).tofile test.npy ) hangs, file is 0 bytes np.zeros 2 32 1, np.int8).tofile test.npy ) works if np.save ) hangs, it takes a while for it to hang np.save test.npy , np.zeros 2 33, np.int8)) hangs, file is 2 32 bytes, npy header is intact np.save test.npy , np.zeros 2 33 1, np.int8)) hangs, file is 2 32 bytes, npy header is intact np.save test.npy , np.zeros 2 32 2 31, np.int8)) hangs, file is 2 31 bytes, npy header is intact np.save test.npy , np.zeros 2 32 2 16, np.int8)) hangs, file is 2 16 bytes, npy header is intact np.save test.npy , np.zeros 2 32 2 12, np.int8)) hangs, file is 2 12 bytes, npy header is intact np.save test.npy , np.zeros 2 32 2 11, np.int8)) works np.save test.npy , np.zeros 2 32 2 10, np.int8)) works np.save test.npy , np.zeros 2 32 2 8, np.int8)) works np.save test.npy , np.zeros 2 32 2 4, np.int8)) works np.save test.npy , np.zeros 2 32 2 2, np.int8)) works np.save test.npy , np.zeros 2 32 2, np.int8)) works np.save test.npy , np.zeros 2 32 1, np.int8)) works np.save test.npy , np.zeros 2 32, np.int8)) works I also generated these 3 large files successfully in 64 bit Linux... np.zeros 2 33, np.int8).tofile test1.npy ) 8 GB np.zeros 2 32, np.int8).tofile test2.npy ) 4 GB np.save test3.npy , np.zeros 2 33, np.int8)) 8 GB ...and then tried reading them back in 64 bit Win7 a = np.fromfile test1.npy , dtype=np.int8) IOError could not seek in file a = np.fromfile test2.npy , dtype=np.int8) IOError could not seek in file a = np.load test3.npy ) strangely enough, this works fine, even though it s 8 GB!
|
|
I have this problem regardless of which of Christoph s amd64 numpy builds 1.4.1, 1.5.0, 1.5.1RC1, MKL or non MKL) I use.
|
|
This happens on both my 64 bit Win7 i7 12 GB machine, and on another machine running 64 bit WinXP.
|
|
Christoph has also confirmed this on Python 2.5 through 3.1 and says there is nothing special about his builds, and that this will affect all 64 bit Windows distributions of numpy EPD, ActiveState, etc).","bug"
|
|
"6013","2257","rint 0.5) = 0 ?
|
|
why?
|
|
Trac 1661).
|
|
Original ticket http projects.scipy.org numpy ticket 1661 on 2010 11 03 by trac user fegdri, assigned to unknown.
|
|
What is wrong with rint, round, around functions?
|
|
>>> a = np.array [ 2.5, 1.5, 1.4, 0.5, 0.5, 1.5, 2.5, 3.4]) >>> np.rint a) array [ 2., 2., 1., 0., 0., 2., 2., 3.])
|
|
>>> np.around a) array [ 2., 2., 1., 0., 0., 2., 2., 3.])
|
|
>>> np.round a) array [ 2., 2., 1., 0., 0., 2., 2., 3.])
|
|
>>> np.round a) array [ 2., 2., 1., 0., 0., 2., 2., 3.])
|
|
How can it be that 2 is the nearest number to both 2.5 and 1.5?
|
|
I was expecting to see array [ 3., 2., 1., 1., 1., 2., 3., 3.])
|
|
Version 1.5.1rc1 win32, python2.7)","bug"
|
|
"6013","2258","nonzero ) requires swap for float types Trac 1662).
|
|
Original ticket http projects.scipy.org numpy ticket 1662 on 2010 11 05 by mwiebe, assigned to unknown.
|
|
The current floating point nonzero ) implementation is incorrect for non native byte order.
|
|
Swapping the order should have no effect, but it can change the result as visible here >>> import numpy as np >>> a = np.array [0x80000000, 0x00000080, 0], dtype=np.uint32) >>> a.dtype = np.float32 >>> print a.nonzero ) array [1]),) >>> print a.byteswap ).newbyteorder ).nonzero ) array [0]),)","bug"
|
|
"6013","2259","Cannot import numpy Trac 1663).
|
|
Original ticket http projects.scipy.org numpy ticket 1663 on 2010 11 05 by trac user mstewart, assigned to unknown.
|
|
I use macports to update python and numpy software.
|
|
I recently accomplished a selfupdate and deliberately executed an install of py27 numpy which appeared to work.
|
|
I still cannot import numpy within python >>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> ImportError No module named numpy Following examples in another bug report, I attempted to append to the sys.path >>> sys.path.append opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy ) ...but the same error appeared.","bug"
|
|
"6013","2260","spacing nextafter failure on OS X PPC Trac 1664).
|
|
Original ticket http projects.scipy.org numpy ticket 1664 on 2010 11 08 by rgommers, assigned to unknown.
|
|
Only the longdouble version fails, as reported by Friedrich Romstedt on list a few days ago.
|
|
It should be checked if this can be fixed, if not it can be marked as knownfail like done in the 1.5.x branch in [516aa4a5d1e] http github.com numpy numpy commit 516aa4a5d1e).","bug"
|
|
"6013","2261","genfromtxt returns inconsistent output if column converters are used Trac 1665).
|
|
Original ticket http projects.scipy.org numpy ticket 1665 on 2010 11 08 by trac user dmoore, assigned to pierregm.
|
|
Disclaimer This was tested on 1.4.1 I don t know if it was addressed in 1.5 I suspect not based on mailing list correspondence).
|
|
Simple example of the issue >>> import numpy, StringIO >>> >>> s=StringIO.StringIO q1,2\nq3,4 ) >>> a=numpy.genfromtxt s,delimiter= , ,converters={0 lambda s float s[1 ])}) >>> a array [ 1.0, 2.0), 3.0, 4.0)], dtype=[ f0 , |O4 ), f1 , <f8 )]) Notice that this is a 1d array of tuples.
|
|
Since the data are effectively the same type, it would have been preferrable to receive 2d output.
|
|
Note that specifying the dtype does not help >>> s=StringIO.StringIO q1,2\nq3,4 ) >>> a=numpy.genfromtxt s,delimiter= , ,converters={0 lambda s float s[1 ])},dtype=float) >>> a array [ 1.0, 2.0), 3.0, 4.0)], dtype=[ f0 , |O4 ), f1 , <f8 )]) dtype appears to be ignored.
|
|
Strangely, however, if the data column could be interpreted without the converter, then the output is produced as desired even if the converter is used >>> s=StringIO.StringIO 1,2\n3,4 ) >>> a=numpy.genfromtxt s,delimiter= , ,converters={0 lambda s float s)}) >>> a array [[ 1., 2.
|
|
], [ 3., 4.]])","bug"
|
|
"6013","2262","Reducing the memory usage of MaskedArrays after DomainedBinaryOperation Trac 1666).
|
|
Original ticket http projects.scipy.org numpy ticket 1666 on 2010 11 08 by trac user lee, assigned to pierregm.
|
|
I noticed that when I create a new masked array with no mask yet) I am given a mask=False.
|
|
After dividing this mask by a nonzero number, the mask becomes the same shape as the array, populated entirely with False .
|
|
I would like to not increase my memory usage unnecessarily after safe operations.
|
|
I attempted to patch the ma core.py file to check the mask If all entries are False, the mask is not applied to the result.
|
|
The patch works for divide, but not for mod ).
|
|
It doesn t seem like mod passes through DomainedBinaryOperation.
|
|
call even though it is decorated as such in ma core.py .
|
|
Is there another location that overrides this mod operation that I can should patch?
|
|
Is this patch even useful or a good idea?
|
|
Chris","enhancement"
|
|
"6013","2263","Provide a context manager for seterr Trac 1667).
|
|
Original ticket http projects.scipy.org numpy ticket 1667 on 2010 11 08 by WarrenWeckesser, assigned to unknown.
|
|
It would be great if there were a context manager for the error settings.
|
|
Something as simple as this import numpy class ErrorSettings object) def enter self, kwds) self.previous = numpy.seterr kwds) def exit self, type, value, traceback) numpy.seterr self.previous) would allow code such as this with ErrorSettings invalid= ignore ) y = sin x) x instead of oldsettings = np.seterr invalid= ignore ) y = sin x) x np.seterr oldsettings)","enhancement"
|
|
"6013","2264","Masked array calculations memory leak Trac 1668).
|
|
Original ticket http projects.scipy.org numpy ticket 1668 on 2010 11 09 by trac user storpipfugl, assigned to pierregm.
|
|
Test code [[BR]] [[BR]] mem leak ma.py[[BR]] Description [[BR]] Use mem leak ma.py.
|
|
At the raw input check mem ) line the array data array and masked array cal data array are the only arrays in scope.
|
|
Base memory consumption before calculation is 95 MB.
|
|
Expected memory consumption at end of test program 3712 2 8 2 1) 10 6 95 = 329 MB.
|
|
Readout from proc <pid> status at line raw input check mem ) VmPeak 848092 kB VmSize 484768 kB This leaves 156 MB unaccounted for VmSize).
|
|
Expected behavior The memory usage at the end of the test program is expected to be comparable to the corresponding example with a regular ndarray plus memory usage for the mask.
|
|
Working code [[BR]] [[BR]] mem leak.py[[BR]] Use mem leak.py.
|
|
Readout from proc <pid> status at line raw input check mem ) VmPeak 417416 kB VmSize 309768 kB as expected VmSize).
|
|
System [[BR]] Ubuntu linux 10.04 64 bit[[BR]] Python 2.6.5[[BR]] Reproduced with numpy 1.3.0 and 1.5.0 [[BR]]","bug"
|
|
"6013","2265","Numpy for symbian s60 Trac 1669).
|
|
Original ticket http projects.scipy.org numpy ticket 1669 on 2010 11 10 by trac user jimver, assigned to unknown.
|
|
Hi, I am making an application in python and I want to install it to my mobile phone that runs Symbian S60 3rd gen.
|
|
The hello world application works perfect, but when I am importing numpy and install the application I get the message in the mobile screen that numpy is not found!
|
|
I there any package like numpy.sis that I can install to the phone?
|
|
I am using Windows 7 Python 2.5.4 Python for S60 Python 2 0 0.sis and PIPS.sis Numpy 1.3.0 The Python and numpy versions are not the latests, but they are the latests that work with Python for S60 .sis file at this moment 10Nov10).
|
|
BR, Dimitrios Ververidis dimitriosververidis.blogspot.com","bug"
|
|
"6013","2266","type casting doesn t raise floating point overflow or underflow Trac 1670).
|
|
Original ticket http projects.scipy.org numpy ticket 1670 on 2010 11 11 by mwiebe, assigned to unknown.
|
|
This leads to some surprising results, observe >>> import numpy as np >>> err = np.seterr all= raise ) >>> a = np.arange 5, dtype= f4 ) >>> a 1e38 Traceback most recent call last) File <stdin> , line 1, in <module> FloatingPointError underflow encountered in divide >>> a 1e40 array [ 0., 0., 0., 0., 0.
|
|
], dtype=float32) >>> np.float32 1e38) 9.9999997e 37 >>> np.float32 1e40) inf >>> a np.inf array [ 0., 0., 0., 0., 0.
|
|
], dtype=float32) I would expect a 1e40 to raise an overflow when 1e40 overflows to inf) instead of silently slipping through and causing underflows to zero.","bug"
|
|
"6013","2267","scalars don t raise overflow underflow divbyzero invalid Trac 1671).
|
|
Original ticket http projects.scipy.org numpy ticket 1671 on 2010 11 11 by mwiebe, assigned to unknown.
|
|
All these cases should be raising, I believe >>> import numpy as np >>> err = np.seterr all= raise ) >>> np.float32 1e 38) np.float32 1e30) 0.0 >>> np.float32 1e 38) np.array 1e38,np.float32) 0.0 >>> np.array 1e 38,np.float32) np.float32 1e30) Traceback most recent call last) File <stdin> , line 1, in <module> FloatingPointError underflow encountered in divide >>> np.float64 1e 300) np.float64 1e300) 0.0 >>> np.float64 1e 300) np.array 1e300, dtype=np.float64) 0.0 >>> np.array 1e 300, dtype=np.float64) np.float64 1e300) Traceback most recent call last) File <stdin> , line 1, in <module> FloatingPointError underflow encountered in divide >>> np.float32 1) np.float32 0) inf >>> np.array 1,dtype=np.float32) np.float32 0) Traceback most recent call last) File <stdin> , line 1, in <module> FloatingPointError divide by zero encountered in divide >>> np.inf np.inf nan >>> np.array np.inf) np.inf Traceback most recent call last) File <stdin> , line 1, in <module> FloatingPointError invalid value encountered in subtract","bug"
|
|
"6013","2268","numpy.allclose raises IndexError on scalar inf Trac 1672).
|
|
Original ticket http projects.scipy.org numpy ticket 1672 on 2010 11 12 by trac user lamblin, assigned to unknown.
|
|
>>> numpy.allclose inf, inf), with inf = float inf ), raises the following exception File usr lib python2.6 dist packages numpy core numeric.py , line 1671, in allclose if not all x[xinf] == y[xinf]) IndexError 0 d arrays can t be indexed The expected behavior is the one of 1 d arrays >>> numpy.allclose [inf], [inf]) True The following work as expected numpy.allclose inf, 2) False numpy.allclose inf, [inf, 2]) False numpy.allclose [inf], [inf]) True The following don t, and raise the mentionned exception numpy.allclose inf, inf) numpy.allclose inf, [inf]) numpy.allclose inf, [inf, inf]) I m using version 1.4.1, but the code does not seem to have change in current git repository.","bug"
|
|
"6013","2269","first nonzero element Trac 1673).
|
|
Original ticket http projects.scipy.org numpy ticket 1673 on 2010 11 13 by trac user tom3118, assigned to unknown.
|
|
The numpy for matlab users suggests using `nonzero A)[0][0]` to find the index of the first nonzero element of array A.
|
|
The problem with this is that A might be a million elements long and the first element might be zero.
|
|
This is an extremely common operation.
|
|
An efficient, built in method for this would be very useful.
|
|
It also would easy people s transition from Matlab in which `find` is so common.","enhancement"
|
|
"6013","2270","Problems outputting elements of ndarray with numeric formatting Trac 1674).
|
|
Original ticket http projects.scipy.org numpy ticket 1674 on 2010 11 15 by trac user pv00, assigned to unknown.
|
|
Hi, what is the best way to print to a file or to stdout) formatted numerical values?
|
|
Numpy Documentation only discusses input from a file, or output of entire arrays.
|
|
I just want tab or space delimited output of formatted values.
|
|
I tried to follow python documentation and find errors.
|
|
Below is ipython pylab transcript In [2] import numpy as np In [3] w = np.arange 1,5,dtype=np.int32).reshape 2,2)) In [4] w Out[4] array [[1, 2], [3, 4]], dtype=int32) In [5] w[0,0] Out[5] 1 In [6] w[0,0].
|
|
class Out[6] <type numpy.int32 > In [7] print {0 2d} .format w[0,0])) ValueError Traceback most recent call last) home p o dev thesis python wavelet <ipython console> in <module> ) ValueError Unknown format code d for object of type str In [8]","bug"
|
|
"6013","2271","No support for ` format ` in scalars arrays Trac 1675).
|
|
Original ticket http projects.scipy.org numpy ticket 1675 on 2010 11 15 by pv, assigned to unknown.
|
|
Numpy scalars and arrays don t work properly in the `format )` string formatting context >>> {0 d} .format np.int16 1)) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError Unknown format code d for object of type str The reason seems to be that the ` format ` method is inherited from `object`, and it converts the item to string via `str` first.
|
|
The correct fix seems to be to implement a properly formatting code aware ` format ` ourselves.
|
|
For most types, this can probably be done directly by calling PyInt s or PyFloat s routine.
|
|
However, for e.g.
|
|
long doubles we might want to implement the formatting ourselves.","bug"
|
|
"6013","2272","dtypes error in recfunctions Trac 1676).
|
|
Original ticket http projects.scipy.org numpy ticket 1676 on 2010 11 16 by WeatherGod, assigned to unknown.
|
|
I was using append fields ) in numpy.lib.recfunctions when I discovered a slight logic mistake in handling the dtypes argument.
|
|
The code first checks to see if dtypes is None.
|
|
If so, it then guesses the dtype info from the input data.
|
|
Then, it goes to see if the dtypes is not a sequence and puts it into a list.
|
|
This is where the logic breaks down.
|
|
It then proceeds to use that dtypes to start merging the data.
|
|
However, if you pass in a sequence of dtype, that condition gets skipped and the rest of the function has to figure out the dtype and ignores the supplied names as well).
|
|
I have attached a patch to fix this.
|
|
I did not check the rest of the module to see if this mis logic crops up anywhere else.","bug"
|
|
"6013","2273","add half float16 data type Trac 1677).
|
|
Original ticket http projects.scipy.org numpy ticket 1677 on 2010 11 18 by mwiebe, assigned to unknown.
|
|
It would be useful to have the float16 data type supported in NumPy.","enhancement"
|
|
"6013","2274","problem in the poisson distribution Trac 1678).
|
|
Original ticket http projects.scipy.org numpy ticket 1678 on 2010 11 18 by trac user gruel, assigned to unknown.
|
|
I tried to use the poissonian distribution but I do think that there are a small problem with it.
|
|
AS I am not a statician I am not sure but the plot attach seems a little bit strange.
|
|
It has been created pylab hist rand.poisson 100, 10000), 100) The regular empty space are not normal.
|
|
The problem is present in 1.5 and also in 2.0.0.dev8597","bug"
|
|
"6013","2275","f2py throws AssertionError on fortran 90 files which declare routines with C bindings Trac 1679).
|
|
Original ticket http projects.scipy.org numpy ticket 1679 on 2010 11 19 by trac user davidpowell, assigned to pearu.
|
|
When running f2py on a fortran 90 file with the following contents interface function my func x) bind c,name= my func ) use iso c binding real c double), intent in) x real c double) my func end function end interface It causes an exception in f2py, ending with the following line File C \Python26\lib\site packages\numpy\f2py\crackfortran.py , line 665, in markoutercomma assert not f,`f,line,l,cc` AssertionError 1, x ) bind c,name= my func , x ) bind c,name= my func , ) ) As far as I understand this is just a bug in the parser, since f2py shouldn t care that my fortran code is internally calling C routines.","bug"
|
|
"6013","2276","constructing an array out of the wrong type fails with SystemError Trac 1680).
|
|
Original ticket http projects.scipy.org numpy ticket 1680 on 2010 11 19 by trac user exarkun, assigned to unknown.
|
|
Python 2.6.4 r264 75706, Dec 7 2009, 18 45 15) [GCC 4.4.1] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> class foo pass ... >>> import numpy >>> numpy.array [foo )], f ) Traceback most recent call last) File <stdin> , line 1, in <module> SystemError error return without exception set >>>","bug"
|
|
"6013","2277","Explain arange floating point issues in docstring Trac 1681).
|
|
Original ticket http projects.scipy.org numpy ticket 1681 on 2010 11 19 by trac user hongdog, assigned to pv.
|
|
Using spyder with python 2.6.2 and numpy 1.3.0 on Windows XP, I found the following problem.
|
|
In [31] arange 12.5,13,0.1) Out[31] array [ 12.5, 12.6, 12.7, 12.8, 12.9]) 13.0 is not here In [32] arange 12.6,13,0.1) Out[32] array [ 12.6, 12.7, 12.8, 12.9, 13. ])
|
|
13.0 is here As shown, the end 13.0 is not shown in 1st array, which is correct based on the document, but shown in 2nd array.","bug"
|
|
"6013","2278","ctypeslib.load library isn t python 3 compatible.
|
|
Trac 1682).
|
|
Original ticket http projects.scipy.org numpy ticket 1682 on 2010 11 20 by drnlm, assigned to unknown.
|
|
PEP 3110 changes the scoping of exceptions in python 3. ctypeslib.load library assumes the old python 2 scoping, and will throw an UnboundLocalError rather than properly re raising the ctypes exception as a result.
|
|
The attached patch fixes this.","bug"
|
|
"6013","2279","numpy does not correctly support sys.getsizeof Trac 1683).
|
|
Original ticket http projects.scipy.org numpy ticket 1683 on 2010 11 22 by trac user ghazel, assigned to unknown.
|
|
>>> sys.getsizeof numpy.array [])) 80 >>> sys.getsizeof numpy.array range 100000))) 80","enhancement"
|
|
"6013","2280","numpy.allclose does not deal well with the minimal value of int dtypes Trac 1684).
|
|
Original ticket http projects.scipy.org numpy ticket 1684 on 2010 11 22 by trac user lamblin, assigned to unknown.
|
|
If I create an array containing the minimal value of an int dtype, for instance an int8 array containing 128, numpy.allclose fails to detect it is close to itself, although it is equal.
|
|
>>> a = numpy.asarray 128, dtype= int8 ) >>> numpy.all a == a) True >>> numpy.allclose a, a) False I would expect >>> numpy.allclose a, a) True This is due to the fact that the tolerance uses numpy.absolute, and in that case, it is negative >>> numpy.absolute a) array [ 128], dtype=int8) I reproduced the same with int16 and 2 15, int32 and 2 31, and int64 and 2 63.","bug"
|
|
"6013","2281","In savetxt check for write and not for unused) seek Trac 1685).
|
|
Original ticket http projects.scipy.org numpy ticket 1685 on 2010 11 24 by trac user naufraghi, assigned to unknown.
|
|
savetxt checks if the first argument is a filename or a filelike object, but it checks for a seek attr, even if in the code below only write and close are used.","bug"
|
|
"6013","2282","relative imports for python 3 Trac 1686).
|
|
Original ticket http projects.scipy.org numpy ticket 1686 on 2010 11 25 by trac user sunqiang, assigned to unknown.
|
|
http docs.python.org dev py3k whatsnew 3.0.html removed syntax said The only acceptable syntax for relative imports is from .
|
|
[module] import name.
|
|
All import forms not starting with .
|
|
are interpreted as absolute imports.
|
|
PEP 0328) line 128 and 129 of numpy init .py from version import git revision as git revision from version import version as version they didn t follow PEP 0328, so it breaks Python 3.x port.
|
|
got this error from Python 3.2a4 and NumPy s git.","bug"
|
|
"6013","2283","Example in documentation on default iteration in incorrect Trac 1687).
|
|
Original ticket http projects.scipy.org numpy ticket 1687 on 2010 11 27 by gerritholl, assigned to pv.
|
|
According to http docs.scipy.org doc numpy reference arrays.classes.html default iteration The default behavior [of for val in arr ] is equivalent to for i in arr.shape[0] val = arr[i] This cannot be correct, because arr.shape[0] is an int and not iterable.
|
|
Correct is for i in xrange arr.shape[0]) val = arr[i]","bug"
|
|
"6013","2284","MaskedArray.
|
|
str fails with ValueError if data are record array Trac 1688).
|
|
Original ticket http projects.scipy.org numpy ticket 1688 on 2010 11 29 by gerritholl, assigned to pierregm.
|
|
When a MaskedArray s data are a record array,, string representation fails with a ValueError as the mask does not contain the mapping of field names that the data does.
|
|
It appears the records are not carried on to the mask.
|
|
python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> Z = np.empty 5,), dtype=[ A , np.float32), B , np.uint32)]) >>> M = np.ma.masked where np.array [True, False, True, False, True]), Z) >>> print M Traceback most recent call last) File <stdin> , line 1, in <module> File storage4 home gerrit .local lib python2.6 site packages numpy ma core.py , line 3558, in str recursive printoption res, m, f) File storage4 home gerrit .local lib python2.6 site packages numpy ma core.py , line 2292, in recursive printoption curdata, curmask) = result[name], mask[name]) ValueError field named A not found.
|
|
>>> print np.version.version 2.0.0.dev 12d0200","bug"
|
|
"6013","2285","fromstring segfaults if reading from text and dtype=bool Trac 1689).
|
|
Original ticket http projects.scipy.org numpy ticket 1689 on 2010 11 30 by gerritholl, assigned to charris.
|
|
The code says it all python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import fromstring, bool , version >>> print version.version 2.0.0.dev 12d0200 >>> fromstring 1 1 , sep= , dtype=bool ) Segmentation fault","bug"
|
|
"6013","2286","numpy.random.randint 2147483648, 2147483647) raises ValueError low >= high Trac 1690).
|
|
Original ticket http projects.scipy.org numpy ticket 1690 on 2010 11 30 by trac user tkf, assigned to charris.
|
|
Following example raises ValueError low >= high .
|
|
It looks like it is due to overflow of diff = hi lo 1 in RandomState.randint.
|
|
import numpy intinfo = numpy.iinfo int) numpy.random.randint 1, intinfo.max) raises nothing numpy.random.randint intinfo.min, intinfo.max) numpy.random.randint 2, intinfo.max) raises ValueError too","bug"
|
|
"6013","2287","PyArray FromArray Properly handle F CONTIGUOUS with 1D arrays Trac 1691).
|
|
Original ticket http projects.scipy.org numpy ticket 1691 on 2010 12 01 by trac user dagss, assigned to unknown.
|
|
When passing F CONTIGUOUS to PyArray FromArray, one is NOT guaranteed an F CONTIGUOUS array if ndim==1 in that case one would have to pass C CONTIGUOUS instead).
|
|
I find this highly confusing and hope my patch is accepted.
|
|
I could not find a way to exercise this from Python, hence no test case the reason is that np.asfortranarray etc.
|
|
seems to always copy, for some reason...perhaps that should be fixed too).
|
|
See a Cython test case that gets fixed by this patch inline below.
|
|
Patch https github.com dagss numpy commit c5f273898ccd4b4bd45715fdc54f629f6ee0532a >>> a = np.zeros 10)[ 2] >>> a = as f contiguous 1d a) >>> a.flags.f contiguous True cimport numpy as np import numpy as np np.import array ) def as f contiguous 1d arr) return np.PyArray FROMANY arr, np.NPY DOUBLE, 1, 1, np.NPY F CONTIGUOUS)","bug"
|
|
"6013","2288","numpy.array segfaults on binary image from Python Imaging Library Trac 1692).
|
|
Original ticket http projects.scipy.org numpy ticket 1692 on 2010 12 03 by gerritholl, assigned to unknown.
|
|
When I try to convert a binary image loaded with the Python Imaging Library version 1.1.7) using numpy version 2.0.0.dev 12d0200), numpy segfaults.
|
|
If I convert the image to 8 bit first, numpy can load it fine.
|
|
14 33 35 14414 gerrit ubuntu10 python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import Image, numpy >>> print Image.VERSION, numpy.version.version 1.1.7 2.0.0.dev 12d0200 >>> im = Image.open local gerrit benchmarking images lsm.png ) >>> print im <PngImagePlugin.PngImageFile image mode=1 size=21601x10801 at 0x1B7BD40> >>> imd = numpy.array im) Segmentation fault If I convert the image to 8 bit first, there is no problem 14 34 14 14414 gerrit ubuntu10 python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import Image, numpy >>> im = Image.open local gerrit benchmarking images lsm.png ) >>> im = im.convert L ) convert to 8 bit >>> print im <Image.Image image mode=L size=21601x10801 at 0x2C63488> >>> imd = numpy.array im) >>> print imd [[ 0 0 0 ..., 0 0 0] [ 0 0 0 ..., 0 0 0] [ 0 0 0 ..., 0 0 0] ..., [255 255 255 ..., 255 255 255] [255 255 255 ..., 255 255 255] [255 255 255 ..., 255 255 255]] If I load a BMP instead of a PNG also written by PIL), the same segfault occurs 14 38 44 14416 gerrit ubuntu10 python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import Image, numpy >>> im = Image.open local gerrit benchmarking images lsm.bmp ) >>> print im <BmpImagePlugin.BmpImageFile image mode=1 size=21601x10801 at 0xDD5D40> >>> numpy.array im) Segmentation fault I have attached lsm.png 784 KiB).","bug"
|
|
"6013","2289","Puzzling difference in building wrapper function Trac 1693).
|
|
Original ticket http projects.scipy.org numpy ticket 1693 on 2010 12 03 by pearu, assigned to pearu.
|
|
http cens.ioc.ee pipermail f2py users 2010 December 002145.html","bug"
|
|
"6013","2290","Adding support to an ARM processor Trac 1694).
|
|
Original ticket http projects.scipy.org numpy ticket 1694 on 2010 12 05 by trac user Snark, assigned to cournape.
|
|
I m trying to compile sage on a toshiba AC100 this device is a netbook with an ARM processor.
|
|
The errors are numpy core include numpy npy cpu.h 49 error error Unknown CPU, please report this to numpy maintainers with information about your platform OS, CPU and compiler) Ok, that s what I ll do.
|
|
The machine is arm linux gnueabi, using gcc 4.4.5. numpy core include numpy npy endian.h 33 error error Unknown CPU can not set endianness That one is strange there is a nice endian.h which numpy could try to detect...","bug"
|
|
"6013","2291","numpy.r [True, False] is not a boolean array Trac 1695).
|
|
Original ticket http projects.scipy.org numpy ticket 1695 on 2010 12 06 by trac user O.C., assigned to unknown.
|
|
Hello, I observe the following behavior numpy.r [True, False] > array [1, 0], dtype=int8) numpy.r [True] > array [ True], dtype=bool) I would expect the first line to give a boolean array array [ True, False], dtype=bool) This may be a bug.
|
|
O.C.
|
|
numpy.
|
|
version = 1.4.1","bug"
|
|
"6013","2292","Nonexistant savez compressed Trac 1696).
|
|
Original ticket http projects.scipy.org numpy ticket 1696 on 2010 12 11 by trac user ipatrol, assigned to pv.
|
|
savez compressed is referenced on the generated savez page.
|
|
Howeverm this simply does not exist.
|
|
Why is it here?","bug"
|
|
"6013","2293","typo in doc of convolve Trac 1697).
|
|
Original ticket http projects.scipy.org numpy ticket 1697 on 2010 12 16 by trac user buschi, assigned to pv.
|
|
.. math f g)[n] = \sum {m = \infty} {\infty} f[m] f[n m] should read .. math f g)[n] = \sum {m = \infty} {\infty} f[m] g[n m]","bug"
|
|
"6013","2294","NumPy should classify itself as supporting Python 3 Trac 1698).
|
|
Original ticket http projects.scipy.org numpy ticket 1698 on 2010 12 20 by trac user loewis, assigned to unknown.
|
|
PyPI has now Trove classifiers to indicate what Python versions are supported.
|
|
It would be good if NumPy would classify itself with these classifiers, in particular Programming Language Python 3 Doing so allows automatic checks whether a package has been ported.","bug"
|
|
"6013","2295","Failure assigning ctypes arrays to numpy arrays of size 1 on Py3 Trac 1699).
|
|
Original ticket http projects.scipy.org numpy ticket 1699 on 2010 12 20 by trac user ulfw, assigned to unknown.
|
|
I have the following problem When I create an array of size 1 using numpy.zeros or numpy.array, then assign a slice using [ ] notation, it fails, possibly when the right hand side is used as a buffer.
|
|
The behavior is inconsistent with earlier NumPy versions.
|
|
The attached example illustrates the problem.
|
|
NOTE This was observed in NumPy 1.5.1.","bug"
|
|
"6013","2296","Registry check fails for MS Install.
|
|
Trac 1700).
|
|
Original ticket http projects.scipy.org numpy ticket 1700 on 2010 12 24 by trac user johnmark54, assigned to rgommers.
|
|
After installing python binary python 2.7.1.amd64.msi from python.org the install of numpy 1.5.0 win32 superpack python2.7.exe fails with error Python 2.7 required, which was not found in the registry.
|
|
Reinstalling python, using different versions of python etc.
|
|
does not help.
|
|
Python runs fine.
|
|
Is this the right place to report this?)
|
|
john.m.agosta intel.com","bug"
|
|
"6013","2297","Accessing Type Conversion Promotion Coercion Rules Trac 1701).
|
|
Original ticket http projects.scipy.org numpy ticket 1701 on 2010 12 25 by trac user pv00, assigned to unknown.
|
|
NumPy dtype arithmetic is the opposite of Python type arithmetic!
|
|
The operators and and or switch roles <type int > or <type float > = <type int > <type int > and <type float > = <type float > int64 or float64 = float64 int64 and float64 = int64 This will be very confusing to users at large.
|
|
Can we make the conventions agree in NumPy 2.0?
|
|
More details I was happy to find that NumPy exposed the numerical datatype promotion rules such as int float = float by means of the logical operator or dtype int64 ), or , dtype float64 ), = , dtype float64 ) But then I was shocked to find that Python had the exact opposite convention !
|
|
I propose that there be a negotiation between the NumPy developers and the Python developers, and that a consistent convention be set.
|
|
Right now I think it s undocumented or not very publically documented) behavior, and it would be better to make a change earlier, and fix a relatively small base of code, rather than carry it along forever, a potential source of confusion when code is written and whenever it is debugged.","bug"
|
|
"6013","2298","dtype truth value is False Trac 1702).
|
|
Original ticket http projects.scipy.org numpy ticket 1702 on 2010 12 26 by pv, assigned to unknown.
|
|
This probably shouldn t be like this >>> bool np.dtype float64 )) False Reported here http permalink.gmane.org gmane.comp.python.numeric.general 41827","bug"
|
|
"6013","2299","fromfile doesn t work with dtype=bool and input being floats Trac 1703).
|
|
Original ticket http projects.scipy.org numpy ticket 1703 on 2010 12 29 by trac user jpeel, assigned to unknown.
|
|
Currently, fromfile doesn t work properly for dtype=bool and with floats in the input file.
|
|
I ve attached a patch to fix it as well as a test.","bug"
|
|
"6013","2300","Patch for utils.py lookfor ) Trac 1704).
|
|
Original ticket http projects.scipy.org numpy ticket 1704 on 2011 01 03 by sebhaase, assigned to unknown.
|
|
lookfor chokes on SWIG globals vars e.g.
|
|
wx module An easy patch is numpy lib utils.py 928,933c928 < try < item name = getattr v, name , s. s name, n)) < except NameError < ref.
|
|
SWIG s global cvars < NameError Unknown C global variable < item name = s. s name, n) > item name = getattr v, name , s. s name, n))","enhancement"
|
|
"6013","2301","Unicode breaks numpy.correlate patch attached) Trac 1705).
|
|
Original ticket http projects.scipy.org numpy ticket 1705 on 2011 01 06 by trac user pyrtsa, assigned to unknown.
|
|
`numpy.correlate` doesn t play nicely with `from future import unicode literals` nor generally a `unicode` string provided as the `mode` argument).
|
|
The following call will result in a `TypeError` >>> numpy.correlate [1,2,3],[1,2,3],unicode full )) TypeError Traceback most recent call last) Users me <ipython console> in <module> ) path to numpy core numeric.pyc in correlate a, v, mode, old behavior) 692 return multiarray.correlate a,v,mode) 693 else > 694 return multiarray.correlate2 a,v,mode) 695 696 def convolve a,v,mode= full ) TypeError an integer is required This patch should fix the problem Python 2.3 and above) diff git a numpy core numeric.py b numpy core numeric.py index dbad73c..519180b 100644 a numpy core numeric.py b numpy core numeric.py 636,7 636,7 mode from name dict = { v 0, f 2} def mode from name mode) if isinstance mode, type )) if isinstance mode, basestring) return mode from name dict[mode.lower )[0]] return mode","bug"
|
|
"6013","2302","Misleading help options Trac 1706).
|
|
Original ticket http projects.scipy.org numpy ticket 1706 on 2011 01 10 by trac user hector1618, assigned to unknown.
|
|
I am a newbie and was using numpy in IPython.
|
|
When I create a matrix object,say A, and look for options available, it shows some options which are not valid for that object like A.shape ),A.size ) in this case.
|
|
This is really mis leading for newbie.
|
|
The code for the same is In[1] import numpy as np In[2] A = np.matrix [[1,2],[3,4]]) In[3] A.shape )","bug"
|
|
"6013","2303","ctypeslib.ndpointer cache ineffective Trac 1707).
|
|
Original ticket http projects.scipy.org numpy ticket 1707 on 2011 01 10 by trac user Colin Hogben, assigned to unknown.
|
|
The cache of pointer types is queried with a tuple dtype, ndim, shape, num) but values stored with a bare dtype as key.
|
|
Thus queries always miss.","bug"
|
|
"6013","2304","pickle.loads ) of pickled numpy.array can corrupt memory Trac 1708).
|
|
Original ticket http projects.scipy.org numpy ticket 1708 on 2011 01 10 by trac user quarl, assigned to unknown.
|
|
We discovered that pickle.loads ) on a pickled numpy.ndarray can lead to memory corruption.
|
|
I can reproduce the problem with arrays of size 1 length=1, dtype=bool or byte) pickled using protocols 1 or 2 pickle.loads pickle.dumps numpy.array [True]), protocol= 1)) In my tests, this affects Solaris and Linux, Sparc, x86, and x86 64, Python versions up to 2.6.6, numpy version 1.4.1.
|
|
After unpickling the size 1 array into data , I replace data[0] with an arbitrary byte 0xbb in the example below) and it overwrites some unrelated data regexp buffers in the example below).
|
|
If the original data contains array [2]) instead of array [1]), then it affects \x02 in the regexp buffer instead of \x01 .
|
|
The second use of pickle.loads ) then gets confused.
|
|
!
|
|
usr bin env python import numpy import pickle import re Unpickle a bool or byte array of length 1 data = numpy.array [1], dtype= b ) similar issue with numpy.array [True]) blob = pickle.dumps data, protocol=1) protocol 1, 1 exhibit bug, but 0 is safe data = pickle.loads blob) print repr re.sub a .)
|
|
, \x01\\1 , a )) \x01 [correct] data[0] = 0xbb print repr re.sub a .)
|
|
, \x01\\1 , a )) \xbb [?!]
|
|
lose = pickle.loads blob) crashes","bug"
|
|
"6013","2305","astype should not raise ComplexWarning Trac 1709).
|
|
Original ticket http projects.scipy.org numpy ticket 1709 on 2011 01 11 by rgommers, assigned to unknown.
|
|
An explicit type cast should not raise a warning In [1] x = np.array [1 1j, 1 1j]) In [2] x.astype np.float64) Library Frameworks Python.framework Versions 2.6 bin ipython 1 ComplexWarning Casting complex values to real discards the imaginary part !
|
|
Library Frameworks Python.framework Versions 2.6 Resources Python.app Contents MacOS Python Out[2] array [ 1., 1.])","bug"
|
|
"6013","2306","output dtype of mean, median, var std should not depend on axis argument Trac 1710).
|
|
Original ticket http projects.scipy.org numpy ticket 1710 on 2011 01 12 by bsouthey, assigned to unknown.
|
|
Keith Goodman reported that axis argument determines the output dtype of mean, median, var std [http mail.scipy.org pipermail numpy discussion 2010 December 054253.html] There should be the expectation the output should be the same regardless of the axis argument.
|
|
For example, for these floating point and integer examples the output dtype should either be float32 or float64 >>> np.array [[1, 2], [3, 4]], dtype=np.float32).mean axis=0).dtype dtype float32 ) >>> np.array [[1, 2], [3, 4]], dtype=np.float32).mean axis=None).dtype dtype float64 ) >>> np.array [[1, 2], [3, 4]], dtype=np.int8).mean axis=None, dtype=np.float32).dtype dtype float64 ) >>> np.array [[1, 2], [3, 4]], dtype=np.int8).mean axis=1, dtype=np.float32).dtype dtype float32 )","bug"
|
|
"6013","2307","Bad slice indices missing function resizebuf in numpy.numarray.functions module Trac 1711).
|
|
Original ticket http projects.scipy.org numpy ticket 1711 on 2011 01 13 by trac user phensley, assigned to unknown.
|
|
Issues 1. some slice indices are floats when they should be ints.
|
|
2. the function resizebuf is missing.
|
|
The bug appears to be present in the versions I ve examined 1.3.0, 1.4.1 and 1.5.1.
|
|
I m attaching a patch which fixed the issue for me.
|
|
I ve tested the patch against 1.3.0.
|
|
It also applies clean against 1.4.1, and against 1.5.1 with 3 lines offset, but I haven t personally tested either of these versions.","bug"
|
|
"6013","2308","concatenate ignores axis argument when given 1 dimensional arrays Trac 1712).
|
|
Original ticket http projects.scipy.org numpy ticket 1712 on 2011 01 14 by trac user chairmanK, assigned to unknown.
|
|
I am surprised that concatenate does not complain when I specify an impossible axis.
|
|
>>> a = numpy.zeros 2,)) >>> b = numpy.ones 3,)) >>> numpy.concatenate a, b), axis=0) array [ 0., 0., 1., 1., 1.])
|
|
>>> numpy.concatenate a, b), axis=20) array [ 0., 0., 1., 1., 1.])
|
|
>>> numpy.concatenate a, b), axis= 20) array [ 0., 0., 1., 1., 1.])
|
|
>>> numpy.concatenate a, b), axis=23423432234234234234L) Traceback most recent call last) File <stdin> , line 1, in <module> OverflowError long int too large to convert to int >>> numpy.concatenate a.reshape a.size, 1)), b), axis= 20) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError bad axis1 argument to swapaxes >>> numpy.
|
|
version 2.0.0.dev f72c605","bug"
|
|
"6013","2309","Solaris segfault Trac 1713).
|
|
Original ticket http projects.scipy.org numpy ticket 1713 on 2011 01 15 by rgommers, assigned to unknown.
|
|
Reported on the mailing list on 3 Jan 2011 by John Hunter.
|
|
johnh udesktop253 > gcc version gcc GCC) 3.4.3 csl sol210 3 4 branch sol rpath) Copyright C) 2004 Free Software Foundation, Inc.
|
|
This is free software; see the source for copying conditions.
|
|
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
|
johnh udesktop253 > uname a SunOS udesktop253 5.10 Generic 142910 17 i86pc i386 i86pc johnh udesktop253 > cat test.py import numpy as np print np.
|
|
version fs = 1000 t = np.linspace 0, 0.3, 301) A = np.array [2, 8]).reshape 1, 1) f = np.array [150, 140]).reshape 1, 1) xn = A np.exp 2j np.pi f t)).sum axis=0) johnh udesktop253 > python test.py 2.0.0.dev 9451260 Segmentation Fault core dumped) johnh udesktop253 > johnh udesktop253 > sudo pstack var core core.python.957 core var core core.python.957 of 9397 python test.py febf1928 cexp 0, 0, 0, 0, 8060ab0, 84321ac) 1b0 fe9657e0 npy cexp 80458e0, 0, 0, 0, 0, 84e2530) 30 fe95064f nc exp 8045920, 84e72a0, 8045978, 8045920, 10, 10) 3f fe937d5b PyUFunc D D 84e2530, 84e20f4, 84e25b0, fe950610, 1, 0) 5b fe95e818 PyUFunc GenericFunction 81e96e0, 807deac, 0, 80460b8, 2, 2) 448 fe95fb10 ufunc generic call 81e96e0, 807deac, 0, fe98a820) 70 feeb2d78 PyObject Call 81e96e0, 807deac, 0, 80a24ec, 8061c08, 0) 28 fef11900 PyEval EvalFrame 80a2394, 81645a0, 8079824, 8079824) 146c fef17708 PyEval EvalCodeEx 81645a0, 8079824, 8079824, 0, 0, 0) 620 fef178af PyEval EvalCode 81645a0, 8079824, 8079824, 8061488, fef3d9ee, 0) 2f fef3d095 PyRun FileExFlags feb91c98, 804687b, 101, 8079824, 8079824, 1) 75 fef3d9ee PyRun SimpleFileExFlags feb91c98, 804687b, 1, 80465a8, fef454a1, 804687b) 172 fef3e4fd PyRun AnyFileExFlags feb91c98, 804687b, 1, 80465a8) 61 fef454a1 Py Main 1, 80466b8, feb1cf35, fea935a1, 29, feb96750) 9d9 08050862 main 2, 80466b8, 80466c4) 22 08050758 start 2, 8046874, 804687b, 0, 8046883, 80468ad) 60","bug"
|
|
"6013","2310","normalized cross correlation Trac 1714).
|
|
Original ticket http projects.scipy.org numpy ticket 1714 on 2011 01 15 by trac user bubla, assigned to unknown.
|
|
Hello, I have made a patch that extends the `correlate` function so it can compute normalized cross correlation now See [http en.wikipedia.org wiki Cross correlation Normalized cross correlation the Wikipedia article].
|
|
I have added documentation and simple doctest too, of course.
|
|
The patch is against the latest master Git","enhancement"
|
|
"6013","2311","numpy.array catches to many errors Trac 1715).
|
|
Original ticket http projects.scipy.org numpy ticket 1715 on 2011 01 15 by trac user nils2, assigned to unknown.
|
|
numpy.array catches to many errors when using the array interface Instead of failing and re raising the exception in the following example, numpy simply constructs a 1 element array out of the image import Image, numpy >>> img = Image.open jane doe.jpg ) >>> img.tobitmap ) IOError decoder jpeg not available ... >>> array = numpy.array img) >>> array array <JpegImagePlugin.JpegImageFile instance at 0x101019dd0>, dtype=object)","bug"
|
|
"6013","2312","ndarray.tofile raises ValueError on IO failure; would IOError be more appropriate?
|
|
Trac 1716).
|
|
Original ticket http projects.scipy.org numpy ticket 1716 on 2011 01 17 by trac user mark.dickinson, assigned to unknown.
|
|
The following exception traceback was reported to me by the user of a numpy based application File C \EPD622\lib\site packages\numpy\lib\io.py , line 345, in save format.write array fid, arr) File C \EPD622\lib\site packages\numpy\lib\format.py , line 401, in write array array.tofile fp) ValueError 135453 requested and 6 written While the cause of the error was never explicitly determined, it seems likely to have been as a result of an overfull disk.
|
|
The type and message of the exception are a little confusing here; would it make sense for the exception type to be something like IOError rather than ValueError?
|
|
This was with the version of numpy that ships with EPD 6.2.2, but I believe that the issue is still present in the current numpy source.
|
|
The error message appears to be coming from numpy core src multiarray convert.c, as a result of a system level fwrite call failing to write all the expected items i.e., encountering a system level IO error).","bug"
|
|
"6013","2313","npma.sum a ma,axis=1) fails when no masked values are found Trac 1717).
|
|
Original ticket http projects.scipy.org numpy ticket 1717 on 2011 01 18 by trac user markhmoulton ..., assigned to pierregm.
|
|
I am using the Enthought distribution, Numpy version 1.4.0.
|
|
Maybe this bug has been fixed already, but I thought it serious enough to mention anyway just in case.
|
|
As the snippet shows, the column of sums of a masked array axis = 1) is valid when there is at least one masked value, but triggers an error when no masked values are present.
|
|
This behavior does not occur when axis = 0 or when the average is used instead.
|
|
This causes Numpy to return an error in a situation that arises routinely, but not uniformly.
|
|
FILE Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages numpy version.py DATA release = True short version = 1.4.0 version = 1.4.0 >>> a = npr.randint 0,5, 4,3)) >>> a array [[0, 3, 1], [1, 0, 2], [1, 4, 4], [3, 0, 3]]) Works when 0 is the value to be masked >>> a0 ma = npma.masked values a,0) >>> np.sum a0 ma,axis=1) masked array data = [4 3 9 6], mask = [False False False False], fill value = 999999) Fails when 9 is the value not present) to be masked >>> a9 ma = npma.masked values a,9) >>> npma.sum a9 ma,axis=1) Traceback most recent call last) File <pyshell 511> , line 1, in <module> np.sum a9 ma,axis=1) File Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages numpy core fromnumeric.py , line 1390, in sum return sum axis, dtype, out) File Library Frameworks EPD64.framework Versions 6.2 lib python2.6 site packages numpy ma core.py , line 4357, in sum newmask = mask.all axis=axis) ValueError axis =1) out of bounds The problem does not appear when we sum along the row axis >>> npma.sum a9 ma,axis=0) masked array data = [ 5 7 10], mask = False, fill value = 999999) Nor does it appear when we take the average >>> npma.average a9 ma,axis=1) array [ 1.33333333, 1. , 3. , 2. ])","bug"
|
|
"6013","2314","shape mismatch error is not informative Trac 1718).
|
|
Original ticket http projects.scipy.org numpy ticket 1718 on 2011 01 22 by trac user andrea, assigned to unknown.
|
|
The best thing about numpy is the support for slicing and indexing of tensors with arbitrary shapes.
|
|
It s a pleasure to work with it.
|
|
However, I often get errors like this ValueError shape mismatch objects cannot be broadcast to a single shape This is not really informative, because it doesn t tell me what I did wrong.
|
|
When you start playing with 5 6 dimensions, and start transposing dimensions, it s not that you can look at the code for an obvious mistake.
|
|
Most of the time I have to put some temporary print statements just to get the size of the arrays involved.
|
|
It would make life easier to have a more informative message such as ValueError shape mismatch objects cannot be broadcast to a single shape.
|
|
Required 5,6,7,8) Provided 5,6,8,7).","enhancement"
|
|
"6013","2315","Array compare function might not work for two byteswapped elements.
|
|
Trac 1719).
|
|
Original ticket http projects.scipy.org numpy ticket 1719 on 2011 01 23 by charris, assigned to unknown.
|
|
From ticket 2087 concerning searchsorted.
|
|
Searchsorted uses the compare function from the key array and it seems to fail when both the searched array and the key array are byteswapped.","bug"
|
|
"6013","2316","logical and ) handles nan incorrectly Trac 1720).
|
|
Original ticket http projects.scipy.org numpy ticket 1720 on 2011 01 24 by trac user joshuaschmitt, assigned to unknown.
|
|
when one types 1 and nan , the return is nan , but when one types logical nan 1,nan) , the return is True .
|
|
This is inconsistent.","bug"
|
|
"6013","2317","genfromtxt does not handle single value missing values Trac 1722).
|
|
Original ticket http projects.scipy.org numpy ticket 1722 on 2011 01 25 by jseabold, assigned to unknown.
|
|
from StringIO import StringIO import numpy as np data = 1, 2, 3\n4, ,5 np.genfromtxt StringIO data), delimiter= , , names= a,b,c , missing values= , filling values=0) array [ 1.0, 2.0, 3.0), 4.0, nan, 5.0)], dtype=[ a , <f8 ), b , <f8 ), c , <f8 )]) np.genfromtxt StringIO data), delimiter= , , names= a,b,c , missing values={ b }, filling values={ b 0}) array [ 1.0, 2.0, 3.0), 4.0, 0.0, 5.0)], dtype=[ a , <f8 ), b , <f8 ), c , <f8 )]) Unless I use the dict for missing values, it doesn t fill them in.
|
|
Without named columns np.genfromtxt StringIO data), delimiter= , , missing values= , filling values=0) array [[ 1., 2., 3.
|
|
], [ 4., nan, 5.]])
|
|
np.genfromtxt StringIO data), delimiter= , , missing values={1 }, filling values={1 0}) array [[ 1., 2., 3.
|
|
], [ 4., 0., 5.]])","bug"
|
|
"6013","2318","Mainly performance improvement in tri and related functions Trac 1723).
|
|
Original ticket http projects.scipy.org numpy ticket 1723 on 2011 01 26 by trac user eat, assigned to unknown.
|
|
First of all this is enhancement request for functions tri, tril, triu and mask indices.
|
|
If these are implemented as demonstrated in the attached code, more than 1.5 speedup old exec time new exec time) may be achieved and as far I m aware the speedup is always >1).
|
|
Thus this is observed on np.
|
|
version == 1.5.1 and sys.version 2.7.1 r271 86832, Nov 27 2010, 18 30 46) [MSC v.1500 32 bit Intel)] .
|
|
The speedup is achieved by avoiding unnecessary copying of arrays in tri and changing the default dtype when tri is called) from int to bool in tril, triu and mask indices.
|
|
Secondly the documentation of tril states that it will return of same shape and data type as input.
|
|
This is not the case with current implementation at least with dtype bool).
|
|
My proposal fixes this trivially.
|
|
Thirdly, naming of rows and cols.
|
|
In tri N, M) while in rest related M, N).
|
|
I propose consistency with M, N) as fixed in attachment.
|
|
My attachment is based on version 1.5.1 twodim base.py.
|
|
Basically just replace f with and remove rows after and including if name ... should give you a working version.
|
|
Additionally all related indices functions in twodim base will gain performance.","enhancement"
|
|
"6013","2319","new iterator branch bus error Trac 1724).
|
|
Original ticket http projects.scipy.org numpy ticket 1724 on 2011 01 27 by rgommers, assigned to unknown.
|
|
After building Mark Wiebe s new iterator branch, commit 527de17, I get a bus error on running the tests In [2] np.test ) Running unit tests for numpy NumPy version 1.6.0.dev 527de17 NumPy is installed in Users rgommers Code numpy numpy Python version 2.6.4 r264 75821M, Oct 27 2009, 19 48 32) [GCC 4.0.1 Apple Inc. build 5493)] nose version 0.11.1 Bus error This on OS X 10.6, Intel 32 bit and Python 2.6 from python.org.
|
|
Build log attached.
|
|
There are many build warnings like numpy core src npymath halffloat.c In function xe2x80x98npy doublebits to halfbitsxe2x80x99 numpy core src npymath halffloat.c 355 warning integer constant is too large for xe2x80x98unsigned longxe2x80x99 type numpy core src multiarray lowlevel strided loops.c.src In function xe2x80x98 aligned swap strided to strided size8xe2x80x99","bug"
|
|
"6013","2320","Solaris Numpy SciPy distutils problem Trac 1725).
|
|
Original ticket http projects.scipy.org numpy ticket 1725 on 2011 01 28 by trac user megamic, assigned to cournape.
|
|
When trying to build SciPy on Solaris using the Sun Studio compiler, the following error occurs compile options Iscipy ndimage src I opt local contrib lib python2.7 site packages numpy core include I opt local contrib lib python2.7 site packages numpy core include I opt local contrib include python2.7 c extra options Wall cc scipy ndimage src ni support.c cc illegal option Wall cc illegal option Wall The problem of course is that Wall is not accepted by the compiler, but there seems to be no way to deactivate it.
|
|
One way of getting around the problem is by disabling the extra options in distutils ccompiler.py in the numpy distro, e.g.
|
|
if extra postargs display = \nextra options s .join extra postargs)) To if extra postargs extra postargs = [] display = \nextra options s .join extra postargs))","bug"
|
|
"6013","2321","Error in tanh for large complex argument Trac 1726).
|
|
Original ticket http projects.scipy.org numpy ticket 1726 on 2011 01 28 by trac user mbakker, assigned to unknown.
|
|
tanh computes nans for large complex argument In [85] tanh 1000 0j) Out[85] nan nan j) while the correct answer is 1.0 Solution Program tanh as tanh z) = 1.0 exp 2.0 z)) 1.0 exp 2.0 z))","bug"
|
|
"6013","2322","MaskedArray object has no attribute log Trac 1727).
|
|
Original ticket http projects.scipy.org numpy ticket 1727 on 2011 01 28 by trac user ianmartin, assigned to pierregm.
|
|
The following snippet import numpy a = numpy.ma.ones 1) a.log ) rasies AttributeError MaskedArray object has no attribute log Why does a masked array not have a log function when numpy.ma.log exists?
|
|
Could this be added or is it intentionally missing?
|
|
Thanks","bug"
|
|
"6013","2323","[Feature request] add a NUFFT module improvement of fftpack current module Trac 1728).
|
|
Original ticket http projects.scipy.org numpy ticket 1728 on 2011 01 28 by trac user ahngeichoa, assigned to unknown.
|
|
Hello, I would like to suggest you to add nufft features to the current fft capabilities of numpy.
|
|
Actually, I found a fortran code which could perfectly fill the task, so the effort would be mainly packaging work with f2py.
|
|
The internet page can be found at [http www.cims.nyu.edu cmcl nufft nufft.html] Thank you very much for numpy.","enhancement"
|
|
"6013","2324","Wrong behavior asarray object with array sets owndata to True Trac 1729).
|
|
Original ticket http projects.scipy.org numpy ticket 1729 on 2011 01 31 by pv, assigned to unknown.
|
|
Is this really intended import numpy as np class Fake object) def init self, data) self.
|
|
data = data def array self) return self.
|
|
data x = np.array [1,2,3]) print np.asarray Fake x)).flags.owndata > True This seems to break the invariant that only one array object can have `owndata` True?","bug"
|
|
"6013","2325","asarray ) with misshaped argument Trac 1730).
|
|
Original ticket http projects.scipy.org numpy ticket 1730 on 2011 02 01 by trac user friedrichromstedt, assigned to unknown.
|
|
consider following arguments to asarray ), with some bogus class X 1a) [X ), numpy.asarray [1, 1])] 1b) [numpy.asarray [1, 1]), X )] 2a) [1, numpy.asarray [1, 1])] 2b) [numpy.asarray [1, 1]), 1] In cases 1ab), it seems to be obvious that the result shape should be 2,).
|
|
It is only in case 1a) 2,), in case 1b it s ).
|
|
Furthermore, the behaviour should not change dramatically when replacing objects by numbers, but cases 2ab) yield ValueError setting an array element with a sequence .
|
|
The behaviour of 2ab) is understandable, but wouldn t it be positive to have consistent results across the dtype=numpy.object ordinarydtype border?
|
|
At least, it seems it should yield the same when reversing the argument list.","bug"
|
|
"6013","2326","allow loadtxt ) to read given number of rows Trac 1731).
|
|
Original ticket http projects.scipy.org numpy ticket 1731 on 2011 02 01 by rc, assigned to unknown.
|
|
IMHO it would be useful to add a numrows or nrows parameter to the loadtxt ) function, so that one can read only a specified number of rows.
|
|
Typical use of this, together with the possibility to pass an open file object to the function, would be to read one after another several array data blocks separated by other information from a single text file.
|
|
See the thread [1] for background explanation.
|
|
This ticket might be related to [2], I have not found other related tickets.
|
|
[1] http thread.gmane.org gmane.comp.python.numeric.general 42342 [2] http projects.scipy.org numpy ticket 1107","enhancement"
|
|
"6013","2327","Set complex128 array with a complex64 element fails Trac 1732).
|
|
Original ticket http projects.scipy.org numpy ticket 1732 on 2011 02 01 by trac user SevenThunders, assigned to unknown.
|
|
Observe the following code snippet import numpy as np z = np.empty 2),dtype=np.complex128) z[0] = np.complex64 1.
|
|
1.j) z[1] = np.complex64 1.
|
|
1.j) print z [ 1.
|
|
0.j 1.
|
|
0.j] Numpy fails to make an obvious conversion from complex64 floats to complex128.
|
|
The imaginary parts are lost, which can lead to surprising and difficult to find bugs.
|
|
Now if the intention is for this conversion to be explicit, then the assignment should have generated an exception, not the bizarre behavior of only keeping the real part.","bug"
|
|
"6013","2328","add.accumulate on an array of shape 0,) causes divide by zero Trac 1733).
|
|
Original ticket http projects.scipy.org numpy ticket 1733 on 2011 02 02 by mdboom, assigned to unknown.
|
|
This works in 1.5.1, fails in git MASTER a7ffc7a) import numpy as np x = np.array [[42, 0]], dtype=np.uint32) np.add.accumulate x[ 1,0]) This is a simple example of something one of pyfits unit tests does).
|
|
Blame suggests this is possibly related to fadddc7.","bug"
|
|
"6013","2329","assignment division not compatible with future division Trac 1734).
|
|
Original ticket http projects.scipy.org numpy ticket 1734 on 2011 02 04 by trac user Daan, assigned to unknown.
|
|
from future import division from numpy import array a = array [1,2,3]) a =max a) gives a = array [0,0,1]) while from future import division from numpy import array a = array [1,2,3]) a = a max a) gives a = array [.333, .667, 1.])
|
|
as it should.
|
|
The assignment division should be identical to division assignment).
|
|
Numpy version 1.5.1","bug"
|
|
"6013","2330","Setting a rank 0 string array with a numpy scalar does not work on 1.6.0.dev Trac 1735).
|
|
Original ticket http projects.scipy.org numpy ticket 1735 on 2011 02 04 by FrancescAlted, assigned to unknown.
|
|
The next script describes the problem.
|
|
import numpy as np print numpy version , np.
|
|
version a = np.zeros ), dtype= S10 ) a[ )] = hello1 works!
|
|
print a a[ )] = np.array hello2 ) fails on 1.6.0.dev print a Running it with 1.5.0 numpy version 1.5.0.dev7987 hello1 hello2 With 1.6.0.dev numpy version 1.6.0.dev 0e60ecb hello1 Traceback most recent call last) File tmp scalar string bug.py , line 8, in <module> a[ )] = np.array hello2 ) ValueError cannot set an array element with a sequence","bug"
|
|
"6013","2331","Setting a rank 0 string array with a numpy scalar does not work on 1.6.0.dev Trac 1736).
|
|
Original ticket http projects.scipy.org numpy ticket 1736 on 2011 02 04 by FrancescAlted, assigned to unknown.
|
|
The next script describes the problem.
|
|
import numpy as np print numpy version , np.
|
|
version a = np.zeros ), dtype= S10 ) a[ )] = hello1 works!
|
|
print a a[ )] = np.array hello2 ) fails on 1.6.0.dev print a Running it with 1.5.0 numpy version 1.5.0.dev7987 hello1 hello2 With 1.6.0.dev numpy version 1.6.0.dev 0e60ecb hello1 Traceback most recent call last) File tmp scalar string bug.py , line 8, in <module> a[ )] = np.array hello2 ) ValueError cannot set an array element with a sequence","bug"
|
|
"6013","2332","numpy.test ) seg faults Trac 1737).
|
|
Original ticket http projects.scipy.org numpy ticket 1737 on 2011 02 08 by stsci sienkiew, assigned to unknown.
|
|
numpy.test ) has been seg faulting for about a week.
|
|
Today, I narrowed it down to the sequence shown here.
|
|
With python 2.7.1 on 32 and 64 bit Red Hat Enterprise 4 >>> import numpy as np >>> shape = 2,4,3) >>> rand = np.random.random >>> x = rand shape).astype np.complex) 1j >>> import numpy as np >>> >>> shape = 2,4,3) >>> rand = np.random.random >>> x = rand shape).astype np.complex) 1j Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 1208502592 LWP 16307)] 0x002a4a94 in ufunc loop matches ) from usr stsci pyssgdev 2.7 numpy core umath.so gdb) where 0 0x002a4a94 in ufunc loop matches ) from usr stsci pyssgdev 2.7 numpy core umath.so 1 0x002a519f in find best ufunc inner loop ) from usr stsci pyssgdev 2.7 numpy core umath.so 2 0x002a74a0 in PyUFunc GenericFunction ) from usr stsci pyssgdev 2.7 numpy core umath.so 3 0x002aa02e in ufunc generic call ) from usr stsci pyssgdev 2.7 numpy core umath.so 4 0x080708a4 in PyObject Call ) 5 0x08070974 in call function tail ) 6 0x08070a46 in PyObject CallFunction SizeT ) 7 0x00f54a7f in PyArray GenericBinaryFunction ) from usr stsci pyssgdev 2.7 numpy core multiarray.so 8 0x00f54bb2 in array multiply ) from usr stsci pyssgdev 2.7 numpy core multiarray.so 9 0x0806df95 in binary op1 ) 10 0x0806e7fa in PyNumber Multiply ) 11 0x080ccc13 in PyEval EvalFrameEx ) 12 0x080d04a3 in PyEval EvalCodeEx ) 13 0x080cc4ec in PyEval EvalCode ) 14 0x080ecac8 in run mod ) 15 0x080eb9c0 in PyRun InteractiveOneFlags ) 16 0x080eb7dc in PyRun InteractiveLoopFlags ) 17 0x080eb6d5 in PyRun AnyFileExFlags ) 18 0x08069892 in Py Main ) 19 0x08068a80 in main ) I got this reproduce by by trimming down one of the setUp functions in core tests test multiarray.py It does not appear to happen on RHE 5 in python 2.7.1, 2.6.1 or 2.5.4) or on a Mac in python 2.7).
|
|
Unfortunately, I do not currently have a python with debugging symbols.
|
|
If I have an opportunity to re compile python, I will post more details.","bug"
|
|
"6013","2333","add find first function Trac 1738).
|
|
Original ticket http projects.scipy.org numpy ticket 1738 on 2011 02 09 by nbecker, assigned to unknown.
|
|
find first would be a performance enhancement.
|
|
The current alternative would be something like argmax some boolean function...) which requires a lot of extra work if all you want is the index of the first occurance.
|
|
This is frequently wanted for 1 d arrays, where it s meaning is unambiguous.
|
|
I think semantics for n d n > 1) arrays is up for discussion.","enhancement"
|
|
"6013","2334","issubdtype <whatever>, bool) malfunctions Trac 1739).
|
|
Original ticket http projects.scipy.org numpy ticket 1739 on 2011 02 10 by pv, assigned to unknown.
|
|
Interestingly, >>> np.issubdtype int, np.bool ) False >>> np.issubdtype str, bool) True The issubdtype routine apparently does not work correctly for this case and never has).
|
|
What it does is if issubclass arg2, generic) return issubclass dtype arg1).type, arg2) mro = dtype arg2).type.mro ) if len mro) > 1 val = mro[1] else val = mro[0] return issubclass dtype arg1).type, val) Using mro[1] makes e.g.
|
|
the Python `int` type behave as the more general `signedinteger`.
|
|
For `bool`, it s directly `generic`, which results into strangeness.
|
|
It might be better to hardcode the special handling of Python types, than just picking mro[1]...","bug"
|
|
"6013","2335","nonzero inconsistency Trac 1740).
|
|
Original ticket http projects.scipy.org numpy ticket 1740 on 2011 02 10 by trac user pch, assigned to unknown.
|
|
ndarrays have a behaviour which is not intuitive.
|
|
>>> bool 0,)), bool [0]), bool array [0])) True, True, False) The result should be True, True, True) imo.
|
|
bool array 0)) is False as expected)","bug"
|
|
"6013","2336","syntax error development numpy in python 2.5 Trac 1741).
|
|
Original ticket http projects.scipy.org numpy ticket 1741 on 2011 02 11 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== ERROR Failure SyntaxError invalid syntax test regression.py, line 1525)) Traceback most recent call last) File usr stsci pyssgdev 2.5.4 nose loader.py , line 379, in loadTestsFromName addr.filename, addr.module) File usr stsci pyssgdev 2.5.4 nose importer.py , line 39, in importFromPath return self.importFromDir dir path, fqname) File usr stsci pyssgdev 2.5.4 nose importer.py , line 86, in importFromDir mod = load module part fqname, fh, filename, desc) File usr stsci pyssgdev 2.5.4 numpy core tests test regression.py , line 1525 a[ )] = b hello1 SyntaxError invalid syntax","bug"
|
|
"6013","2337","cholesky error ?
|
|
Trac 1742).
|
|
Original ticket http projects.scipy.org numpy ticket 1742 on 2011 02 11 by trac user pcrs, assigned to pv.
|
|
from numpy import from numpy.linalg import cholesky a=array [[1.0, 0.0],[0.0, 1.0]]) k=0.3; b=dot a,array [[cos k), sin k)],[ sin k), cos k)]])) cholesky b) array [[ 0.97741316, 0.
|
|
], [ 0.30234932, 0.92947371]]) dot c,c.transpose )) array [[ 0.95533649, 0.29552021], [ 0.29552021, 0.95533649]]) b array [[ 0.95533649, 0.29552021], [ 0.29552021, 0.95533649]]) Isn t dot c,c.transpose )) not supposed to be equal to b?
|
|
the sign of b[0,1] is sign inverted","bug"
|
|
"6013","2338","can not pickle a ufunc Trac 1743).
|
|
Original ticket http projects.scipy.org numpy ticket 1743 on 2011 02 16 by trac user chairmanK, assigned to unknown.
|
|
I can not pickle a ufunc.
|
|
I would like to be able to do so.
|
|
>>> pickle.dumps numpy.sin) TypeError Traceback most recent call last) home pythonuser sandbox <ipython input 4 1b00731f1a7d> in <module> ) > 1 pickle.dumps numpy.sin) usr lib python2.6 pickle.pyc in dumps obj, protocol) 1364 def dumps obj, protocol=None) 1365 file = StringIO ) > 1366 Pickler file, protocol).dump obj) 1367 return file.getvalue ) 1368 usr lib python2.6 pickle.pyc in dump self, obj) 222 if self.proto >= 2 223 self.write PROTO chr self.proto)) > 224 self.save obj) 225 self.write STOP) 226 usr lib python2.6 pickle.pyc in save self, obj) 304 reduce = getattr obj, reduce ex , None) 305 if reduce > 306 rv = reduce self.proto) 307 else 308 reduce = getattr obj, reduce , None) home steve sandbox lib python2.6 copy reg.pyc in reduce ex self, proto) 68 else 69 if base is self.
|
|
class > 70 raise TypeError, can t pickle s objects base.
|
|
name 71 state = base self) 72 args = self.
|
|
class , base, state) TypeError can t pickle ufunc objects","enhancement"
|
|
"6013","2339","savetxt doesn t handle recarrays with substructures.
|
|
Trac 1744).
|
|
Original ticket http projects.scipy.org numpy ticket 1744 on 2011 02 17 by fengy research, assigned to unknown.
|
|
test case ar = zeros dtype=[ pos , f4 , 3)), vel , f4 , 3))], shape=10) savetxt ar , ar) TypeError float argument required, not numpy.ndarray.","bug"
|
|
"6013","2340","Documentation of ndarray.clip is misleading Trac 1745).
|
|
Original ticket http projects.scipy.org numpy ticket 1745 on 2011 02 17 by fengy research, assigned to unknown.
|
|
clip doesn t accept a min and a max as keyword arguments.","bug"
|
|
"6013","2341","array ), asarray ) fail Trac 1746).
|
|
Original ticket http projects.scipy.org numpy ticket 1746 on 2011 02 18 by trac user pa.basso, assigned to unknown.
|
|
I can t make a numpy array with this list mtx=[[ 0., 0., 0.
|
|
], [ 0., 1., 2.
|
|
], [ 1., 1., 1.]]
|
|
A=np.array mtx) doesn t work A=np.asarray mtx) doesn t work both return me the following warnings Warning invalid value encountered in double scalars Warning invalid value encountered in double scalars Warning invalid value encountered in double scalars Warning divide by zero encountered in double scalars Warning invalid value encountered in double scalars Warning invalid value encountered in double scalars Warning invalid value encountered in double scalars Warning invalid value encountered in double scalars And at the end... A = [[ nan nan nan] [ nan nan nan] [ nan nan nan]] If I try instead A=np.asmatrix mtx) this works!
|
|
!","bug"
|
|
"6013","2342","numpy.median ) and numpy.mean ) shoud be more predictable Trac 1747).
|
|
Original ticket http projects.scipy.org numpy ticket 1747 on 2011 02 19 by trac user eriko, assigned to unknown.
|
|
numpy.mean ) and numpy.median ) behave differnetly when presented one dimensional input and a given axis argument.
|
|
While mean ) returns a result, median ) raises an IndexError.
|
|
I think, both ways are fine as long as both functions behave simmilar.
|
|
Example code >>> import numpy >>> numpy.mean 1, axis=0) 1.0 >>> numpy.median 1, axis=0) Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 6.2 lib python2.6 site packages numpy lib function base.py , line 2805, in median index = int sorted.shape[axis] 2) IndexError tuple index out of range >>>","bug"
|
|
"6013","2343","astype str ) behavior is different in 1.6.0 vs 1.5.1 Trac 1748).
|
|
Original ticket http projects.scipy.org numpy ticket 1748 on 2011 02 21 by jseabold, assigned to unknown.
|
|
>>> import numpy as np >>> np.
|
|
version 1.6.0.dev c50af53 >>> tmp arr = np.array [[ black ],[ white ],[ other ]]) >>> tmp arr.astype str ) array [[ b ], [ w ], [ o ]], dtype= |S1 ) The old behavior >>> import numpy as np >>> np.
|
|
version 1.5.1 >>> tmp arr = np.array [[ black ],[ white ],[ other ]]) >>> tmp arr.astype str ) array [[ black ], [ white ], [ other ]], dtype= |S5 )","bug"
|
|
"6013","2344","SOABI problem with external libraries and python3.2 Trac 1749).
|
|
Original ticket http projects.scipy.org numpy ticket 1749 on 2011 02 24 by trac user don.fanucci, assigned to cournape.
|
|
Hi!
|
|
It seems that from python 3.2 and on, distutils.sysconfig.get config vars SO ) behaves differently than for python < 3.2.
|
|
Before PEP3149 it returned .so or, I guess, .dll under windows), while with version 3.2 of python, you ll get .cpython 32m.so .
|
|
Since numpy is using get config vars SO ) to determine library filename extensions, numpy with python3.2 fails to properly find external libraries i.e.
|
|
liblapack.so, or libmkl blabla.so, etc).
|
|
As a workaround for this problem, I made some changes to the numpy distutils stuff see attached patch).
|
|
The patch seems to work both on 1.5.1 and latest git HEAD. )
|
|
at least in a works for me kind of way. )","bug"
|
|
"6013","2345","f2py breaks complex kind statements Trac 1750).
|
|
Original ticket http projects.scipy.org numpy ticket 1750 on 2011 02 26 by trac user lorenz, assigned to pearu.
|
|
f2py apparently tries to translate kind statements it found earlier in the file, this seems to fail if they are not just a simple number.
|
|
Consider the following example module precision integer, parameter rk = selected real kind 8) end module subroutine test x, n, res) use precision implicit none real kind=rk), intent in) x n) integer n real kind=rk), intent out) res integer i print , x = , x res = 0.0 do i = 1, size x) res = x i) enddo end subroutine Compiling this with f2py user host f2py kind test> f2py build dir build c fcompiler=gnu95 m test test.f90 running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building extension test sources f2py options [] f2py > build src.linux i686 2.6 testmodule.c Reading fortran codes... Reading file test.f90 format free) Post processing... Block test Block precision In test test.f90 precision get parameters got name selected is not defined on selected Block test In test test.f90 test get parameters got name selected is not defined on selected Post processing stage 2)... Block test Block unknown interface Block precision Block test Building modules... Building module test ... Constructing wrapper function test ... getctype real kind=selected) not supported use .f2py f2cmap).
|
|
getctype real kind=selected) not supported use .f2py f2cmap).
|
|
getctype real kind=selected) not supported use .f2py f2cmap).
|
|
getctype real kind=selected) not supported use .f2py f2cmap).
|
|
getctype real kind=selected) not supported use .f2py f2cmap).
|
|
getctype real kind=selected) not supported use .f2py f2cmap).
|
|
res = test x,[n]) Constructing F90 module support for precision ... Variables rk Wrote C API module test to file build src.linux i686 2.6 testmodule.c Fortran 90 wrappers are saved to build src.linux i686 2.6 test f2pywrappers2.f90 adding build src.linux i686 2.6 fortranobject.c to sources.
|
|
adding build src.linux i686 2.6 to include dirs.
|
|
adding build src.linux i686 2.6 test f2pywrappers2.f90 to sources.
|
|
build src building npy pkg config files running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize Gnu95FCompiler Found usr bin gfortran customize Gnu95FCompiler using build ext building test extension compiling C sources C compiler gcc pthread fno strict aliasing DNDEBUG g fwrapv O3 Wall Wstrict prototypes fPIC compile options Ibuild src.linux i686 2.6 I afs mpa home lorenz sys lib python2.6 site packages numpy core include I usr common pdsoft appl pyapps include python2.6 c gcc build src.linux i686 2.6 testmodule.c compiling Fortran 90 module sources Fortran f77 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore fPIC O3 funroll loops Fortran f90 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall fno second underscore fPIC O3 funroll loops Fortran fix compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore Wall fno second underscore fPIC O3 funroll loops compile options Ibuild src.linux i686 2.6 I afs mpa home lorenz sys lib python2.6 site packages numpy core include I usr common pdsoft appl pyapps include python2.6 c extra options Jbuild Ibuild gfortran f90 test.f90 compiling Fortran sources Fortran f77 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore fPIC O3 funroll loops Fortran f90 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall fno second underscore fPIC O3 funroll loops Fortran fix compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore Wall fno second underscore fPIC O3 funroll loops compile options Ibuild src.linux i686 2.6 I afs mpa home lorenz sys lib python2.6 site packages numpy core include I usr common pdsoft appl pyapps include python2.6 c extra options Jbuild Ibuild gfortran f90 build src.linux i686 2.6 test f2pywrappers2.f90 afs mpa sys system MPA 6.00 usr bin gfortran Wall Wall shared build build src.linux i686 2.6 testmodule.o build build src.linux i686 2.6 fortranobject.o build test.o build build src.linux i686 2.6 test f2pywrappers2.o lgfortran o .
|
|
test.so running scons Some parsing obviously fails, looking at the message getctype real kind=selected) not supported use .f2py f2cmap).
|
|
Running the example produces wrong results, as apparently the arguments are interpreted as floats user host f2py kind test> python c import test; print test.test [1.0,2.0,3.0]) x = 2.0000004731118679 4.27053725193275668E 312 0.0000000000000000 1.0 Changing selected real kind 8) to 8 produces the expected result user host f2py kind test> python c import test; print test.test [1.0,2.0,3.0]) x = 1.0000000000000000 2.0000000000000000 3.0000000000000000 6.0","bug"
|
|
"6013","2346","recfunctions.append fields fails on arrays containing objects Trac 1751).
|
|
Original ticket http projects.scipy.org numpy ticket 1751 on 2011 02 28 by trac user aickley, assigned to unknown.
|
|
The problem is that append fields calls array.view ) which does not accept arrays containing objects See [http projects.scipy.org numpy ticket 674]) Test case to include in numpy lib tests recfunctions.py) class TestAppendFieldsObj TestCase) Test append fields with arrays containing objects def setUp self) from datetime import date self.data = dict obj=date 2000, 1, 1)) def test append to objects self) Test append fields when the base array contains objects obj = self.data[ obj ] x = np.array [ obj, 1.
|
|
), obj, 2.
|
|
)], dtype=[ A , object), B , float)]) y = np.array [10, 20], dtype=int) test = append fields x, C , data=y, usemask=False) control = np.array [ obj, 1.0, 10), obj, 2.0, 20)], dtype=[ A , object), B , float), C , int)]) assert equal test, control) def test append with objects self) Test append fields when the appended data contains objects obj = self.data[ obj ] x = np.array [ 10, 1.
|
|
), 20, 2.
|
|
)], dtype=[ A , int), B , float)]) y = np.array [obj, obj], dtype=object) test = append fields y, C , data=y, dtypes=object, usemask=False) control = np.array [ 10, 1.0, obj), 20, 2.0, obj)], dtype=[ A , int), B , float), C , object)]) assert equal test, control) ====================================================================== ERROR Test append fields when the base array contains objects Traceback most recent call last) File home kolpakov lib python2.6 site packages numpy lib tests test recfunctions.py , line 406, in test append to objects test = append fields x, C , data=y, usemask=False) File home kolpakov lib python2.6 site packages numpy lib recfunctions.py , line 629, in append fields base = merge arrays base, usemask=usemask, fill value=fill value) File home kolpakov lib python2.6 site packages numpy lib recfunctions.py , line 399, in merge arrays return seqarrays.view dtype=seqdtype, type=seqtype) TypeError Cannot change data type for object array.
|
|
====================================================================== ERROR Test append fields when the appended data contains objects Traceback most recent call last) File home kolpakov lib python2.6 site packages numpy lib tests test recfunctions.py , line 416, in test append with objects test = append fields y, C , data=y, dtypes=object, usemask=False) File home kolpakov lib python2.6 site packages numpy lib recfunctions.py , line 627, in append fields for a, n, d) in zip data, names, dtypes)] TypeError Cannot change data type for object array.
|
|
Ran 33 tests in 0.062s Regarding the first failure, append fields calls merge arrays on the base array base = merge arrays base, usemask=usemask, fill value=fill value) Actually this line can be completely removed without breaking existing unit tests.
|
|
I don t like this call because if we are calling merge array on a single array we are are not actually merging anything, but rather making sure that the array conforms to certain criteria.
|
|
Maybe this code should be refactored out from merge arrays...","bug"
|
|
"6013","2347","np.loadtxt fails on empty files Trac 1752).
|
|
Original ticket http projects.scipy.org numpy ticket 1752 on 2011 03 02 by WeatherGod, assigned to unknown.
|
|
loadtxt ) should return an empty array instead of raising an exception for an empty but existing!)
|
|
file.
|
|
This behavior would be consistent with Matlab.
|
|
import numpy as np from StringIO import StringIO a = np.loadtxt StringIO )) Traceback most recent call last) File <stdin> , line 1, in <module> File home bvr Programs numpy numpy lib npyio.py , line 748, in loadtxt raise IOError End of file reached before encountering data. )
|
|
IOError End of file reached before encountering data.
|
|
This has occurred since version 1.2, I believe.
|
|
The current behavior is an improvement over previous versions, where numpy would hang on empty files.
|
|
Therefore, I am filing this as an enhancement request.","enhancement"
|
|
"6013","2348","empty like not passing across attributes of memmap objects Trac 1753).
|
|
Original ticket http projects.scipy.org numpy ticket 1753 on 2011 03 03 by trac user bryancole, assigned to unknown.
|
|
If I create a memmap object >>> a = numpy.memmap foo , mode= w , shape=10) then later I create a new array thus >>> b = numpy.empty like a) then b is also a memmap object.
|
|
However, even though the type of b is memmap , it has no mmap attribute as you d expect, since empty like cannot know what file object to use).
|
|
The broken memmap object b causes failures elsewhere in array finalise as it tries to copy attributes from b filename, offset, mode) which do not exist.
|
|
In summary there are TWO bugs here 1) empty like and friends) should not return objects of type memmap 2) the memmap.
|
|
array finalise method is checking for the existence of the mmap attribute.
|
|
In non memmap objects, this may exist but be set to None, therefore this method needs to check that both mmap exists as an attribute and is not None.","bug"
|
|
"6013","2349","Feature Request Trac 1754).
|
|
Original ticket http projects.scipy.org numpy ticket 1754 on 2011 03 03 by trac user tomrod, assigned to unknown.
|
|
Howdy!
|
|
Newb python programmer here.
|
|
Could there be a feature implemented to handle array multiplication to make it similar to matrix object multiplication besides dot A,B)?
|
|
Perhaps creating a new keyword like or or something to that effect?","enhancement"
|
|
"6013","2350","Raise FPE test failures on Windows and PPC Trac 1755).
|
|
Original ticket http projects.scipy.org numpy ticket 1755 on 2011 03 04 by rgommers, assigned to unknown.
|
|
np.power fails to raise an exception on Windows tested on py 3.1 and py 2.6).
|
|
Test failure ====================================================================== FAIL test floating exceptions test numeric.TestFloatExceptions) Test basic arithmetic function errors Traceback most recent call last) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\core\tests\test numeric.py , line 320, in test floating exceptions np.power, ftype 2), ftype 2 fi.nexp)) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\core\tests\test numeric.py , line 261, in assert raises fpe Type s did not raise fpe error s .
|
|
ftype, fpeerr)) File Z \Users\rgommers\.wine\drive c\Python31\lib\site packages\numpy\testing\utils.py , line 34, in assert raise AssertionError msg) AssertionError Type <class numpy.float64 > did not raise fpe error overflow .
|
|
Result on Windows >>> 2.0 2048 Traceback most recent call last) File <stdin> , line 1, in <module> OverflowError 34, Result too large ) >>> np.power 2.0, 2048) inf Result on OS X as expected) >>> 2.0 2048 Traceback most recent call last) File <ipython console> , line 1, in <module> OverflowError 34, Result too large ) >>> np.power 2.0, 2048) Warning overflow encountered in power inf","bug"
|
|
"6013","2351","Segfault when recasting string arrays Trac 1756).
|
|
Original ticket http projects.scipy.org numpy ticket 1756 on 2011 03 05 by FrancescAlted, assigned to unknown.
|
|
The next makes the segfault to happen import numpy as np arr = np.array [ 1 , 12 , 123 , 1234 , 12345 ]) nparr = np.array arr, dtype= |S3 ) Valgrind shows this ==27003== Invalid write of size 1 ==27003== at 0x4C2860D memcpy mc replace strmem.c 402) ==27003== by 0x62746A7 strided to contig lowlevel strided loops.c.src 311) ==27003== by 0x62B8D09 PyArray CopyInto ctors.c 2645) ==27003== Address 0x6e58522 is 3 bytes after a block of size 15 alloc d ==27003== at 0x4C279AE malloc vg replace malloc.c 207) ==27003== by 0x62ABA66 PyArray NewFromDescr ctors.c 1013) ==27003== by 0x61E84FF ???
|
|
==27003== by 0x93B18CF ???
|
|
==27003== by 0x61E84AF ???
|
|
==27003== by 0x6206FCF ???
|
|
This happens with current master in numpy 1.6.0.dev 5aa65d9","bug"
|
|
"6013","2352","Segfault in pandas.
|
|
Trac 1757).
|
|
Original ticket http projects.scipy.org numpy ticket 1757 on 2011 03 06 by charris, assigned to unknown.
|
|
The following code segfaults.
|
|
>>> import pandas >>> import pandas.util.testing as common >>> df = common.makeTimeDataFrame ) >>> objs = [df, df] >>> s = pandas.Series objs, index=[0, 1])","bug"
|
|
"6013","2353","Assiging to structured array rows with list array fails Trac 1758).
|
|
Original ticket http projects.scipy.org numpy ticket 1758 on 2011 03 07 by rgommers, assigned to unknown.
|
|
Assigning to a structured array with a tuple works, with a list results in an error.
|
|
Would be good if this could be made to work this is an enhancement request).
|
|
With an array gives no warning or exception, but results in garbage.
|
|
This is a bug.
|
|
See also http thread.gmane.org gmane.comp.python.numeric.general 30793 focus=30793 >>> arr = np.zeros 5,), dtype=[ var1 , f8 ), var2 , f8 )]) >>> arr[ var1 ] = np.arange 5) >>> arr array [ 0.0, 0.0), 1.0, 0.0), 2.0, 0.0), 3.0, 0.0), 4.0, 0.0)], dtype=[ var1 , <f8 ), var2 , <f8 )]) >>> arr[0] = 10,20) >>> arr array [ 10.0, 20.0), 1.0, 0.0), 2.0, 0.0), 3.0, 0.0), 4.0, 0.0)], dtype=[ var1 , <f8 ), var2 , <f8 )]) >>> arr[0] = np.array [10,20]) >>> arr array [ 4.2439915824246103e 313, 0.0), 1.0, 0.0), 2.0, 0.0), 3.0, 0.0), 4.0, 0.0)], dtype=[ var1 , <f8 ), var2 , <f8 )]) If this is fixed, please update doc structured arrays.py section Filling structured arrays ).","bug"
|
|
"6013","2354","add docs for ldexp and frexp Trac 1759).
|
|
Original ticket http projects.scipy.org numpy ticket 1759 on 2011 03 07 by rgommers, assigned to pv.
|
|
The ldexp and frexp docstrings are out of date.
|
|
It s not clear to me if entries have to be added to core code generators generate umath.py, or it has to be done in the C source directly.","bug"
|
|
"6013","2355","Copying strided slices does not work on structured arrays with MD fields Trac 1760).
|
|
Original ticket http projects.scipy.org numpy ticket 1760 on 2011 03 07 by FrancescAlted, assigned to unknown.
|
|
The next code snippet shows the problem import numpy as np print numpy version , np.
|
|
version r=np.fromstring abcd 16 5, dtype= i4, 2,3)u2 ) start, stop, step = 0, 5, 1 step == 1 works start, stop, step = 0, 5, 2 r2 = r[start stop step] print r > , r[start][ f1 ] print r2 > , r2[0][ f1 ] assert np.all r[start][ f1 ] == r2[0][ f1 ]) Using numpy 1.5.x numpy version 1.5.0.dev r > [[25185 25699 25185] [25699 25185 25699]] r2 > [[25185 25699 25185] [25699 25185 25699]] Using current 1.6 in git master numpy version 1.6.0.dev c081ad7 r > [[25185 25699 25185] [25699 25185 25699]] r2 > [[25185 25185 25185] [25699 25699 25699]] Traceback most recent call last) File tmp copy recarr.py , line 12, in <module> assert np.all r[start][ f1 ] == r2[0][ f1 ]) AssertionError","bug"
|
|
"6013","2356","Inconsistency with type in scalar operations Trac 1761).
|
|
Original ticket http projects.scipy.org numpy ticket 1761 on 2011 03 08 by FrancescAlted, assigned to unknown.
|
|
I m getting different behaviour with type comparison in scalar operations between 1.5.1 and 1.6.
|
|
Here it is code that shows the issue import numpy as np print numpy > , np.
|
|
version r1 = np.array 8L) print r1 > , `r1`, r1.dtype r1 = r1[ )] print r1 2) > , `r1`, r1.dtype f = long 3) g = np.int16 2) r2 = f g assertion fails only with 1.6 r2 = g f assertion fails with 1.5.1 too print r2 > , `r2`, r2.dtype print types > , type r1), type r2) assert type r1) is type r2) Output when using 1.5.1 numpy > 1.5.1 r1 > array 8L) int64 r1 2) > 8 int64 r2 > 5 int64 types > <type numpy.int64 > <type numpy.int64 > When using 1.6 master) numpy > 1.6.0.dev c081ad7 r1 > array 8L) int64 r1 2) > 8 int64 r2 > 5 int64 types > <type numpy.int64 > <type numpy.int64 > Traceback most recent call last) File tmp scalar types.py , line 17, in <module> assert type r1) is type r2) AssertionError Interestingly, if we change the order of the operation r2 = g f the resulting type also differs with 1.5.1 numpy > 1.5.1 r1 > array 8L) int64 r1 2) > 8 int64 r2 > 5 int64 types > <type numpy.int64 > <type numpy.int64 > Traceback most recent call last) File tmp scalar types.py , line 17, in <module> assert type r1) is type r2) AssertionError","bug"
|
|
"6013","2357","Error message prints dimensions in wrong order?
|
|
Trac 1762).
|
|
Original ticket http projects.scipy.org numpy ticket 1762 on 2011 03 10 by charris, assigned to unknown.
|
|
ERROR test setitem test panel.TestWidePanel) Traceback most recent call last) File usr lib64 python2.7 site packages pandas core tests test panel.py , line 295, in test setitem self.panel[ ItemE ] = df File usr lib64 python2.7 site packages pandas core panel.py , line 330, in setitem self.values[loc] = mat ValueError non broadcastable output operand with shape 30,4) doesn t match the broadcast shape 4,30,1) It appears to me that 4,30,1) should be 1,30,4)","bug"
|
|
"6013","2358","f2py support for size array, dim) syntax Trac 1765).
|
|
Original ticket http projects.scipy.org numpy ticket 1765 on 2011 03 10 by trac user lorenz, assigned to pearu.
|
|
In the discussion on issue 2129 it was suggested to open another issue requesting support for the 2 argument size ) syntax, as in subroutine foo a, n, m, b) implicit none real, intent in) a n, m) integer, intent in) n, m real, intent out) b size a, 1)) integer i do i = 1, size b) b i) = sum a i, )) enddo end subroutine Of course, in this trivial example there is no real point in using `size a,1)` instead of `n` directly however, when issue 2129 is done assumed shape support), this could be very useful!
|
|
Kind regards, Lorenz","enhancement"
|
|
"6013","2359","Nested recarray crashes numpy 1.5.x on Python 3.x Trac 1766).
|
|
Original ticket http projects.scipy.org numpy ticket 1766 on 2011 03 10 by cgohlke, assigned to unknown.
|
|
This issue was reported at <http mail.scipy.org pipermail numpy discussion 2011 March 055269.html>.
|
|
The following code crashes with numpy 1.5.x on Windows using official or self compiled binaries for Python 3.x import numpy numpy.recarray 1,), [ i , [ j , i4 )], 2)]).data The code works correctly with numpy 1.5.x on Python 2.x and also with numpy 1.6.0dev on Python 3.x.
|
|
The crash is during the data attribute lookup.
|
|
The i attribute can be accessed without crash, e.g.
|
|
the following works numpy.recarray 1,), [ i , [ j , i4 )], 2)]).i numpy.recarray 1,), [ i , [ j , i4 )], 2)]).i.j","bug"
|
|
"6013","2360","f2py assumed shape merge test failure Trac 1767).
|
|
Original ticket http projects.scipy.org numpy ticket 1767 on 2011 03 11 by rgommers, assigned to pearu.
|
|
The merge of the branch that closed 2129 and 2345 introduced a test failure ====================================================================== FAIL test kind.TestKind.test all Traceback most recent call last) File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages nose 0.11.1 py2.6.egg nose case.py , line 183, in runTest self.test self.arg) File Users rgommers Code numpy numpy f2py tests test kind.py , line 27, in test all assert selected int kind i)==selectedintkind i),`i, selected int kind i), selectedintkind i)` AssertionError 19, 16, 1) Ran 3279 tests in 60.489s This fails with Python 2.6 on OS X but works with Python 3.2.
|
|
Also, some print statements have been left in the code ..............This module test ext module 5403 is auto generated with f2py version 2).
|
|
Functions res = sum x) res = fsum x) res = sum with use x) Fortran 90 95 modules precision ik,rk.
|
|
..This module test ext module 5405 is auto generated with f2py version 2).
|
|
Functions res = selectedrealkind p,r=0) res = selectedintkind p)","bug"
|
|
"6013","2361","test random.TestRandomDist tests too aggressive?
|
|
Trac 1768).
|
|
Original ticket http projects.scipy.org numpy ticket 1768 on 2011 03 11 by stsci sienkiew, assigned to rgommers.
|
|
These tests are failing on 32 bit Red Hat Enterprise 4 linux.
|
|
They are passing on 64 bit RHE4, 64 bit RHE 5, and 32 bit Mac Leopard.
|
|
I have not looked at every single number in this result, but of those that I have, the printed representation of the number is identical.
|
|
I suspect the problem may be that 15 decimal places is too high a standard.
|
|
assert array almost equal does not actually print all the digits it is comparing, so there is no way to know how far off the actual values are.
|
|
====================================================================== FAIL test chisquare test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 154, in test chisquare np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 50.0 ) x array [[ 63.87858176, 68.68407749], [ 65.77116117, 47.09686762], [ 72.38284032, 74.18408615]]) y array [[ 63.87858176, 68.68407749], [ 65.77116117, 47.09686762], [ 72.38284032, 74.18408615]]) ====================================================================== FAIL test gamma test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 190, in test gamma np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 16.6666666667 ) x array [[ 24.60509189, 28.54993563], [ 26.1347611 , 12.56988483], [ 31.71863276, 33.30143303]]) y array [[ 24.60509189, 28.54993563], [ 26.1347611 , 12.56988483], [ 31.71863276, 33.30143303]]) ====================================================================== FAIL test lognormal test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 238, in test lognormal np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 66.6666666667 ) x array [[ 16.50698632, 36.54846706], [ 22.678866 , 0.71617561], [ 65.72798502, 86.84341601]]) y array [[ 16.50698632, 36.54846706], [ 22.678866 , 0.71617561], [ 65.72798502, 86.84341601]]) ====================================================================== FAIL test noncentral chisquare test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 287, in test noncentral chisquare np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 16.6666666667 ) x array [[ 23.91905354, 13.35324693], [ 31.22452661, 16.60047399], [ 5.03461598, 17.94973089]]) y array [[ 23.91905354, 13.35324693], [ 31.22452661, 16.60047399], [ 5.03461598, 17.94973089]]) ====================================================================== FAIL test noncentral f test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 296, in test noncentral f np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 16.6666666667 ) x array [[ 1.405981 , 0.34207973], [ 3.57715069, 7.92632663], [ 0.43741599, 1.17742088]]) y array [[ 1.405981 , 0.34207973], [ 3.57715069, 7.92632663], [ 0.43741599, 1.17742088]]) ====================================================================== FAIL test rayleigh test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 344, in test rayleigh np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 16.6666666667 ) x array [[ 13.88824965, 13.38331834], [ 20.95413364, 21.08285016], [ 11.06066537, 17.35468506]]) y array [[ 13.88824965, 13.38331834], [ 20.95413364, 21.08285016], [ 11.06066537, 17.35468506]]) ====================================================================== FAIL test standard gamma test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 368, in test standard gamma np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 66.6666666667 ) x array [[ 5.50841531, 6.6295347 ], [ 5.93988485, 2.31044849], [ 7.54838614, 8.01275609]]) y array [[ 5.50841531, 6.6295347 ], [ 5.93988485, 2.31044849], [ 7.54838614, 8.01275609]]) ====================================================================== FAIL test triangular test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 393, in test triangular np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 33.3333333333 ) x array [[ 12.68117179, 12.41292061], [ 16.20131377, 16.25692139], [ 11.20400691, 14.49781448]]) y array [[ 12.68117179, 12.41292061], [ 16.20131377, 16.25692139], [ 11.20400691, 14.49781448]]) ====================================================================== FAIL test wald test random.TestRandomDist) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy random tests test random.py , line 418, in test wald np.testing.assert array almost equal actual, desired, decimal=15) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 791, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 627, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 15 decimals mismatch 16.6666666667 ) x array [[ 3.82935266, 5.13125249], [ 0.35045404, 1.50832397], [ 0.2412432 , 0.22031101]]) y array [[ 3.82935266, 5.13125249], [ 0.35045404, 1.50832397], [ 0.2412432 , 0.22031101]])","bug"
|
|
"6013","2362","assert almost equal and assert array almost equal raise ValueError Trac 1769).
|
|
Original ticket http projects.scipy.org numpy ticket 1769 on 2011 03 11 by kwgoodman, assigned to unknown.
|
|
assert almost equal ) and assert array almost equal ) raise a ValueError instead of AssertionError when the array contains np.inf >> a = np.array [[1., 2.
|
|
], [3., 4.]])
|
|
>> b = a.copy ) >> np.testing.assert almost equal a, b) >> b[0,0] = np.inf >> np.testing.assert almost equal a, b) <snip> ValueError Arrays are not almost equal x array [[ 1., 2.
|
|
], [ 3., 4.]])
|
|
y array [[ inf, 2.
|
|
], [ 3., 4.]])
|
|
>> np.testing.assert array almost equal a, b) <snip> ValueError Arrays are not almost equal x array [[ 1., 2.
|
|
], [ 3., 4.]])
|
|
y array [[ inf, 2.
|
|
], [ 3., 4.]])","bug"
|
|
"6013","2363","Segfault with python 3.2 structured array non existent field Trac 1770).
|
|
Original ticket http projects.scipy.org numpy ticket 1770 on 2011 03 13 by matthew brett, assigned to unknown.
|
|
python3.2 Python 3.2 r32 88452, Feb 20 2011, 11 12 31) [GCC 4.2.1 Apple Inc. build 5664)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> a = np.zeros 1,), dtype=[ f1 , f )]) >>> a[ f1 ] = 1 >>> a[ f2 ] = 1 Segmentation fault All tests pass with np.test ) Expected behavior with same code on python2.6 >>> a[ f2 ] = 1 Traceback most recent call last) File <stdin> , line 1, in <module> ValueError field named f2 not found.
|
|
Discussion on the mailing list with email of same title as this ticket.
|
|
Christoph Gohlke suggested the following patch diff git a numpy core src multiarray mapping.c b numpy core src multiarray mapping.c index 8db85bf..3a72811 100644 a numpy core src multiarray mapping.c b numpy core src multiarray mapping.c 812,10 812,16 array ass sub PyArrayObject self, PyObject index, PyObject op) } } } if defined NPY PY3K) PyErr Format PyExc ValueError, field named S not found.
|
|
, index); else PyErr Format PyExc ValueError, field named s not found.
|
|
, PyString AsString index)); endif return 1; }","bug"
|
|
"6013","2364","Support for nagfor Fortran compiler Trac 1771).
|
|
Original ticket http projects.scipy.org numpy ticket 1771 on 2011 03 14 by trac user matcross, assigned to cournape.
|
|
Hi.
|
|
Are you interested in adding FCompiler info for NAG 5.2 nagfor)?
|
|
I believe the fields different to the existing nag ones should look like compiler type = nagfor description = NAG Fortran Compiler executables[ compiler f77 ] = [ nagfor , fixed ] executables[ compiler f90 ] = [ nagfor ] executables[ compiler fix ] = [ nagfor , fixed ] version pattern = r NAG Fortran Compiler Release ?P<version>[ \s] )","enhancement"
|
|
"6013","2365","autosummary output for classes broken Trac 1772).
|
|
Original ticket http projects.scipy.org numpy ticket 1772 on 2011 03 15 by rgommers, assigned to pv.
|
|
The methods and attributes sections are wrong, because sphinx tries to import numpy.methodname instead of numpy.classname.methodname.
|
|
See numpy.chararray for example.
|
|
This is probably a Sphinx bug in autosummary generate.py), but currently we duplicate extend the autosummary class template in doc source reference templates.
|
|
That class template should probably be completely removed anyway, this is fixed in Sphinx.
|
|
I tried removing it, and the generated pdf got 20 pages longer while the number of warnings went from 1600 down to 800.","bug"
|
|
"6013","2366","corrcoef warnings Trac 1773).
|
|
Original ticket http projects.scipy.org numpy ticket 1773 on 2011 03 16 by trac user senyai, assigned to unknown.
|
|
Hi.
|
|
I m pretty sure that corrcoef should not give any warnings with empty arrays >>> numpy.corrcoef []) Warning invalid value encountered in divide 1 >>> numpy.corrcoef [],[]) Warning invalid value encountered in divide Warning invalid value encountered in divide array [[ nan, nan], [ nan, nan]])","bug"
|
|
"6013","2367","assert array almost equal prints low precision Trac 1774).
|
|
Original ticket http projects.scipy.org numpy ticket 1774 on 2011 03 16 by stsci sienkiew, assigned to unknown.
|
|
If you do this, a=numpy.array [ 16.50698631688883822 ] ) b=numpy.array [ 16.50698632 ]) numpy.testing.assert array almost equal a,b,decimal=15) it says AssertionError Arrays are not almost equal to 15 decimals mismatch 100.0 ) x array [ 16.50698632]) y array [ 16.50698632]) In discussion on the mailing list, Robert Kern gave a good explanation of why this happens and suggested that instead of repr ), It should probably use np.array repr x, precision=16) Instead of fixing the precision at 16, maybe it should use the decimal parameter to choose the precision to print.
|
|
The point being that if the arrays are not almost equal, we should be able to see which elements did and did not match.","enhancement"
|
|
"6013","2368","frompyfunc function segfault Trac 1775).
|
|
Original ticket http projects.scipy.org numpy ticket 1775 on 2011 03 18 by trac user borisaqua, assigned to unknown.
|
|
from numpy import fv = vectorize lambda a tuple [a] 50)) fv [2,3]) Program received signal SIGSEGV, Segmentation fault.
|
|
0x00007ffff6252acf in ?? )
|
|
from usr lib pymodules python2.6 numpy core umath.so gdb) bt 0 0x00007ffff6252acf in ?? )
|
|
from usr lib pymodules python2.6 numpy core umath.so 1 0x00007ffff6254734 in ?? )
|
|
from usr lib pymodules python2.6 numpy core umath.so 2 0x00007ffff6255928 in ?? )
|
|
from usr lib pymodules python2.6 numpy core umath.so 3 0x0000000000000000 in ?? )
|
|
gdb)","bug"
|
|
"6013","2369","Complex divide by zero returns NAN Trac 1776).
|
|
Original ticket http projects.scipy.org numpy ticket 1776 on 2011 03 18 by trac user ksmith13, assigned to unknown.
|
|
Dividing a complex numpy array by another with zero values returns nan nani instead of inf infi EX np.array [1 4i, 2 3i]) np.array [0.0, 0.0])","bug"
|
|
"6013","2370","fh is not defined in npyio.py fromregex Trac 1777).
|
|
Original ticket http projects.scipy.org numpy ticket 1777 on 2011 03 22 by trac user citizenwad, assigned to unknown.
|
|
NameError global name fh is not defined File Users williamdawson Programs fat wip.py , line 263, in <module> header gal = readheader gal cat) File Users williamdawson Programs tools.py , line 96, in readheader [ column ,numpy.int16), name , S10 )]) File Library Frameworks EPD64.framework Versions 7.0 lib python2.7 site packages numpy lib npyio.py , line 972, in fromregex Note that this code works perfectly fine with the python 2.6 EPD release.
|
|
The readheader program just uses header = numpy.fromregex catalog,r ttype [0 9] ) ?
|
|
\s)?= ?
|
|
\s)?
|
|
\w ) , [ column ,numpy.int16), name , S10 )]) where catalog = [string], Name perhaps including path) of the catalog that contains all of the data e.g.
|
|
x,y,e1,e2,...).
|
|
Must include ttype header designations for the columns e.g.
|
|
ttype0 = objid ttype1 = x","bug"
|
|
"6013","2371","Object arrays of complex arrays can t be multiplied by numpy floats Trac 1778).
|
|
Original ticket http projects.scipy.org numpy ticket 1778 on 2011 03 22 by inducer, assigned to unknown.
|
|
>>> import numpy as np >>> z = np.array [1j]) >>> zz = np.zeros 2,), dtype=object) >>> zz.fill z) >>> zz array [[ 0.
|
|
1.j], [ 0.
|
|
1.j]], dtype=object) >>> zz np.float64 5) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError unsupported operand type s) for numpy.ndarray and numpy.float64 >>> zz 5. array [[ 0.
|
|
5.j], [ 0.
|
|
5.j]], dtype=object) >>> In my humble estimation, that TypeError shouldn t be happening.","bug"
|
|
"6013","2372","array.tolist ) speed enhancement Trac 1779).
|
|
Original ticket http projects.scipy.org numpy ticket 1779 on 2011 03 23 by trac user Han, assigned to unknown.
|
|
Hi, For a while, a small issue has been bugging me, where array.tolist ) takes a huge amount of time, compared to the speed of Python.
|
|
To illustrate, here are some timings on Windows XP) >>> timeit.timeit a.tolist ) , from numpy import arange; a = arange 1e5) , number=500) 19.07303038231646 The conversion of 500 x 100000 elements takes up to 20 !)
|
|
seconds.
|
|
>>> timeit.timeit a = arange 1e5) , from numpy import arange , number=500) 0.4194663546483639 While creating a NumPy arrays with the same amount of items takes .5 seconds.
|
|
>>> timeit.timeit a = range int 1e5)) , number=500) 0.97656364422391562 And creating a Python list takes no more than 1 second, this is 20x faster than array.tolist ).
|
|
So where is this discrepancy coming from?
|
|
To find out, I did some runs with valgrind on NumPy 1.4.1 dbg Debian), and used kcachegrind to produce a few calling graphs.
|
|
The first thing I noticed was the amount of array alloc and consequent array dealloc calls.
|
|
The number of calls amount up to the number of elements in the array!
|
|
PyArray NewFromDescr is called per array element, which creates a lot of overhead.
|
|
In NumPy 1.6.1b1, this overhead still exists; it stems from the PyArray ToList function in convert.c NPY NO EXPORT PyObject PyArray ToList PyArrayObject self) { PyObject lp; PyArrayObject v; intp sz, i; if !PyArray Check self)) { return PyObject )self; } if self >nd == 0) { return self >descr >f >getitem self >data,self); } sz = self >dimensions[0]; lp = PyList New sz); for i = 0; i < sz; i ) { v = PyArrayObject )array big item self, i); if PyArray Check v) v >nd >= self >nd)) { PyErr SetString PyExc RuntimeError, array item not returning smaller \ dimensional array ); Py DECREF v); Py DECREF lp); return NULL; } PyList SetItem lp, i, PyArray ToList v)); Py DECREF v); } return lp; } For every element in the array, a array big item call is made to create a new array with that element and given recursively to PyArray ToList to get the actual element item.
|
|
I added an extra clause to the function to account for 1 dimensional arrays if self >nd == 1) { sz = self >dimensions[0]; lp = PyList New sz); for i = 0; i < sz; i ) { PyList SetItem lp, i, self >descr >f >getitem index2ptr self, i),self)); } return lp; } Which gets the time down to 2 seconds on Windows.
|
|
I m not sure about the patch, though, because it does not account for errors, and might be more optimized streamlined.
|
|
Anyway, hope it can go in at 1.6.0 in some way or another, because it really helps in NumPy >Python conversions!
|
|
[attached calling graphs]","enhancement"
|
|
"6013","2373","nditer broadcasting test failure under python 2.4 Trac 1780).
|
|
Original ticket http projects.scipy.org numpy ticket 1780 on 2011 03 24 by rgommers, assigned to mwiebe.
|
|
Reported on OS X and Linux against 1.6.0b1.
|
|
FAIL test iterator.test iter broadcasting errors Traceback most recent call last) File sw lib python2.4 site packages nose case.py , line 187, in runTest self.test self.arg) File sw lib python2.4 site packages numpy core tests test iterator.py , line 639, in test iter broadcasting errors Message s doesn\ t contain operand shape 2,3) msg) File sw lib python2.4 site packages numpy testing utils.py , line 34, in assert raise AssertionError msg) AssertionError Message non broadcastable output operand with shape zd, zd) doesn t match the broadcast shape zd, zd, zd) doesn t contain operand shape 2,3)","bug"
|
|
"6013","2374","long double representation not recognized Trac 1781).
|
|
Original ticket http projects.scipy.org numpy ticket 1781 on 2011 03 24 by trac user hoytak, assigned to cournape.
|
|
In attempting to install numpy from source commit e6f35b), it seems that whatever long double representation my computer has is not picked up by long double representation ) in core setup common.py.
|
|
I attached a pickle file of the lines variable that gets scanned to detect it.
|
|
My machine is an intel i7 2600 running ubuntu 10.10.
|
|
The compiler I m using is intel icc 12.3), which is also what python was compiled with.
|
|
I m using python version 2.7.1.
|
|
I d be happy to give any more information about my system that s needed.","bug"
|
|
"6013","2375","KeyError in crackfortran.py Trac 1782).
|
|
Original ticket http projects.scipy.org numpy ticket 1782 on 2011 03 24 by fonnesbeck, assigned to pearu.
|
|
I get the following when I try to build extensions with f2py Traceback most recent call last) File setupegg.py , line 7, in <module> execfile setup.py ) File setup.py , line 131, in <module> config dict)) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils core.py , line 186, in setup return old setup new attr) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils core.py , line 152, in setup dist.run commands ) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 975, in run commands self.run command cmd) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File System Library Frameworks Python.framework Versions 2.6 Extras lib python setuptools command develop.py , line 27, in run self.install for development ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command develop.py , line 15, in install for development old develop.install for development self) File System Library Frameworks Python.framework Versions 2.6 Extras lib python setuptools command develop.py , line 85, in install for development self.run command egg info ) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils cmd.py , line 333, in run command self.distribution.run command command) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command egg info.py , line 8, in run self.run command build src ) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils cmd.py , line 333, in run command self.distribution.run command command) File System Library Frameworks Python.framework Versions 2.6 lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command build src.py , line 152, in run self.build sources ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command build src.py , line 169, in build sources self.build extension sources ext) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command build src.py , line 334, in build extension sources sources = self.f2py sources sources, ext) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy distutils command build src.py , line 593, in f2py sources [ m ,ext name] f sources) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py f2py2e.py , line 359, in run main postlist=callcrackfortran files,options) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py f2py2e.py , line 281, in callcrackfortran postlist=crackfortran.crackfortran files) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 2721, in crackfortran postlist=postcrack grouplist[0]) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 1554, in postcrack g=postcrack g,tab=tab \t ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 1573, in postcrack block[ body ]=analyzebody block,args,tab=tab) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 1722, in analyzebody b=postcrack b,as ,tab=tab \t ) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 1573, in postcrack block[ body ]=analyzebody block,args,tab=tab) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 1719, in analyzebody b[ saved interface ] = crack2fortrangen b, \n 6, as interface=True) File Library Python 2.6 site packages numpy 2.0.0.dev 2e96d91 20110324 py2.6 macosx 10.6 universal.egg numpy f2py crackfortran.py , line 2511, in crack2fortrangen al = [a for a in block[ args ] if not isintent callback vars[a])] KeyError n","bug"
|
|
"6013","2376","Support full long double precision in format Trac 1783).
|
|
Original ticket http projects.scipy.org numpy ticket 1783 on 2011 03 25 by mwiebe, assigned to unknown.
|
|
For 2271, support was added for the format ) function to numpy scalars.
|
|
This support downcasts long doubles to doubles, so cannot print long doubles at their full precision.
|
|
This should be fixed with custom formatting code in that case.","bug"
|
|
"6013","2377","Support zero dimensional arrays in format Trac 1784).
|
|
Original ticket http projects.scipy.org numpy ticket 1784 on 2011 03 25 by mwiebe, assigned to unknown.
|
|
For 2271, support was added for the format ) function in numpy scalars.
|
|
It would be useful for zero dimensional arrays to support this as well.","bug"
|
|
"6013","2378","fastCopyAndTranspose test error Trac 1785).
|
|
Original ticket http projects.scipy.org numpy ticket 1785 on 2011 03 27 by rgommers, assigned to mwiebe.
|
|
In 1.6.x branch ====================================================================== ERROR test api.test fastCopyAndTranspose Traceback most recent call last) File Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages nose 1.0.0 py2.6.egg nose case.py , line 187, in runTest self.test self.arg) File Users zouzoujing Code numpy numpy core tests test api.py , line 11, in test fastCopyAndTranspose b = np.fastCopyAndTranspose a) ValueError only 2 d arrays are allowed","bug"
|
|
"6013","2379","Backport f2py related commits to 1.6.x Trac 1786).
|
|
Original ticket http projects.scipy.org numpy ticket 1786 on 2011 03 27 by pearu, assigned to rgommers.
|
|
The following commits ought to backported to 1.6.x maintenance branch 225f28c e46ed8f 2e96d91 In addition, I am waiting feedback from a PowerPC user about one failing f2py test and when received, another bug fix commit is anticipated.","bug"
|
|
"6013","2380","ctypeslib test error on python 3.2 Trac 1787).
|
|
Original ticket http projects.scipy.org numpy ticket 1787 on 2011 03 29 by matthew brett, assigned to unknown.
|
|
nosetests tmp junk lib python3.2 site packages numpy tests test ctypeslib.py gives ====================================================================== ERROR Failure OSError home mb312 tmp junk lib python3.2 site packages numpy core multiarray.pyd cannot open shared object file No such file or directory) Traceback most recent call last) File usr local lib python3.2 dist packages nose 1.0.0 py3.2.egg nose failure.py , line 37, in runTest raise self.exc class self.exc val).with traceback self.tb) File usr local lib python3.2 dist packages nose 1.0.0 py3.2.egg nose loader.py , line 390, in loadTestsFromName addr.filename, addr.module) File usr local lib python3.2 dist packages nose 1.0.0 py3.2.egg nose importer.py , line 39, in importFromPath return self.importFromDir dir path, fqname) File usr local lib python3.2 dist packages nose 1.0.0 py3.2.egg nose importer.py , line 86, in importFromDir mod = load module part fqname, fh, filename, desc) File home mb312 tmp junk lib python3.2 site packages numpy tests test ctypeslib.py , line 8, in <module> cdll = load library multiarray , np.core.multiarray.
|
|
file ) File home mb312 tmp junk lib python3.2 site packages numpy ctypeslib.py , line 122, in load library raise exc File home mb312 tmp junk lib python3.2 site packages numpy ctypeslib.py , line 119, in load library return ctypes.cdll[libpath] File usr lib python3.2 ctypes init .py , line 415, in getitem return getattr self, name) File usr lib python3.2 ctypes init .py , line 410, in getattr dll = self.
|
|
dlltype name) File usr lib python3.2 ctypes init .py , line 340, in init self.
|
|
handle = dlopen self.
|
|
name, mode) OSError home mb312 tmp junk lib python3.2 site packages numpy core multiarray.pyd cannot open shared object file No such file or directory This is because the compiled file is now multiarray.cpython 32mu.so not multiarray.so or multiarray.pyd as the ctypeslib was previously expecting.
|
|
One fix is to add in ctypeslib around line 100 py exts = [suf for suf, mode, mtype in imp.get suffixes ) if mtype == imp.C EXTENSION] to the extensions to be tested which will pick up .so along with cpython 32mu.so and modulecpython 32mu.so).
|
|
This would mean, for non python libraries, the number of exts to search for would increase, presumably not very usefully because we don t often use ctypeslib for python extensions, I suppose.","bug"
|
|
"6013","2381","histogramdd doesn t like bins with infinite edges Trac 1788).
|
|
Original ticket http projects.scipy.org numpy ticket 1788 on 2011 03 30 by trac user eschlafly, assigned to unknown.
|
|
The following code causes an Overflow Error and doesn t need to h,e = histogramdd randn 100,2), bins= 100,[ inf,0,inf])) The problem is the section of code which looks for points on the right edge of the rightmost edge.
|
|
Specifically, it looks for the amount that is close enough to the edge via line 314 in lib function base.py, v1.5.1) decimal = int log10 dedges[i].min ))) 6 but in this case dedges[i].min ) == inf, triggering the overflow.
|
|
I would propose looking for the case dedges[i].min ) == inf explicitly and not trying to find such points in that case.
|
|
Of course, I can get around this, but it s really convenient when the library is robust against such situations.
|
|
Thanks a lot.","bug"
|
|
"6013","2382","Support half floats via buffer interface Trac 1789).
|
|
Original ticket http projects.scipy.org numpy ticket 1789 on 2011 04 02 by trac user wickedgrey, assigned to unknown.
|
|
I am working on adding half floats to the struct module for CPython 3.3. http bugs.python.org issue11734 I would like to get support for this into 1.6.0b2 if possible.
|
|
I realize that that s putting the cart before the horse a little bit, since the change hasn t been formally accepted into CPython yet.
|
|
Pull request here https github.com numpy numpy pull 68 More discussion on the mailing list here http mail.scipy.org pipermail numpy discussion 2011 March 055795.html","enhancement"
|
|
"6013","2383","dtype ..., align=True) does not insert trailing padding Trac 1790).
|
|
Original ticket http projects.scipy.org numpy ticket 1790 on 2011 04 02 by pv, assigned to mwiebe.
|
|
Note the following >>> dt = np.dtype [ a , b ), b , i ), c , b )], align=True) >>> dt.itemsize 9 whereas include <stdio.h> typedef struct { char a; int b; char c; } xxx t; int main ) { printf d\n , sizeof xxx t)); return 0; } > sizeof xxx t) == 12 Changing this might cause problems in third party code.
|
|
On the other hand, align=True is probably not that often used.","bug"
|
|
"6013","2384","Handle SWIG Cython Pyrex files correctly during install Trac 1791).
|
|
Original ticket http projects.scipy.org numpy ticket 1791 on 2011 04 02 by rgommers, assigned to rgommers.
|
|
The files under doc swig, doc cython.
|
|
doc pyrex should not be installed by default through setup.py because this makes Linux packagers unhappy.
|
|
In the Windows and OS X installers these files could be put somewhere, because they are useful for users.","bug"
|
|
"6013","2385","absolute complex inf)) raises invalid warning py 3.2, win32) Trac 1792).
|
|
Original ticket http projects.scipy.org numpy ticket 1792 on 2011 04 03 by rgommers, assigned to unknown.
|
|
Happens only on win32, not on OS X with python 3.2 >>> np.abs np.inf) inf >>> np.abs np.complex np.inf)) Warning invalid value encountered in absolute inf","bug"
|
|
"6013","2386","genfromtxt should probably return an empty array if there is no data.
|
|
Trac 1793).
|
|
Original ticket http projects.scipy.org numpy ticket 1793 on 2011 04 04 by charris, assigned to unknown.
|
|
genfromtxt currently raises an IOError if it reaches the eof without finding a line of data.","bug"
|
|
"6013","2387","Test for loadtxt with a data line of tab delimited spaces Trac 1794).
|
|
Original ticket http projects.scipy.org numpy ticket 1794 on 2011 04 05 by bsouthey, assigned to unknown.
|
|
This implements a test based on the corner case of a line of spaces separated by a tab delimiter reported by derek in Ticket 1071.
|
|
Prior to this, loadtxt would skip any lines composed of whitespace regardless of the delimiter even if the whitespace could be considered valid elements.
|
|
dt=np.dtype [ label , S4 ), comment , S4 )]) a=np.loadtxt StringIO aa,bb\n , \ncc,dd ), delimiter= , , dtype=dt)","enhancement"
|
|
"6013","2388","Sort out the NPY FMT printf formatting macros Trac 1795).
|
|
Original ticket http projects.scipy.org numpy ticket 1795 on 2011 04 06 by mwiebe, assigned to unknown.
|
|
These macros are supposed to help portably print different types, but they re currently not consistent.
|
|
The points are PyString FromFormat and PyOS snprintf support different sets of formatting codes the first is intended to be portable and consistent, the second calls the OS vsnprintf routine, according to the 2.7 documentation).
|
|
Within the NumPy code base, NPY INTP FMT is used with PyString Format, while the rest of them are used with PyOS snprintf.
|
|
In 2.6, PyString Format doesn t support a long long formatting character, so on 64 bit windows, there is no mechanism to format the npy intp type.
|
|
The code also casts to long and uses ld in some places.
|
|
This would produce the wrong value when using extremely large arrays.","bug"
|
|
"6013","2389","Zen of numpy required Trac 1796).
|
|
Original ticket http projects.scipy.org numpy ticket 1796 on 2011 04 06 by bsouthey, assigned to unknown.
|
|
Travis Oliphant provided the version 0.1 of the Zen of NumPy in his [http fperez.org events 2011 siam cse ]SIAM 2011 talk.
|
|
This is provided here using the same coding as this.py but really should have a numpy context.","enhancement"
|
|
"6013","2390","Better generator tests from IPython Trac 1797).
|
|
Original ticket http projects.scipy.org numpy ticket 1797 on 2011 04 09 by rgommers, assigned to unknown.
|
|
See http thread.gmane.org gmane.comp.python.numeric.general 43437 focus=43438 Issue now is that generator tests using yield) show no useful info when failing and are very hard to debug.
|
|
IPython has a better implementation that can be adapted for use in numpy.testing.","enhancement"
|
|
"6013","2391","Type promotion rules regression Trac 1798).
|
|
Original ticket http projects.scipy.org numpy ticket 1798 on 2011 04 11 by mwiebe, assigned to mwiebe.
|
|
[scalar float64] [array int8] should produce [array float64], but when the value of the scalar float64 is small, it may produce [array float16] or [array float32].
|
|
Here is the current behavior >>> 3.1 np.ones 2,dtype=np.int8) array [ 3.09960938, 3.09960938], dtype=float16) >>> 1e10 np.ones 2,dtype=np.int8) array [ 1.00000000e 10, 1.00000000e 10], dtype=float32) >>> 1e40 np.ones 2,dtype=np.int8) array [ 1.00000000e 40, 1.00000000e 40]) I propose a fix in two parts.
|
|
1.
|
|
Fix this in the PyArray ResultType numpy.result type function of the type promotion API.
|
|
In particular, the algorithm should track the promoted array and scalar types separately, then if the kind of the promoted scalar type is greater than the kind of the promoted array type, continue the promotion of the promoted array type with the scalars without using minscalartype.
|
|
2.
|
|
Change the ufunc to detect binary operators during creation, and set a flag for it.
|
|
Then, during evaluation, use PyArray ResultType to do the type promotion for binary operators instead of the existing slow linear search.
|
|
The ufunc doesn t have a flags field, but it does have a field called check return which is unused, and could be repurposed as a flags field without changing the size of the python type.","bug"
|
|
"6013","2392","assert approx equal against 0.0 always fails Trac 1799).
|
|
Original ticket http projects.scipy.org numpy ticket 1799 on 2011 04 15 by trac user jaety, assigned to unknown.
|
|
The following lines import numpy as np np.testing.assert approx equal 0.0, 0.000000000000001) returns that the two values are not equal to 7 significant digits.
|
|
Basically, assert approx equal will fail incorrectly if either of the two numbers is identically 0.0, regardless of the number of significant digits you specify, because of the way the floating point numbers are scaled prior to testing them.
|
|
I think you need an explicit test for 0.0, something like Specially handle desired or actual identically zero if desired==0.0 or actual==0.0) and np.abs desired actual) np.power 10,significant)>=1.0) raise AssertionError msg) else return","bug"
|
|
"6013","2393","add license file to Windows installers Trac 1800).
|
|
Original ticket http projects.scipy.org numpy ticket 1800 on 2011 04 17 by rgommers, assigned to rgommers.
|
|
This should be added in tools win32build nsis scripts numpy superinstaller.nsi.in Documentation on how to do this can be found at http nsis.sourceforge.net Docs Modern 20UI 202 Readme.html","bug"
|
|
"6013","2394","Failure test noncentral f in random Trac 1801).
|
|
Original ticket http projects.scipy.org numpy ticket 1801 on 2011 04 17 by rgommers, assigned to unknown.
|
|
Reported by Christoph Gohlke against 1.6.0b2 on 64 bit Windows with MKL.
|
|
Unlike 2361 this seems to be a real bug, not just about test precision.
|
|
====================================================================== FAIL test noncentral f test random.TestRandomDist) Traceback most recent call last) File X \Python26 x64\lib\site packages\numpy\random\tests\test random.py , line 297, in test noncentral f np.testing.assert array almost equal actual, desired, decimal=14) File X \Python26 x64\lib\site packages\numpy\testing\utils.py , line 800, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File X \Python26 x64\lib\site packages\numpy\testing\utils.py , line 636, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 14 decimals mismatch 100.0 ) x array [[ 1.62003345, 1.7253997 ], [ 0.96735921, 0.42933718], [ 0.71714872, 6.24979552]]) y array [[ 1.405981 , 0.34207973], [ 3.57715069, 7.92632663], [ 0.43741599, 1.17742088]]) Christoph, could you please have a look at this?
|
|
I haven t got a 64 bit Windows machine available.","bug"
|
|
"6013","2395","Problem pickling length 1 boolean arrays Trac 1802).
|
|
Original ticket http projects.scipy.org numpy ticket 1802 on 2011 04 18 by trac user eewallace, assigned to unknown.
|
|
If I pickle a dictionary containing two length one numpy arrays with dtype bool and the same value e.g.
|
|
dict a1=numpy.array [False]),a2=numpy.array [False]), when the data is unpickled, the resulting arrays contain references to the same boolean objects, so that changing the value of one of them changes both.
|
|
The attached text file contains a transcript of a short IPython session illustrating the problem.
|
|
This seems to come up specifically with the standard pickle module, not with cPickle.
|
|
It also occurs with arrays that are data members of separate class instances, which is where I first encountered it.
|
|
I ve produced this on two Linux installations, one Arch Linux with python 2.7.1, and numpy 1.5.1, both built under gcc 4.5, the other with Red Hat 4.1, python 2.6.6, numpy 1.5.1 and gcc 4.1.","bug"
|
|
"6013","2396","ndarray dump function and straight cPickle) fails for large arrays Trac 1803).
|
|
Original ticket http projects.scipy.org numpy ticket 1803 on 2011 04 19 by trac user meawoppl, assigned to unknown.
|
|
a = zeros 300000, 1000)) f = open test.pkl , w ) cPickle.dump a, f) SystemError Traceback most recent call last) home kddcup code matt svd projection take5 <ipython console> in <module> ) SystemError error return without exception set Or using the .dump function a.dump test.pkl ) SystemError Traceback most recent call last) home kddcup code matt svd projection take5 <ipython console> in <module> ) SystemError NULL result without error in PyObject Call I am not sure if this is a numpy or Pickle cPickle glitch.
|
|
In either case, a more instructive error message would certainly help.
|
|
I think the problem only happens for arrays larger than 2 32 1) bytes but I would have to experiment more to be sure.","bug"
|
|
"6013","2397","numpy.apply along axis returns strange dimensionless array Trac 1804).
|
|
Original ticket http projects.scipy.org numpy ticket 1804 on 2011 04 20 by trac user Mark ..., assigned to unknown.
|
|
In [18] x=numpy.apply along axis numpy.sum, 0, numpy.ones 1,))); print x.shape, type x), x, x==1 In [19] In [20] ) <type numpy.ndarray > 1.0 True","bug"
|
|
"6013","2398","Test errors when the char type is unsigned.
|
|
Trac 1805).
|
|
Original ticket http projects.scipy.org numpy ticket 1805 on 2011 04 21 by trac user manowar, assigned to unknown.
|
|
Hi!
|
|
I m trying to port the NumPy to the armv5tel platform.
|
|
While most of the code seem to work, there is a problem in initialize casting tables ) function, numpy core src multiarray scalartypes.c.src 3642 the table index update operation k = npy next larger type table[k] results in an infinite loop, because npy next larger type table[0] is 255 and npy next larger type table[255] is 0.
|
|
No loop escape break) condition is either satisfied.
|
|
So, the initialize casting tables ) function does never return and the loading of the numpy module hangs forever in dl load multiarray.so taking 99 of CPU time.
|
|
I m not sure assuming that the problem is ARM specific.
|
|
What steps should I perform to expose the roots of the problem?
|
|
TIA, Paul.","bug"
|
|
"6013","2399","Datetime needs docs and updated NEP Trac 1806).
|
|
Original ticket http projects.scipy.org numpy ticket 1806 on 2011 04 22 by rgommers, assigned to unknown.
|
|
API docs as well as docstrings for the few Python level functions are needed.
|
|
Furthermore, the datetime NEP needs to be updated so it s clear what is implemented and what not.
|
|
Also, the other datetime related NEP needs to be updated too.
|
|
A small tutorial would be a good bonus, could be easily extracted from the NEP.","bug"
|
|
"6013","2400","Resizing string array to zero length causes datatype to change to 1 character.
|
|
Trac 1807).
|
|
Original ticket http projects.scipy.org numpy ticket 1807 on 2011 04 25 by trac user bellamyj, assigned to unknown.
|
|
If a string array is resized to zero length, and then to non zero length, the data type will change to a single character |S1).
|
|
See the attached code for an example.
|
|
I m running 32 bit Python 2.7.1 on Windows and !NumPy 1.5.1.","bug"
|
|
"6013","2401","numpy.load segfaults Trac 1808).
|
|
Original ticket http projects.scipy.org numpy ticket 1808 on 2011 04 26 by jpaalasm, assigned to unknown.
|
|
An array with dtype numpy.dtype [ time , object), value , |S1 )]) was saved to a file with numpy.save on a 32 bit Ubuntu.
|
|
The loading of the file using numpy.load crashes on 64 bit Ubuntu, but succeeds with 32 bit Ubuntu.
|
|
Numpy version is 1.5.1.
|
|
== 64 bit Ubuntu 10.04 == jpaalasm ubuntu gdb args python c import numpy; numpy.load file seqfaulting sleep stage data.npy )) GNU gdb GDB) 7.1 ubuntu Copyright C) 2010 Free Software Foundation, Inc. License GPLv3 GNU GPL version 3 or later <http gnu.org licenses gpl.html> This is free software you are free to change and redistribute it.
|
|
There is NO WARRANTY, to the extent permitted by law.
|
|
Type show copying and show warranty for details.
|
|
This GDB was configured as x86 64 linux gnu .
|
|
For bug reporting instructions, please see <http www.gnu.org software gdb bugs >... Reading symbols from usr bin python...Reading symbols from usr lib debug usr bin python2.6...done.
|
|
done.
|
|
gdb) c The program is not being run.
|
|
gdb) run Starting program usr bin python c import\ numpy\;\ numpy.load\ file\ \ seqfaulting sleep stage data.npy\ \)\) [Thread debugging using libthread db enabled] Program received signal SIGSEGV, Segmentation fault.
|
|
0x00007ffff6755778 in PyArray Item XDECREF data=0xe3f960 xxdexae , descr=0x7ffff3bdc198) at numpy core src arrayobject.c 205 205 numpy core src arrayobject.c No such file or directory.
|
|
in numpy core src arrayobject.c gdb) bt 0 0x00007ffff6755778 in PyArray Item XDECREF data=0xe3f960 xxdexae , descr=0x7ffff3bdc198) at numpy core src arrayobject.c 205 1 0x00007ffff6755749 in PyArray Item XDECREF data=0xe3f960 xxdexae , descr=0x7ffff3bdc4f8) at numpy core src arrayobject.c 221 2 0x00007ffff675ddcb in PyArray XDECREF mp=0xb29040) at numpy core src arrayobject.c 308 3 0x00007ffff67719bc in array dealloc self=0xb29040) at numpy core src arrayobject.c 2127 4 0x00000000004a3631 in PyEval EvalFrameEx f=Frame 0x91c180, for file <string>, line 1, in <module> ), throwflag=<value optimized out>) at .. Python ceval.c 1010 5 0x00000000004a9671 in PyEval EvalCodeEx co=0x7ffff7ef6af8, globals=<value optimized out>, locals=<value optimized out>, args=0x0, argcount=<value optimized out>, kws=<value optimized out>, kwcount=0, defs=0x0, defcount=0, closure=0x0) at .. Python ceval.c 3000 6 0x00000000004a9742 in PyEval EvalCode co=0xe3f960, globals=<numpy.dtype at remote 0x7ffff3bdc198>, locals=0x0) at .. Python ceval.c 541 7 0x00000000004caae7 in run mod command=<value optimized out>, flags=<value optimized out>) at .. Python pythonrun.c 1339 8 PyRun StringFlags command=<value optimized out>, flags=<value optimized out>) at .. Python pythonrun.c 1302 9 PyRun SimpleStringFlags command=<value optimized out>, flags=<value optimized out>) at .. Python pythonrun.c 961 10 0x000000000041a268 in Py Main argc=1, argv=<value optimized out>) at .. Modules main.c 521 11 0x00007ffff69d8c4d in libc start main ) from lib libc.so.6 12 0x00000000004199f9 in start ) == 32 bit Ubuntu 10.04 == In [5] numpy.load file seqfaulting sleep stage data.npy )) Out[5] array [ 0.0, A ), 2.0, A ), 4.0, A ), ..., 64794.0, L ), 64796.0, L ), 64798.0, L )], dtype=[ time , |O4 ), value , |S1 )]) The offending file is attached.","bug"
|
|
"6013","2402","Instance attributes of ndarray subtypes get lost during pickling Trac 1809).
|
|
Original ticket http projects.scipy.org numpy ticket 1809 on 2011 04 27 by trac user tfmoraes, assigned to unknown.
|
|
Note see post below this for underlying problem.
|
|
Running the following code https gist.github.com 929168 in Windows I ve tested only in XP) a AttributeError exception happens, here a more detailed traceback https gist.github.com 944144 .
|
|
That code creates a new python process via multiprocessing to run a function, the parameter to this function is a memmap array created by the main process.
|
|
This code works with the last numpy stable version in Linux and Mac OS X, but in Windows.
|
|
Numpy version 1.4.1 runs correctly that code in Windows, Linux and Mac OS X.","bug"
|
|
"6013","2403","masked array with named components does not behave as expected Trac 1810).
|
|
Original ticket http projects.scipy.org numpy ticket 1810 on 2011 04 27 by trac user risa2000, assigned to pierregm.
|
|
In the included file is the log from IPython.
|
|
According to doc, assigning valid value should unmask the item > this is not happening see In Out 4,5 and 6,7 in the file) The mask is intact In Out 8) Manually changing mask In Out 9 10) seems to change the mask, but the unmasked value is arbitrary and not possible to change In Out 11 12 and 13 14).","bug"
|
|
"6013","2404","test fails test array too big test regression.TestRegression) Trac 1811).
|
|
Original ticket http projects.scipy.org numpy ticket 1811 on 2011 04 27 by stsci sienkiew, assigned to unknown.
|
|
====================================================================== ERROR test array too big test regression.TestRegression) Ticket 1080.
|
|
Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test regression.py , line 1157, in test array too big assert raises MemoryError, np.zeros, [2 30] 2, np.int8) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 1007, in assert raises return nose.tools.assert raises args, kwargs) File usr stsci pyssgdev Python 2.7 lib python2.7 unittest case.py , line 456, in assertRaises callableObj args, kwargs) ValueError array is too big.
|
|
This is in Python 2.7 on 32 bit Mac and 32 bit Red Hat Enterprise 4.
|
|
Works on Python 2.5, 2.6, 2.7 on 64 bit RHE 5.
|
|
Works on Python 2.7 on 64 bit Mac Snow Leopard.
|
|
Works on Python 2.7 on 64 bit RHE 4.
|
|
The line of the test that is erroring is assert raises MemoryError, np.zeros, [2 30] 2, np.int8)","bug"
|
|
"6013","2405","FAIL string test on win 32 Py 3.2 has dtype U5 instead of S5 Trac 1812).
|
|
Original ticket http projects.scipy.org numpy ticket 1812 on 2011 04 27 by rgommers, assigned to unknown.
|
|
====================================================================== FAIL Ticket 1748 Traceback most recent call last) File Z \Users\zouzoujing\.wine\drive c\Python32\lib\site packages\numpy\core\tests\test regression.py , line 1550, in test string astype assert equal b.dtype, np.dtype S5 )) File Z \Users\zouzoujing\.wine\drive c\Python32\lib\site packages\numpy\testing\utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL dtype <U5 ) DESIRED dtype |S5 ) Test issue or actual issue?","bug"
|
|
"6013","2406","percentile examples Trac 1813).
|
|
Original ticket http projects.scipy.org numpy ticket 1813 on 2011 04 28 by trac user nzarate, assigned to pv.
|
|
The percentile function do not give the posted answer; e.g.
|
|
np.percentile a, 0.5) should be np.percentile a, 50) Thanks nz","bug"
|
|
"6013","2407","dtype field names cannot be unicode Trac 1814).
|
|
Original ticket http projects.scipy.org numpy ticket 1814 on 2011 04 29 by jonovik, assigned to unknown.
|
|
Is there a reason why Unicode strings are not accepted as field names for record arrays?
|
|
>>> np.dtype [ a , int)]) dtype [ a , <i4 )]) >>> np.dtype [ u a , int)]) TypeError data type not understood A workaround is .encode ascii ).
|
|
>>> np.dtype [ u a .encode ascii ), int)]) dtype [ a , <i4 )]) It is okay for the type specification to be Unicode.
|
|
>>> np.dtype [ a , u S2 )]) dtype [ a , |S2 )])}}} I came across this while building a record array from data in a Microsoft Excel spreadsheet using pythoncom and win32com.
|
|
Converting to ascii isn t too much of a hassle, but maybe it wouldn t be difficult to allow Unicode strings as field names?","enhancement"
|
|
"6013","2408","Segmentation Fault when running numpy test Trac 1815).
|
|
Original ticket http projects.scipy.org numpy ticket 1815 on 2011 04 30 by trac user alex lai, assigned to unknown.
|
|
Hi experts, Ixe2x80x99m wondering if anyone would look into this issue.
|
|
While testing numpy package, Python crashed with the following error sbtorsvr391 cd home dcottr local tests sbtorsvr391 local tests export PYTHONPATH= PYTHONPATH local lib python3.1 site packages sbtorsvr391 local tests usr local bin python3 c import numpy; numpy.test ) Running unit tests for numpy NumPy version 1.5.1 NumPy is installed in home dcottr local lib python3.1 site packages numpy Python version 3.1.2 r312 79147, Mar 23 2010, 02 42 06) [GCC 3.4.6] nose version 1.0.0 ..S...............................................................................................................................................S........Warning invalid value encountered in isfinite ............................................................................................................................................................................................Warning invalid value encountered in isinf Warning invalid value encountered in isinf ......................................................................................................SSS........................K...........................................................K.....................................................................................................................................................F.F....................K.............................................................................................K...SK.F.......S...........................................................................................Segmentation Fault core dumped) Below is the stack trace from the core dump mdb core sbtorsvr391 python3 10439 5000 1304101376 11246 Loading modules [ libc.so.1 libavl.so.1 libuutil.so.1 ld.so.1 ] > > stack libc.so.1`strlen 0x50 fba55238, ffbf9ef8, ffbf9761, 0, 0, 0) libc.so.1`sprintf 0x40 ffbf9f18, 7fffffff, 7ffffc00, 2, 2, 1b74cc) test array from pyobj ext.so`array from pyobj 0x4e0 6, 17d3ec0, 1, 7, 16176c0, ff13a5a0) test array from pyobj ext.so`f2py rout wrap call 0xbc 1, 1, 0, 17f000, fffffffe, fba65b94) PyCFunction Call 0x90 15ff350, 1841900, 1844e90, 0, 1, 183df38) PyEval EvalFrameEx 0x4f10 0, ffbfa1c0, 1615770, 1, 16491b0, 15ff350) PyEval EvalCodeEx 0x874 1615770, 161ff60, 0, 160cf3c, 5, 0) function call 0x8c 162e270, 160cf30, 0, 17f000, fffffffe, 40) PyObject Call 0x44 162e270, 160cf30, 0, 160cf3c, 4, 16176c0) method call 0x8c 162e270, 18418d0, 0, 1849f60, 1, 3f) PyObject Call 0x44 176b620, 18418d0, 0, 2ed5c, 2efe0, 176b620) slot tp init 0x7c 176b620, 18418d0, 0, 2, 1, 16296d0) type call 0xdc 16521a0, 18418d0, 0, 17f000, fffffffe, 3e) PyObject Call 0x44 16521a0, 18418d0, 0, 18418d8, 1820030, 183ddbc) PyEval EvalFrameEx 0x37b4 0, ffbfa688, 162e9b0, 1, 16259c0, 16521a0) PyEval EvalFrameEx 0x5c98 0, ffbfa788, 16154e8, 1, 1626030, 162eb28) PyEval EvalFrameEx 0x5c98 0, ffbfa888, 52b260, 1, 57ec00, 162e468) PyEval EvalCodeEx 0x874 52b260, 5229c0, 0, 1808834, 2, 1ac03c) function call 0x8c 5a36a8, 1808828, 0, 17f000, fffffffe, 3a) PyObject Call 0x44 5a36a8, 1808828, 1825a50, 19b000, 163ab30, fe36c4) PyEval EvalFrameEx 0x13bc 0, fe36c8, 52b2f0, 1, 526d50, 1808828) PyEval EvalCodeEx 0x874 52b2f0, 1, 0, 1808c1c, 2, 0) function call 0x8c 5a3738, 1808c10, 0, 17f000, fffffffe, 38) PyObject Call 0x44 5a3738, 1808c10, 0, 1808c10, 1, 1849e50) method call 0x8c 5a3738, 1630450, 0, 1630440, 9, 37) PyObject Call 0x44 1612850, 1630450, 0, 193618, 2, 1612850) slot tp call 0x7c 163ab30, 1630450, 0, 17f000, fffffffe, 36) PyObject Call 0x44 163ab30, 1630450, 0, 1630458, 1849e50, fe355c) PyEval EvalFrameEx 0x37b4 0, ffbfaf78, b1e0f8, 1, a873a0, 163ab30) PyEval EvalFrameEx 0x5c98 0, ffbfb078, b1e0b0, 1, b1db58, b1ed20) PyEval EvalCodeEx 0x874 b1e0b0, b0e810, 0, 180b67c, 2, 1ac03c) function call 0x8c b1ecd8, 180b670, 0, 17f000, fffffffe, 33) PyObject Call 0x44 b1ecd8, 180b670, 18258a0, 19b000, 163abf0, fe2f3c) PyEval EvalFrameEx 0x13bc 0, fe2f40, b1ccc8, 1, b1a350, 180b670) PyEval EvalCodeEx 0x874 b1ccc8, 1, 0, 1844dac, 2, 0) function call 0x8c b1ea50, 1844da0, 0, 17f000, fffffffe, 31) PyObject Call 0x44 b1ea50, 1844da0, 0, 1844da0, 1, e4ddf0) method call 0x8c b1ea50, 16306d0, 0, 16306c0, b, 30) PyObject Call 0x44 176b5a8, 16306d0, 0, 1ab728, 18f708, 176b5a8) slot tp call 0x7c 163abf0, 16306d0, 0, 17f000, fffffffe, 2f) PyObject Call 0x44 163abf0, 16306d0, 0, 16306d8, e4ddf0, 18374d8) PyEval EvalFrameEx 0x37b4 0, ffbfb768, b2e4e8, 1, ad9ce8, 163abf0) PyEval EvalCodeEx 0x874 b2e4e8, b1b4b0, 0, 18087bc, 2, 1ac03c) function call 0x8c a00390, 18087b0, 0, 17f000, fffffffe, 2d) PyObject Call 0x44 a00390, 18087b0, 176c810, 19b000, 165c450, 182ea24) PyEval EvalFrameEx 0x13bc 0, 182ea28, b2e380, 1, b1ad50, 18087b0) PyEval EvalCodeEx 0x874 b2e380, 1, 0, 176b8ac, 2, 0) function call 0x8c a002b8, 176b8a0, 0, 17f000, fffffffe, 2b) PyObject Call 0x44 a002b8, 176b8a0, 0, 176b8a0, 1, e4ddf0) method call 0x8c a002b8, 16141b0, 0, 16141a0, 9, 2a) PyObject Call 0x44 1612c10, 16141b0, 0, 1ab728, 18f708, 1612c10) slot tp call 0x7c 165c450, 16141b0, 0, 17f000, fffffffe, 29) PyObject Call 0x44 165c450, 16141b0, 0, 16141b8, e4ddf0, 1645cf0) PyEval EvalFrameEx 0x37b4 0, ffbfbe58, b2e4e8, 1, ad9ce8, 165c450) PyEval EvalCodeEx 0x874 b2e4e8, b1b4b0, 0, 16296a4, 2, 1ac03c) function call 0x8c a00390, 1629698, 0, 17f000, fffffffe, 27) PyObject Call 0x44 a00390, 1629698, 161f8a0, 19b000, 176a270, 124868c) PyEval EvalFrameEx 0x13bc 0, 1248690, b2e380, 1, b1ad50, 1629698) PyEval EvalCodeEx 0x874 b2e380, 1, 0, 16127bc, 2, 0) > quit The C library used by Python is as follows sbtorsvr391 local tests ldd usr local bin python3|grep libc libc.so.1 => lib libc.so.1 platform SUNW,Sun Fire V490 lib libc psr.so.1 sbtorsvr391 ls l lib libc.so.1 rwxr xr x 1 root bin 1640776 Aug 10 2010 lib libc.so.1 sbtorsvr391 pkgchk l p lib libc.so.1 NOTE Couldn t lock the package database.
|
|
Pathname lib libc.so.1 Type regular file Expected mode 0755 Expected owner root Expected group bin Expected file size bytes) 1640776 Expected sum 1) of contents 50250 Expected last modification Aug 10 13 55 34 2010 Referenced by the following packages SUNWcslr Current status installed sbtorsvr391 pkginfo l SUNWcslr PKGINST SUNWcslr NAME Core Solaris Libraries Root) CATEGORY system ARCH sparc VERSION 11.10.0,REV=2005.01.21.15.53 BASEDIR VENDOR Sun Microsystems, Inc. DESC core software for a specific instruction set architecture PSTAMP on10 patch20100511083333 INSTDATE Jan 22 2011 16 10 HOTLINE Please contact your local service provider STATUS completely installed FILES 245 installed pathnames 2 shared pathnames 5 directories 133 executables 34303 blocks used approx) The same problem doesn t occur when the command is run as root.
|
|
I opened a ticket with Python support.
|
|
they responsed with the comment The issue is actually in numpy.
|
|
In numpy f2py src fortranobject.c, function array from pyobj ), there is sprintf mess, failed to initialize intent inout|inplace|cache) array input must be array but got s , PyString AsString PyObject Str PyObject Type obj))) ); In numpy headers, PyString AsString is aliased to PyBytes AsString.)
|
|
Besides the reference leaks, this function will fail when given the unicode string returned by PyObject Str ), and sprintf receives the null string...
|
|
Please report this to numpy.
|
|
I m hoping I will find answer to this issue here...
|
|
Thanks in advance, Alex","bug"
|
|
"6013","2409","numpy.double comparison with list is confusing Trac 1816).
|
|
Original ticket http projects.scipy.org numpy ticket 1816 on 2011 05 02 by trac user boz, assigned to unknown.
|
|
Comparison of a scalar numpy.double with iterables is a bit confusing.
|
|
Consider an example f = numpy.double 1.
|
|
3.)
|
|
if f == [] print expect this code ) But the print statement never runs, since comparison returns [] which is False.
|
|
At least for me this is counterintuitive.","bug"
|
|
"6013","2410","Slice assignment of chararrays broken in Python 3 Trac 1817).
|
|
Original ticket http projects.scipy.org numpy ticket 1817 on 2011 05 03 by trac user embray, assigned to unknown.
|
|
I m not sure if this is by design or not, but it s certainly surprising !python >>> charar = np.chararray 3, 3), itemsize=5) >>> charar[ ] = abcde >>> charar chararray [[b \xa8 \x15 8 , b \x00\x00\x00\xa8 , b \x15 8 ], [b \x00\xd0\xbc\x9b , b \x00\x00\x00\x00\xb0 , b \xac\x9b ], [b \x00\x003 , , b \x10=\xd2 ]], dtype= |S5 ) Works as intended on Python 2.x.","bug"
|
|
"6013","2411","gfortran missing shared flag when building extensions with numpy.distutils and fPIC missing for gfortran on x64 Trac 1818).
|
|
Original ticket http projects.scipy.org numpy ticket 1818 on 2011 05 04 by trac user winterstream, assigned to cournape.
|
|
I attempted to compile PyMC, which makes use of numpy.distutils to build its Fortran extensions.
|
|
The build first failed because gfortran was executed without the shared flag so there was a missing main).
|
|
I hacked this into numpy distutils fcompiler gnu.py 248 by changing linker so [ <F90> , Wall ], to linker so [ <F90> , Wall , shared ] I also hacked fPIC into the file by adding it to EXTRAFLAGS.
|
|
Neither of these hacks are solutions but I hope that it helps to pinpoint a fix.","bug"
|
|
"6013","2412","unexpected behaviour for numpy sum Trac 1819).
|
|
Original ticket http projects.scipy.org numpy ticket 1819 on 2011 05 05 by gerritholl, assigned to unknown.
|
|
numpy sum does not work as expected on a list of elements.
|
|
I think it is good if numpy s functions work where builtins work, but here it obviously doesn t In [97] a = rand 5, 1, 1); b = rand 1, 5, 1); c = rand 1, 5, 5) In [98] builtins .sum [a, b, c]) Out[98] array [[[ 1.34984882, 1.42768038, 0.7970086 , 1.26296369, 1.21887566], [ 0.74989493, 0.58476975, 0.78491321, 1.19311338, 0.4112359 ], [ 1.8563255 , 1.23005 , 1.8958019 , 1.90095001, 1.42403426], [ 1.22509658, 1.05595136, 1.06340261, 0.63567017, 1.30473052], [ 0.38608382, 0.67873255, 0.83798171, 0.81364562, 0.25934601]], [[ 1.4516518 , 1.52948336, 0.89881158, 1.36476667, 1.32067864], [ 0.85169791, 0.68657273, 0.88671619, 1.29491636, 0.51303888], [ 1.95812848, 1.33185298, 1.99760488, 2.00275299, 1.52583724], [ 1.32689956, 1.15775434, 1.16520559, 0.73747315, 1.4065335 ], [ 0.4878868 , 0.78053553, 0.93978469, 0.9154486 , 0.36114899]], [[ 2.23280273, 2.3106343 , 1.67996252, 2.1459176 , 2.10182958], [ 1.63284885, 1.46772367, 1.66786713, 2.0760673 , 1.29418982], [ 2.73927942, 2.11300391, 2.77875582, 2.78390393, 2.30698818], [ 2.1080505 , 1.93890528, 1.94635653, 1.51862408, 2.18768444], [ 1.26903773, 1.56168646, 1.72093562, 1.69659954, 1.14229993]], [[ 1.69520205, 1.77303361, 1.14236183, 1.60831692, 1.56422889], [ 1.09524816, 0.93012298, 1.13026644, 1.53846661, 0.75658913], [ 2.20167873, 1.57540323, 2.24115513, 2.24630324, 1.76938749], [ 1.57044981, 1.40130459, 1.40875584, 0.9810234 , 1.65008375], [ 0.73143704, 1.02408578, 1.18333494, 1.15899885, 0.60469924]], [[ 1.31299812, 1.39082968, 0.7601579 , 1.22611299, 1.18202496], [ 0.71304424, 0.54791905, 0.74806251, 1.15626269, 0.3743852 ], [ 1.8194748 , 1.1931993 , 1.85895121, 1.86409932, 1.38718357], [ 1.18824588, 1.01910067, 1.02655192, 0.59881947, 1.26787983], [ 0.34923312, 0.64188185, 0.80113101, 0.77679493, 0.22249532]]]) In [99] numpy.sum [a, b, c]) ValueError Traceback most recent call last) storage4 home gerrit svn faculteitsstuff spacemaster round 7 assistance <ipython console> in <module> ) storage4 home gerrit .local lib python2.6 site packages numpy core fromnumeric.pyc in sum a, axis, dtype, out) 1426 sum = a.sum 1427 except AttributeError > 1428 return wrapit a, sum , axis, dtype, out) 1429 return sum axis, dtype, out) 1430 storage4 home gerrit .local lib python2.6 site packages numpy core fromnumeric.pyc in wrapit obj, method, args, kwds) 35 except AttributeError 36 wrap = None > 37 result = getattr asarray obj),method) args, kwds) 38 if wrap 39 if not isinstance result, mu.ndarray) storage4 home gerrit .local lib python2.6 site packages numpy core numeric.pyc in asarray a, dtype, order) 282 283 > 284 return array a, dtype, copy=False, order=order) 285 286 def asanyarray a, dtype=None, order=None) ValueError setting an array element with a sequence.
|
|
In [101] numpy.version.version Out[101] 2.0.0.dev 12d0200 Should numpy s sum here behave like builtin sum?","bug"
|
|
"6013","2413","Memory leak in numpy.lib.npyio.NpzFile Trac 1820).
|
|
Original ticket http projects.scipy.org numpy ticket 1820 on 2011 05 05 by trac user benanne, assigned to unknown.
|
|
The del method of NpzFile is prone to causing memory leaks.
|
|
This problem just bit me pretty hard as I have a generator that loads up close to a million .npz files in turn.
|
|
I think it is related to what is mentioned here http utcc.utoronto.ca cks space blog python DebuggingPythonMemoryLeaks There are more obscure ways to hold references alive, including a reference cycle involving an object with a del method.
|
|
As the gc module mentions, Python can t pick an order to destroy things in so it just stuffs them into a holding list for your program to look at.
|
|
Someone on the python IRC channel on FreeNode also told me objects that implement del are never collected if they re in a reference cycle.
|
|
I think I ve identified this reference cycle to be self.f = BagObj self) in the init method of NpzFile.
|
|
The problem occurs in numpy 1.5.1, and also with the NpzFile BagObj code from trunk that I tried https github.com numpy numpy blob fd2630366581049af6c9acc4b41315363b47b55a numpy lib npyio.py I also found another version that doesn t have the del method, so it doesn t have the problem http svn.scipy.org svn numpy trunk numpy lib npyio.py I m presuming this is an old version from before git was used?
|
|
Getting rid of the del method in the current code fixes the problem, but then the own fid flag no longer works of course.
|
|
Luckily it is an acceptable fix in my case since I don t use this flag.
|
|
I have attached some code that demonstrates the problem.
|
|
I used it on an npz file of about 70kB containing 3 matrices also attached), but if my assumptions are correct the problem should occur with any npz file.","bug"
|
|
"6013","2414","numpy string arrays truncate trailing null bytes Trac 1821).
|
|
Original ticket http projects.scipy.org numpy ticket 1821 on 2011 05 06 by trac user saffsd, assigned to unknown.
|
|
>>> numpy.
|
|
version 1.3.0 >>> map len,numpy.array [ \x00 , \x00X , X\x00 , \x00\x00 ])) [0, 2, 1, 0] This bug prevents round trip behaviour for conversion from list >array >list >>> list numpy.array [ \x00 ])) == [ \x00 ] False","bug"
|
|
"6013","2415","loadtxt option to ensure 2D arrays returned from files with one line Trac 1822).
|
|
Original ticket http projects.scipy.org numpy ticket 1822 on 2011 05 06 by trac user brendanarnold, assigned to unknown.
|
|
See this discussion on the mailing list http mail.scipy.org pipermail numpy discussion 2010 June 051146.html The basic gist is that if you use loadtxt on a file that has one line of data it returns a 1D Numpy array np.squeeze is applied) but if the file has more than one lines of data a 2D array is returned.
|
|
This means operations such as ... data = np.loadtxt filename) x = data[ ,0] y = data[ ,1] Cannot be reliably performed.
|
|
How about a keyword argument, i.e.
|
|
data = np.loadtxt filename, squeeze=False)","enhancement"
|
|
"6013","2416","AttributeError module object has no attribute scipy Trac 1823).
|
|
Original ticket http projects.scipy.org numpy ticket 1823 on 2011 05 08 by trac user alex lai, assigned to unknown.
|
|
Hi experts, I m new to Python.
|
|
I ve just successfully compiled and installed NumPy and SciPy.
|
|
When I run the tests, Numpy shows no issues python3 c import numpy; numpy.test ) Running unit tests for numpy NumPy version 1.5.1 NumPy is installed in usr local lib python3.1 site packages numpy Python version 3.1.2 r312 79147, Mar 23 2010, 02 42 06) [GCC 3.4.6] nose version 1.0.0 Ran 0 tests in 0.027s However, test for Scipy gave the following error python3 c import numpy; import scipy; numpy.scipy ) Traceback most recent call last) File <string> , line 1, in <module> AttributeError module object has no attribute scipy Any idea what might be missing here?
|
|
Note to moderator I posted the question on this board since the module in this context is numpy) Thanks in advance, Alex","bug"
|
|
"6013","2417","Arctan2 failures on Solaris for special values Trac 1824).
|
|
Original ticket http projects.scipy.org numpy ticket 1824 on 2011 05 08 by rgommers, assigned to unknown.
|
|
Reported by Ilan Schell on Solaris ====================================================================== FAIL test zero nzero test umath.TestArctan2SpecialValues) Traceback most recent call last) File home demo master lib python2.7 site packages numpy core tests test umath.py , line 322, in test zero nzero assert almost equal ncu.arctan2 np.PZERO, np.NZERO), np.pi) File home demo master lib python2.7 site packages numpy testing utils.py , line 468, in assert almost equal raise AssertionError msg) AssertionError Arrays are not almost equal to 7 decimals ACTUAL 0.0 DESIRED 3.141592653589793 ====================================================================== FAIL test zero pzero test umath.TestArctan2SpecialValues) Traceback most recent call last) File home demo master lib python2.7 site packages numpy core tests test umath.py , line 328, in test zero pzero assert arctan2 isnzero np.NZERO, np.PZERO) File home demo master lib python2.7 site packages numpy core tests test umath.py , line 310, in assert arctan2 isnzero assert ncu.arctan2 x, y) == 0 and np.signbit ncu.arctan2 x, y))), arctan s, s) is s, not 0 x, y, ncu.arctan2 x, y)) AssertionError arctan 0.0, 0.0) is 0.0, not 0","bug"
|
|
"6013","2418","assert array less behavior with inf nan Trac 1825).
|
|
Original ticket http projects.scipy.org numpy ticket 1825 on 2011 05 08 by rgommers, assigned to unknown.
|
|
This passed with 1.5.1 but fails with 1.6.0 due to Inf fixes to assert array compare for functions testing near )equality.
|
|
np.testing.assert array less 2, np.inf) np.testing.assert array less np.array [ 0.911, 1.065, 1.325, 1.587]), np.inf) These functions pass butit s not clear that they should assert array less [1, np.inf], [2, np.inf]) assert array less [1, np.nan], [2, np.nan])","bug"
|
|
"6013","2419","numpy dtype incorrectly reports duplicate elements Trac 1826).
|
|
Original ticket http projects.scipy.org numpy ticket 1826 on 2011 05 10 by trac user shegler, assigned to unknown.
|
|
When using custom named data types, the wrong element is reported as the conflicting one.
|
|
The following snippet dt=numpy.dtype [ a , f8 ), a , f8 ), b , f8 ), ]) reports Traceback most recent call last) File dtype bug.py , line 8, in <module> b , f8 ), ValueError two fields with the same name It should, however, report the name which occurs twice, a in this case.","bug"
|
|
"6013","2420","Inconsistent output dtype in mixed scalar array operations Trac 1827).
|
|
Original ticket http projects.scipy.org numpy ticket 1827 on 2011 05 10 by trac user tihocan, assigned to unknown.
|
|
It looks like the resulting dtype of a operation between a scalar and an array can depend on the order of the arguments.
|
|
This seems wrong.
|
|
Tested with numpy 1.5.1 under Ubuntu 11.04 64 bit).
|
|
In [4] numpy.array 0, dtype= uint16 ) numpy.array [1], dtype= int8 )).dtype Out[4] dtype int16 ) In [6] numpy.array [1], dtype= int8 ) numpy.array 0, dtype= uint16 )).dtype Out[6] dtype int8 )","bug"
|
|
"6013","2421","f2py fails to compile KeyError coutput ) Trac 1828).
|
|
Original ticket http projects.scipy.org numpy ticket 1828 on 2011 05 12 by trac user Iceman, assigned to pearu.
|
|
When I try to compile any fortran file with f2py I get the next error message f2py c control.f90 m control running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building extension control sources f2py options [] f2py > tmp tmps7XyGJ src.linux i686 2.6 controlmodule.c creating tmp tmps7XyGJ creating tmp tmps7XyGJ src.linux i686 2.6 Reading fortran codes... Reading file control.f90 format free) Post processing... Block control Block control fortran Block control 1 Post processing stage 2)... Block control Block unknown interface Block control fortran Block control 1 Building modules... Building module control ... Constructing F90 module support for control fortran ... Constructing wrapper function control fortran.control 1 ... salidas = control 1 ent analogicas,ent digitales,[n analogicas,n digitales]) Traceback most recent call last) File usr local bin f2py , line 24, in <module> main ) File usr local lib python2.6 dist packages numpy f2py f2py2e.py , line 561, in main run compile ) File usr local lib python2.6 dist packages numpy f2py f2py2e.py , line 547, in run compile setup ext modules = [ext]) File usr local lib python2.6 dist packages numpy distutils core.py , line 186, in setup return old setup new attr) File usr lib python2.6 distutils core.py , line 152, in setup dist.run commands ) File usr lib python2.6 distutils dist.py , line 975, in run commands self.run command cmd) File usr lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File usr local lib python2.6 dist packages numpy distutils command build.py , line 37, in run old build.run self) File usr lib python2.6 distutils command build.py , line 135, in run self.run command cmd name) File usr lib python2.6 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File usr local lib python2.6 dist packages numpy distutils command build src.py , line 152, in run self.build sources ) File usr local lib python2.6 dist packages numpy distutils command build src.py , line 169, in build sources self.build extension sources ext) File usr local lib python2.6 dist packages numpy distutils command build src.py , line 334, in build extension sources sources = self.f2py sources sources, ext) File usr local lib python2.6 dist packages numpy distutils command build src.py , line 593, in f2py sources [ m ,ext name] f sources) File usr local lib python2.6 dist packages numpy f2py f2py2e.py , line 369, in run main ret=buildmodules postlist) File usr local lib python2.6 dist packages numpy f2py f2py2e.py , line 316, in buildmodules dict append ret[mnames[i]],rules.buildmodule modules[i],um)) File usr local lib python2.6 dist packages numpy f2py rules.py , line 1254, in buildmodule fn = os.path.join options[ buildpath ],vrd[ coutput ]) KeyError coutput Why I get this error message?
|
|
I am under Ubuntu 10.04 and I am compiling with gfortran.","bug"
|
|
"6013","2422","ComplexWarning exceptions swallowed Trac 1829).
|
|
Original ticket http projects.scipy.org numpy ticket 1829 on 2011 05 13 by pv, assigned to unknown.
|
|
This should raise ComplexWarning as an error import numpy as np, warnings warnings.simplefilter error , np.ComplexWarning) x = np.ones 2) x = 1j Instead, it only prints the warning message.
|
|
Apparently, something swallows the exceptions along this code path.
|
|
The following works as intended, though >>> x[...] = np.array 1j) Traceback most recent call last) File <stdin> , line 1, in <module> numpy.core.numeric.ComplexWarning Casting complex values to real discards the imaginary part","bug"
|
|
"6013","2423","example on numpy.histogram2d is broken Trac 1830).
|
|
Original ticket http projects.scipy.org numpy ticket 1830 on 2011 05 14 by samtygier, assigned to pv.
|
|
the matplotlib example in http docs.scipy.org doc numpy reference generated numpy.histogram2d.html plots the data with the axis the wrong way.
|
|
i think this is mostly the fault of imshow working in backwards coordinates.
|
|
the lines should be extent = [xedges[0], xedges[ 1], yedges[0], yedges[ 1]] plt.imshow H.transpose ), extent=extent, interpolation= nearest , origin= lower ) test script broken, as per numpy docs) x = np.array [2, 2, 1, 1, 0.2]) y = np.array [1, 1.1, 1, 1, 0.5]) H, xedges, yedges = np.histogram2d x,y, bins= 50, 50)) extent = [yedges[0], yedges[ 1], xedges[ 1], xedges[0]] plt.imshow H, extent=extent, interpolation= nearest ) plt.show ) test script fixed) x = np.array [2, 2, 1, 1, 0.2]) y = np.array [1, 1.1, 1, 1, 0.5]) H, xedges, yedges = np.histogram2d x,y, bins= 50, 50)) extent = [xedges[0], xedges[ 1], yedges[0], yedges[ 1]] plt.imshow H.transpose ), extent=extent, interpolation= nearest , origin= lower ) plt.show )","bug"
|
|
"6013","2424","linalg.eig and linalg.eigh yields different results on a symmetric matrix randomized signs) Trac 1831).
|
|
Original ticket http projects.scipy.org numpy ticket 1831 on 2011 05 15 by trac user tommy carstensen, assigned to pv.
|
|
The eigenvalues are the same, but some of the signs of the values in the eigenvectors are reversed.
|
|
Is it a bug, or am I not using it correctly?
|
|
symmetric matrix = [[ 92.2308943 0.66200473 77.97395776] [ 0.66200473 22.32583547 6.30294705] [ 77.97395776 6.30294705 168.44327022]] linalg.eig symmetric matrix) = [ 217.29080501 43.90403054 21.80516445] [[ 0.52896445 0.84363444 0.09207356] [ 0.02921532 0.12653319 0.99153206] [ 0.84814095 0.52179525 0.09157861]] linalg.eigh symmetric matrix) = [ 21.80516445 43.90403054 217.29080501] [[ 0.09207356 0.84363444 0.52896445] [ 0.99153206 0.12653319 0.02921532] [ 0.09157861 0.52179525 0.84814095]]","bug"
|
|
"6013","2425","python W error does not work in some cases Trac 1832).
|
|
Original ticket http projects.scipy.org numpy ticket 1832 on 2011 05 17 by trac user ondrej.marsalek, assigned to unknown.
|
|
I am trying to turn Python warnings to errors, but it does not work as expected with numpy.
|
|
The following two piece of code work python W error c import warnings; import numpy; warnings.warn test , numpy.core.numeric.ComplexWarning) Traceback most recent call last) File <string> , line 1, in <module> numpy.core.numeric.ComplexWarning test python W error c import warnings; import numpy; raise numpy.core.numeric.ComplexWarning Traceback most recent call last) File <string> , line 1, in <module> numpy.core.numeric.ComplexWarning while this does not i.e.
|
|
still produces just a warning) python W error c import numpy; x=numpy.ones 2); x =1j numpy.core.numeric.ComplexWarning Casting complex values to real discards the imaginary part This is with Python 2.6.6 and numpy 1.5.1, but also confirmed with Python 2.7.","bug"
|
|
"6013","2426","[Feature Request] Add Polyval2 and Polyfitweighted routines.
|
|
Trac 1833).
|
|
Original ticket http projects.scipy.org numpy ticket 1833 on 2011 05 17 by trac user jsmidt, assigned to unknown.
|
|
Numpy has the capacity to fit 1d polynomials with poly1d and polyval.
|
|
It would be nice if numpy could add 2D polynomial fitting like exists in Matlab with these Polyval2 and Polyfitweighted routines http www.mathworks.com matlabcentral fileexchange 13719 2d weighted polynomial fitting and evaluation Adding 2D polynomial fitting to weighted data would be very helpful.","enhancement"
|
|
"6013","2427","einsum fails for specific shapes Trac 1834).
|
|
Original ticket http projects.scipy.org numpy ticket 1834 on 2011 05 18 by trac user wieland, assigned to unknown.
|
|
Given two arrays d = 3 T = random.random 2,d,d)) W = random.random 2,d,d,i)) the following equality should always hold up to a rounding arrow) all einsum ij...,j... >i... ,T[0],W[0],dtype=float64) einsum ij...,j... >i... ,T[1],W[1],dtype=float64) einsum pij,pjts >its ,T,W,dtype=float64) < 1e 10) However, for d > 2 you always find i s such that the equality is not fulfilled.
|
|
More specifically d = 3, i = 911...1365 d = 8, I = 129..204, 257..341 Used Numpy 1.6 and Python 2.6.6.
|
|
Mailing list discussion http mail.scipy.org pipermail numpy discussion 2011 May 056340.html","bug"
|
|
"6013","2428","f2py error when using assumed shape arrays and parametrized kinds Trac 1835).
|
|
Original ticket http projects.scipy.org numpy ticket 1835 on 2011 05 18 by trac user lorenz, assigned to pearu.
|
|
I have a small problem when using both the new assumed shape support see ticket 2129) and kind statements via a parameter see ticket 2345).
|
|
Sorry it didn t spot this earlier!
|
|
Consider the following two files `precision.f90` module precision integer, parameter rk = 8 end module `test.f90` module test contains subroutine testfunc x, res) use precision implicit none real kind=rk), intent in) x ) real kind=rk), intent out) res integer i print , x = , x res = 0.0 do i = 1, size x) res = res x i) enddo end subroutine end module The compilation fails with > gfortran c precision.f90 > f2py build dir build c m test test.f90 precision.o fcompiler=gnu95 Reading .f2py f2cmap ... Mapping real kind=rk) to double Mapping integer kind=ik) to int Succesfully applied user defined changes from .f2py f2cmap running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building extension test sources f2py options [] adding build src.linux x86 64 2.7 fortranobject.c to sources.
|
|
adding build src.linux x86 64 2.7 to include dirs.
|
|
adding build src.linux x86 64 2.7 test f2pywrappers2.f90 to sources.
|
|
build src building npy pkg config files running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize Gnu95FCompiler Found executable afs mpa sys system MPA 6.00 usr bin gfortran customize Gnu95FCompiler using build ext building test extension compiling C sources C compiler gcc pthread fno strict aliasing g O2 DNDEBUG g fwrapv O3 Wall Wstrict prototypes fPIC compile options Ibuild src.linux x86 64 2.7 I usr common pdsoft appl python 2.7.1 amd64 lib python2.7 site packages numpy core include I usr common pdsoft appl python 2.7.1 amd64 include python2.7 c gcc build src.linux x86 64 2.7 fortranobject.c gcc build src.linux x86 64 2.7 testmodule.c compiling Fortran 90 module sources Fortran f77 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore fPIC O3 funroll loops Fortran f90 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall fno second underscore fPIC O3 funroll loops Fortran fix compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore Wall fno second underscore fPIC O3 funroll loops compile options Ibuild src.linux x86 64 2.7 I usr common pdsoft appl python 2.7.1 amd64 lib python2.7 site packages numpy core include I usr common pdsoft appl python 2.7.1 amd64 include python2.7 c extra options Jbuild Ibuild gfortran f90 test.f90 gfortran f90 precision.f90 compiling Fortran sources Fortran f77 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore fPIC O3 funroll loops Fortran f90 compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall fno second underscore fPIC O3 funroll loops Fortran fix compiler afs mpa sys system MPA 6.00 usr bin gfortran Wall ffixed form fno second underscore Wall fno second underscore fPIC O3 funroll loops compile options Ibuild src.linux x86 64 2.7 I usr common pdsoft appl python 2.7.1 amd64 lib python2.7 site packages numpy core include I usr common pdsoft appl python 2.7.1 amd64 include python2.7 c extra options Jbuild Ibuild gfortran f90 build src.linux x86 64 2.7 test f2pywrappers2.f90 build src.linux x86 64 2.7 test f2pywrappers2.f90 17.16 real kind=rk) res 1 Error Parameter rk at 1) has not been declared or is a variable, which does not reduce to a constant expression build src.linux x86 64 2.7 test f2pywrappers2.f90 19.16 real kind=rk) x f2py x d0) 1 Error Parameter rk at 1) has not been declared or is a variable, which does not reduce to a constant expression build src.linux x86 64 2.7 test f2pywrappers2.f90 17.16 real kind=rk) res 1 Error Parameter rk at 1) has not been declared or is a variable, which does not reduce to a constant expression build src.linux x86 64 2.7 test f2pywrappers2.f90 19.16 real kind=rk) x f2py x d0) 1 Error Parameter rk at 1) has not been declared or is a variable, which does not reduce to a constant expression The cause is the omission of an use precision statement in the interface declaration for `f2pywrap test testfunc` in the generated routine `f2pyinittest )` in `test f2pywrappers2.f90` !
|
|
f90 !
|
|
This file is autogenerated with f2py version 2) !
|
|
It contains Fortran 90 wrappers to fortran functions.
|
|
subroutine f2pywrap test testfunc x, res, f2py x d0) use test, only testfunc use precision real kind=rk) res integer f2py x d0 real kind=rk) x f2py x d0) call testfunc x, res) end subroutine f2pywrap test testfunc subroutine f2pyinittest f2pysetupfunc) interface subroutine f2pywrap test testfunc x, res, f2py x d0) real kind=rk) res integer f2py x d0 real kind=rk) x f2py x d0) end subroutine f2pywrap test testfunc end interface external f2pysetupfunc call f2pysetupfunc f2pywrap test testfunc) end subroutine f2pyinittest I.e.
|
|
a patch to it would look like build src.linux x86 64 2.7 test f2pywrappers2.f90 2011 05 18 15 33 47.000000000 0200 test f2pywrappers2.f90 2011 05 18 15 33 43.000000000 0200 14,6 14,7 subroutine f2pyinittest f2pysetupfunc) interface subroutine f2pywrap test testfunc x, res, f2py x d0) use precision real kind=rk) res integer f2py x d0 real kind=rk) x f2py x d0) I would suggest that all modules {{{use}}d by the original function should to be on the safe side also be stated in any interface declarations.
|
|
Note that this error is only present when using both assumed shape support and parametrized `kind` statements.
|
|
If `test.f90` is changed to ... subroutine testfunc x, n, res) use precision implicit none integer, intent in) n real kind=rk), intent in) x n) ... everything works as expected.
|
|
Regards, Lorenz","bug"
|
|
"6013","2429","dot ) fails for complex arguments Trac 1836).
|
|
Original ticket http projects.scipy.org numpy ticket 1836 on 2011 05 19 by trac user aickley, assigned to unknown.
|
|
Pre multiplying a complex matrix with a complex vector doesn t work right, namely x = array [1j, 0]) A = array [[1j, 0], [0, 0]]) print x.dot A)[0] yields essentially zero.
|
|
I also noticed that x.dot A.T) works correctly while A.T).dot x) also fails.
|
|
I m running an 32 bit Ubuntu 10.04.2 LTS with a numpy 1.6.0 built from source.
|
|
Actually I get the same error with the prepackaged ubuntu package 1.3.0) and if I build numpy 1.7.0 from source.","bug"
|
|
"6013","2430","einsum throws nan or freezes python for specific array shapes Trac 1837).
|
|
Original ticket http projects.scipy.org numpy ticket 1837 on 2011 05 19 by trac user wieland, assigned to unknown.
|
|
The examples below throw np.nan first three expressions) or freeze python.
|
|
However, it depends on the dimensions again.
|
|
If you put 2 instead of five for the first axis of A and B, the first two expressions evaluate correctly.
|
|
The last expression also evaluates in this case but gives the wrong result 8600832 instead of 11059200).
|
|
A = ones 2, 12, 4, 2, 200)) B = ones 2, 12, 12)) print einsum ijklm,ijklm,ijn,ijn > ,A,A,B,B) print einsum ijklm,ijklm,ijn > ,A,A,B) print einsum A,[0,2] range 5,8),A,[1,3] range 5,8),B,[0,3,4],B,[0,4,2],[])","bug"
|
|
"6013","2431","object < > structured type arrays regression Trac 1838).
|
|
Original ticket http projects.scipy.org numpy ticket 1838 on 2011 05 21 by mwiebe, assigned to unknown.
|
|
Reported by Mark Dickinson here http mail.scipy.org pipermail numpy discussion 2011 May 056375.html","bug"
|
|
"6013","2432","cannot access masked array rows with np.object dtype Trac 1839).
|
|
Original ticket http projects.scipy.org numpy ticket 1839 on 2011 05 24 by trac user risa2000, assigned to pierregm.
|
|
Using masked array with np.object type makes the rows with that type, which also contain masked values raise an exception.
|
|
Slightly modified sample from documentation follows In [22] from numpy import ma In [23] y = ma.masked array [ 1, 2 ), 3, 4 )], mask=[ 0, 0), 0, 1)], dtype=[ a , int), b , np.object)]) In [24] y[0] Out[24] 1, 2 ) In [25] y[1] ValueError Traceback most recent call last) f \Work\python\apps\icao\icao results.py in <module> ) > 1 2 3 4 5 C \Python27\lib\site packages\numpy\ma\core.pyc in getitem self, indx) 2946 return mvoid dout, mask=mask) 2947 if flatten mask mask).any ) > 2948 dout = mvoid dout, mask=mask) 2949 else 2950 return dout C \Python27\lib\site packages\numpy\ma\core.pyc in new self, data, mask, dtype, fill value) 5506 dtype = dtype or data.dtype 5507 data = ndarray ), dtype=dtype) > 5508 data[ )] = data 5509 data = data.view self) 5510 if mask is not nomask ValueError Setting void array with object members using buffer.
|
|
In [26] y Out[26] masked array data = [ 1, 2 ) 3, )], mask = [ False, False) False, True)], fill value = 999999, ?
|
|
), dtype = [ a , <i4 ), b , |O4 )])","bug"
|
|
"6013","2433",".fill ) writes to array even if flags.writable = False Trac 1840).
|
|
Original ticket http projects.scipy.org numpy ticket 1840 on 2011 05 25 by gerritholl, assigned to unknown.
|
|
Using a recent bleeding edge numpy I didn t test with older versions), an array can be written to with .fill ) even if X.flags.writeable is set to False.
|
|
That doesn t look like something that should happen on purpose.
|
|
In [1] X = rand 2, 2) In [2] X.flags.writeable = False In [3] X.fill 42) In [4] print X [[ 42.
|
|
42.]
|
|
[ 42.
|
|
42.]]
|
|
In [6] print numpy.version.full version 2.0.0.dev 3071eab In [7] print numpy.version.git revision 3071eab84b81ef6e0d157d46404c631547fed763","bug"
|
|
"6013","2434","numpy installation thru install requires directive issue Trac 1841).
|
|
Original ticket http projects.scipy.org numpy ticket 1841 on 2011 05 25 by trac user ohe, assigned to unknown.
|
|
Hello there, In a virtualenv, with python v2.7.1, i can t install numpy thru the install requires directive of setuptools distribute.
|
|
from setuptools import setup setup ... install requires=[ numpy ], ...) A `python setup.py install` gives me the following traceback running install running bdist egg running egg info writing requirements to UNKNOWN.egg info requires.txt writing UNKNOWN.egg info PKG INFO writing top level names to UNKNOWN.egg info top level.txt writing dependency links to UNKNOWN.egg info dependency links.txt reading manifest file UNKNOWN.egg info SOURCES.txt writing manifest file UNKNOWN.egg info SOURCES.txt installing library code to build bdist.macosx 10.6 intel egg running install lib warning install lib build lib does not exist no Python modules to install creating build bdist.macosx 10.6 intel creating build bdist.macosx 10.6 intel egg creating build bdist.macosx 10.6 intel egg EGG INFO copying UNKNOWN.egg info PKG INFO > build bdist.macosx 10.6 intel egg EGG INFO copying UNKNOWN.egg info SOURCES.txt > build bdist.macosx 10.6 intel egg EGG INFO copying UNKNOWN.egg info dependency links.txt > build bdist.macosx 10.6 intel egg EGG INFO copying UNKNOWN.egg info requires.txt > build bdist.macosx 10.6 intel egg EGG INFO copying UNKNOWN.egg info top level.txt > build bdist.macosx 10.6 intel egg EGG INFO zip safe flag not set; analyzing archive contents... creating dist UNKNOWN 0.0.0 py2.7.egg and adding build bdist.macosx 10.6 intel egg to it removing build bdist.macosx 10.6 intel egg and everything under it) Processing UNKNOWN 0.0.0 py2.7.egg Copying UNKNOWN 0.0.0 py2.7.egg to Users olivier tc karmatest lib python2.7 site packages Adding UNKNOWN 0.0.0 to easy install.pth file Installed Users olivier tc karmatest lib python2.7 site packages UNKNOWN 0.0.0 py2.7.egg Processing dependencies for UNKNOWN==0.0.0 Searching for numpy Reading http pypi.python.org simple numpy Reading http numpy.scipy.org Reading http sourceforge.net project showfiles.php?group id=1369 package id=175103 Reading http numeric.scipy.org Best match numpy 1.6.0 Downloading http sourceforge.net projects numpy files NumPy 1.6.0 numpy 1.6.0.tar.gz download Processing download Running numpy 1.6.0 setup.py q bdist egg dist dir var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 egg dist tmp Zr2lMr Running from numpy source directory.Warning distutils distribution has been initialized, it may be too late to add a subpackage commandWarning distutils distribution has been initialized, it may be too late to add a subpackage fcompilernon existing path in private var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils site.cfg Warning distutils distribution has been initialized, it may be too late to add a subpackage distutilsWarning distutils distribution has been initialized, it may be too late to add a subpackage testingWarning distutils distribution has been initialized, it may be too late to add a subpackage f2pyWarning distutils distribution has been initialized, it may be too late to add an extension sortWarning distutils distribution has been initialized, it may be too late to add an extension multiarrayWarning distutils distribution has been initialized, it may be too late to add an extension umathWarning distutils distribution has been initialized, it may be too late to add an extension scalarmathWarning distutils distribution has been initialized, it may be too late to add an extension dotblasWarning distutils distribution has been initialized, it may be too late to add an extension umath testsWarning distutils distribution has been initialized, it may be too late to add an extension multiarray testsWarning distutils distribution has been initialized, it may be too late to add a subpackage coreWarning distutils distribution has been initialized, it may be too late to add an extension compiled baseWarning distutils distribution has been initialized, it may be too late to add a subpackage libWarning distutils distribution has been initialized, it may be too late to add a subpackage oldnumericWarning distutils distribution has been initialized, it may be too late to add an extension capiWarning distutils distribution has been initialized, it may be too late to add a subpackage numarrayWarning distutils distribution has been initialized, it may be too late to add an extension fftpack liteWarning distutils distribution has been initialized, it may be too late to add a subpackage fftWarning distutils distribution has been initialized, it may be too late to add an extension lapack liteWarning distutils distribution has been initialized, it may be too late to add a subpackage linalgWarning distutils distribution has been initialized, it may be too late to add an extension mtrandWarning distutils distribution has been initialized, it may be too late to add a subpackage randomWarning distutils distribution has been initialized, it may be too late to add a subpackage maWarning distutils distribution has been initialized, it may be too late to add a subpackage matrixlibWarning distutils distribution has been initialized, it may be too late to add a subpackage compatWarning distutils distribution has been initialized, it may be too late to add a subpackage polynomialWarning distutils distribution has been initialized, it may be too late to add a subpackage docWarning distutils distribution has been initialized, it may be too late to add a subpackage numpyCould not locate executable f95 Could not locate executable f90 Could not locate executable f77 Could not locate executable xlf90 Could not locate executable xlf Could not locate executable ifort Could not locate executable ifc Could not locate executable g77 Found executable usr local bin gfortran configtest.c 1 warning conflicting types for built in function xe2x80x98expxe2x80x99 configtest.c 1 warning conflicting types for built in function xe2x80x98expxe2x80x99 configtest.c 1 20 error Python.h No such file or directory configtest.c 1 20 error Python.h No such file or directory lipo can t figure out the architecture type of var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp ccLgEqLk.out configtest.c 1 20 error Python.h No such file or directory configtest.c 1 20 error Python.h No such file or directory lipo can t figure out the architecture type of var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp ccLgEqLk.out Traceback most recent call last) File setup.py , line 4, in <module> install requires=[ numpy ] File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command install.py , line 76, in run File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command install.py , line 104, in do egg install File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 211, in run File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 427, in easy install File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 478, in install item File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 519, in process distribution File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg pkg resources.py , line 563, in resolve File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg pkg resources.py , line 799, in best match File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg pkg resources.py , line 811, in obtain File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 446, in easy install File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 476, in install item File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 655, in install eggs File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 930, in build and install File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command easy install.py , line 919, in run setup File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools sandbox.py , line 62, in run setup File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools sandbox.py , line 105, in run File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools sandbox.py , line 64, in <lambda> File setup.py , line 196, in <module> File setup.py , line 189, in setup package File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils core.py , line 186, in setup File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File Users olivier tc karmatest lib python2.7 site packages setuptools 0.6c11 py2.7.egg setuptools command bdist egg.py , line 167, in run File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils command egg info.py , line 8, in run File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils command build src.py , line 152, in run File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils command build src.py , line 169, in build sources File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils command build src.py , line 328, in build extension sources File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils command build src.py , line 385, in generate sources File private var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy core setup.py , line 410, in generate config h File private var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy core setup.py , line 41, in check types File private var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy core setup.py , line 271, in check types SystemError Cannot compile Python.h .
|
|
Perhaps you need to install python dev|python devel.
|
|
var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils misc util.py 251 RuntimeWarning Parent module numpy.distutils not found while handling absolute import Error in atexit.
|
|
run exitfuncs Traceback most recent call last) File Library Frameworks Python.framework Versions 2.7 lib python2.7 atexit.py , line 24, in run exitfuncs func targs, kargs) File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils misc util.py , line 251, in clean up temporary directory ImportError No module named numpy.distutils Error in sys.exitfunc Traceback most recent call last) File Library Frameworks Python.framework Versions 2.7 lib python2.7 atexit.py , line 24, in run exitfuncs func targs, kargs) File var folders Iy Iyse1OVUE38 IcPmyTb65E TI Tmp easy install W5KwbN numpy 1.6.0 numpy distutils misc util.py , line 251, in clean up temporary directory ImportError No module named numpy.distutils I have no probleme to install numpy with ` pip install numpy ` or ` easy install numpy ` inside or outside any virtualenv.
|
|
OS Mac OS 10.6.7","bug"
|
|
"6013","2435","inconsistency in RandomState Trac 1842).
|
|
Original ticket http projects.scipy.org numpy ticket 1842 on 2011 05 25 by trac user gtakacs, assigned to unknown.
|
|
The following code gives strange results import numpy d = 10 alpha = 1 d 0.5 mu = numpy.ones d) R = alpha numpy.ones d, d)) 1 alpha) numpy.eye d) rs = numpy.random.RandomState 587482) print rs.multivariate normal mu, R, 1)[0][0] Sometimes it prints 1.4183503917, other times 0.571882720736.
|
|
It should always print the same value, since RandomState is initialized with the seed value 587482.","bug"
|
|
"6013","2436","numpy.rec.array fails with open file on Py3 Trac 1843).
|
|
Original ticket http projects.scipy.org numpy ticket 1843 on 2011 05 26 by cgohlke, assigned to unknown.
|
|
As discussed on the [http mail.scipy.org pipermail numpy discussion 2011 May 056446.html mailing list], the following code raises a ValueError on Python 3, but works as expected with Python 2 import sys, numpy fd = open sys.executable, rb ) numpy.rec.array fd, dtype=[ , b )]) Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python32\lib\site packages\numpy\core\records.py , line 799, in array raise ValueError Unknown input type ) ValueError Unknown input type The problem seems to be the test for file objects in numpy compat py3k.py def isfileobj f) return isinstance f, io.FileIO) If this is updated to include io.BufferedReader, the type of `fd`, numpy.rec.array works with file objects opened in rb mode def isfileobj f) return isinstance f, io.FileIO, io.
|
|
io.BufferedReader)) I am not sure this is the best correct fix.
|
|
Tested with numpy 1.6.0 on Windows.","bug"
|
|
"6013","2437","1 D masked array numpy.ma) is corrupted by numpy.atleast 2d Trac 1844).
|
|
Original ticket http projects.scipy.org numpy ticket 1844 on 2011 05 26 by trac user doylejg, assigned to pierregm.
|
|
I construct a 1 dimensional masked array then create a second 2 D array from the original array and the original array is modified.
|
|
Best explained with an example shown below.
|
|
ar = numpy.ma.masked where [False,False,False,True],[1,2,3,99]) ar2 = numpy.atleast 2d ar) print ar Below is the traceback from the failed print line Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib pymodules python2.7 numpy ma core.py , line 3570, in repr data=str self), mask=str self.
|
|
mask), File usr lib pymodules python2.7 numpy ma core.py , line 3554, in str res[m] = f ValueError boolean index array should have 1 dimension","bug"
|
|
"6013","2438","The memory layout is wrong after sliced Trac 1845).
|
|
Original ticket http projects.scipy.org numpy ticket 1845 on 2011 05 27 by trac user ddev, assigned to unknown.
|
|
When you get a slice of an ndarray, it is a view into the memory of the array, so doing a setitem to the slice changes the parent array.
|
|
If however you pickle out a tuple of the parent array and the slice, the slice comes back as its own array, no longer pointing at the parent.
|
|
I ve seen this happen on Linux with Python 2.6.4 with several different versions of numpy up to 1.6.0.
|
|
Perhaps the pickling should throw an error if arrays s OWNDATA flag is not set?
|
|
The following code import numpy,pickle array = numpy.arange 2) slice = array[ 2] array ,slice = pickle.loads pickle.dumps array,slice))) slice[0] = 7 slice [0] = 7 print original , array , slice print pickled , array , slice Prints original [7 1] [7] pickled [0 1] [7]","bug"
|
|
"6013","2439","np.linalg.cond gives incorrect results for large ill conditioned matrices Trac 1846).
|
|
Original ticket http projects.scipy.org numpy ticket 1846 on 2011 05 27 by trac user rgrout, assigned to pv.
|
|
Hi, I notice the following behaviour.
|
|
In [31] np.linalg.cond np.vander np.arange 1,16))) Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Warning invalid value encountered in power Out[31] 1166991362111.4319 Where as Matlab and Octave agree on octave 2> cond vander 1 15)) ans = 2.5824e 21 Look at the source code, both numpy and matlab octave attempt to find the condition number in the same way by calculating the SVD).
|
|
However, the SVD that numpy returns is completely different than matlab octave s SVD.
|
|
When I give np.vander a small range np.arange 10)), numpy returns a condition number that agrees with octave.","bug"
|
|
"6013","2440","Inconsistent indexing of ndarray with lists Trac 1847).
|
|
Original ticket http projects.scipy.org numpy ticket 1847 on 2011 05 28 by trac user gmotron, assigned to unknown.
|
|
In a multi dimensional array, numpy crashes in iPython and python) when I use two lists as indexes.
|
|
The following code demonstrates this, first by showing many ways that work, and then combining the working methods to create a failure.
|
|
from numpy import a = 2 b = 50 c = 1000 data = zeros a,b,c) ) data[0, 0 b 5, ] data[0, , 0 c 10 ] data[0, 0 b 5, 0 c 10 ] bSub = range 0, b, 5) cSub = range 0, c, 10) data[0, bSub, ] data[0, , cSub] data[0, bSub, cSub] The crash output in iPython is home pcuser trunk python autotune bugExample.py in <module> ) 21 data[0, bSub, ] 22 data[0, , cSub] > 23 data[0, bSub, cSub] 24 25 ValueError shape mismatch objects cannot be broadcast to a single shape or in normal python Traceback most recent call last) File bugExample.py , line 23, in <module> data[0, bSub, cSub] ValueError shape mismatch objects cannot be broadcast to a single shape","bug"
|
|
"6013","2441","tril and triu return value has the wrong dtype Trac 1848).
|
|
Original ticket http projects.scipy.org numpy ticket 1848 on 2011 05 28 by WarrenWeckesser, assigned to unknown.
|
|
tril and triu, which defers to tril for documentation of its return value), says that the return value is the same type as the input.
|
|
This fails for float32 and complex64 input arrays In [16] a = array [[1,2],[3,4]], dtype=float32) In [17] tril a).dtype Out[17] dtype float64 ) In [18] b = array [[1 1j,2],[3,4]], dtype=complex64) In [19] tril b).dtype Out[19] dtype complex128 ) tril looks like this def tril m, k=0) m = asanyarray m) out = multiply tri m.shape[0], m.shape[1], k=k, dtype=int),m) return out and `tri m.shape[0], m.shape[1], k=k, dtype=int)` is an array of integer zeros and ones.
|
|
The problem is that the multiplication of a float32 array by an array of integers result in a float64 array In [168] a = array [[1.0, 2.0], [3.0, 4.0]], dtype=float32) In [169] m = array [[1, 1], [0, 1]]) In [170] a m Out[170] array [[ 1., 2.
|
|
], [ 0., 4.]])
|
|
In [171] m a Out[171] array [[ 1., 2.
|
|
], [ 0., 4.]])
|
|
Note that multiplying a float32 array by an integer scalar does not change the type In [172] a 3 Out[172] array [[ 3., 6.
|
|
], [ 9., 12.
|
|
]], dtype=float32) In [173] a array 3) Out[173] array [[ 3., 6.
|
|
], [ 9., 12.
|
|
]], dtype=float32) which seems inconsistent.
|
|
The problem with tril and triu was also pointed out by eat in 2318.","bug"
|
|
"6013","2442","nose 1.0.0 test failures on Mac OS X 10.6.7 and Xcode 4.0.2 using numpy 1.6.0 Trac 1849).
|
|
Original ticket http projects.scipy.org numpy ticket 1849 on 2011 05 29 by trac user elvisjohndowson, assigned to unknown.
|
|
Hi, I just installed numpy 1.6.0 and nose 1.0.0 on Mac OS X 10.6.7, using Xcode 4.0.2 and gfortran 4.2 5664 from the R tools site).
|
|
When I run the tests, I get 1 failure and 1 error.
|
|
====================================================================== ERROR Tests the min max functions with explicit outputs Traceback most recent call last) File System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy ma tests test core.py , line 699, in test minmax funcs with output result = npfunc xm,axis=0,out=nout) File System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy core fromnumeric.py , line 1569, in amin return amin axis, out) File System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy ma core.py , line 3119, in min np.putmask out, newmask, np.nan) ValueError cannot convert float NaN to integer ====================================================================== FAIL test umath.TestComplexFunctions.test against cmath Traceback most recent call last) File Library Python 2.6 site packages nose 1.0.0 py2.6.egg nose case.py , line 187, in runTest self.test self.arg) File System Library Frameworks Python.framework Versions 2.6 Extras lib python numpy core tests test umath.py , line 268, in test against cmath assert abs a b) < atol, s s s; cmath s fname,p,a,b) AssertionError arcsin 2 1.57079632679 1.31695789692j); cmath 1.57079632679 1.31695789692j) Ran 1740 tests in 11.679s FAILED KNOWNFAIL=1, errors=1, failures=1) <nose.result.TextTestResult run=1740 errors=1 failures=1> Here are the commands that I used to build numpy and nose Step 36.00 Install Numerical Python library.
|
|
Step 36.01 Download numpy 1.6.0 for python 2.6 from http sourceforge.net projects numpy files NumPy 1.6.0 numpy 1.6.0.tar.gz download cd numpy 1.6.0 env ARCHFLAGS= arch i386 arch x86 64 python setup.py build sudo env ARCHFLAGS= arch i386 arch x86 64 python setup.py install cd ..
|
|
Step 36.02 Download nose 1.0.0 to test the numpy installation from http somethingaboutorange.com mrl projects nose nose 1.0.0.tar.gz cd nose 1.0.0 env ARCHFLAGS= arch i386 arch x86 64 python setup.py build sudo env ARCHFLAGS= arch i386 arch x86 64 python setup.py install cd ..
|
|
Step 36.03 Test the numpy installation.
|
|
export VERSIONER PYTHON PREFER 32 BIT=yes python >>> import numpy >>> numpy.test 1 , 10 ) >>> exit )","bug"
|
|
"6013","2443","nose 1.0.0 error ld symbol s) not found for architecture x86 64 on Mac OS X 10.6.7, gfortran 4.2 5664, Xcode 4.0.2 5666 Trac 1850).
|
|
Original ticket http projects.scipy.org numpy ticket 1850 on 2011 05 29 by trac user elvisjohndowson, assigned to unknown.
|
|
Hi, I installed python 2.7 installer from the following location http python.org ftp python 2.7.1 python 2.7.1 macosx10.6.dmg rebuilt numpy 1.6.0 and nose 1.0.0, and get the following errors creating var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmp6HplH7 compile options I var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 I Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core include I Library Frameworks Python.framework Versions 2.7 include python2.7 c gfortran f90 var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmp6HplH7 foo.f90 gfortran f77 var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 test ext module 5403 f2pywrappers.f usr bin gfortran Wall var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 test ext module 5403module.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 fortranobject.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmp6HplH7 foo.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 test ext module 5403 f2pywrappers.o lgfortran o .
|
|
test ext module 5403.so Undefined symbols for architecture x86 64 PyNumber Int , referenced from int from pyobj in test ext module 5403module.o PyComplex Type , referenced from int from pyobj in test ext module 5403module.o PyType IsSubtype , referenced from int from pyobj in test ext module 5403module.o array from pyobj in fortranobject.o PyErr Occurred , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o PyErr SetString , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o ... PyObject GetAttrString , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o fortran repr in fortranobject.o PyErr Clear , referenced from int from pyobj in test ext module 5403module.o fortran repr in fortranobject.o F2PyDict SetItemString in fortranobject.o PySequence Check , referenced from int from pyobj in test ext module 5403module.o PySequence GetItem , referenced from int from pyobj in test ext module 5403module.o Py InitModule4 64 , referenced from init test ext module 5403 in test ext module 5403module.o PyType Type , referenced from init test ext module 5403 in test ext module 5403module.o PyImport ImportModule , referenced from init test ext module 5403 in test ext module 5403module.o PyCObject Type , referenced from init test ext module 5403 in test ext module 5403module.o F2PyCapsule Check in fortranobject.o PyExc RuntimeError , referenced from init test ext module 5403 in test ext module 5403module.o fortran call in fortranobject.o PyExc ImportError , referenced from init test ext module 5403 in test ext module 5403module.o PyErr Print , referenced from init test ext module 5403 in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o PyCObject AsVoidPtr , referenced from init test ext module 5403 in test ext module 5403module.o F2PyCapsule AsVoidPtr in fortranobject.o PyErr Format , referenced from init test ext module 5403 in test ext module 5403module.o fortran call in fortranobject.o PyExc AttributeError , referenced from init test ext module 5403 in test ext module 5403module.o fortran setattr in fortranobject.o PyModule GetDict , referenced from init test ext module 5403 in test ext module 5403module.o PyString FromString , referenced from init test ext module 5403 in test ext module 5403module.o fortran repr in fortranobject.o fortran getattr in fortranobject.o PyDict SetItemString , referenced from init test ext module 5403 in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o fortran setattr in fortranobject.o fortran getattr in fortranobject.o PyFortranObject New in fortranobject.o PyErr NewException , referenced from init test ext module 5403 in test ext module 5403module.o PyDict GetItemString , referenced from init test ext module 5403 in test ext module 5403module.o fortran getattr in fortranobject.o PyObject SetAttrString , referenced from init test ext module 5403 in test ext module 5403module.o Py NoneStruct , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o fortran getattr in fortranobject.o PyArg ParseTupleAndKeywords , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o Py BuildValue , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o PyExc TypeError , referenced from fortran call in fortranobject.o array from pyobj in fortranobject.o PyString AsString , referenced from fortran repr in fortranobject.o array from pyobj in fortranobject.o PyString FromFormat , referenced from fortran repr in fortranobject.o PyMem Free , referenced from fortran dealloc in fortranobject.o PyCObject FromVoidPtr , referenced from F2PyCapsule FromVoidPtr in fortranobject.o fortran getattr in fortranobject.o PyExc ValueError , referenced from array from pyobj in fortranobject.o PyObject Type , referenced from array from pyobj in fortranobject.o PyObject Str , referenced from array from pyobj in fortranobject.o PyObject New , referenced from PyFortranObject NewAsAttr in fortranobject.o PyFortranObject New in fortranobject.o PyDict New , referenced from PyFortranObject NewAsAttr in fortranobject.o fortran setattr in fortranobject.o PyFortranObject New in fortranobject.o PyDict DelItemString , referenced from fortran setattr in fortranobject.o Py FindMethod , referenced from fortran getattr in fortranobject.o PyString ConcatAndDel , referenced from fortran getattr in fortranobject.o MAIN , referenced from main in libgfortranbegin.a fmain.o) ld symbol s) not found for architecture x86 64 collect2 ld returned 1 exit status Undefined symbols for architecture x86 64 PyNumber Int , referenced from int from pyobj in test ext module 5403module.o PyComplex Type , referenced from int from pyobj in test ext module 5403module.o PyType IsSubtype , referenced from int from pyobj in test ext module 5403module.o array from pyobj in fortranobject.o PyErr Occurred , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o PyErr SetString , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o ... PyObject GetAttrString , referenced from int from pyobj in test ext module 5403module.o init test ext module 5403 in test ext module 5403module.o fortran repr in fortranobject.o PyErr Clear , referenced from int from pyobj in test ext module 5403module.o fortran repr in fortranobject.o F2PyDict SetItemString in fortranobject.o PySequence Check , referenced from int from pyobj in test ext module 5403module.o PySequence GetItem , referenced from int from pyobj in test ext module 5403module.o Py InitModule4 64 , referenced from init test ext module 5403 in test ext module 5403module.o PyType Type , referenced from init test ext module 5403 in test ext module 5403module.o PyImport ImportModule , referenced from init test ext module 5403 in test ext module 5403module.o PyCObject Type , referenced from init test ext module 5403 in test ext module 5403module.o F2PyCapsule Check in fortranobject.o PyExc RuntimeError , referenced from init test ext module 5403 in test ext module 5403module.o fortran call in fortranobject.o PyExc ImportError , referenced from init test ext module 5403 in test ext module 5403module.o PyErr Print , referenced from init test ext module 5403 in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o PyCObject AsVoidPtr , referenced from init test ext module 5403 in test ext module 5403module.o F2PyCapsule AsVoidPtr in fortranobject.o PyErr Format , referenced from init test ext module 5403 in test ext module 5403module.o fortran call in fortranobject.o PyExc AttributeError , referenced from init test ext module 5403 in test ext module 5403module.o fortran setattr in fortranobject.o PyModule GetDict , referenced from init test ext module 5403 in test ext module 5403module.o PyString FromString , referenced from init test ext module 5403 in test ext module 5403module.o fortran repr in fortranobject.o fortran getattr in fortranobject.o PyDict SetItemString , referenced from init test ext module 5403 in test ext module 5403module.o F2PyDict SetItemString in fortranobject.o fortran setattr in fortranobject.o fortran getattr in fortranobject.o PyFortranObject New in fortranobject.o PyErr NewException , referenced from init test ext module 5403 in test ext module 5403module.o PyDict GetItemString , referenced from init test ext module 5403 in test ext module 5403module.o fortran getattr in fortranobject.o PyObject SetAttrString , referenced from init test ext module 5403 in test ext module 5403module.o Py NoneStruct , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o fortran getattr in fortranobject.o PyArg ParseTupleAndKeywords , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o Py BuildValue , referenced from f2py rout test ext module 5403 flatten in test ext module 5403module.o f2py rout test ext module 5403 trans in test ext module 5403module.o f2py rout test ext module 5403 foo in test ext module 5403module.o PyExc TypeError , referenced from fortran call in fortranobject.o array from pyobj in fortranobject.o PyString AsString , referenced from fortran repr in fortranobject.o array from pyobj in fortranobject.o PyString FromFormat , referenced from fortran repr in fortranobject.o PyMem Free , referenced from fortran dealloc in fortranobject.o PyCObject FromVoidPtr , referenced from F2PyCapsule FromVoidPtr in fortranobject.o fortran getattr in fortranobject.o PyExc ValueError , referenced from array from pyobj in fortranobject.o PyObject Type , referenced from array from pyobj in fortranobject.o PyObject Str , referenced from array from pyobj in fortranobject.o PyObject New , referenced from PyFortranObject NewAsAttr in fortranobject.o PyFortranObject New in fortranobject.o PyDict New , referenced from PyFortranObject NewAsAttr in fortranobject.o fortran setattr in fortranobject.o PyFortranObject New in fortranobject.o PyDict DelItemString , referenced from fortran setattr in fortranobject.o Py FindMethod , referenced from fortran getattr in fortranobject.o PyString ConcatAndDel , referenced from fortran getattr in fortranobject.o MAIN , referenced from main in libgfortranbegin.a fmain.o) ld symbol s) not found for architecture x86 64 collect2 ld returned 1 exit status error Command usr bin gfortran Wall var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 test ext module 5403module.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 fortranobject.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmp6HplH7 foo.o var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG var folders kE kEV99wuDFGadcsS84CS4mE TI Tmp tmpaAXktG src.macosx 10.6 intel 2.7 test ext module 5403 f2pywrappers.o lgfortran o .
|
|
test ext module 5403.so failed with exit status 1 Ran 3548 tests in 42.107s FAILED KNOWNFAIL=3, SKIP=1, errors=17) <nose.result.TextTestResult run=3548 errors=17 failures=0> Elvis Dowson","bug"
|
|
"6013","2444","regression for SWIG based code in 1.6.0 Trac 1851).
|
|
Original ticket http projects.scipy.org numpy ticket 1851 on 2011 05 29 by rgommers, assigned to pearu.
|
|
Reported on the mailing list line 332, in build extension sources sources = self.swig sources sources, ext) File home bsesar usr pydebug lib python2.7 site packages numpy distutils command build src.py , line 646, in swig sources extension.swig opts.remove c ) AttributeError str object has no attribute remove The reason is that `swig opts` is a string.
|
|
This bug was introduced in [ff0822c4] http github.com numpy numpy commit ff0822c4) and can be fixed by inserting if isinstance extension.swig opts, basestring) extension.swig opts = extension.swig opts.split ) just above the if block where the error occurs.
|
|
Pearu, since you made that change, can you please check this fix is correct and complete?","bug"
|
|
"6013","2445","Complex number confusion Trac 1852).
|
|
Original ticket http projects.scipy.org numpy ticket 1852 on 2011 05 30 by trac user anirudh215, assigned to unknown.
|
|
If I run the following code Please don t analyze the code.
|
|
It isn t anything useful, but I ve only used it to illustrate the problem.)
|
|
import numpy as np h = np.ones 5, dtype=np.complex64) for p in range 5) h[p] = 1 5 j print h) I get an error.
|
|
It tells me the global variable j is not defined.
|
|
I obviously mean the complex number j.
|
|
It should recognize this.
|
|
However if I change the code to import numpy as np h = np.ones 5, dtype=np.complex64) for p in range 5) h[p] = 1 np.complex 0,5) print h) I get what I want.","bug"
|
|
"6013","2446","Complex Number error Trac 1853).
|
|
Original ticket http projects.scipy.org numpy ticket 1853 on 2011 05 31 by trac user anirudh215, assigned to unknown.
|
|
Hey guys, Sorry about this, but last night I was so sleepy that I made a mistake in what I wanted to say.
|
|
After posting a ticket, I just fell flat on my face to find the ticket has been closed already.
|
|
Thus, I m opening a new ticket with what I intended to say.
|
|
If I run the following code Please don t analyze the code.
|
|
It isn t anything useful, but I ve only used it to illustrate the problem.)
|
|
import numpy as np h = np.ones 5, dtype=np.complex64) for p in range 5) h[p] = 1 p j print h) I get an error.
|
|
It tells me the global variable j is not defined.
|
|
I obviously mean the complex number j.
|
|
It should recognize this.
|
|
However if I change the code to import numpy as np h = np.ones 5, dtype=np.complex64) for p in range 5) h[p] = 1 np.complex 0,p) print h) I get what I want.
|
|
Sorry about the confusion.
|
|
Thanks!","bug"
|
|
"6013","2447","Problem of convergence depending on platform Intel or Amd Trac 1854).
|
|
Original ticket http projects.scipy.org numpy ticket 1854 on 2011 05 31 by trac user aquassarim, assigned to unknown.
|
|
Hello to everyone, I expect my post to be relevant on this site since I am not sure of the source of my problem.
|
|
If not, please forgive me.
|
|
I face a problem with linalg in solving a non linear problem with a Newton Raphson scheme A x) x = b, with A an array of shape 40,40), x and b arrays of shape 40,).
|
|
Basically I do r x) = A x) x b dx = inv At x)) r x) x = dx and so on until r x) is small enaugth At is the tangent matrix of my Newton Raphson scheme).
|
|
On my work computer, my scheme achieves his goal it converges and find the solution) whereas on my personnal computer it diverges.
|
|
The code is exactly the same.
|
|
I use numpy.arrays with dtype= float64 , linalg.inv, np.dot, la.norm and that s all for my Newton Raphson.
|
|
The problem is of small dimension 40x40 matrix) and I obtain the same results with the use of a preconditionnor before solving.
|
|
Both my computers have 64 bits processors and OS professional Intel R) Core TM)2 Quad CPU Q9400 2.66GHz Linux 2.6.32 31 generic 61 Ubuntu SMP Fri Apr 8 18 25 51 UTC 2011 x86 64 GNU Linux Ubuntu 10.04 LTS) python 2.6.5, gcc 4.4.3 numpy 1.3.0 personnal AMD Phenom tm) II X4 955 Processor Linux 2.6.38 8 generic 42 Ubuntu SMP Mon Apr 11 03 31 24 UTC 2011 x86 64 x86 64 x86 64 GNU Linux Ubuntu 11.04) Python 2.7.1 , GCC 4.5.2 numpy 1.5.1 I do not understand where it can come from.
|
|
I also have to say that on another computer, with Amd 64 processor to and also running under Ubuntu 11.04, the calculation also fails...
|
|
Does someone have any idea on how to fix it on my Amd computer ?
|
|
Can it be a floating point precision difference ?
|
|
I checked the tangent matrix on both my computers, they are very closed the one from each other on the first iterations...
|
|
But it appears to be enought to make my calculation diverges on Amd...) Thanks ) !
|
|
When I try to use float32 instead of float64 both calculations fail.
|
|
When I try to use float128 I have an exception about a unsupported type in linalg.","bug"
|
|
"6013","2448","Numerical stable sum similar to math.fsum) Trac 1855).
|
|
Original ticket http projects.scipy.org numpy ticket 1855 on 2011 06 02 by trac user ling, assigned to unknown.
|
|
For some applications, having a numerical stable sum is critical.
|
|
What I mean can be well illustrated in the examples below >>> math.fsum [1, 1e 10] 1000), numpy.sum [1, 1e 10] 1000) 1000.0000001, 1000.0000001000171) >>> math.fsum [1, 1e100, 1, 1e100] 10000), numpy.sum [1, 1e100, 1, 1e100] 10000) 20000.0, 0.0) Here math.fsum is a numerical stable sum introduced in Python 2.6, which uses the Shewchuk algorithm seems to be described in http code.activestate.com recipes 393090 ).
|
|
Other complaints about the plain vanilla numpy.sum could be found in a recent thread http article.gmane.org gmane.comp.python.numeric.general 42756.
|
|
I hope numpy.sum can also implement something similar e.g., Shewchuk algorithm or Kahan summation algorithm or partial sum).
|
|
We could add an optional argument stable=True False to numpy.sum and let the user decides whether they want a more accuracy but potentially slower) version.","bug"
|
|
"6013","2449","Shift operator cycles for int32 and int64 Trac 1856).
|
|
Original ticket http projects.scipy.org numpy ticket 1856 on 2011 06 02 by trac user klasj, assigned to unknown.
|
|
The bitshift operations, both right and left cycles when they are shifted more bits than the length of the datatype.
|
|
This happens for both int32 and int64 but not for int16, unsigned integers works as expected.
|
|
In the example below version 1.6.1 is used but this is also present in at least 1.5.0.
|
|
In [1] import numpy as np In [2] np.right shift np.int32 8),32) Out[2] 8 In [3] np.right shift np.int64 8),64) Out[3] 8 In [4] np.right shift np.int16 8),16) Out[4] 0 In [5] np.left shift np.int32 8),32) Out[5] 8 In [6] np.left shift np.int64 8),64) Out[6] 8 In [7] np.left shift np.int16 8),16) Out[7] 524288 In [8] np.version.full version Out[8] 1.6.1.dev 08953ea","bug"
|
|
"6013","2450","Pip install into virtualenv doesn t work Trac 1857).
|
|
Original ticket http projects.scipy.org numpy ticket 1857 on 2011 06 03 by rgommers, assigned to rgommers.
|
|
From Matthew Brett I ve appended the result of a fresh python3.2 virtualenv and a pip install numpy .
|
|
RefactoringTool home mb312 .virtualenvs bare 32 build numpy build py3k numpy core defchararray.py Running from numpy source directory.Traceback most recent call last) File <string> , line 14, in <module> File home mb312 .virtualenvs bare 32 build numpy setup.py , line 211, in <module> setup package ) File home mb312 .virtualenvs bare 32 build numpy setup.py , line 204, in setup package configuration=configuration ) File home mb312 .virtualenvs bare 32 build numpy build py3k numpy distutils core.py , line 152, in setup config = configuration ) File home mb312 .virtualenvs bare 32 build numpy setup.py , line 151, in configuration config.add subpackage numpy ) File home mb312 .virtualenvs bare 32 build numpy build py3k numpy distutils misc util.py , line 972, in add subpackage caller level = 2) File home mb312 .virtualenvs bare 32 build numpy build py3k numpy distutils misc util.py , line 941, in get subpackage caller level = caller level 1) File home mb312 .virtualenvs bare 32 build numpy build py3k numpy distutils misc util.py , line 878, in get configuration from setup py config = setup module.configuration args) File numpy setup.py , line 5, in configuration config = Configuration numpy ,parent package,top path) File home mb312 .virtualenvs bare 32 build numpy build py3k numpy distutils misc util.py , line 713, in init raise ValueError r is not a directory package path,)) ValueError build py3k numpy is not a directory","bug"
|
|
"6013","2451","a function einmean based on einsum Trac 1858).
|
|
Original ticket http projects.scipy.org numpy ticket 1858 on 2011 06 07 by trac user wieland, assigned to unknown.
|
|
I suggest a new function einmean that is almost equal to einsum but divides by the number of elements it summed over.
|
|
Example >>> a = np.arange 25).reshape 5,5) >>> np.einsum ii , a) 60 >>> np.einsum ii , a) 5.
|
|
12 >>> np.einmean ii , a) 12","enhancement"
|
|
"6013","2452","Broadcasting fails for A = B Trac 1859).
|
|
Original ticket http projects.scipy.org numpy ticket 1859 on 2011 06 07 by gerritholl, assigned to unknown.
|
|
The shape of the rhs in an operation A = B is wrongly read, and thus the operation fails.
|
|
See code below.
|
|
python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import zeros >>> A1 = zeros 2, 1, 1)) >>> A2 = zeros 2,)) >>> A1 A2 works array [[[ 0., 0.
|
|
]], [[ 0., 0.]]])
|
|
>>> A1 = A2 fails Traceback most recent call last) File <stdin> , line 1, in <module> ValueError non broadcastable output operand with shape 2,1,1) doesn t match the broadcast shape 0,140018678770184,2) >>> A1 = A2 fails Traceback most recent call last) File <stdin> , line 1, in <module> ValueError non broadcastable output operand with shape 2,1,1) doesn t match the broadcast shape 0,0,2) >>> A1 = A2 fails Traceback most recent call last) File <stdin> , line 1, in <module> ValueError non broadcastable output operand with shape 2,1,1) doesn t match the broadcast shape 14885712,6881,2) >>> from numpy.version import full version, git revision >>> print full version 2.0.0.dev 3071eab >>> print git revision 3071eab84b81ef6e0d157d46404c631547fed763 The same for =, =, etc.","bug"
|
|
"6013","2453","einsum ... based broadcasting Trac 1860).
|
|
Original ticket http projects.scipy.org numpy ticket 1860 on 2011 06 07 by trac user wieland, assigned to unknown.
|
|
I and M. Wiebe) suggest a new function that generalizes the broadcasting that is so nicely implemented into einsum .
|
|
Consider the following example >>> A = np.arange 25).reshape 5,5) >>> B = np.arange 5) >>> np.einsum ij,j , A, B) array [ 30, 80, 130, 180, 230]) Here, einsum takes the product of every element A {ij}, multiplies with B {j} and then sums over j leaving i fixed.
|
|
So, two binary operations are at the heart of einsum, np.add and np.multiply.
|
|
In this logic we could rewrite einsum using a more general function broadcast op , einsum ij,j , A, B) = broadcast op ij,j , A, B, oper=[np.add, np.multiply]) With this notation we can consider any kind of binary operation to replace np.add and np.multiply.
|
|
As an example, consider np.add instead of np.multiply, >>> broadcast op ij,j , A, B, oper=[np.add, np.add]) array [ 20, 45, 70, 95, 120]) Equivalent but more cumbersome to write especially in higher dimensions !))
|
|
is >>> sum A B.reshape 1,5)),axis=1) array [ 20, 45, 70, 95, 120]) Note that you are spared from any reshapes.
|
|
One could also image to use np.power instead of np.multiply such that B is the vector of powers and we take all elements of A {ij} to the power of B j.
|
|
In some sense, you can see this broadcasting as a generalization of reduce to higher dimensions.","enhancement"
|
|
"6013","2454","documentation incorrect about python 3 support Trac 1861).
|
|
Original ticket http projects.scipy.org numpy ticket 1861 on 2011 06 07 by stsci sienkiew, assigned to pv.
|
|
The release notes for numpy 1.5 and 1.6 both claim python 3 support.
|
|
http new.scipy.org faq.html python version support says that neither numpy nor scipy are compatible with python 3.
|
|
This is the second google hit for numpy python3 and the first google hit for scipy python3 .
|
|
http scipy.org FAQ head fbdfa57ebeb7a730265d5281c32ce493e298171c says that numpy works only with python 2.4 and 2.5.
|
|
I was not able to find a statement about python 3 on numpy.scipy.org I think a clear statement about python 3 support is a good idea.
|
|
It is not clear to me where exactly it belongs in the organization of the web site, but Download, Getting Started, and Documentation are all good candidates.","bug"
|
|
"6013","2455","improvement of index notation in einsum Trac 1862).
|
|
Original ticket http projects.scipy.org numpy ticket 1862 on 2011 06 08 by trac user wieland, assigned to unknown.
|
|
I suggest some small and very natural improvements on the exceptional index notation of einsum.
|
|
1) Given an N dimensional array A and a 1 dimensional array B, the following code throws an error einsum ij...,j >ij... ,A,B) Instead one has to write an empty) ellipsis for B, einsum ij...,j... >ij... ,A,B) which just obscures the real index summation.
|
|
2) At the moment, all indices to sum over have to be made explicit.
|
|
However, I claim that einsum ij...,j... >ij ,A,B) is.
|
|
as an example, naturally equivalent to einsum ijts,jts >ij ,A,B) if the dimensions of A and B are appropriate).
|
|
In the same way, in einsum ij,j... >ij ,A,B) you sum over the other axes of B.","enhancement"
|
|
"6013","2456","Buggy results when operating on array copied with astype ) Trac 1863).
|
|
Original ticket http projects.scipy.org numpy ticket 1863 on 2011 06 08 by trac user embray, assigned to unknown.
|
|
I m getting really buggy results from any mathematical operation I do on an array of float64 that was obtained by calling `arr.astype float64 )` on an existing array of float64s.
|
|
Here s an example that illustrates the problem the fft stuff is irrelevant; it just creates an array similar to the one my code is bugging out on) !python import numpy from numpy import fft shape = 32, 16) type = numpy.complex64 true = numpy.arange shape[0] shape[1], dtype=type) true.shape = shape y = fft.fft true, shape[0], 0) y = y.imag z = y.astype numpy.float64) print y 2)[3] print z 2)[3] This particular example also throws a `RuntimeWarning overflow encountered in square` when squaring `z`, the copy of the original array `y`.
|
|
This just started happening recently using the latest from git.
|
|
I m betting it has something to do with [https github.com numpy numpy commit a17a4996e4ed63d1b855a0917fb5fcdd5855a7d0 this commit] that made changes to astype ), but I haven t tracked down the exact issue yet.","bug"
|
|
"6013","2457","arange using float for step and integer dtype results in an array with all values being the same Trac 1864).
|
|
Original ticket http projects.scipy.org numpy ticket 1864 on 2011 06 10 by bsouthey, assigned to unknown.
|
|
When using a float step but a integer dtype, np.arange appears to full the array with a constant value first value in the range) rather than an integer array of different integers or giving an error.
|
|
The following code illustrates the problem >>> import numpy as np >>> np.
|
|
version 2.0.0.dev a1e7be3 >>> np.arange 0.,5.,0.5) array [ 0. , 0.5, 1. , 1.5, 2. , 2.5, 3. , 3.5, 4. , 4.5]) >>> np.arange 0.,5.,0.5, dtype=int) array [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) >>> np.asarray np.arange 0.,5.,0.5), dtype=int) array [0, 0, 1, 1, 2, 2, 3, 3, 4, 4])","bug"
|
|
"6013","2458","test datetime.TestDateTime errors Trac 1865).
|
|
Original ticket http projects.scipy.org numpy ticket 1865 on 2011 06 13 by stsci sienkiew, assigned to unknown.
|
|
Started doing this on RHE 5 64 bit, python 2.7, on Jun 8.
|
|
====================================================================== ERROR test datetime add test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 459, in test datetime add np.datetime64 2012 12 21 , [D] ), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime as string test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 880, in test datetime as string assert equal np.datetime as string np.datetime64 date, Y )), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime divide test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 635, in test datetime divide np.datetime64 2012 12 21 , [D] ), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime multiply test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 594, in test datetime multiply np.datetime64 2012 12 21 , [D] ), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime nat casting test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 163, in test datetime nat casting b = np.datetime64 NaT , [D] ) SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime scalar construction test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 78, in test datetime scalar construction assert equal np.datetime64 1950 03 12 , D ), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test datetime subtract test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 532, in test datetime subtract np.datetime64 2012 12 21 , [D] ), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test different unit comparison test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 371, in test different unit comparison assert equal np.datetime64 1945 , unit1), SystemError NULL result without error in PyObject Call ====================================================================== ERROR test string parser variants test datetime.TestDateTime) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy core tests test datetime.py , line 787, in test string parser variants assert equal np.datetime64 1977 03 02T12 30 0230 ), SystemError NULL result without error in PyObject Call Checked out on Jun 8 20 00 EDT.
|
|
Version 76ca55f8f399a9eb8081b75c3c9c2ac075a4af9a through a1e7be3cb49577d77831e2246f39bb80acd5a22f","bug"
|
|
"6013","2459","Cannot vectorize fortran function Trac 1866).
|
|
Original ticket http projects.scipy.org numpy ticket 1866 on 2011 06 14 by trac user aragilar, assigned to unknown.
|
|
When I attempt to call numpy.vectorize on a function in a module generated by f2py, I get the following error Traceback most recent call last) File .
|
|
graph.py , line 40, in <module> print calc alpha data) File .
|
|
graph.py , line 31, in calc alpha lambdas = np.vectorize flambda) m mss, star cat) File usr lib pymodules python2.6 numpy lib function base.py , line 1769, in init nin, ndefault = get nargs pyfunc) File usr lib pymodules python2.6 numpy lib function base.py , line 1701, in get nargs failed to determine the number of arguments for s obj)) ValueError failed to determine the number of arguments for <fortran flambda> The version of numpy I am using is from debian 1.5.1 2.b1 unstable).","bug"
|
|
"6013","2460","numpy.absolute ) can return a negative value Trac 1867).
|
|
Original ticket http projects.scipy.org numpy ticket 1867 on 2011 06 14 by trac user phil.hodge, assigned to unknown.
|
|
On a 32 bit Linux machine, when the input to the np.absolute ) function is a big endian, floating point array, the output is the same as the input, even if the input array contains negative values.
|
|
Here is an example {{{>>> print np.
|
|
version 2.0.0.dev Unknown >>> x = np.array [ 1., 0., 1.
|
|
], dtype= >f8 ) >>> print np.absolute x) [ 1.00000000e 00 0.00000000e 00 1.00000000e 00] >>> print np.absolute x) x [ 0.
|
|
0.
|
|
0.]}}}","bug"
|
|
"6013","2461","datetime fixes break numpy on python2.4 Trac 1868).
|
|
Original ticket http projects.scipy.org numpy ticket 1868 on 2011 06 15 by dhomeier, assigned to unknown.
|
|
Since merging of the datetime fixes [commits a827436 does not build f986fd4] numpy fails to load in python2.4, at least on MacOS X i386 ppc Python 2.4.4 1, Jan 5 2011, 03 05 41) [GCC 4.0.1 Apple Inc. build 5493)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File Users derek lib python2.4 site packages numpy init .py , line 137, in ?
|
|
import add newdocs File Users derek lib python2.4 site packages numpy add newdocs.py , line 9, in ?
|
|
from numpy.lib import add newdoc File Users derek lib python2.4 site packages numpy lib init .py , line 4, in ?
|
|
from type check import File Users derek lib python2.4 site packages numpy lib type check.py , line 8, in ?
|
|
import numpy.core.numeric as nx File Users derek lib python2.4 site packages numpy core init .py , line 5, in ?
|
|
import multiarray ImportError Failure linking new module Users derek lib python2.4 site packages numpy core multiarray.so Symbol not found PyErr WarnEx Referenced from Users derek lib python2.4 site packages numpy core multiarray.so Expected in dynamic lookup I reckon the solution can be found in this changeset [http projects.scipy.org numpy changeset 7996]","bug"
|
|
"6013","2462","np.place produce bad data with structured array columns as input Trac 1869).
|
|
Original ticket http projects.scipy.org numpy ticket 1869 on 2011 06 16 by trac user ged, assigned to unknown.
|
|
It silently produce bad results, using data from all the array columns, instead of just the one column I pass as argument.
|
|
with a being a structured array with several columns let s say x and y), np.place arr, mask, a[ x ]) is roughly equal to np.place arr, mask, a) The problem is present in at least versions 1.4.1 to 1.6.0.
|
|
See attached test case for details.","bug"
|
|
"6013","2463","Cannot assign a list of lists to dtype=object ndarray if sublists are equal length Trac 1870).
|
|
Original ticket http projects.scipy.org numpy ticket 1870 on 2011 06 16 by trac user mjuric, assigned to unknown.
|
|
This snippet illustrates the problem x = np.zeros 4, dtype=object) y = [[1], [2], [3], [4]] x[ ] = y In versions prior to 1.6, this code as expected) assigned the sublists of y to x so that, for example, x[2] would become [2]).
|
|
In 1.6, this code raises an exception ValueError output operand requires a reduction, but reduction is not enabled However, if y is not a list of lists with equal number of elements x = np.zeros 4, dtype=object) y = [[1], [2, 3], [3], [4]] x[ ] = y no exception is raised.","bug"
|
|
"6013","2464","f2py failures with empty strings on Python3 MacOS X ppc Trac 1871).
|
|
Original ticket http projects.scipy.org numpy ticket 1871 on 2011 06 19 by dhomeier, assigned to pearu.
|
|
1.6.1rc1 fails the following tests on 10.5 ppc gfortran 4.2 4.4) >>> numpy.test full ) Running unit tests for numpy NumPy version 1.6.1rc1 NumPy is installed in sw lib python3.2 site packages numpy Python version 3.2 r32 88445, Mar 1 2011, 18 28 16) [GCC 4.0.1 Apple Inc. build 5493)] nose version 1.0.0 ... ====================================================================== FAIL test return character.TestF77ReturnCharacter.test all Traceback most recent call last) File sw lib python3.2 site packages nose case.py , line 188, in runTest self.test self.arg) File sw lib python3.2 site packages numpy f2py tests test return character.py , line 78, in test all self.check function getattr self.module, name)) File sw lib python3.2 site packages numpy f2py tests test return character.py , line 12, in check function r = t array ab ));assert r==asbytes a ),repr r)) File sw lib python3.2 site packages numpy testing utils.py , line 34, in assert raise AssertionError msg) AssertionError b ====================================================================== FAIL test return character.TestF90ReturnCharacter.test all Traceback most recent call last) File sw lib python3.2 site packages nose case.py , line 188, in runTest self.test self.arg) File sw lib python3.2 site packages numpy f2py tests test return character.py , line 136, in test all self.check function getattr self.module.f90 return char, name)) File sw lib python3.2 site packages numpy f2py tests test return character.py , line 12, in check function r = t array ab ));assert r==asbytes a ),repr r)) File sw lib python3.2 site packages numpy testing utils.py , line 34, in assert raise AssertionError msg) AssertionError b Both errors are raised on the first function tested, t0 I suggest to change the test to r = t ab );assert r==asbytes a ), s ab )= s tname,`r`)) for better diagnostic.","bug"
|
|
"6013","2465","Problem with matrix dtype= complex ) in Python 2.5 Trac 1872).
|
|
Original ticket http projects.scipy.org numpy ticket 1872 on 2011 06 20 by trac user asmeurer, assigned to unknown.
|
|
If I put the following in a file called numpyfloattest.py import numpy class A object) def float self) return 2.0 def complex self) return 5.0j a = A ) print numpy.matrix [[a]], dtype= complex ) I get python2.5 numpyfloattest.py [[ 2.
|
|
0.j]] python2.6 numpyfloattest.py [[ 0.
|
|
5.j]] python2.7 numpyfloattest.py [[ 0.
|
|
5.j]] In other words, for some reason, in Python 2.5 it is calling float ) on the data even though the dtype is complex.
|
|
As you can imagine, this causes problems when the data is an object that represents a complex number.
|
|
In particular, this bug is causing problems with SymPy code that tries to convert the SymPy matrix [[0, I], [I, 0]] to a numpy matrix I is sympy.I, i.e., sqrt 1)).
|
|
See https groups.google.com d topic sympy 9FOPjxC0D6s discussion.","bug"
|
|
"6013","2466","Recarrays do not work with numpy.lib.stride tricks.as strided Trac 1873).
|
|
Original ticket http projects.scipy.org numpy ticket 1873 on 2011 06 21 by trac user irwin, assigned to unknown.
|
|
It seems that as strided does not work as it should for recarrays.
|
|
I think the issue is that the named fields of a recarray do not survive being run through as strided.
|
|
For example, the second print statement below is not able to output the array using numpy 1.6 import numpy import numpy.lib.stride tricks t = numpy.dtype { names [ a , b ], formats [numpy.float64, numpy.float64]}, align = True) a = numpy.zeros 10, dtype = t) print a print numpy.lib.stride tricks.as strided a) I had a look at stride tricks.py, but I am not familiar enough with array interface to suggest a fix.
|
|
I would be grateful if someone else is able to resolve this.","bug"
|
|
"6013","2467","f2py include path option not functional Trac 1874).
|
|
Original ticket http projects.scipy.org numpy ticket 1874 on 2011 06 21 by pearu, assigned to rgommers.
|
|
See http cens.ioc.ee pipermail f2py users 2011 June 002231.html for the bug report.
|
|
A fix is coming up soon.","bug"
|
|
"6013","2468","hex output format changed in Python3 Trac 1875).
|
|
Original ticket http projects.scipy.org numpy ticket 1875 on 2011 06 21 by dhomeier, assigned to unknown.
|
|
test arrayprint.py fails on python3.x with FAIL Test custom format function for each element in array.
|
|
Traceback most recent call last) File Users derek lib python3.2 site packages numpy core tests test arrayprint.py , line 86, in test format function [0x0L 0x1L 0x2L] ) File Users derek lib python3.2 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL [0x0 0x1 0x2] DESIRED [0x0L 0x1L 0x2L] Apparently the hex ) and oct ) functions have been decided to print out a format without the trailing L as of python3 [http mail.python.org pipermail python dev 2006 June 065946.html], so I suggest to define the expected result depending on the version see patch).","bug"
|
|
"6013","2469","dot ) non symmetrical for scalar matrix of objects Trac 1876).
|
|
Original ticket http projects.scipy.org numpy ticket 1876 on 2011 06 21 by lebigot, assigned to unknown.
|
|
NumPy s dot ) function has a non symmetrical behavior for scalar multipliers, when one of the objects is a matrix of objects >>> arr = matrix [1, 2], dtype=object) >>> dot arr, 3) matrix [[3, 6]], dtype=object) >>> dot 3, arr) NotImplemented My implementation of NumPy uses BLAS, but I think I remember also seeing this problem with the basic implementation.)
|
|
A consequence of this is the following failed simple multiplication >>> arr 3 matrix [[3, 6]], dtype=object) >>> 3 arr Traceback most recent call last) File <ipython console> , line 1, in <module> TypeError unsupported operand type s) for int and matrix The final TypeError looks definitely like a bug.
|
|
Mathematically, one also wants to have a symmetrical dot ) product, for scalars both dot 3, arr) and dot arr, 3) should work), no?
|
|
A symmetrical dot ) product for matrices of objects would solve both problems.","bug"
|
|
"6013","2470","Incorrect RuntimeWarning on first call to numpy.where Trac 1877).
|
|
Original ticket http projects.scipy.org numpy ticket 1877 on 2011 06 21 by trac user deil, assigned to unknown.
|
|
On my Mac OS 10.6 with python 2.7 and numpy 1.6.0 installed via macports I get the following RunTimeWarning on the first call to numpy.where, but not on subsequent calls python Python 2.7.1 r271 86832, May 15 2011, 23 17 21) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.where [False, True],np.log [0, 1]), 42) main 1 RuntimeWarning divide by zero encountered in log array [ 42., 0.])
|
|
>>> np.where [False, True],np.log [0, 1]), 42) array [ 42., 0.])
|
|
>>> np.
|
|
version 1.6.0 This warning occurs when calling np.log 0), which here shouldn t have been invoked, right?
|
|
On a Linux box I have I don t get the warning using the same commands.","bug"
|
|
"6013","2471","numpy.argmin ) fails to find the minimum for the most negative integer Trac 1878).
|
|
Original ticket http projects.scipy.org numpy ticket 1878 on 2011 06 22 by trac user tlatorre, assigned to unknown.
|
|
!python >>> a = np.array [ 2 7, 2 7 1, 2 7 2], dtype=np.int8) >>> a array [ 128, 127, 126], dtype=int8) >>> np.argmin a) 1 >>> b = np.array [ 2 15, 2 15 1, 2 15 2], dtype=np.int16) >>> b array [ 32768, 32767, 32766], dtype=int16) >>> np.argmin b) 1 >>> c = np.array [ 2 31, 2 31 1, 2 31 2], dtype=np.int32) >>> c array [ 2147483648, 2147483647, 2147483646], dtype=int32) >>> np.argmin c) 1","bug"
|
|
"6013","2472","bug of numpy.ndarray.ndim Trac 1879).
|
|
Original ticket http projects.scipy.org numpy ticket 1879 on 2011 06 25 by trac user coneagoe, assigned to unknown.
|
|
1) It is OK, an array is created without argument dtype .
|
|
While ndarray.ndim will be zero if the argument is specified.
|
|
>>> y = numpy.array [1, 2, 3]) >>> y array [1, 2, 3]) >>> y.ndim 1 >>> y = numpy.array 1, 2, 3), dtype = [ a , numpy.int), b , numpy.int), c , numpy.int)]) >>> y array 1, 2, 3), dtype=[ a , <i4 ), b , <i4 ), c , <i4 )]) >>> y.ndim 0","bug"
|
|
"6013","2473","FAIL test timedelta scalar construction Trac 1880).
|
|
Original ticket http projects.scipy.org numpy ticket 1880 on 2011 06 26 by nilswagner01, assigned to unknown.
|
|
====================================================================== FAIL test timedelta scalar construction test datetime.TestDateTime) Traceback most recent call last) File home nwagner local lib64 python2.6 site packages numpy core tests test datetime.py , line 189, in test timedelta scalar construction assert equal str np.timedelta64 3, s )), 3 seconds ) File home nwagner local lib64 python2.6 site packages numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL lld seconds DESIRED 3 seconds","bug"
|
|
"6013","2474","Segfault during garbage collection after some operations with arrays Trac 1881).
|
|
Original ticket http projects.scipy.org numpy ticket 1881 on 2011 06 28 by trac user Bogdan, assigned to unknown.
|
|
System OSX 10.6.7 Python 2.6.1 the one provided with system) Numpy 2.0.0.dev 3071eab installed via SciPy Superpack) After running the file attached) I get a segfault.
|
|
Can t find any other lines to remove, because the bug disappears.
|
|
GDB trace Program received signal EXC BAD ACCESS, Could not access memory.
|
|
Reason KERN INVALID ADDRESS at address 0x000000000000000a 0x00000001000b03eb in Py Main ) gdb) bt 0 0x00000001000b03eb in Py Main ) 1 0x0000000100039751 in PyDict GetItemString ) 2 0x00000001000b0bc4 in PyObject GC Track ) 3 0x00000001000b118a in PyGC Collect ) 4 0x00000001000a30ca in Py Finalize ) 5 0x00000001000b03b1 in Py Main ) 6 0x0000000100000e6c in ?? )","bug"
|
|
"6013","2475","power function returns wrong values for int64 within the dtype range Trac 1882).
|
|
Original ticket http projects.scipy.org numpy ticket 1882 on 2011 06 28 by trac user abx numpy, assigned to unknown.
|
|
numpy.power a,i) returns wrong values even when they are within the range determined by dtype.
|
|
Example >>> import math, sys >>> import numpy as np >>> a = range 41,56,2) >>> b = np.array a); b.dtype dtype int64 ) >>> for x in a ... print x, , x 10 math.pow x,10) ... 41 0.0 43 0.0 45 0.0 47 0.0 49 0.0 51 0.0 53 0.0 55 0.0 >>> for x in b ... print x, , x 10 np.power x,10) ... 41 1 43 1 45 1 47 1 49 1 51 7 53 7 55 15 >>> the resuls are wrong only for odd integers >>> c = np.array [56],dtype=np.int64) >>> print c[0], , 56 10 np.power c,10))[0] 56, , 0) >>> d = np.array [57],dtype=np.int64) >>> print d[0], , 57 10 np.power d,10))[0] 57, , 15) >>> I am on 64 bit machine MacBookPro, Intel, 10.6) >>> 57 10 < sys.maxint True >>> sys.maxint 2 63 1) >>> 0L >>> i = int d[0]) >>> i 10 d[0] 10 0 >>> i 10 d 10)[0] 15 The last two outputs indicate that this is likely the ufunc problem.","bug"
|
|
"6013","2476","unpickling a pickled empty masked array error Trac 1883).
|
|
Original ticket http projects.scipy.org numpy ticket 1883 on 2011 06 28 by trac user mauro, assigned to pierregm.
|
|
When unpickling an empty masked array the following error is thrown TypeError object pickle not returning list Code import numpy as np import cPickle as pickle import pickle as pickle ma = np.ma.core.MaskedArray ) with open tmp.pkl , wb ) as fil pickle.dump ma, fil, protocol=pickle.HIGHEST PROTOCOL) with open tmp.pkl ) as fil pickle.load fil)","bug"
|
|
"6013","2477","polyval turns masked array into array Trac 1884).
|
|
Original ticket http projects.scipy.org numpy ticket 1884 on 2011 06 28 by trac user mauro, assigned to pierregm.
|
|
When using a masked array as input for polyval, then the return is an array.
|
|
I think it should be a masked array again or at least document it that it is not).
|
|
The fix seems easy, remove the line x = NX.asarray x) from np.lib.polyval.
|
|
But presumably that line is there for a reason...","bug"
|
|
"6013","2478","einsum wrong result crash Trac 1885).
|
|
Original ticket http projects.scipy.org numpy ticket 1885 on 2011 06 28 by trac user royf, assigned to unknown.
|
|
numpy.einsum sometimes returns wrong results.
|
|
Consider the following example >>> p = numpy.array [.5, .5]) >>> q = numpy.array [[0.26894142, 1 0.26894142], [1 0.26894142, 0.26894142]]) >>> r = numpy.identity 2) ... >>> p, q, r array [ 0.5, 0.5]), array [[ 0.26894142, 0.73105858], [ 0.73105858, 0.26894142]]), array [[ 1., 0.
|
|
], [ 0., 1.]]))
|
|
>>> numpy.einsum z,mz,zm > , p, q, r) 3.7305612660634556e 177 >>> numpy.einsum z,mz,zm > , p, q, r) 0.26894141999999999 >>> numpy.einsum z,mz,zm > , p, q, r) 3.7305612660634556e 177 >>> numpy.einsum z,mz,zm > , p, q, r) 0.26894141999999999 >>> numpy.einsum z,mz,zm > , p, q, r) 3.7305612660634556e 177 >>> numpy.einsum z,mz,zm > , p, q, r) 0.26894141999999999 >>> numpy.einsum z,mz,zm > , p, q, r) 0.26894141999999999 >>> numpy.einsum z,mz,zm > , p, q, r) 0.13447070999999999 In other runs this same example crashes the process.
|
|
The correct answer is 0.26894142.
|
|
Interestingly, one of the answers returned is half the correct one.
|
|
This probably has to do with the specific value of p in the example.
|
|
My guess is it s something like an off by one bug.
|
|
I d try to fix it myself, but the code is a scary mess Good luck!","bug"
|
|
"6013","2479","ndarray.data not compatible between python 2.7 and 3.2 versions Trac 1886).
|
|
Original ticket http projects.scipy.org numpy ticket 1886 on 2011 06 29 by trac user axh, assigned to unknown.
|
|
I am running python 3.2 2.7.2 with the standards from Fedora 15 numpy.x86 64 1 1.5.1 1.fc15 python3 numpy f2py.x86 64 1 1.5.1 1.fc15 Here is the issue with i4 = np.ndarray [1],dtype=np.int32) in 2,7 I have type i4.data) <type buffer > len i4.data[ ]) 4 so the buffer is byte size in 3.2 the data buffer comes in chunks of the size of the basic type len i4.data[0]) 4 len i4.data) 1 type i4.data) <class memoryview > It would be most desirable to switch this back to a mode that provides byte size buffer in the same way in both versions.
|
|
Especially, it would be good if the buffer type memory layout does not depend on the type of objects it stores.
|
|
It should be just a data buffer.
|
|
Alexander","bug"
|
|
"6013","2480","broken int64 format in timedelta64 for Python < 2.7 Trac 1887).
|
|
Original ticket http projects.scipy.org numpy ticket 1887 on 2011 06 29 by dhomeier, assigned to unknown.
|
|
With Python versions 2.4 2.6, the output of np.timedelta only produces the format string since probably) merging the datetime branch, like this >>> np.
|
|
version 2.0.0.dev f7c16d7 >>> np.timedelta64 3) numpy.timedelta64 lld, generic ) >>> str np.timedelta64 3, s )) lld seconds or in the most recent version >>> np.
|
|
version 2.0.0.dev 192ac74 >>> np.timedelta64 3) numpy.timedelta64 lld) making the corresponding tests fail, whereas with 2.7 and 3.2 the intended output is produced >>> np.timedelta64 3) numpy.timedelta64 3) >>> str np.timedelta64 3, s )) 3 seconds Tested on MacOS X 10.5 ppc and 10.6 x86 64)","bug"
|
|
"6013","2481","Multi dimensionial array support for savetxt loadtxt Trac 1888).
|
|
Original ticket http projects.scipy.org numpy ticket 1888 on 2011 06 30 by brickZA, assigned to unknown.
|
|
I think it would be feasible to support multi dimensional saving loading quite simply by using multiple line breaks for higher dimensional linebreaks.
|
|
d3D = numpy.array [[[111, 112], [121, 122]], [[211, 212], [221, 222]]]) could be written to text quite unambigiously as 111 112 121 122 < two line breaks indicate break in the 3rd dimension 211 212 221 222 Similarly, 3 line breaks could indicate a break in the 4th dimension.
|
|
Another option would be to have the option to write out text in nested list format.
|
|
I.e., simply write [[[111, 112], [121, 122]], [[211, 212], [221, 222]]] to the file.","enhancement"
|
|
"6013","2482","Numpy links wrong lapack blas and atlas in lapack lite.so Trac 1889).
|
|
Original ticket http projects.scipy.org numpy ticket 1889 on 2011 06 30 by trac user maldun, assigned to cournape.
|
|
I m working on Ubuntu 10.10 11.04, and Opensuse 11.4.
|
|
I installed Gentoo prefix on both systems and compiled the blas lapack and atlas versions from.
|
|
No I installed numpy linked with the libraries from prefix On Ubuntu it worked fine.
|
|
On openSuse I got the following error message when I try to import numpy parmenides math146 gentoo python Python 2.7.1 r271 86832, Jun 30 2011, 08 50 27) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File home parmenides gentoo usr lib python2.7 site packages numpy init .py , line 136, in <module> import add newdocs File home parmenides gentoo usr lib python2.7 site packages numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File home parmenides gentoo usr lib python2.7 site packages numpy lib init .py , line 13, in <module> from polynomial import File home parmenides gentoo usr lib python2.7 site packages numpy lib polynomial.py , line 17, in <module> from numpy.linalg import eigvals, lstsq File home parmenides gentoo usr lib python2.7 site packages numpy linalg init .py , line 48, in <module> from linalg import File home parmenides gentoo usr lib python2.7 site packages numpy linalg linalg.py , line 23, in <module> from numpy.linalg import lapack lite ImportError liblapack.so.0 cannot open shared object file No such file or directory >>> quit ) but blas, lapack and atlas should have been built properly.
|
|
Then after some other failed experiments I tried my build of lapack lite.so from ubuntu in the opensuse version and got this .... from numpy.linalg import lapack lite ImportError liblapack.so.3gf cannot open shared object file No such file or directory when I opended with emacs the lapack lite.so with my emacs I found this entry within the assembler code liblapack.so.3gf...libblas.so.3gf The gag behind this is that this doesn t appear in my directory where I installed blas, lapack and atlas but in my usr lib directory.
|
|
But my site cfg holds the following [atlas] include dirs = home maldun Gentoo usr include library dirs = home maldun Gentoo usr lib blas threaded atlas home maldun Gentoo usr lib blas reference home maldun Gentoo usr lib lapack atlas home maldun Gentoo usr lib atlas libs = cblas, atlas, blas lapack libs = lapack, blas, cblas, atlas [blas opt] include dirs = home maldun Gentoo usr include library dirs = home maldun Gentoo usr lib blas threaded atlas home maldun Gentoo usr lib blas reference home maldun Gentoo usr lib libraries = cblas, atlas, blas [lapack opt] library dirs = home maldun Gentoo usr lib lapack atlas home maldun Gentoo usr lib blas reference home maldun Gentoo usr lib blas threaded atlas home maldun Gentoo usr lib libraries = lapack, blas, cblas, atlas and those files are not located there!
|
|
My PATH, LIBRARY PATH, and LD LIBRARY PATH also don t point there maldun math112 Gentoo echo LD LIBRARY PATH home maldun Gentoo usr lib home maldun Gentoo lib maldun math112 Gentoo echo LIBRARY PATH home maldun Gentoo usr lib home maldun Gentoo lib maldun math112 Gentoo echo PATH home maldun Gentoo usr bin home maldun Gentoo bin home maldun Gentoo opt bin home maldun Gentoo usr sbin home maldun Gentoo sbin home maldun Gentoo usr x86 64 pc linux gnu binutils bin 2.21.51.0.7 home maldun Gentoo usr x86 64 pc linux gnu gcc bin 4.4.5 This shouldn t be happening if I want to build exclusively with one version of BLAS LAPACK ATLAS, but I don t know how to fix this issue!
|
|
Please help me!
|
|
greez, maldun","bug"
|
|
"6013","2483","piecewise throws TypeError for int input and sin Trac 1890).
|
|
Original ticket http projects.scipy.org numpy ticket 1890 on 2011 07 01 by trac user deil, assigned to unknown.
|
|
I would have expected the following np.piecewise call to work >>> x = np.arange 0, 10) >>> np.piecewise x, [x < 5, x >= 5], [lambda x x, lambda x np.sin x)]) Traceback most recent call last) File <stdin> , line 1, in <module> File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib function base.py , line 715, in piecewise y[condlist[k]] = item vals, args, kw) TypeError array cannot be safely cast to required type >>> x.dtype dtype int64 ) It works for an array of floats though >>> x = np.arange 0, 10, dtype=float) >>> np.piecewise x, [x < 5, x >= 5], [lambda x x, lambda x np.sin x)]) array [ 0. , 1. , 2. , 3. , 4. , 0.95892427, 0.2794155 , 0.6569866 , 0.98935825, 0.41211849]) Is this a bug?","bug"
|
|
"6013","2484","Inconsisitent error when assigning rhs floats to lhs ints.
|
|
Trac 1891).
|
|
Original ticket http projects.scipy.org numpy ticket 1891 on 2011 07 02 by charris, assigned to unknown.
|
|
Examples a = np.zeros 10, dtype=int) a[[1,4]] = np.array [1.1, 4]) TypeError a[1 3] = np.array [1.1, 4]) Casts to int a[[1,4]] = [1.1, 4] Casts to int","bug"
|
|
"6013","2485","object arrays converted to string arrays of S dtype have default length.
|
|
Trac 1892).
|
|
Original ticket http projects.scipy.org numpy ticket 1892 on 2011 07 04 by trac user ehiggs, assigned to unknown.
|
|
I m sure this must have been reported before, but I can t find it in the bug tracker.
|
|
The bug is that calling a vectorized function on a string type truncates the string to |S8 regardless of the original length.
|
|
Example In [1] import numpy In [2] def f x) ... return x ...
|
|
In [3] f = numpy.vectorize f) In [4] s = 2011 07 04 In [5] y = f s) In [6] y Out[6] array 2011 07 04 , dtype= |S10 ) In [8] z = f numpy.array [s,s,s,s])) In [9] z Out[9] array [ 2011 07 , 2011 07 , 2011 07 , 2011 07 ], dtype= |S8 ) If this has been raised, please merge this in so I can track progress on the original.","bug"
|
|
"6013","2486","Broadcasting shape mismatch exception Trac 1893).
|
|
Original ticket http projects.scipy.org numpy ticket 1893 on 2011 07 04 by trac user smif1984, assigned to unknown.
|
|
The exception which is currently, v1.6), raised when two non broadcastable arrays are summed is a ValueError exception.
|
|
I propose to create a specific exception class, e.g.
|
|
BroadcastError, to be more specific and give better control in exception catching?","enhancement"
|
|
"6013","2487","DeprecationWarning Setting NumPy dtype names is deprecated in several tests Trac 1894).
|
|
Original ticket http projects.scipy.org numpy ticket 1894 on 2011 07 05 by stsci sienkiew, assigned to unknown.
|
|
`DeprecationWarning Setting NumPy dtype names is deprecated, the dtype will become immutable in a future version` numpy core tests test regression.py line 1016, 1425 numpy lib iotools.py line 832, 839, 842 numpy lib npyio.py line 1619 numpy core tests test multiarray.py line 1319","bug"
|
|
"6013","2488","segfault in numpy.subtract on MacOS X 10.5 Trac 1895).
|
|
Original ticket http projects.scipy.org numpy ticket 1895 on 2011 07 05 by r owen, assigned to rgommers.
|
|
Using the attached pickled data file and script I m reliably getting a segfault in numpy.subtract on MacOS X 10.5.8 with a fresh Mac Python 2.7.2 32 bit from python.org and a fresh install of the official numpy 1.6.0 Mac binary.
|
|
The self test reports no errors 3 known fail, 2 skip, 3179 tests in all).
|
|
Other tid bits that may be relevant I started with a larger data file and kept trying different subregions.
|
|
The attached data file is about the smallest subregion I could find that showed the problem.
|
|
Below that the crash went away.
|
|
If I convert it to float32 I still get the crash.
|
|
I don t see the crash on MacOS X 10.6.2 using the same python and numpy binary installer.","bug"
|
|
"6013","2489","Another crash in numpy.subtract Trac 1896).
|
|
Original ticket http projects.scipy.org numpy ticket 1896 on 2011 07 06 by r owen, assigned to unknown.
|
|
I ve seen this crash on MacOS X 10.6 using python.org Python 2.7.2 32 bit backward compatible version) and three flavors of numpy numpy 1.6.0 standard Mac binary numpy 1.6.0 built from source numpy 1.6.1rc2 built from source I have attached a data file and script.
|
|
Text in the script shows how to run it.
|
|
You can choose to use Tkinter or not, but I find the bug shows up more reliably if you use Tkinter the configure event triggers numpy.subtract).","bug"
|
|
"6013","2490","Pickling a numpy.ma.MaskedConstant fails Trac 1897).
|
|
Original ticket http projects.scipy.org numpy ticket 1897 on 2011 07 07 by jpaalasm, assigned to pierregm.
|
|
Python 2.7.1 r271 86832, Apr 11 2011, 18 13 53) [GCC 4.5.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import pickle; import numpy; numpy.
|
|
version 1.5.1 >>> ma = numpy.ma.MaskedArray [0, 0, 0], mask=True) >>> m = ma[0] >>> m masked >>> pickle.dumps m) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.7 pickle.py , line 1374, in dumps Pickler file, protocol).dump obj) File usr lib python2.7 pickle.py , line 224, in dump self.save obj) File usr lib python2.7 pickle.py , line 306, in save rv = reduce self.proto) File usr lib pymodules python2.7 numpy ma core.py , line 5494, in reduce self.
|
|
getstate )) File usr lib pymodules python2.7 numpy ma core.py , line 5467, in getstate self.
|
|
fill value, AttributeError MaskedConstant object has no attribute fill value >>>","bug"
|
|
"6013","2491","Casting complex numpy array scalars to Python scalars Trac 1898).
|
|
Original ticket http projects.scipy.org numpy ticket 1898 on 2011 07 08 by inducer, assigned to unknown.
|
|
Contrast this >>> x = np.array 17 3j, dtype=np.complex128) >>> complex x) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError can t convert complex to float with this >>> x = np.array 17, dtype=np.float64) >>> float x) 17.0 There s no obvious reason why one should work, but not the other.
|
|
In other words, numpy implements .
|
|
float ) but not .
|
|
complex ).)","bug"
|
|
"6013","2492","histogramdd improperly handles empty arrays with known bins Trac 1899).
|
|
Original ticket http projects.scipy.org numpy ticket 1899 on 2011 07 09 by trac user WeatherGod, assigned to unknown.
|
|
Subject says it all.
|
|
In particular, this impacts histogram2d.
|
|
histogramdd appears to attempt to handle this case, but doesn t do so correctly.
|
|
>>> np.histogram2d [], [], bins=4) array [ 0., 0.
|
|
]), array [ 0. , 0.25, 0.5 , 0.75, 1.
|
|
]), array [ 0. , 0.25, 0.5 , 0.75, 1. ]))
|
|
The first element of the return tuple is a 1 D array with two elements.
|
|
What is expected is a 2 D 4x4 array.
|
|
Consider the following with non empty inputs >>> np.histogram2d [0.3], [0.7], bins=4) array [[ 0., 0., 0., 0.
|
|
], [ 0., 0., 0., 0.
|
|
], [ 0., 0., 1., 0.
|
|
], [ 0., 0., 0., 0.
|
|
]]), array [ 0.2 , 0.05, 0.3 , 0.55, 0.8 ]), array [ 0.2 , 0.45, 0.7 , 0.95, 1.2 ])) For comparison, consider how histogram behaves with empty inputs >>> np.histogram [], bins=4) array [0, 0, 0, 0]), array [ 0. , 0.25, 0.5 , 0.75, 1. ]))","bug"
|
|
"6013","2493","weave raise invalid version number 4. error Trac 1900).
|
|
Original ticket http projects.scipy.org numpy ticket 1900 on 2011 07 11 by trac user hyry0088, assigned to unknown.
|
|
When updated to NumPy 1.6.0, weave can t compile c code by using mingw.
|
|
It raise ValueError invalid version number 4. .
|
|
Following is the test program from scipy.weave import inline a = 0 inline a =1; , [ a ], compiler= gcc ) To fix it open file numpy distutils mingw32compiler.py replace two 4. by 4.0","bug"
|
|
"6013","2494","Speed up np.copyto with where= , and any routines using nditer in its masked mode Trac 1901).
|
|
Original ticket http projects.scipy.org numpy ticket 1901 on 2011 07 11 by mwiebe, assigned to unknown.
|
|
As part of the missing value functionality, I have implemented a masked data copying and casting mechanism.
|
|
It has all the hooks in place to be fast, but I have only implemented a slow wrapper around the unmasked routines.
|
|
One example where it currently performs poorly is filling based on a mask In [1] import numpy as np In [2] a = np.zeros 100,100,100)) In [3] m = np.random.rand 100,100,100) > 0.5 In [4] timeit np.copyto a, 1, where=m) 100 loops, best of 3 9.22 ms per loop In [5] timeit np.putmask a, m, 1) 100 loops, best of 3 6.02 ms per loop To do this optimization, 1) Learn how dtype transfer.c works, by reading its code and the documentation in comments.
|
|
In particular, the function PyArray GetDTypeTransferFunction is the main thing to understand.
|
|
Its arguments are documented in private lowlevel strided loops.h.
|
|
2) Learn how lowlevel stided loops.c.src works, just as for 1).
|
|
In particular, understanding how zero strides, contiguous strides, and alignment affect the nature of the inner loop functions.
|
|
3) Create specialized inner loops for various masked transfer functions.
|
|
Just specializing aligned data is probably ok.
|
|
The important cases are likely contiguous src, dst, and mask zero stride src, contiguous dst and mask zero stride src, general strided dst and mask This should be done for both straight data copies and cast operations.
|
|
4) Edit PyArray GetMaskedDTypeTransferFunction to return these specialized masked loops where appropriate, analogously to how PyArray GetDTypeTransferFunction does it.
|
|
5) Demonstrate that it s working with some before after benchmarks of the different cases.","enhancement"
|
|
"6013","2495","numpy.ma.compress fails because of the argument order Trac 1902).
|
|
Original ticket http projects.scipy.org numpy ticket 1902 on 2011 07 11 by efiring, assigned to pierregm.
|
|
The frommethod class used to make ma.compress assumes the first argument is the object on which the function is acting, but the compress signature is compress cond, arr).
|
|
Example of the resulting failure In [14] arr = np.arange 8) In [15] arr.shape = 4,2 In [16] cond = np.array [True, False, True, True]) In [17] np.ma.compress cond, arr, axis=0) ValueError Traceback most recent call last) home efiring <ipython input 17 50a13b754f15> in <module> ) > 1 np.ma.compress cond, arr, axis=0) usr local lib python2.7 dist packages numpy ma core.pyc in call self, a, args, params) 5931 method = getattr a, method name, None) 5932 if method is not None > 5933 return method args, params) 5934 Still here ?
|
|
Then a is not a MaskedArray 5935 method = getattr MaskedArray, method name, None) ValueError condition must be 1 d array","bug"
|
|
"6013","2496","sphinxext issue with field names Trac 1903).
|
|
Original ticket http projects.scipy.org numpy ticket 1903 on 2011 07 11 by astrofrog, assigned to pv.
|
|
I am having an issue with sphinxext that can also be seen on the following page http docs.scipy.org doc scipy reference generated scipy.integrate.quad.html scipy.integrate.quad The after Parameters is on the next line.
|
|
I think there should be no space before the colons, which would solve this issue.","bug"
|
|
"6013","2497","ndarray derived classes inconsistent return type of min ) Trac 1904).
|
|
Original ticket http projects.scipy.org numpy ticket 1904 on 2011 07 12 by trac user cstraehl, assigned to unknown.
|
|
.
|
|
we expect the type to be the same because we did not overload the min ) function... assert type a1) == type a2), inconsitent type","bug"
|
|
"6013","2498","numpy core numeric allclose ) fails to test for NaN inputs Trac 1905).
|
|
Original ticket http projects.scipy.org numpy ticket 1905 on 2011 07 12 by trac user robertpyle, assigned to rgommers.
|
|
Despite a comment that says NaNs in either input cause it to return False, routine allclose ) in core numeric.py fails to test its inputs for NaN.
|
|
This causes a warning to be issued by numpy.test ), at least on MacOSX 10.6.8 with python.org python2.7.1, numpy 1.6.1.rc3.
|
|
A 2 line addition fixes this.
|
|
Here s the diff 1915a1916,1917 > if any isnan x)) or any isnan y)) > return False","bug"
|
|
"6013","2499","FAIL test two keys two vars test recfunctions.TestJoinBy2) Trac 1906).
|
|
Original ticket http projects.scipy.org numpy ticket 1906 on 2011 07 14 by stsci sienkiew, assigned to unknown.
|
|
Observed in 32 bit python 2.7 on mac leopard and red hat enterprise 4 linux.
|
|
Not observed in any of my 64 bit systems.
|
|
FAIL test two keys two vars test recfunctions.TestJoinBy2) Traceback most recent call last) File usr stsci pyssgdev 2.7 numpy lib tests test recfunctions.py , line 677, in test two keys two vars assert equal test, control) File usr stsci pyssgdev 2.7 numpy ma testutils.py , line 120, in assert equal return assert array equal actual, desired, err msg) File usr stsci pyssgdev 2.7 numpy ma testutils.py , line 193, in assert array equal header= Arrays are not equal ) File usr stsci pyssgdev 2.7 numpy ma testutils.py , line 186, in assert array compare verbose=verbose, header=header) File usr stsci pyssgdev 2.7 numpy testing utils.py , line 635, in assert array compare raise AssertionError msg) AssertionError Arrays are not equal mismatch 100.0 ) x array [ 10, 0, 50, 65, 10, 0), 11, 0, 51, 66, 11, 1), 10, 1, 52, 67, 12, 2), 11, 1, 53, 68, 13, 3), 10, 2, 54, 69, 14, 4), 11, 2, 55, 70, 15, 5),... y array [ 10L, 0L, 50L, 65L, 10L, 0L), 11L, 0L, 51L, 66L, 11L, 1L), 10L, 1L, 52L, 67L, 12L, 2L), 11L, 1L, 53L, 68L, 13L, 3L), 10L, 2L, 54L, 69L, 14L, 4L), 11L, 2L, 55L, 70L, 15L, 5L),...","bug"
|
|
"6013","2500","treat an array with a single value as a scaler Trac 1907).
|
|
Original ticket http projects.scipy.org numpy ticket 1907 on 2011 07 14 by trac user rgrout, assigned to unknown.
|
|
If I have an array [3]) an array with a single value), I should be able to take the dot product with another vector or matrix as though I were multiplying by a scalar.
|
|
For example.
|
|
[[[ a=array [3]) b = array [3,5,6,2,3]).reshape 2,2)) c = rand 5,5) np.dot a,b)==3 b np.dot a,c)==3 c ]]]","bug"
|
|
"6013","2501","Incorrect parsing of dtype comma separated strings that have datatime64 Trac 1908).
|
|
Original ticket http projects.scipy.org numpy ticket 1908 on 2011 07 16 by trac user rocha, assigned to unknown.
|
|
The following dtypes should use the same datetime64 time resultions, however one is in [us] and the other in [s] >>> print numpy.dtype i, datetime64[s] ); [ f0 , <i4 ), f1 , <M8[us] , {}))] >>> print numpy.dtype [ f1 , i ), f2 , datetime64[s] )]) [ f1 , <i4 ), f2 , <M8[s] , {}))] Python 2.6.6 numpy 1.6.1rc3","bug"
|
|
"6013","2502","Datetime related build error on Windows Trac 1909).
|
|
Original ticket http projects.scipy.org numpy ticket 1909 on 2011 07 17 by rgommers, assigned to mwiebe.
|
|
Only in master.
|
|
Builds okay on OS X. creating build\temp.win32 2.6\Release\numpy\core\src\multiarray compile options Inumpy\core\include Ibuild\src.win32 2.6\numpy\core\include numpy Inumpy\core\src\private Inumpy\core\src Inumpy\core Inumpy\core\src\npymath Inumpy\core\src\multiarray Inumpy\core\src\umath Inumpy\core\src\npysort Inumpy\core\include IC \Python26\include IC \Python26\PC Ibuild\src.win32 2.6\numpy\core\src\multiarray Ibuild\src.win32 2.6\numpy\core\src\umath c gcc mno cygwin O2 Wall Wstrict prototypes Inumpy\core\include Ibuild\src.win32 2.6\numpy\core\include numpy Inumpy\core\src\private Inumpy\core\src Inumpy\core Inumpy\core\src\npymath Inumpy\core\src\multiarray Inumpy\core\src\umath Inumpy\core\src\npysort Inumpy\core\include IC \Python26\include IC \Python26\PC Ibuild\src.win32 2.6\numpy\core\src\multiarray Ibuild\src.win32 2.6\numpy\core\src\umath c numpy\core\src\multiarray\multiarraymodule onefile.c o build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o g mno cygwin shared build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o LC \Python26\libs LC \Python26\PCbuild Lbuild\temp.win32 2.6 lnpymath lnpysort lpython26 lmsvcr90 o build\lib.win32 2.6\numpy\core\multiarray.pyd build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o multiarraymodule onefile.c .text 0x7b15) undefined reference to `localtime s build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o multiarraymodule onefile.c .text 0x8139) undefined reference to `gmtime s build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o multiarraymodule onefile.c .text 0x98d1) undefined reference to `localtime s collect2 ld returned 1 exit status error Command g mno cygwin shared build\temp.win32 2.6\Release\numpy\core\src\multiarray\multiarraymodule onefile.o LC \Python26\libs LC \Python26\PCbuild Lbuild\temp.win32 2.6 lnpymath lnpysort lpython26 lmsvcr90 o build\lib.win32 2.6\numpy\core\multiarray.pyd failed with exit status 1","bug"
|
|
"6013","2503","Check buffer bounds in ndarray creation Trac 1910).
|
|
Original ticket http projects.scipy.org numpy ticket 1910 on 2011 07 18 by pv, assigned to unknown.
|
|
Forwarded from Stefan Krah With a recent NumPy clone Python3.3, I get this invalid access Python 3.3.0a0 default 1dd6908df8f5, Jul 16 2011, 11 16 00) [GCC 4.4.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> x = ndarray buffer=bytearray [1,2,3,4,5,6,7,8,9,10]), shape=[2,3], strides=[ 3, 2], dtype= B , offset=5) >>> x array [[ 6, 4, 2], [ 3, 1, 251]], dtype= uint8 ) >>> x[1][2] 251 >>> I think NumPy should probably refuse to create such an array, like here >>> x = ndarray buffer=bytearray [1,2,3,4,5,6,7,8,9,10]), shape=[2,3], strides=[ 3, 2], dtype= B , offset=0) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError strides is incompatible with shape of requested array and size of buffer","bug"
|
|
"6013","2504","recarray fromfile doesn t support file like objects Trac 1911).
|
|
Original ticket http projects.scipy.org numpy ticket 1911 on 2011 07 18 by trac user sbreuss, assigned to unknown.
|
|
The current implementation of the record array fromfile method doesn t support file like objects e.g.
|
|
StringIO, io.BytesIO, ...) By two minor changes in the source this would be possible 1) change the get remaining size method to use seek 2) not using the readinto method of the file object I m open to any suggestions, a patch for the current master version showing my changes is attached.
|
|
Cheers, Simon Breuss","enhancement"
|
|
"6013","2505","datetime and record arrays masked arrays Trac 1912).
|
|
Original ticket http projects.scipy.org numpy ticket 1912 on 2011 07 18 by chrisjordansquire, assigned to unknown.
|
|
I m using the 2.0.0 development version of numpy on ububtu 10.10 .
|
|
When string arguments for the dtype are given for a record array, and one of the individual dtypes is datetime64[D], the day part is lost and the datetime become generic.
|
|
This makes it impossible to assign anything to it.
|
|
For example, look at np.zeros 4, dtype = , .join [ datetime64[D] , i4 ])) This gives an array with dtype dtype=[ f0 , <M8), f1 , <i4 )], where the <M8 should be <M8[D].
|
|
A different bug happens when putting datetimes into masked arrays.
|
|
For example, np.ma.zeros 4,dtype= datetime64[D],i4 ) gives a ValueError mismatch in size of old and new data descriptor with traceback home intern my26 lib python2.6 site packages numpy ma core.pyc in call self, a, args, params) 7142 extras[p] = params.pop p) 7143 Get the result > 7144 result = self.
|
|
func.
|
|
call a, args, params).view MaskedArray) 7145 if fill value in common params 7146 result.fill value = extras.get fill value , None) home intern my26 lib python2.6 site packages numpy ma core.pyc in array finalize self, obj) 2794 if self.dtype.names 2795 if self.
|
|
fill value is None > 2796 self.
|
|
fill value = check fill value None, self.dtype) 2797 return 2798 home intern my26 lib python2.6 site packages numpy ma core.pyc in check fill value fill value, ndtype) 392 if fields 393 descr = ndtype.descr > 394 fill value = np.array recursive set default fill value descr), 395 dtype=ndtype,) 396 else home intern my26 lib python2.6 site packages numpy ma core.pyc in recursive set default fill value dtypedescr) 360 deflist.append tuple recursive set default fill value currenttype))) 361 else > 362 deflist.append default fill value np.dtype currenttype))) 363 return tuple deflist) 364 ValueError mismatch in size of old and new data descriptor","bug"
|
|
"6013","2506","PyUFunc RegisterLoopForType doesn t work for custom data types Trac 1913).
|
|
Original ticket http projects.scipy.org numpy ticket 1913 on 2011 07 21 by trac user jvporter, assigned to unknown.
|
|
I upgraded from NumPy 1.5 to 1.6, and found that ufuncs on my custom array data type stopped working specifically == and !=).
|
|
I don t think I m doing anything wrong, but it s possible that something about the API has changed here.
|
|
For reference, the code where I m attempting this looks like so minus some preprocessor magic) HANDLE NPYTYPE = PyArray RegisterDataType descr); types[0] = types[1] = HANDLE NPYTYPE; types[2] = NPY BOOL; PyUFunc RegisterLoopForType eq, HANDLE NPYTYPE, Arr equal,types,0); If you need more context, the full source file is here http trac.mcs.anl.gov projects ITAPS browser python trunk iBase handleTempl.def","bug"
|
|
"6013","2507","Delayed ValueError when using count nonzero with object array Trac 1914).
|
|
Original ticket http projects.scipy.org numpy ticket 1914 on 2011 07 25 by trac user rcl33, assigned to unknown.
|
|
Using count nonzero on an array with dtype=numpy.object causes an error on the next logical test.
|
|
Example import numpy x = numpy.zeros 3,), dtype=numpy.object) x[0] = numpy.zeros 30,)) This sets up the error numpy.count nonzero x) A for loop or truth test triggers error for i in [1] print i Result is exception ValueError The truth value of an array with more than one element is ambiguous.
|
|
Use a.any ) or a.all )","bug"
|
|
"6013","2508","Incomplete Cholesky Decomposition Trac 1915).
|
|
Original ticket http projects.scipy.org numpy ticket 1915 on 2011 07 26 by trac user goretkin, assigned to pv.
|
|
PyMC [http code.google.com p pymc ] provides an ichol function.
|
|
The Fortran code already f2py d) is here [https github.com pymc devs pymc blob master pymc gp incomplete chol.f]","enhancement"
|
|
"6013","2509","genfromtxt strips brackets from names Trac 1916).
|
|
Original ticket http projects.scipy.org numpy ticket 1916 on 2011 07 28 by trac user mboquien, assigned to unknown.
|
|
genfromtxt systematically strips any bracket []) from the name of the columns.
|
|
Here is a short snippet exhibitting the problem.
|
|
In [1] from StringIO import StringIO In [2] s = StringIO 1 2\n3 4 ) In [3] d = genfromtxt s,names=[ a[1] , a[2] ]) In [4] d Out[4] array [ 1.0, 2.0), 3.0, 4.0)], dtype=[ a1 , <f8 ), a2 , <f8 )]) Note that a[1] and a[2] have unexpectedly been substituted with a1 and a2 .
|
|
Thank you very much, regards, Mxc3xa9dxc3xa9ric","bug"
|
|
"6013","2510","Memory leak in 2.0.0.dev Unknown Trac 1917).
|
|
Original ticket http projects.scipy.org numpy ticket 1917 on 2011 07 28 by trac user jiffyclub, assigned to unknown.
|
|
The following code overruns my system s memory in a fraction of a second >>> import numpy as np >>> np.
|
|
version 2.0.0.dev Unknown >>> def test ) ... a = np.zeros 10000,10000),dtype=np.float) ... >>> while True ... test ) ...
|
|
I believe this is the most recent numpy, checked out and compiled last night 7 28 2011).
|
|
I m using Python 2.7.1 on Mac OS 10.6.8.","bug"
|
|
"6013","2511","Failed to compile numpy PyArrayObject has no member ob type Trac 1918).
|
|
Original ticket http projects.scipy.org numpy ticket 1918 on 2011 07 30 by trac user v l, assigned to unknown.
|
|
Hi.
|
|
[[BR]] I m on Archlinux x86 64, gcc 4.6.1, python 2.7.2 and 3.2.1.
|
|
[[BR]] I try to compile numpy using python of 2 and 3 versions but it fails on python3 with the following a bit reformatted) [[BR]] creating build temp.linux x86 64 3.2 numpy lib creating build temp.linux x86 64 3.2 numpy lib src compile options Inumpy core include Ibuild src.linux x86 64 3.2 numpy core include numpy \ Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath \ Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include \ I usr include python3.2mu Ibuild src.linux x86 64 3.2 numpy core src multiarray \ Ibuild src.linux x86 64 3.2 numpy core src umath c gcc numpy lib src compiled base.c numpy lib src compiled base.c In function pack or unpack bits numpy lib src compiled base.c 1317 51 error PyArrayObject has no member named ob type numpy lib src compiled base.c 1357 43 error PyArrayObject has no member named ob type numpy lib src compiled base.c In function pack or unpack bits numpy lib src compiled base.c 1317 51 error PyArrayObject has no member named ob type numpy lib src compiled base.c 1357 43 error PyArrayObject has no member named ob type error Command gcc pthread DDYNAMIC ANNOTATIONS ENABLED=1 DNDEBUG g fwrapv O3 Wall \ Wstrict prototypes march=x86 64 mtune=generic O2 pipe fPIC Inumpy core include \ Ibuild src.linux x86 64 3.2 numpy core include numpy Inumpy core src private Inumpy core src \ Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath \ Inumpy core src npysort Inumpy core include I usr include python3.2mu \ Ibuild src.linux x86 64 3.2 numpy core src multiarray \ Ibuild src.linux x86 64 3.2 numpy core src umath c numpy lib src compiled base.c \ o build temp.linux x86 64 3.2 numpy lib src compiled base.o failed with exit status 1 For both versions the build options are setup.py config fc fcompiler=gnu95 build","bug"
|
|
"6013","2512","import umath macro has return type incompatible with Python 3 Trac 1919).
|
|
Original ticket http projects.scipy.org numpy ticket 1919 on 2011 07 30 by trac user Arfrever, assigned to unknown.
|
|
import umath macro generated by numpy core code generators generate ufunc api.py) currently always returns void, which causes compilation failure when trying to use this function in an init function in extension module with Python 3.
|
|
I suggest that the same solution be implemented, which is used for import array macro generated by numpy core code generators generate numpy api.py).
|
|
I m attaching the patch.","bug"
|
|
"6013","2513","Improve numpy.interp running time Trac 1920).
|
|
Original ticket http projects.scipy.org numpy ticket 1920 on 2011 08 01 by trac user tkluck, assigned to unknown.
|
|
copy pasted from numpy discussion, but extra comment below) The current implementation of numpy.interp x,xp,fp) comes down to first calculating all the slopes of the linear interpolant these are len xp) 1), then use a binary search to find where x is in xp running time log len xp)).
|
|
So we obtain a running time of O len xp) len x) log len xp) ) We could improve this to just O len x) log len xp) ) by not caching the slopes.
|
|
The point is, of course, that this is slightly slower in the common use case where x is is refinement of xp, and where you will have to compute all the slopes anyway.
|
|
In my personal use case, however, I needed the value of the interp x0,xp,fp) in order to calculate the next point x1 where I wanted to calculate interp x1,xp,fp).
|
|
The current implementation gave a severe running time penalty.
|
|
After a short discussion on the mailinglist, it was thought to be a good idea to offer both implementations to cache the slopes when len xp) <= len x), and to not do so otherwise.
|
|
In the first cases caching gives about 10 increase in speed in my not so thorough testing), which I feel justifies the minor) extra code complexity.
|
|
Patch attached.","enhancement"
|
|
"6013","2514","method index and coords of flatiter object are one off Trac 1921).
|
|
Original ticket http projects.scipy.org numpy ticket 1921 on 2011 08 02 by trac user rvdb, assigned to unknown.
|
|
Somewhere i read that the flatiter iterator object returned by a.flat where a is some ndarray have methods index and coords.
|
|
And indeed they have but they seem to be one off In [71] a = ones 3,3)) In [72] fla = a.flat In [73] for va in fla .... print fla.index, fla.coords, va .... 1 0, 1) 1.0 2 0, 2) 1.0 3 1, 0) 1.0 4 1, 1) 1.0 5 1, 2) 1.0 6 2, 0) 1.0 7 2, 1) 1.0 8 2, 2) 1.0 9 3, 0) 1.0 Otherwise it seems a quite fast way to iterate over all elements...","bug"
|
|
"6013","2515","Indexing arrays with arrays having redundant entries Trac 1922).
|
|
Original ticket http projects.scipy.org numpy ticket 1922 on 2011 08 02 by trac user gfabbian, assigned to unknown.
|
|
I m experiencing a problem with array indexing.
|
|
Suppose you have an array a and another array b you want to use to use as index for a in order to assign some values to the position pointed by b elements.
|
|
>>>a=numpy.zeros 5) >>> print a >>>[ 0.
|
|
0.
|
|
0.
|
|
0.
|
|
0.]
|
|
Now I would like to increase the second element twice using this technique >>>b=numpy.array [1,1]) >>>a[b] =1.
|
|
>>> print a >>>[ 0.
|
|
1.
|
|
0.
|
|
0.
|
|
0.]
|
|
while I expected to have >>>[ 0.
|
|
2.
|
|
0.
|
|
0.
|
|
0.]
|
|
There are no problems if the array b has no redundancies all values of its elements are different).
|
|
Is it a bug or I m missing something?
|
|
Thanks in advance","bug"
|
|
"6013","2516","Datetime related failures under python 3 Trac 1923).
|
|
Original ticket http projects.scipy.org numpy ticket 1923 on 2011 08 02 by rgommers, assigned to unknown.
|
|
These are found with Python 3.2 in a virtualenv on OS X.
|
|
====================================================================== ERROR test datetime array str test datetime.TestDateTime) Traceback most recent call last) File Users rgommers Code numpy build py3k numpy core tests test datetime.py , line 510, in test datetime array str assert equal str a), [ 2011 03 16 1920 01 01 2013 05 19 ] ) File Users rgommers Code numpy build py3k numpy core numeric.py , line 1385, in array str return array2string a, max line width, precision, suppress small, , , str) File Users rgommers Code numpy build py3k numpy core arrayprint.py , line 437, in array2string separator, prefix, formatter=formatter) File Users rgommers Code numpy build py3k numpy core arrayprint.py , line 313, in array2string summaryEdgeItems, summary insert)[ 1] File Users rgommers Code numpy build py3k numpy core arrayprint.py , line 480, in formatArray word = format function a[ i]) separator File Users rgommers Code numpy build py3k numpy core arrayprint.py , line 726, in call casting=self.casting) TypeError Cannot create a local timezone based date string from a NumPy datetime without forcing unsafe casting ====================================================================== ERROR test datetime divide test datetime.TestDateTime) Traceback most recent call last) File Users rgommers Code numpy build py3k numpy core tests test datetime.py , line 926, in test datetime divide assert equal tda tdb, 6.0 9.0) TypeError internal error could not find appropriate datetime inner loop in true divide ufunc ====================================================================== FAIL test datetime as string test datetime.TestDateTime) Traceback most recent call last) File Users rgommers Code numpy build py3k numpy core tests test datetime.py , line 1166, in test datetime as string 1959 ) File Users rgommers Code numpy build py3k numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL b 1959 DESIRED 1959 ====================================================================== FAIL test datetime as string timezone test datetime.TestDateTime) Traceback most recent call last) File Users rgommers Code numpy build py3k numpy core tests test datetime.py , line 1277, in test datetime as string timezone 2010 03 15T06 30Z ) File Users rgommers Code numpy build py3k numpy testing utils.py , line 313, in assert equal raise AssertionError msg) AssertionError Items are not equal ACTUAL b 2010 03 15T06 30Z DESIRED 2010 03 15T06 30Z","bug"
|
|
"6013","2517","Support floating point hex representation in loadtxt Trac 1924).
|
|
Original ticket http projects.scipy.org numpy ticket 1924 on 2011 08 03 by trac user claumann, assigned to unknown.
|
|
Python supports a hexadecimal representation for floating numbers in float.hex ) and float.fromhex ) >>> a = 3.14 >>> a.hex ) 0x1.91eb851eb851fp 1 >>> float.fromhex a.hex )) 3.14 Numpy s loadtxt does not support this hex text format but it seems like it would be a useful enhancement.
|
|
For example, a simple implementation might come from changing the default converter returned by getconv for columns of type float def floatconv x) try return float x) except ValueError pass return float.fromhex x) And then change the return float in getconv to return floatconv.
|
|
This converter obeys essentially the same semantics and exceptions) as float ) except that it falls back on fromhex before giving up.
|
|
A slightly less permissive change would be to check that the string x begins with 0x before calling fromhex and throwing a ValueError if not.
|
|
The code above would accept bad and convert it to 2989, which might be unnecessarily permissive.
|
|
Best, Chris","enhancement"
|
|
"6013","2518","output of b \x00 is an empty bytes object Trac 1925).
|
|
Original ticket http projects.scipy.org numpy ticket 1925 on 2011 08 04 by trac user skrah, assigned to unknown.
|
|
I m not sure if packing a buffer directly via the struct module is supported.
|
|
I came across this corner case >>> import struct >>> buf = bytearray 2) >>> struct.pack into c , buf, 0, b \x00 ) >>> struct.pack into c , buf, 1, b \x01 ) >>> nd = ndarray buffer=buf, shape=[2], dtype= c ) >>> nd.tolist ) [b , b \x01 ] >>> struct.unpack from c , buf, 0) b \x00 ,) >>> struct.unpack from c , buf, 1) b \x01 ,) The first character in the tolist ) representation should be b \x00 .","bug"
|
|
"6013","2519","test umath complex fails under OS X Lion Trac 1926).
|
|
Original ticket http projects.scipy.org numpy ticket 1926 on 2011 08 05 by trac user RuiPereira, assigned to unknown.
|
|
8 tests related to umath complex fail when testing numpy 1.6.1 under OS X Lion.
|
|
The same tests fail with numpy 1.5.1 shipped with the system as well as with the most recent HEAD from github.
|
|
python c import numpy; numpy.test ) Running unit tests for numpy NumPy version 1.6.1 NumPy is installed in Library Python 2.7 site packages numpy Python version 2.7.1 r271 86832, Jun 16 2011, 16 59 05) [GCC 4.2.1 Based on Apple Inc. build 5658) LLVM build 2335.15.00)] nose version 1.0.0 ...) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 1, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 0.0, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, 0.0, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, nan, inf, inf, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array inf infj)) ====================================================================== FAIL test umath complex.TestCsqrt.test special values <ufunc sqrt >, inf, 1, 0.0, inf) Traceback most recent call last) File Library Python 2.7 site packages nose case.py , line 187, in runTest self.test self.arg) File Library Python 2.7 site packages numpy core tests test umath complex.py , line 578, in check complex value assert equal f z1), z2) File Library Python 2.7 site packages numpy testing utils.py , line 256, in assert equal return assert array equal actual, desired, err msg, verbose) File Library Python 2.7 site packages numpy testing utils.py , line 707, in assert array equal verbose=verbose, header= Arrays are not equal ) File Library Python 2.7 site packages numpy testing utils.py , line 605, in assert array compare chk same position x id, y id, hasval= nan ) File Library Python 2.7 site packages numpy testing utils.py , line 588, in chk same position raise AssertionError msg) AssertionError Arrays are not equal x and y nan location mismatch x array [ nan infj]) y array infj) Ran 3533 tests in 22.955s FAILED KNOWNFAIL=3, SKIP=2, failures=8)","bug"
|
|
"6013","2520","datetime masked array and record array repr bugs Trac 1927).
|
|
Original ticket http projects.scipy.org numpy ticket 1927 on 2011 08 08 by chrisjordansquire, assigned to unknown.
|
|
I have a record array with the following elements np.array [ datetime.date 1991, 11,3), np.datetime64 NaT )), np.datetime64 NaT ),datetime.date 2001, 3, 10))], dtype=[ date1 , <M8[D] ), date2 , <M8[D] )]) np.array repr of the above array gives array [ datetime.date 1991, 11, 3), None), None, datetime.date 2001, 3, 10))], dtype=[ date1 , <M8[D] ), date2 , <M8[D] )]) So the np.datetime64 NaT ) are printed as None.
|
|
This certainly isn t correct for a repr.
|
|
Additionally, for a masked record array the repr drops the datetime.date if there is a missing entry in the row.
|
|
For example, if you let z be the array given above, and then do w = np.ma.array z) w.mask[0] = False, True) Then np.array repr w) will yield MaskedArray [ 1991 11 03, ), None, datetime.date 2001, 3, 10))], \n dtype=[ date1 , <M8[D] ), date2 , <M8[D] )]) This is wrong because it gives 1991 11 03 instead of datetime.date 1991, 11, 3) in the [0,0] entry.","bug"
|
|
"6013","2521","PyArray API is not a unique symbol.
|
|
Trac 1928).
|
|
Original ticket http projects.scipy.org numpy ticket 1928 on 2011 08 10 by trac user phil ..., assigned to pearu.
|
|
numpy makes the API to its array type available through the symbol PyArray API.
|
|
But it also allows users to make it available through some other symbol by defining PY ARRAY UNIQUE SYMBOL to be whatever other symbol is desired.
|
|
Why anyone would want to use a different symbol name never became clear to me, but some do.)
|
|
multiarray api.h has some C preprocessor magic that defines the symbol one way if PY ARRAY UNIQUE SYMBOL is defined, and another way if PY ARRAY UNIQUE SYMBOL is not defined.
|
|
It all works well, except...
|
|
...In the numpy header fortranobject.h, PY ARRAY UNIQUE SYMBOL is defined as PyArray API.
|
|
This makes it look as if a different symbol is being used, but the different symbol being used is the same as the real symbol for the API.
|
|
This results in anyone including fortranobject.h scipy does) having a definition of PyArray API that does not play well at all with the real global definition of PyArray API.
|
|
I was sometimes getting a crash when invoking import scipy.signal I tracked it down to signal imports special directly or indirectly) import of special results in call of C function initspecfun ) initspecfun ) calls import array import array executes this line of code PyArray API = void )PyCObject AsVoidPtr c api); It was this statement that was causing trouble.
|
|
My debugger was telling me that the value being returned from PyCObject AsVoidPtr was the same as the address of PyArray API.
|
|
In other words, after the above statement, PyArray API == PyArray API is true.
|
|
I m pretty sure that s not good.
|
|
It resulted in the first attempt to access the contents of PyArray API a call to PyArray GetNDArrayCVersion ) a few lines later) crashing.
|
|
The crash only happens if I do the import scipy.signal after I import my own C python module that uses numpy arrays.
|
|
I suppose it could be a bug in my code, but I can t see how.
|
|
My source file just has a include <numarray libnumarray.h> , without define any optional symbols no PY ARRAY UNIQUE SYMBOL, no NO IMPORT or NO IMPORT ARRAY).
|
|
It does not anywhere directly) reference PyArray API.
|
|
The only thing the init routine does with numpy is call import array.
|
|
Of course, my module has routines that do stuff with numpy arrays, but I don t call any of those between importing my module and importing scipy.signal to get the crash.
|
|
If I am doing something wrong, then consider this to be a bug report on the docs, which have not made it clear what I am doing wrong.
|
|
I ve looked.
|
|
A bit.)
|
|
I tried to fix the problem by removing the define of PY ARRAY UNIQUE SYMBOL in fortranobject.h.
|
|
That did indeed prevent the crash when I import my module followed by scipy.signal.
|
|
However, if I import scipy.signal without first importing my module, the attempt to import scipy.signal raises in ImportError with complaint of undefined symbol PyArray API.
|
|
I then tried to fix it by changing the define of PY ARRAY UNIQUE SYMBOL in fortranobject.h to something anything) other than PyArray API.
|
|
That worked.
|
|
I am now able to import scipy.signal whether or not I first import my module.
|
|
Is that the right fix?
|
|
You decide.
|
|
Python 2.5.4 numpy 1.5.1rc2 scipy 0.8.0","bug"
|
|
"6013","2522","numpy.diff fails on unsigned integers Trac 1929).
|
|
Original ticket http projects.scipy.org numpy ticket 1929 on 2011 08 10 by trac user tlatorre, assigned to unknown.
|
|
It appears numpy.diff insists on returning an array of the same datatype as the input which is incorrect for unsigned integers.
|
|
>>> np.diff np.array [3,2], dtype=np.uint8)) array [255], dtype=uint8) >>> np.diff np.array [1,0], dtype=np.uint16)) array [65535], dtype=uint16) >>> np.diff np.array [3,2,0], dtype=np.uint32)) array [4294967295, 4294967294], dtype=uint32) >>> np.diff np.array [3,2,0], dtype=np.uint64)) array [18446744073709551615, 18446744073709551614], dtype=uint64)","bug"
|
|
"6013","2523","np.left shift and np.right shift fail on np.uint64 scalar types Trac 1930).
|
|
Original ticket http projects.scipy.org numpy ticket 1930 on 2011 08 10 by trac user tlatorre, assigned to unknown.
|
|
>>> np.uint64 5) << 1 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError ufunc left shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe >>> np.uint64 5) >> 3 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError ufunc right shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe","bug"
|
|
"6013","2524","np.left shift and np.right shift fail on np.uint64 scalar types Trac 1931).
|
|
Original ticket http projects.scipy.org numpy ticket 1931 on 2011 08 10 by trac user tlatorre, assigned to unknown.
|
|
>>> np.uint64 5) << 1 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError ufunc left shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe >>> np.uint64 5) >> 3 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError ufunc right shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe","bug"
|
|
"6013","2525","f2py crashes with UnboundLocalError exception Trac 1932).
|
|
Original ticket http projects.scipy.org numpy ticket 1932 on 2011 08 12 by trac user josdekloe, assigned to pearu.
|
|
I have been trying to use the f2py routinely to build an interface around the ECMWF BUFR library.
|
|
See http code.google.com p pybufr ecmwf for details.
|
|
This works fine for the python 2.x versions.
|
|
However, I also tried to add the possibility to use python3 and here I encountered some trouble.
|
|
My python code converts properly to python3 using to 2to3 tool, but when I try to use the python3 version of f2py I encounter this error somewhere deep inside the f2py internals Post processing stage 2)...
|
|
Saving signatures to file f2py build signatures.pyf Traceback most recent call last) File .
|
|
run f2py tool.py , line 11, in <module> main ) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 563, in main run main sys.argv[1 ]) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 342, in run main postlist=callcrackfortran files,options) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 279, in callcrackfortran pyf=crackfortran.crack2fortran postlist) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2693, in crack2fortran pyf=crack2fortrangen block) \n File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2505, in crack2fortrangen body=crack2fortrangen block[ body ],tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2505, in crack2fortrangen body=crack2fortrangen block[ body ],tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2506, in crack2fortrangen vars=vars2fortran block,block[ vars ],al,tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2655, in vars2fortran lst = true intent list vars[a]) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2557, in true intent list if c UnboundLocalError local variable c referenced before assignment ERROR build of python wrapper failed the signatures file could not be found Traceback most recent call last) File .
|
|
build interface.py , line 1803, in <module> BI.build ) File .
|
|
build interface.py , line 543, in build self.generate python wrapper source dir) File .
|
|
build interface.py , line 1667, in generate python wrapper raise InterfaceBuildError main .InterfaceBuildError > I any additional information is needed I would be happy to provide it.","bug"
|
|
"6013","2526","format strings native size used instead of standard size Trac 1933).
|
|
Original ticket http projects.scipy.org numpy ticket 1933 on 2011 08 12 by trac user skrah, assigned to unknown.
|
|
I think this array should fit into a buffer of size 4, even on a 64 bit platform >>> import struct >>> struct.calcsize =l ) 4 >>> ndarray buffer=bytearray 4), shape=[1], dtype= =l ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError buffer is too small for requested array >>> ndarray buffer=bytearray 8), shape=[1], dtype= =l ) array [0])","bug"
|
|
"6013","2527","Improve assignment error message when output can broadcast to the input Trac 1934).
|
|
Original ticket http projects.scipy.org numpy ticket 1934 on 2011 08 15 by mwiebe, assigned to unknown.
|
|
The code a = numpy.zeros 5,5)) v = numpy.ones 5,1)) a[ ,3] = v raises the exception ValueError?
|
|
output operand requires a reduction, but reduction is not enabled which is not very enlightening as to the nature of the error.
|
|
This illustrates why it errors >>> a = numpy.zeros 5,5)) >>> v = numpy.ones 5,1)) >>> a[ ,3] array [ 0., 0., 0., 0., 0.])
|
|
>>> v array [[ 1.
|
|
], [ 1.
|
|
], [ 1.
|
|
], [ 1.
|
|
], [ 1.]])
|
|
>>> a[ ,3].shape 5,) >>> v.shape 5, 1) The 5,) and 5,1) get broadcast to a 5,5) shape, which is what is really then being assigned into the 5,) array.
|
|
To fix it, write the assignment like a[ ,3] = v[ ,0].
|
|
Example taken from ticket 2463 comments.","enhancement"
|
|
"6013","2528","numpydoc latex longtables error Trac 1935).
|
|
Original ticket http projects.scipy.org numpy ticket 1935 on 2011 08 15 by matthew brett, assigned to unknown.
|
|
I ve run into an odd problem that seems to be an interaction of numpydoc and autosummary and large classes.
|
|
In summary, large classes and numpydoc lead to large tables of class methods, and there seems to be an error in the creation of the large tables in latex.
|
|
Specifically, if I run make latexpdf with the attached minimal sphinx setup, I get a pdflatex error ending thus ... l.118 \begin{longtable}{LL} and this is because longtable does not accept LL as an argument, but needs |l|l| bar el bar el bar).
|
|
I see in sphinx.writers.latex.py, around line 657, that sphinx knows about this in general, and long tables in standard ReST work fine with the el bar arguments passed to longtable.
|
|
if self.table.colspec self.body.append self.table.colspec) else if self.table.has problematic colwidth = 0.95 self.table.colcount colspec = p{ .3f\\linewidth}| colwidth) \ self.table.colcount self.body.append {| colspec }\n ) elif self.table.longtable self.body.append {| l| self.table.colcount) }\n ) else self.body.append {| L| self.table.colcount) }\n ) However, using numpydoc and autosummary see the conf.py file), what seems to happen is that, when we reach the self.table.colspec test at the beginning of the snippet above, self.table.colspec is defined In [1] self.table.colspec Out[1] {LL}\n and thus the LL gets written as the arg to longtable \begin{longtable}{LL} and the pdf build breaks.
|
|
I m using the numpydoc out of the current numpy source tree.","bug"
|
|
"6013","2529","loadtxt with nested dtype fails with ValueError setting an array element with a sequence Trac 1936).
|
|
Original ticket http projects.scipy.org numpy ticket 1936 on 2011 08 16 by gerritholl, assigned to unknown.
|
|
I m sure this used to work.
|
|
It also works for someone else on IRC.
|
|
python Python 2.6.6 r266 84292, Sep 15 2010, 16 22 56) [GCC 4.4.5] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import loadtxt, uint8 >>> from StringIO import StringIO >>> from numpy.version import version >>> print version 2.0.0.dev 5cf0a07 >>> loadtxt StringIO 0 1 2 3 ), dtype=[ a , uint8, 2), b , uint8, 2)]) Traceback most recent call last) File <stdin> , line 1, in <module> File storage4 home gerrit .local lib python2.6 site packages numpy lib npyio.py , line 806, in loadtxt X = np.array X, dtype) ValueError setting an array element with a sequence.
|
|
>>> loadtxt StringIO 0 1 2 3 ), dtype=[ a , uint8, 4)]) Traceback most recent call last) File <stdin> , line 1, in <module> File storage4 home gerrit .local lib python2.6 site packages numpy lib npyio.py , line 806, in loadtxt X = np.array X, dtype) ValueError setting an array element with a sequence.
|
|
What s going on here?","bug"
|
|
"6013","2530","Segfault triggered by importing print coercion tables.py Trac 1937).
|
|
Original ticket http projects.scipy.org numpy ticket 1937 on 2011 08 16 by charris, assigned to unknown.
|
|
Here array scalar ?
|
|
b h i l q p B H I L Q P e f d g F D G S U V O M m ?
|
|
?
|
|
b h i l l l B H I L L L e f d g F D G O O O !
|
|
m b b b b b b b b b b b b b b e f d g F D G O O O !
|
|
m h h h h h h h h h h h h h h f f d g F D G O O O !
|
|
m i i i i i i i i i i i i i i d d d g D D G O O O !
|
|
m l l l l l l l l l l l l l l d d d g D D G O O O !
|
|
m q l l l l l l l l l l l l l d d d g D D G O O O !
|
|
m p l l l l l l l l l l l l l d d d g D D G O O O !
|
|
m B B B B B B B B B B B B B B e f d g F D G O O O !
|
|
m H H H H H H H H H H H H H H f f d g F D G O O O !
|
|
m I I I I I I I I I I I I I I d d d g D D G O O O !
|
|
m L L L L L L L L L L L L L L d d d g D D G O O O !
|
|
m Q L L L L L L L L L L L L L d d d g D D G O O O !
|
|
m P L L L L L L L L L L L L L d d d g D D G O O O !
|
|
m e e e e e e e e e e e e e e e e e e F F F O O O !
|
|
f f f f f f f f f f f f f f f f f f F F F O O O !
|
|
d d d d d d d d d d d d d d d d d d D D D O O O !
|
|
g g g g g g g g g g g g g g g g g g G G G O O O !
|
|
F F F F F F F F F F F F F F F F F F F F F O O O !
|
|
D D D D D D D D D D D D D D D D D D D D D O O O !
|
|
G G G G G G G G G G G G G G G G G G G G G O O O !
|
|
S O O O O O O O O O O O O O O O O O O O O O O O !
|
|
O U O O O O O O O O O O O O O O O O O O O O O O O !
|
|
O Segmentation fault core dumped)","bug"
|
|
"6013","2531","nditer reusing reduce loop inappropriately Trac 1938).
|
|
Original ticket http projects.scipy.org numpy ticket 1938 on 2011 08 18 by mwiebe, assigned to mwiebe.
|
|
In changing how reductions work in my missingdata branch, the following caused a memory corruption a = np.random.rand 100,100,100) np.minimum.reduce a) I ve tracked this down to the logic for reusing the reduction double loop structure when the buffer is smaller than the innermost loop.","bug"
|
|
"6013","2532","Datetime tests incorrectly require pytz module Trac 1939).
|
|
Original ticket http projects.scipy.org numpy ticket 1939 on 2011 08 19 by bsouthey, assigned to unknown.
|
|
With the new datetime features, Numpy tests require the use of the pytz module.
|
|
This breaks the Numpy policy of requiring no other dependencies.
|
|
The tests in test datetime.py need to rewritten to avoid this module.","bug"
|
|
"6013","2533","uint64 modulo arithmetic broken Trac 1940).
|
|
Original ticket http projects.scipy.org numpy ticket 1940 on 2011 08 22 by trac user AngerMCS, assigned to unknown.
|
|
Running numpy 1.6.0 on python 3.2.1 both compiled from source.
|
|
uint64 1) 2 is returning 0.0 which is obviously incorrect.
|
|
Also when using uint64, modulo is returning a a float64 which is illogical since it can never return anything other than a uint64.
|
|
Python 3.2.1 default, Jul 20 2011, 15 59 19) [GCC 4.3.2 20081105 Red Hat 4.3.2 7)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import uint64 >>> uint64 1) 18446744073709551615 >>> a = uint64 1) >>> a 18446744073709551615 >>> hex a) 0xffffffffffffffff >>> a 2 0.0","bug"
|
|
"6013","2534","ENH array equal array equiv for structured arrays Trac 1941).
|
|
Original ticket http projects.scipy.org numpy ticket 1941 on 2011 08 23 by dhomeier, assigned to unknown.
|
|
Could be considered a defect as well the array comparison functions do not operate on structured record arrays.
|
|
This could be fixed using a simple bool all [ logical and.reduce equal a1[n],a2[n]).ravel )) for n in a1.dtype.names ]) ) in numeric.py","enhancement"
|
|
"6013","2535","fortranobject.h not found on Debian Wheezy testing with kernel 3.0.0 1 486 Trac 1942).
|
|
Original ticket http projects.scipy.org numpy ticket 1942 on 2011 08 25 by trac user SergeStroobandt, assigned to pearu.
|
|
What steps will reproduce the problem?
|
|
[[BR]] easy install pymc[[BR]] or[[BR]] sudo easy install pymc[[BR]] What is the expected output?
|
|
What do you see instead?
|
|
Constructing wrapper function aniso geo rad ...[[BR]] aniso geo rad d,x,y,inc,ecc,[cmin,cmax,symm]) [[BR]]Wrote C API module distances to file build src.linux i686 2.6 pymc gp cov funs distancesmodule.c Could not locate executable g77[[BR]] Could not locate executable f77[[BR]] Could not locate executable ifort[[BR]] Could not locate executable ifc[[BR]] Could not locate executable lf95[[BR]] Could not locate executable pgf90[[BR]] Could not locate executable pgf77[[BR]] Could not locate executable f90[[BR]] Found executable usr bin f95[[BR]] Could not locate executable fort[[BR]] Could not locate executable efort[[BR]] Could not locate executable efc[[BR]] Found executable usr bin gfortran[[BR]] In file included from build src.linux i686 2.6 fortranobject.c 2 0 [[BR]] build src.linux i686 2.6 fortranobject.h 7 20 fatal error [[BR]]Python.h No such file or directory[[BR]] compilation terminated.
|
|
[[BR]] In file included from build src.linux i686 2.6 fortranobject.c 2 0 [[BR]] build src.linux i686 2.6 fortranobject.h 7 20 fatal error [[BR]]Python.h No such file or directory compilation terminated.
|
|
[[BR]] error Setup script exited with error Command gcc pthread fno strict aliasing DNDEBUG g fwrapv O2 Wall Wstrict prototypes fPIC DNO ATLAS INFO=1 Ibuild src.linux i686 2.6 I usr lib pymodules python2.6 numpy core include I usr include python2.6 c build src.linux i686 2.6 fortranobject.c o build temp.linux i686 2.6 build src.linux i686 2.6 fortranobject.o failed with exit status 1 usr lib pymodules python2.6 numpy distutils misc util.py 251 RuntimeWarning Parent module numpy.distutils not found while handling absolute import[[BR]] from numpy.distutils import log[[BR]] [[BR]] What version of the product are you using?
|
|
On what operating system?
|
|
[[BR]] [[BR]] Linux Mint Debian = Debian Wheezy testing)[[BR]] kernel 3.0.0 1 486[[BR]] on a VirtualBox machine[[BR]] [[BR]] Please provide any additional information below.
|
|
[[BR]] [[BR]] dpkg S fortranobject.h[[BR]] python numpy usr share pyshared numpy f2py src fortranobject.h[[BR]]","bug"
|
|
"6013","2536","numpy.frombuffer fails when given an empty buffer Trac 1943).
|
|
Original ticket http projects.scipy.org numpy ticket 1943 on 2011 08 25 by trac user hniksic, assigned to unknown.
|
|
numpy.frombuffer raises an exception when given an empty buffer, where one would expect it to return an empty array of given dtype.
|
|
For example >>> import array, numpy as np >>> np.frombuffer array.array d )) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError offset must be non negative and smaller than buffer lenth 0) This requires code converting buffers to numpy arrays to special case empty buffers.
|
|
Furthermore, the problem is often not discovered until someone actually passes a zero sized buffer to the code that calls frombuffer, and trips over this exception at run time.","bug"
|
|
"6013","2537","ieee754.c in numpy 1.6.1 doesnt compile under cygwin Trac 1944).
|
|
Original ticket http projects.scipy.org numpy ticket 1944 on 2011 08 25 by trac user sebastian sohr, assigned to unknown.
|
|
At least compile and install works with the following change In numpy core src npymath ieee754.c replace elif defined CYGWIN ) include fenv fenv.c endif by elif defined CYGWIN ) include numpy fenv fenv.h endif dont know whether functionality is ok","bug"
|
|
"6013","2538","matrix [1,2,3]) and matrix [[1,2,3]]) gives the same result Trac 1945).
|
|
Original ticket http projects.scipy.org numpy ticket 1945 on 2011 08 28 by trac user smichr, assigned to unknown.
|
|
It would be convenient if the former representation gave a column vector matrix [[1],[2],[3]]) instead of being equivalent to the row vector matrix [[1,2,3]])).","bug"
|
|
"6013","2539","warning ISO C when creating an array in C API Trac 1946).
|
|
Original ticket http projects.scipy.org numpy ticket 1946 on 2011 08 31 by trac user kcharpen, assigned to unknown.
|
|
Every array creation function gives me a warning when compiling with the option pedantic, with gcc 4.4.3.
|
|
Even the simplest creation such as PyObject pnewArray = PyArray SimpleNew 2, dim, 1); will give me the following warning at compilation warning ISO C forbids casting between pointer to function and pointer to object Is there any way to get rid of it ?
|
|
Since that despite the warning, the array creation seems to work, and i can see no problem at execution when further manipulating the array.","bug"
|
|
"6013","2540","Numerical error in Numpy causes different eigenvector matrices to be returned when close to identity.
|
|
Trac 1947).
|
|
Original ticket http projects.scipy.org numpy ticket 1947 on 2011 09 05 by trac user shyuep, assigned to unknown.
|
|
Using numpy to find eigenvectors, I have found that different results are returned for matrices very close to identity.
|
|
An example of the output >>> import numpy as np >>> mat = np.array [[ 1, 0, 5.55111512e 17], [ 0, 1, 0], [ 5.55111512e 17, 0, 1]]) >>> np.linalg.eig mat) array [ 1., 1., 1.
|
|
]), array [[ 0.70710678, 0.70710678, 0.
|
|
], [ 0. , 0. , 1.
|
|
], [ 0.70710678, 0.70710678, 0. ]]))
|
|
>>> mat = np.array [[ 1, 0, 5.55111512e 19], [ 0, 1, 0], [ 5.55111512e 17, 0, 1]]) >>> np.linalg.eig mat) array [ 1., 1., 1.
|
|
]), array [[ 0.09950372, 0.15094684, 0.
|
|
], [ 0. , 0. , 1.
|
|
], [ 0.99503719, 0.98854188, 0. ]]))
|
|
>>> mat = np.array [[ 1, 0, 5.55111512e 29], [ 0, 1, 0], [ 5.55111512e 17, 0, 1]]) >>> np.linalg.eig mat) array [ 1., 1., 1.
|
|
]), array [[ 1.00000000e 06, 9.09494745e 07, 0.00000000e 00], [ 0.00000000e 00, 0.00000000e 00, 1.00000000e 00], [ 1.00000000e 00, 1.00000000e 00, 0.00000000e 00]])) >>> mat = np.array [[ 1, 0, 0], [ 0, 1, 0], [ 0, 0, 1]]) >>> np.linalg.eig mat) array [ 1., 1., 1.
|
|
]), array [[ 1., 0., 0.
|
|
], [ 0., 1., 0.
|
|
], [ 0., 0., 1.]]))
|
|
The first three matrices only differ in one element in the exponent, going from 5.55111512e 17 to 5.55111512e 19 to 5.55111512e 29, i.e., very small numbers and the matrices are essentially the identiy matrix.
|
|
However, in all cases, different eigenvectors are returned.
|
|
While all the eigenvectors are valid of course, the different answers creates issues in comparing the eigenvectors obtained via automatically generated matrices.
|
|
This issue does not appear in other numerical packages in other languages.
|
|
I have tested the Jama package in Java and consistent eigenvectors are obtained.","bug"
|
|
"6013","2541","chararray subscription returns empty string instead of empty list Trac 1948).
|
|
Original ticket http projects.scipy.org numpy ticket 1948 on 2011 09 06 by trac user jazzer, assigned to unknown.
|
|
Current behaviour 1.5.1, 1.6.1) >>> import numpy >>> a=numpy.rec.array [ a ,1), bb ,2), cc ,3)],dtype=[ s , |S4 ), n , i4 )]) >>> print a.s[a.s> b ] [ bb cc ] >>> print a.s[a.s> c ] [ cc ] >>> print a.s[a.s> d ] >>> a.s[a.s> d ] The correct behaviour 1.2.1) >>> print a.s[a.s> d ] [] >>> a.s[a.s> d ] chararray [], dtype= |S4 ) The bug was introduced in the [7deb4d5db4282e1e1c86] http github.com numpy numpy commit 7deb4d5db4282e1e1c86), file numpy core defchararray.py I m not sure to which NumPy release this commit corresponds, it s somewhere between 1.2.1 and 1.5.1) def getitem self, obj) val = ndarray.
|
|
getitem self, obj) if isinstance val, string , unicode )) if issubclass val.dtype.type, character) temp = val.rstrip ) Reverting the change in this line fixes the bug.","bug"
|
|
"6013","2542","Upcasting a float32 array to complex64 instead of complex128 when adding a complex128 scalar Trac 1949).
|
|
Original ticket http projects.scipy.org numpy ticket 1949 on 2011 09 06 by trac user tihocan, assigned to unknown.
|
|
This is with numpy 1.6.0 and 1.6.1 under Linux x86 64, testing the upcast mechanism of scalar array >>> import numpy; print numpy.array 2, dtype=numpy.complex128) numpy.ones 3, dtype=numpy.float32)).dtype complex64 Since it has to upcast my array float32 is not compatible enough with complex128), why does it upcast it to complex64 instead of complex128?
|
|
As far as I can tell 1.4.x and 1.5.x versions of numpy are indeed upcasting to complex128.
|
|
Also note that When replacing numpy.ones with numpy.array it yields complex128 expected upcast of scalar addition of complex128 with float32) The behavior is similar if instead of 2 I use a number which cannot be represented exactly with a complex64 so it s not a rule about picking the smallest data type able to exactly represent the result) This issue was reported on the mailing list in http mail.scipy.org pipermail numpy discussion 2011 August 057981.html","bug"
|
|
"6013","2543","Redundant numeric type classes lead to unreliable behavior of isinstance ) Trac 1950).
|
|
Original ticket http projects.scipy.org numpy ticket 1950 on 2011 09 09 by trac user stereotype441, assigned to unknown.
|
|
This bug was discovered during the development of the Piglit OpenGL test suite see https bugs.freedesktop.org show bug.cgi?id=40697) Numpy seems to define multiple classes called numpy.int32 , which are considered distinct by Python s isinstance ) function.
|
|
This makes it unreliable to use isinstance ) to figure out the type of a numpy numeric value.
|
|
For example, this is the behavior I see on Windows using using numpy 1.6.1 Python 2.7.2 default, Jun 12 2011, 15 08 59) [MSC v.1500 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.version.version 1.6.1 >>> x = numpy.int32 5) >>> y = numpy.abs x) >>> type x) <type numpy.int32 > >>> type y) <type numpy.int32 > >>> isinstance x, numpy.int32) True >>> isinstance y, numpy.int32) False >>> type x) == type y) False I would have expected the last two False outputs to be True .
|
|
And when I run the same experiment on Fedora Linux which is currently shipping numpy 1.5.1), that is indeed what I see.
|
|
Since I m not running the same version of numpy on Linux and Windows, I m not sure whether this problem is linked to numpy version or to platform.
|
|
I have seen similar problems with uint32.
|
|
You can see further evidence that there are multiple redundant numeric type classes by typing help numpy) from the command prompt.
|
|
When I do this on Windows, I see duplicate copies of the following numeric classes complex128, complex192, complex64, float16, float32, float64, float96, int16, int32, int64, int8, datetime64, timedelta64, uint16, uint32, uint64, and uint8.
|
|
Surprisingly, I see similar duplication even on Fedora Linux, even though it doesn t exhibit this bug.
|
|
This arose in our project because we were trying to use isinstance ) to tell the type of a value that might be one of several numpy numeric types, or might by a Python built in type.
|
|
We are planning to work around the problem by using isinstance value, numpy.numeric) to detect whether the value is a numpy numeric type, and then using value.dtype to figure out which numeric type it is.
|
|
But it seems like this extra work shouldn t be necessary isinstance value, numpy.int32) ought to be sufficient to determine whether a value is a numpy 32 bit int.","enhancement"
|
|
"6013","2544","Numpy doesn t build with llvm gcc on OS X 10.7 Lion) Trac 1951).
|
|
Original ticket http projects.scipy.org numpy ticket 1951 on 2011 09 10 by fonnesbeck, assigned to unknown.
|
|
The current main repository code no longer builds on OSX 10.7.
|
|
Sometime during the last 3 weeks, something in multiarray was broken.
|
|
Here is the relevant part of the build log In file included from numpy core src multiarray multiarraymodule onefile.c 45 numpy core src multiarray lowlevel strided loops.c.src At top level numpy core src multiarray lowlevel strided loops.c.src 314 error expected xe2x80x98=xe2x80x99, xe2x80x98,xe2x80x99, xe2x80x98;xe2x80x99, xe2x80x98asmxe2x80x99 or xe2x80x98 attribute xe2x80x99 before xe2x80x98 xe2x80x99 token numpy core src multiarray lowlevel strided loops.c.src 469 error expected xe2x80x98=xe2x80x99, xe2x80x98,xe2x80x99, xe2x80x98;xe2x80x99, xe2x80x98asmxe2x80x99 or xe2x80x98 attribute xe2x80x99 before xe2x80x98 xe2x80x99 token numpy core src multiarray lowlevel strided loops.c.src 469 error expected xe2x80x98=xe2x80x99, xe2x80x98,xe2x80x99, xe2x80x98;xe2x80x99, xe2x80x98asmxe2x80x99 or xe2x80x98 attribute xe2x80x99 before xe2x80x98 xe2x80x99 token numpy core src multiarray lowlevel strided loops.c.src 851 error expected xe2x80x98=xe2x80x99, xe2x80x98,xe2x80x99, xe2x80x98;xe2x80x99, xe2x80x98asmxe2x80x99 or xe2x80x98 attribute xe2x80x99 before xe2x80x98 xe2x80x99 token numpy core src multiarray lowlevel strided loops.c.src 932 error expected declaration specifiers or xe2x80x98...xe2x80x99 before xe2x80x98PyArray StridedTransferFnxe2x80x99 numpy core src multiarray lowlevel strided loops.c.src 934 error conflicting types for xe2x80x98PyArray TransferNDimToStridedxe2x80x99 numpy core src private lowlevel strided loops.h 317 error previous declaration of xe2x80x98PyArray TransferNDimToStridedxe2x80x99 was here numpy core src multiarray lowlevel strided loops.c.src In function xe2x80x98PyArray TransferNDimToStridedxe2x80x99 numpy core src multiarray lowlevel strided loops.c.src 943 warning implicit declaration of function xe2x80x98stransferxe2x80x99 numpy core src multiarray lowlevel strided loops.c.src At top level numpy core src multiarray lowlevel strided loops.c.src 1051 error expected declaration specifiers or xe2x80x98...xe2x80x99 before xe2x80x98PyArray StridedTransferFnxe2x80x99 numpy core src multiarray lowlevel strided loops.c.src 1053 error conflicting types for xe2x80x98PyArray TransferStridedToNDimxe2x80x99 numpy core src private lowlevel strided loops.h 327 error previous declaration of xe2x80x98PyArray TransferStridedToNDimxe2x80x99 was here numpy core src multiarray lowlevel strided loops.c.src 1171 error expected declaration specifiers or xe2x80x98...xe2x80x99 before xe2x80x98PyArray MaskedStridedTransferFnxe2x80x99 numpy core src multiarray lowlevel strided loops.c.src 1173 error conflicting types for xe2x80x98PyArray TransferMaskedStridedToNDimxe2x80x99 numpy core src private lowlevel strided loops.h 338 error previous declaration of xe2x80x98PyArray TransferMaskedStridedToNDimxe2x80x99 was here In file included from numpy core src multiarray multiarraymodule onefile.c 47 numpy core src multiarray einsum.c.src In function xe2x80x98PyArray EinsteinSumxe2x80x99 numpy core src multiarray einsum.c.src 3016 warning implicit declaration of function xe2x80x98PyArray FillWithZeroxe2x80x99 In file included from numpy core src multiarray multiarraymodule onefile.c 49 numpy core src multiarray array assign scalar.c At top level numpy core src multiarray array assign scalar.c 335 error conflicting types for xe2x80x98PyArray AssignRawScalarxe2x80x99 numpy core src multiarray array assign scalar.c 335 note an argument type that has a default promotion canxe2x80x99t match an empty parameter name list declaration numpy core src multiarray convert.c 407 error previous implicit declaration of xe2x80x98PyArray AssignRawScalarxe2x80x99 was here In file included from numpy core src multiarray multiarraymodule onefile.c 50 numpy core src multiarray array assign array.c 392 error conflicting types for xe2x80x98PyArray AssignArrayxe2x80x99 numpy core src multiarray array assign array.c 392 note an argument type that has a default promotion canxe2x80x99t match an empty parameter name list declaration numpy core src multiarray ctors.c 2907 error previous implicit declaration of xe2x80x98PyArray AssignArrayxe2x80x99 was here numpy core src multiarray array assign array.c In function xe2x80x98PyArray AssignArrayxe2x80x99 numpy core src multiarray array assign array.c 403 warning initialization makes pointer from integer without a cast numpy core src multiarray array assign array.c 407 warning implicit declaration of function xe2x80x98PyArray AssignNAxe2x80x99 In file included from numpy core src multiarray multiarraymodule onefile.c 51 numpy core src multiarray reduction.c At top level numpy core src multiarray reduction.c 801 error conflicting types for xe2x80x98PyArray ReduceWrapperxe2x80x99 numpy core src multiarray item selection.c 2028 error previous implicit declaration of xe2x80x98PyArray ReduceWrapperxe2x80x99 was here In file included from numpy core src multiarray multiarraymodule onefile.c 52 numpy core src multiarray na mask.c 55 error static declaration of xe2x80x98PyArray ContainsNAxe2x80x99 follows non static declaration numpy core src multiarray common.c 104 error previous implicit declaration of xe2x80x98PyArray ContainsNAxe2x80x99 was here numpy core src multiarray na mask.c 243 error conflicting types for xe2x80x98PyArray AssignMaskNAxe2x80x99 numpy core src multiarray na mask.c 243 note an argument type that has a default promotion canxe2x80x99t match an empty parameter name list declaration numpy core src multiarray ctors.c 2670 error previous implicit declaration of xe2x80x98PyArray AssignMaskNAxe2x80x99 was here numpy core src multiarray na mask.c 331 error conflicting types for xe2x80x98PyArray AllocateMaskNAxe2x80x99 numpy core src multiarray na mask.c 331 note an argument type that has a default promotion canxe2x80x99t match an empty parameter name list declaration numpy core src multiarray flagsobject.c 286 error previous implicit declaration of xe2x80x98PyArray AllocateMaskNAxe2x80x99 was here numpy core src multiarray na mask.c 468 error conflicting types for xe2x80x98PyArray AssignNAxe2x80x99 numpy core src multiarray na mask.c 468 note an argument type that has a default promotion canxe2x80x99t match an empty parameter name list declaration numpy core src multiarray array assign array.c 407 error previous implicit declaration of xe2x80x98PyArray AssignNAxe2x80x99 was here numpy core src multiarray na mask.c In function xe2x80x98PyArray IsNAxe2x80x99 numpy core src multiarray na mask.c 500 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) In file included from numpy core src multiarray multiarraymodule onefile.c 53 numpy core src multiarray na object.c In function xe2x80x98na callxe2x80x99 numpy core src multiarray na object.c 84 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c In function xe2x80x98na richcomparexe2x80x99 numpy core src multiarray na object.c 164 warning assignment makes pointer from integer without a cast numpy core src multiarray na object.c In function xe2x80x98NpyNA CombineNAxe2x80x99 numpy core src multiarray na object.c 285 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c In function xe2x80x98NpyNA CombineNAWithObjectxe2x80x99 numpy core src multiarray na object.c 328 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c At top level numpy core src multiarray na object.c 424 error conflicting types for xe2x80x98NpyNA FromObjectxe2x80x99 numpy core src multiarray scalarapi.c 824 error previous implicit declaration of xe2x80x98NpyNA FromObjectxe2x80x99 was here numpy core src multiarray na object.c In function xe2x80x98NpyNA FromObjectxe2x80x99 numpy core src multiarray na object.c 426 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c At top level numpy core src multiarray na object.c 489 error conflicting types for xe2x80x98NpyNA FromDTypeAndPayloadxe2x80x99 numpy core src multiarray mapping.c 1806 error previous implicit declaration of xe2x80x98NpyNA FromDTypeAndPayloadxe2x80x99 was here numpy core src multiarray na object.c In function xe2x80x98NpyNA FromDTypeAndPayloadxe2x80x99 numpy core src multiarray na object.c 493 error xe2x80x98Npy NAxe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c 497 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c In function xe2x80x98na binaryopxe2x80x99 numpy core src multiarray na object.c 551 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c In function xe2x80x98na andxe2x80x99 numpy core src multiarray na object.c 577 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) numpy core src multiarray na object.c In function xe2x80x98na orxe2x80x99 numpy core src multiarray na object.c 615 error xe2x80x98NpyNA Typexe2x80x99 undeclared first use in this function) error Command llvm gcc 4.2 DNDEBUG g fwrapv Os Wall Wstrict prototypes arch x86 64 pipe Inumpy core include Ibuild src.macosx 10.7 x86 64 2.7 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I System Library Frameworks Python.framework Versions 2.7 include python2.7 Ibuild src.macosx 10.7 x86 64 2.7 numpy core src multiarray Ibuild src.macosx 10.7 x86 64 2.7 numpy core src umath c numpy core src multiarray multiarraymodule onefile.c o build temp.macosx 10.7 x86 64 2.7 numpy core src multiarray multiarraymodule onefile.o failed with exit status 1","bug"
|
|
"6013","2545","Float not allowed in std ddof Trac 1952).
|
|
Original ticket http projects.scipy.org numpy ticket 1952 on 2011 09 11 by pv, assigned to unknown.
|
|
This should also take a floating point value >>> np.array [1, 1, 2]).std ddof=1.4) TypeError integer argument expected, got float In earlier versions of Numpy, this raised a DeprecationWarning which is also not correct behavior.","bug"
|
|
"6013","2546","fill diagonal fail on rectangular arrays Trac 1953).
|
|
Original ticket http projects.scipy.org numpy ticket 1953 on 2011 09 18 by trac user dimasad, assigned to unknown.
|
|
For rectangular arrays that are more tall than large, fill diagonal wraps up.
|
|
Example >>> a = np.zeros 4,2)) >>> np.fill diagonal a,4) >>> a array [[ 4., 0.
|
|
], [ 0., 4.
|
|
], [ 0., 0.
|
|
], [ 4., 0.]])
|
|
When it should be just >>> a array [[ 4., 0.
|
|
], [ 0., 4.
|
|
], [ 0., 0.
|
|
], [ 0., 0.]])","bug"
|
|
"6013","2547","f2py no lower does not work in one step compilation Trac 1954).
|
|
Original ticket http projects.scipy.org numpy ticket 1954 on 2011 09 26 by trac user azrael, assigned to pearu.
|
|
In order to impliment exception handling for Fortran errors into a f2py wrapped function, I followed this advice [http mail.scipy.org pipermail numpy discussion 2009 January 039672.html] I wanted this to work without editing the .pyf file, and got that by adding the following to the Fortran code !f2py callstatement f2py func) ... ); if ierr == 1) PyErr SetString PyExc ValueError, msg ) !f2py callprotoargument ...
|
|
The ... is just a replacement for actual code) When I tried to compile this with f2py c m m src.f I got several warnings errors like this warning implicit declaration of function xe2x80x98pyerr setstringxe2x80x99 Obviously the case lowering acted on this line.
|
|
As far as I understand the User Guide this should not happen, since I didn t call f2py with h or lower .
|
|
Nevertheless I added the no lower , but still got the same warnings errors.
|
|
Then I tried whether the added line will show up in the .pyf correctly, which it did, when I used no lower like this f2py no lower m m h m.pyf src.f After generating the .pyf I can compile the module without further editing in the .pyf!
|
|
f2py c m.pyf src.f 1.
|
|
Is this intended behaviour?
|
|
Should it make a difference whether I generate the .pyf first and then compile or use the one step way?
|
|
2.
|
|
Is there a way to circumvent using callprotoargument ?
|
|
I d much rather have f2py do it s magic to guess the first line in callstatement and the callprotoargument, and then just add my piece of extra code.
|
|
Can this be done?
|
|
Something like a addtocallstatement directive comes to mind ...","bug"
|
|
"6013","2548","Assignment into sub array fields functions wrong Trac 1955).
|
|
Original ticket http projects.scipy.org numpy ticket 1955 on 2011 09 29 by pv, assigned to unknown.
|
|
Consider assignment of an array to a sub array field >>> import numpy as np >>> np.
|
|
version 2.0.0.dev fac8674 >>> dt = np.dtype a3, 2,2)u2 ) >>> x = np.zeros 5, dtype=dt) >>> x[1][ f1 ] = np.array [[1,1],[1,1]]) >>> x[1][ f1 ] array [[1, 0], [0, 0]], dtype=uint16) This is clearly wrong behavior.
|
|
The problem has probably to do with assigning an array into an array scalar field.
|
|
This works, however >>> x[ f1 ][1] = np.array [[2,2],[2,2]]) >>> x[1][ f1 ] array [[2, 2], [2, 2]], dtype=uint16)","bug"
|
|
"6013","2549","inner1d sometimes dumps core on arrays whose shape contains 0 Trac 1956).
|
|
Original ticket http projects.scipy.org numpy ticket 1956 on 2011 09 29 by trac user tbarnetlamb, assigned to unknown.
|
|
The following appears to be an error In [1] import numpy.version In [2] numpy.version.version Out[2] 1.4.1 In [3] from numpy.core.umath tests import inner1d In [4] inner1d numpy.zeros 0,)), numpy.zeros 0,))) okay Out[4] 0.0 In [5] inner1d numpy.zeros 0, 2)), numpy.zeros 0, 2))) crashes Fatal Python error PyEval RestoreThread NULL tstate Abort","bug"
|
|
"6013","2550","column mean is not invariant to row permutation Trac 1957).
|
|
Original ticket http projects.scipy.org numpy ticket 1957 on 2011 09 29 by trac user john200p4, assigned to unknown.
|
|
Hello.
|
|
I noticed this behavior and do not know if it is a bug.
|
|
Given a large matrix, the column means will be slightly different depending on the permutation of the rows.
|
|
For example >>> a = random.rand 2300,20) >>> ii = random.permutation 2300) >>> b = a[ii].copy ) >>> abs a.mean 0) b.mean 0)).max ) 1.3322676295501878e 15 >>> abs a.mean 0) b[argsort ii)].mean 0)).max ) 0.0","bug"
|
|
"6013","2551","SWIG target filetype bug Trac 1958).
|
|
Original ticket http projects.scipy.org numpy ticket 1958 on 2011 10 03 by trac user woosong, assigned to cournape.
|
|
When SWIG is used in distutils with swig opts c target filetype is set to be .cpp .
|
|
However, when .i file has a header c in front, target filetype is incorrectly set to .c causing error.
|
|
https github.com numpy numpy blob master numpy distutils command build src.py L642 This line sets the target ext to .c but does not change it according to typ in the following several lines.
|
|
Should be a simple fix.","bug"
|
|
"6013","2552","assert equal should test for equality, not for inequality Trac 1959).
|
|
Original ticket http projects.scipy.org numpy ticket 1959 on 2011 10 07 by trac user ezander, assigned to unknown.
|
|
I just had the strange indicent that I had two objects, call them a and b, that compared fine when I used `self.assertEqual a, b)` in a `unittest.TestCase` derived class, but failed when I used `assert equal a, b)` from `numpy.testing`.
|
|
The reason for this strange behaviour was that `assertEqual` uses something like `not a==b)` in its test, while `numpy.assert equal` uses `a!=b`.
|
|
Therefore, the standard unittest uses ` eq ` to compare the objects, while numpy uses ` ne `, which I hadn t overwritten.
|
|
Sure, it was partly my fault here, because I had forgotten to overwrite ` ne ` appropriately, but there can also be cases, where this is on purpose that e.g.
|
|
both return True just consider NaN).
|
|
So, the question is which of both test is more appropriate for `assert equal`.
|
|
I think, `not a==b)` would be the right one, as I am asking for the objects to be equal, and not for them to be not unequal then the function should be called something like `assert not unequal` or `assert not not equal`).","bug"
|
|
"6013","2553","chararray assignment doesn t work as expected Trac 1960).
|
|
Original ticket http projects.scipy.org numpy ticket 1960 on 2011 10 08 by pv, assigned to unknown.
|
|
Pasted from http projects.scipy.org scipy ticket 1532 by faheem .
|
|
Seems to be also in current git master like this.
|
|
The following code is from the Numpy 1.6.0 manual, [http docs.scipy.org doc numpy 1.6.0 reference generated numpy.chararray.html numpy.chararray].
|
|
I ve tested with 1.6.1 on Debian squeeze.
|
|
There is a similar link in the development docs [http docs.scipy.org doc numpy reference generated numpy.chararray.html numpy.chararray].
|
|
import numpy as np charar = np.chararray 3, 3)) charar[ , ] = abc print charar According to the manual, this is supposed to output chararray [[ a , a , a ], [ a , a , a ], [ a , a , a ]], dtype= |S1 ) Instead I get [[ \xd8 3 U ] [ \xb7 \x18 \xc2 ] [ \xbd \x10 ]] This looks like uninitialized values.
|
|
I haven t tested to see if I can reproduce this in the dev version.","bug"
|
|
"6013","2554","numpy.log ) doesn t accept very large numbers Trac 1961).
|
|
Original ticket http projects.scipy.org numpy ticket 1961 on 2011 10 13 by trac user staticfloat, assigned to unknown.
|
|
If I type numpy.log 2 63), I get the expected answer of 42.975.
|
|
If I type numpy.log 2 64), I get AttributeError Traceback most recent call last) Users sabae <ipython input 40 d6bd9063dbbc> in <module> ) > 1 np.log 2 63) AttributeError log My numpy.version.version says 1.6.1 , but that is not an option in the dropdown box, so I ve set it to devel","bug"
|
|
"6013","2555","import multi array DLL load failed a dynamic link library <DLL> initialization routine failed Trac 1962).
|
|
Original ticket http projects.scipy.org numpy ticket 1962 on 2011 10 14 by trac user linden paul, assigned to unknown.
|
|
I am trying to embed a Python2.6 also using numpy and scipy) into a large C program.
|
|
Simple programs worked fine.
|
|
However, if the python module that I am calling from C includes the line from numpy import The whole thing fails and I get an error in visual studios when the python module runs.
|
|
R6034 an application has made an attempt to load the c runtime library incorrectly.
|
|
Using dependency walker I narrowed down the problem to GetProcAddress 0x73C10000 [DWMAPI.DLL], DwmGetWindowAttribute ) called from UXTHEME.DLL at address 0x7404C3DB and returned 0x73C11C76.
|
|
GetProcAddress 0x75520000 [GDI32.DLL], GdiIsMetaPrintDC ) called from USP10.DLL at address 0x75AA8306 and returned 0x75528AB9.
|
|
DllMain 0x5C7B0000, DLL PROCESS ATTACH, 0x00000000) in MSVCR90.DLL returned 0 0x0).
|
|
DllMain 0x5C7B0000, DLL PROCESS DETACH, 0x00000000) in MSVCR90.DLL called.
|
|
DllMain 0x5C7B0000, DLL PROCESS DETACH, 0x00000000) in MSVCR90.DLL returned 0 0x0).
|
|
Unloaded MULTIARRAY.PYD at address 0x69B80000.
|
|
Unloaded MSVCR90.DLL at address 0x5C7B0000.
|
|
LoadLibraryExA C \Python26\Lib\site packages\numpy\core\multiarray.pyd , 0x00000000, LOAD WITH ALTERED SEARCH PATH) returned NULL.
|
|
Error A dynamic link library DLL) initialization routine failed 1114).
|
|
Second chance exception 0xC0000005 Access Violation) occurred in PYTHON26.DLL at address 0x1E01A2FA.
|
|
Exited PYTHON TEST.EXE process 0x1064) with code 1073741819 0xC0000005).
|
|
Then I just adjusted the Python module to just import numarray from numpy import numarray def my eig a,b,c,d) print hello world return x In the command window the error displayed is Import multiarray ImportError DLL load failed A dynamic link library DLL) initialization routine failed.
|
|
I should note that numpy and scipy work fine if run from IDLE.","bug"
|
|
"6013","2556","hNumPy Example List wiki page is corrupted Trac 1963).
|
|
Original ticket http projects.scipy.org numpy ticket 1963 on 2011 10 16 by trac user AlanFrankel, assigned to unknown.
|
|
I ve been editing the Tentative NumPy Tutorial and occasionally referring to the NumPy Example List http www.scipy.org Numpy Example List ).
|
|
In the process, I think I mistakenly corrupted the NumPy Example List.
|
|
Since the website does not offer any wiki type functionality for reverting changes or referring to a history of changes, there doesn t seem to be a way for me to fix the problem.
|
|
Nor is there a Contact Us link on the page, so this was the only way I could find to report the problem.","bug"
|
|
"6013","2557","followup to 1962 Trac 1964).
|
|
Original ticket http projects.scipy.org numpy ticket 1964 on 2011 10 18 by trac user linden paul, assigned to unknown.
|
|
I narrowed the problem down to a group of pyd files includeing multiarray.pyd in the numpy core.
|
|
I have embedded python in MATLAB.
|
|
However, a call to these pyd files in the numpy core cause MATLAB to fail.
|
|
Is this a MATLAB problem or a Numpy problem?
|
|
follow up to ticket 2555","bug"
|
|
"6013","2558","unify names of numpy.distutils options Trac 1965).
|
|
Original ticket http projects.scipy.org numpy ticket 1965 on 2011 10 19 by rc, assigned to cournape.
|
|
rationale I have now spent several hours hunting down a major slowdown of my code caused apparently) by using config.add library ) for a reusable part of C source files instead of just config.add extension ).
|
|
The reason of the slowdown was different, but hard to discern, naming of options and silent ignoring of non existing ones add library ) extra compiler args add extension ) extra compile args Other build keys used for the same purpose also differ.","bug"
|
|
"6013","2559","numpy distutils fails to recognize current gfortran from mingw Trac 1966).
|
|
Original ticket http projects.scipy.org numpy ticket 1966 on 2011 10 22 by trac user jarrodwilcox, assigned to cournape.
|
|
There is no gfortran.py file in the directory.
|
|
There is a gnu file that in theory recognizes gfortran as an f77 compiler.
|
|
However, it does not appear to recognize the current gfortran.exe version distributed by MinGW.
|
|
Ideally, a gfortran.py file is needed.
|
|
however, at least F77 code could be compiled in f2py if gnu were fixed.
|
|
if you can not replicate the error in five minutes with a Windows 32 machine, please email me for further details.","bug"
|
|
"6013","2560","Buffer test error on python 3.
|
|
Trac 1967).
|
|
Original ticket http projects.scipy.org numpy ticket 1967 on 2011 10 23 by charris, assigned to pv.
|
|
ERROR test multiarray.TestFromBuffer.test empty , array [], dtype=float64), {}) Traceback most recent call last) File usr lib python3.2 site packages nose case.py , line 198, in runTest self.test self.arg) File home charris .local lib python3.2 site packages numpy core tests test multiarray.py , line 1446, in tst basic assert array equal np.frombuffer buffer, kwargs),expected) AttributeError str object has no attribute buffer","bug"
|
|
"6013","2561","PY3K bug in numpy.info numpy lib utils.py) Trac 1968).
|
|
Original ticket http projects.scipy.org numpy ticket 1968 on 2011 10 24 by WarrenWeckesser, assigned to unknown.
|
|
The problem was reported in a scipy ticket http projects.scipy.org scipy ticket 1541 but it is a Py3K issue in numpy.info.
|
|
Specifically, the code in numpy lib utils.py uses types.InstanceType, but that name no longer exists in Python 3.x.","bug"
|
|
"6013","2562","Port to SL4A Android) Trac 1969).
|
|
Original ticket http projects.scipy.org numpy ticket 1969 on 2011 10 28 by trac user encukou, assigned to unknown.
|
|
Matthew Turk has successfully hacked Numpy to run on Android already http groups.google.com group android scripting browse thread thread 2378edd0dc5048fb He was told to open an issue, but apparently never did xe2x80x93 searching for Android and SL4A yields nothing.
|
|
The main problem seems to be that Numpy doesn t cross compile well.","enhancement"
|
|
"6013","2563","have array list number of omitted items rather than ellipsis Trac 1970).
|
|
Original ticket http projects.scipy.org numpy ticket 1970 on 2011 10 28 by trac user AlanFrankel, assigned to unknown.
|
|
Currently, the display for an array looks like this [[ 0.
|
|
0.64 0.64 ..., 61.
|
|
278.
|
|
1. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
..., 6.
|
|
21.
|
|
1. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
..., 9.
|
|
23.
|
|
1. ]
|
|
..., [ 0.
|
|
0.
|
|
0.
|
|
..., 5.
|
|
11.
|
|
0. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
..., 11.
|
|
79.
|
|
0. ]
|
|
[ 0.
|
|
0.
|
|
0.65 ..., 5.
|
|
40.
|
|
0. ]]
|
|
It would be helpful if instead of the ellipses, we had the number of missing items [[ 0.
|
|
0.64 0.64 {12x} 61.
|
|
278.
|
|
1. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
{12x} 6.
|
|
21.
|
|
1. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
{12x} 9.
|
|
23.
|
|
1. ]
|
|
{205x} [ 0.
|
|
0.
|
|
0.
|
|
{12x} 5.
|
|
11.
|
|
0. ]
|
|
[ 0.
|
|
0.
|
|
0.
|
|
{12x} 11.
|
|
79.
|
|
0. ]
|
|
[ 0.
|
|
0.
|
|
0.65 {12x} 5.
|
|
40.
|
|
0. ]]
|
|
Here, you can tell at a glance that the array is 211 rows by 18 columns.
|
|
Otherwise, you need a separate command to list the array size.
|
|
It could be debated whether {12x} or {..12..} or <..12..> or something else would be best for symbolizing the omitted 12 items, but any of those choices would be fine for me.
|
|
I personally would always prefer this style of display to the one with ellipses.
|
|
However, even if you made this a nondefault option, that would make me happy.","enhancement"
|
|
"6013","2564","genfromtxt comments do not work on first line when names option is used Trac 1971).
|
|
Original ticket http projects.scipy.org numpy ticket 1971 on 2011 11 02 by trac user mesmith, assigned to unknown.
|
|
In genfromtxt when the names option is used a comment character will cause an error.
|
|
I think the error is on this line 1274 in version 1.6.1, the version drop down on the trac interface does not go up to 1.6.1) first line = asbytes ).join first line.split comments)[1 ]) I think the line shoud read first line = asbytes ).join first line.split comments)[ 1])","bug"
|
|
"6013","2565","numpy.fromstring , sep = ) returns array [ 1]) instead of array []) Trac 1972).
|
|
Original ticket http projects.scipy.org numpy ticket 1972 on 2011 11 03 by trac user nitro, assigned to unknown.
|
|
>>> import numpy >>> numpy.fromstring , sep = ) array [ 1.])
|
|
>>> numpy.version.version 1.6.1 using prebuilt windows binaries.
|
|
I d except numpy.fromstring to return array []) there instead of the arbitrary 1 value.
|
|
There s no way to tell if the string was really 1 except parsing the data myself.","bug"
|
|
"6013","2566","The .astype method output array loses the maskna property.
|
|
Trac 1973).
|
|
Original ticket http projects.scipy.org numpy ticket 1973 on 2011 11 04 by bsouthey, assigned to pierregm.
|
|
I get an error when displaying a masked array that contains the new missing value, np.NA under 64 bit Linux.
|
|
But the operations seem to work as expected.
|
|
>>> import numpy as np >>> np.
|
|
version 2.0.0.dev 7297785 >>> b=np.ma.array [3,5,np.NA,6], mask=[0,0,1,0]) >>> b Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib64 python2.7 site packages numpy ma core.py , line 3551, in repr data=str self), mask=str self.
|
|
mask), File usr lib64 python2.7 site packages numpy ma core.py , line 3534, in str res = self.
|
|
data.astype O ) ValueError Cannot assign NA to an array which does not support NAs >>> b. data array [3, 5, NA, 6]) >>> b.sum ) 14 >>> b.mask=[0,0,0,0] >>> b Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib64 python2.7 site packages numpy ma core.py , line 3551, in repr data=str self), mask=str self.
|
|
mask), File usr lib64 python2.7 site packages numpy ma core.py , line 3534, in str res = self.
|
|
data.astype O ) ValueError Cannot assign NA to an array which does not support NAs >>> b.sum ) NA dtype= int64 )","bug"
|
|
"6013","2567","Subscript assignment on `hard mask` masked array produces wrong result whenever the subscripted mask returns all False i.e.
|
|
`nomask`) Trac 1974).
|
|
Original ticket http projects.scipy.org numpy ticket 1974 on 2011 11 08 by trac user codewarrior, assigned to pierregm.
|
|
Using subscript assignment on a `hard mask` masked array with a slice or other index that selects a fully unmasked section of the masked array will result in the second item of the section being assigned to instead of the entire section.
|
|
Upon investigating, it is because mask or reduces the fully unmasked section of the mask to a single nomask value, which is negated and used to subscript assign the section of the data array.
|
|
This is incorrect because negating nomask produces a single True value, which turns into a 1 and selects the second item of the section.
|
|
This is apparently due to a brain fart on line 3038 of numpy.ma.core.py, in the final else clause of MaskedArray.
|
|
setitem ) The original code reads mindx = mask or mask[indx], mval, copy=True) dindx = self.
|
|
data[indx] if dindx.size > 1 dindx[ mindx] = dval elif mindx is nomask dindx = dval It seems like it should be checking mindx.size and not dindx.size because mask or is free to shrink the return value down to nomask which would have size 1.
|
|
When I use this corrected code, I no longer observe the problem mindx = mask or mask[indx], mval, copy=True) dindx = self.
|
|
data[indx] if mindx.size > 1 dindx[ mindx] = dval elif mindx is nomask dindx = dval This was apparently fixed in https github.com numpy numpy commit a6e869b70b09df9381d341ed0d2b18f88d8fe3d6 but that fix can t be backported to 1.6 because it uses np.copyto ).
|
|
Here is code that demonstrates the error.
|
|
>>> from numpy import >>> a = arange 30) >>> a.shape=5,6 >>> b = zeros like a) >>> m = ma.masked array a,b,hard mask=True) only happens when hard mask is True >>> m masked array data = [[0 1 2 3 4 5] [6 7 8 9 10 11] [12 13 14 15 16 17] [18 19 20 21 22 23] [24 25 26 27 28 29]], mask = [[False False False False False False] [False False False False False False] [False False False False False False] [False False False False False False] [False False False False False False]], fill value = 999999) >>> m[ ] = 333 >>> m entire array should be 333 now masked array data = [[0 1 2 3 4 5] [333 333 333 333 333 333] uh oh, only the second element is set [12 13 14 15 16 17] [18 19 20 21 22 23] [24 25 26 27 28 29]], mask = [[False False False False False False] [False False False False False False] [False False False False False False] [False False False False False False] [False False False False False False]], fill value = 999999)","bug"
|
|
"6013","2568","[Mixing integers and floats] wrong results in algebraic equation Trac 1975).
|
|
Original ticket http projects.scipy.org numpy ticket 1975 on 2011 11 09 by trac user bde, assigned to unknown.
|
|
I m experiencing a strange behavior of numpy when using a ndarray filled with integers in an algebraic equation with floats.
|
|
See code snippet below >>>import numpy >>> def A ar) return 12200 2) ar 4)) ar 2 20.6 2) ar 2 12200 2) ar 2 107.7 2) 0.5 ar 2 737.9 2) 0.5 ) ... >>> A numpy.array [1,2,3])) array with integers array [ 2.95808751e 08, 3.77586449e 07, 1.61632406e 07]) negative numbers are wrong A is never negative) >>> A numpy.array [1,2.0,3.0])) array of int and floats array [ 2.95808751e 08, 4.69917244e 07, 2.35099089e 06]) correct!
|
|
>>> A numpy.array [1.,2,3.0])) array [ 2.95808751e 08, 4.69917244e 07, 2.35099089e 06]) correct!
|
|
>>> A 1) 2.9580875078524171e 08 >>> A 1.)
|
|
2.9580875078524171e 08 >>> A 2) 4.6991724410274925e 07 >>> A 2.0) 4.6991724410274925e 07 Is it a problem related to the way numpy makes the cast from int to float ?
|
|
This is strange because there is no need to have ALL the numbers of ar to be float in order to make it work, one can remain int as you can see below...","bug"
|
|
"6013","2569","linalg.qr returns incorrect shape for nx1 arrays Trac 1976).
|
|
Original ticket http projects.scipy.org numpy ticket 1976 on 2011 11 10 by trac user bransfop, assigned to pv.
|
|
Scipy s linalg.qr returns the correct shape when decomposing the vector [[1],[0]] while Numpy s linalg.qr does not.
|
|
The correct shape should be 2x2 for Q and 2x1 for R. Numpy s linalg.qr returns a 2x1 and 1x1 array while Scipy s linalg.qr returns a 2x2 and 2x1 array.
|
|
Scipy is correct.","bug"
|
|
"6013","2570","assignment of scalar subclass of int to a slice of of a recarray field of np.object silently upcasts Trac 1977).
|
|
Original ticket http projects.scipy.org numpy ticket 1977 on 2011 11 11 by trac user hughsw, assigned to unknown.
|
|
Demonstrates a Numpy bug whereby assigning from a scalar that is a subclass of int into a slice of a recarray field of type np.object will lose the subclass info import numpy as np def numpy bug ) >>> print np.version.version , np.version.version np.version.version 1.6.1 >>> numpy bug ) [ foo , |O8 )] [<class main .MyInt >] 52,) [<class main .MyInt >] 53,) [<type int >] 100,) [<type int >] 100,) named field of dtype np.object rec object t = np.dtype [ foo , np.object)]) subclass of int class MyInt int) pass set the items, using two forms of broadcasting items = np.empty 4,), dtype=rec object t) this works items[ foo ][ 2] = map MyInt, xrange 52, 54)) this fails it loses the MyInt type, silently casting to int even though the foo field slot is of dtype np.object items[ foo ][2 ] = MyInt 100) print rec object t for item in items print map type, item), , item import doctest doctest.testmod )","bug"
|
|
"6013","2571","numpy.int64 returns OverflowError but numpy.int does not Trac 1978).
|
|
Original ticket http projects.scipy.org numpy ticket 1978 on 2011 11 11 by trac user JoeCM, assigned to unknown.
|
|
Hi, I am currently running python 2.7.2 compiled using the C compiler GCC 4.2.1 Compatible Apple Clang 3.0 tags Apple clang 211.10.1) and fortran compiler gfortran on a new MacBook Pro with Mac OSX 10.7 Lion).
|
|
I successfully installed numpy 2.0.0.dev 7297785 from the git archive.
|
|
However, when trying to import another python package Atpy) it fails with invalid[np.int64] = np.int64 2 63 1) OverflowError Python int too large to convert to C long After a short test, I ve found that the error comes from simply trying numpy.int64 2 63 1) However, if you try numpy.int 2 63 1) The error does not occur.
|
|
This same error occurred under the latest stable release 1.6.2)","bug"
|
|
"6013","2572","empty like does not copy mask Trac 1979).
|
|
Original ticket http projects.scipy.org numpy ticket 1979 on 2011 11 14 by trac user mauro, assigned to unknown.
|
|
If empty like and also zeros like, ones like) is called with a masked array, then the returned array s mask is identical to the originals i.e.
|
|
shares same reference).
|
|
I think what should happen is that a new mask is returned set to False.","bug"
|
|
"6013","2573","Matrix A from zeros, after changing values, yields strange A A.T behavior Trac 1980).
|
|
Original ticket http projects.scipy.org numpy ticket 1980 on 2011 11 15 by trac user ehassler, assigned to unknown.
|
|
Apologies if this is expected behavior, or if dupe.
|
|
I searched trac for zeros and multiply and found nothing that described this.
|
|
I m making a permutation matrix P, starting with zeros and setting the 1 s. When I do P.T P or P P.T I get all 0 s back.
|
|
But if I multiply P by an identical matrix manually constructed, or if I use asmatrix to redefine P as itself, I get the expected behavior that is, I get an identity matrix back).
|
|
I ve included example code at the bottom.
|
|
It seems like either one of the following should happen P should zero out everything it s multiplied into and raise errors when I change a value, forcing it to be all zeros all the time.
|
|
P P.T = I import numpy; P0 = numpy.zeros 6,6),dtype=numpy.float64); for coord in zip range 6), 1,3,5,0,2,4)) P0[coord[0],coord[1]] = 1.; P1 = numpy.matrix [ [ 0., 1., 0., 0., 0., 0.
|
|
], [ 0., 0., 0., 1., 0., 0.
|
|
], [ 0., 0., 0., 0., 0., 1.
|
|
], [ 1., 0., 0., 0., 0., 0.
|
|
], [ 0., 0., 1., 0., 0., 0.
|
|
], [ 0., 0., 0., 0., 1., 0.]
|
|
], dtype=numpy.float64); print numpy.all P0 == P1); True print numpy.all P0 P0.T == P1 P1.T) False print numpy.all P0 P1.T == P1 P1.T) True print numpy.all P1 P0.T == P1 P1.T) True P0 = numpy.asmatrix P0,dtype=numpy.float64); print numpy.all P0 P0.T == P1 P1.T) True","bug"
|
|
"6013","2574","numpy crash under windows with a very big array Trac 1981).
|
|
Original ticket http projects.scipy.org numpy ticket 1981 on 2011 11 15 by trac user bred, assigned to unknown.
|
|
I m using [[BR]] python 3.2 32bit) [[BR]] numpy 1.6.1 32bit) [[BR]] Windows 7 64 bit) [[BR]] In some case numpy crash if we use a very big array.
|
|
for example try this import numpy y = numpy.ones 40393836, 2) ) y 1 As a result python crash while it s computing y 1","bug"
|
|
"6013","2575","In place arithmetic operations wrong calculation order?
|
|
Trac 1982).
|
|
Original ticket http projects.scipy.org numpy ticket 1982 on 2011 11 18 by trac user ling, assigned to unknown.
|
|
In place operation like subtract b[ 1], b[1 ], b[ 1]) used to work fine in Numpy 1.4.1, but sometimes fails with Numpy 1.6.0.
|
|
See the example below the problematic output is Out[6]) In [1] import numpy; numpy.
|
|
version Out[1] 1.6.0 In [2] a = numpy.arange 12).reshape 3, 4)) In [3] b = a[ 1].copy ); numpy.subtract b[ 1], b[1 ]) Out[3] array [[4, 4, 4, 4], [4, 4, 4, 4]]) In [4] numpy.subtract b[ 1], b[1 ], b[ 1]) Out[4] array [[4, 4, 4, 4], [4, 4, 4, 4]]) In [5] b = a[ 1]; numpy.subtract b[ 1], b[1 ]) Out[5] array [[4, 4, 4, 4], [4, 4, 4, 4]]) In [6] numpy.subtract b[ 1], b[1 ], b[ 1]) Out[6] array [[4, 5, 6, 7], [4, 4, 4, 4]]) Similar problem happens to other in place operations e.g., multiply, divide).","bug"
|
|
"6013","2576","segmentation fault with dot product of sparse matrices Trac 1983).
|
|
Original ticket http projects.scipy.org numpy ticket 1983 on 2011 11 18 by trac user fp4code, assigned to unknown.
|
|
The dot product of two sparse matrices crashes python.
|
|
The defect has appeared with commit 867cabefe92b127b765580432b4c05d92342e275 Author Mark Wiebe <mwiebe enthought.com> Date Fri Jul 22 10 24 14 2011 0500 Here an example import sys, numpy, scipy, scipy.sparse md = numpy.matrix numpy.zeros 2,2))); md[0,1]= 1; md[1,0]=1 ms = scipy.sparse.lil matrix md) print sys.version) 2.7.2 [GCC 4.5.2] print numpy.version.version) 2.0.0.dev 867cabe to 7297785 print scipy.version.version) 0.11.0.dev badad3f or earlier print md md) OK print numpy.dot md,md)) OK print ms ms) OK print numpy.dot ms,ms)) Segmentation fault","bug"
|
|
"6013","2577","Bento build script does not populate version.py Trac 1984).
|
|
Original ticket http projects.scipy.org numpy ticket 1984 on 2011 11 20 by stefanv, assigned to unknown.
|
|
Currently, building by Bento does not generate versions inside of version.py, breaking other projects such as matplotlib.","bug"
|
|
"6013","2578","histogram2d Trac 1985).
|
|
Original ticket http projects.scipy.org numpy ticket 1985 on 2011 11 22 by trac user gsiisg, assigned to unknown.
|
|
I notice that when I plot two variables, the x and y axis are flipped according to the online example, so I did a simple distribution of points x=np.array [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1]) y=np.array [ 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 100]) H, yedges, xedges = np.histogram2d y,x) extent = [xedges[0], xedges[ 1], yedges[0], yedges[ 1]] pylab.imshow H, extent=extent, interpolation= nearest ,origin= lower ,aspect= auto ) I had to enter into histgram2d ) the y before the x to get the graph to come out as expected.","bug"
|
|
"6013","2579","Fix bento build with macport gfortran Trac 1986).
|
|
Original ticket http projects.scipy.org numpy ticket 1986 on 2011 11 22 by cournape, assigned to cournape.
|
|
See https github.com cournape Bento issues 62","bug"
|
|
"6013","2580","polyfit and basic arithmetic function errors with clang on Mac OS X Lion Trac 1987).
|
|
Original ticket http projects.scipy.org numpy ticket 1987 on 2011 11 23 by trac user deil, assigned to unknown.
|
|
I get two test failures of 2.0.0.dev 7f302cc with clang.
|
|
Furthermore I get a MismatchCAPIWarning when building numpy, which results in a ` Using deprecated NumPy API ` when building scipy.
|
|
The full build and test logs are attached.
|
|
See http mail.scipy.org pipermail scipy user 2011 November 031042.html System info Mac OS X Lion 10.7.2 11C74) XCode 4.2.1 4D502) which python; python opt local bin python Python 2.7.2 default, Nov 23 2011, 11 40 08) [GCC 4.2.1 Based on Apple Inc. build 5658) LLVM build 2335.15.00)] on darwin which clang; clang version usr bin clang Apple clang version 3.0 tags Apple clang 211.12) based on LLVM 3.0svn) Target x86 64 apple darwin11.2.0 which gfortran; gfortran version opt local bin gfortran GNU Fortran GCC) 4.4.6 Test warning and failures Users deil Library Python 2.7 lib python site packages numpy ma core.py 4778 RuntimeWarning invalid value encountered in power np.power out, 0.5, out=out, casting= unsafe ) ====================================================================== FAIL Test basic arithmetic function errors Traceback most recent call last) File Users deil Library Python 2.7 lib python site packages numpy core tests test numeric.py , line 319, in test floating exceptions lambda a,b a b, ft tiny, ft tiny) File Users deil Library Python 2.7 lib python site packages numpy core tests test numeric.py , line 270, in assert raises fpe Type s did not raise fpe error s .
|
|
ftype, fpeerr)) File Users deil Library Python 2.7 lib python site packages numpy testing utils.py , line 34, in assert raise AssertionError msg) AssertionError Type <type numpy.complex64 > did not raise fpe error .
|
|
====================================================================== FAIL Tests polyfit Traceback most recent call last) File Users deil Library Python 2.7 lib python site packages numpy ma tests test extras.py , line 622, in test polyfit assert almost equal a, a ) File Users deil Library Python 2.7 lib python site packages numpy ma testutils.py , line 155, in assert almost equal err msg=err msg, verbose=verbose) File Users deil Library Python 2.7 lib python site packages numpy ma testutils.py , line 221, in assert array almost equal header= Arrays are not almost equal ) File Users deil Library Python 2.7 lib python site packages numpy ma testutils.py , line 186, in assert array compare verbose=verbose, header=header) File Users deil Library Python 2.7 lib python site packages numpy testing utils.py , line 677, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal mismatch 100.0 ) x array [ 4.25134878, 1.14131297, 0.20519666, 0.01701 ]) y array [ 1.9345248 , 0.49711011, 0.10202554, 0.00928034]) MismatchCAPIWarning in numpy build numpy core setup common.py 86 MismatchCAPIWarning API mismatch detected, the C API version numbers have to be updated.
|
|
Current C api version is 6, with checksum eb54c77ff4149bab310324cd7c0cb176, but recorded checksum for C API version 6 in codegen dir cversions.txt is e61d5dc51fa1c6459328266e215d6987.
|
|
If functions were added in the C API, you have to update C API VERSION in numpy core setup common.py.
|
|
MismatchCAPIWarning) Warning in scipy build Users deil Library Python 2.7 lib python site packages numpy core include numpy npy deprecated api.h 11 2 warning warning Using deprecated NumPy API, disable it by defining NPY NO DEPRECATED API [ W warnings] I also see this C API Warning on a Linux machine.
|
|
Is this a bug in numpy or an installation problem on my side?
|
|
Does it have any consequences on numpy scipy working properly?
|
|
The polyfit failure apparently is known, on the mailing list Charles Harris wrote Don t worry about this one, it comes from Travis changing the unmasked version of polyfit which changes the singular values, etc.","bug"
|
|
"6013","2581","ndarray.min failure with dtype= O on Python 3.1 Trac 1988).
|
|
Original ticket http projects.scipy.org numpy ticket 1988 on 2011 11 25 by wesm, assigned to unknown.
|
|
This little fun bug came up when I was running my pandas test suite on Python 3.1 arr = np.array [ 0.05, 0, 0.04], dtype= O ) arr.min ) returns 0.04 on NumPy 1.6.1, Python 3.1.4.
|
|
Will test on 3.2 also...","bug"
|
|
"6013","2582","apply along axis error when applied function returns something other than scalar or array Trac 1989).
|
|
Original ticket http projects.scipy.org numpy ticket 1989 on 2011 11 26 by trac user pjohn, assigned to unknown.
|
|
When using a function that returns something that is not a scalar or an array e.g.
|
|
scipy.interpolate.UnivariateSpline returns an object), numpy.apply along axis gives an error.
|
|
xefxbbxbfxefxbbxbfxefxbbxbfTypeError object of type LSQUnivariateSpline has no len ) function apply along axis in shape base.py at line 104 outshape[axis] = len res) However, when using the masked version numpy.ma.apply along axis), this error does not occur.
|
|
The masked version in ma extras.py) includes this bit of code res = func1d arr[tuple i.tolist ))], args, kwargs) if res is a number, then we have a smaller output array asscalar = np.isscalar res) if not asscalar try len res) except TypeError asscalar = True Note we shouldn t set the dtype of the output from the first result... ...so we force the type to object, and build a list of dtypes ...we ll just take the largest, to avoid some downcasting dtypes = [] if asscalar ... else ... outshape[axis] = res.shape whereas numpy.apply along axis in lib shape base.py) does if isscalar res) ... else ... outshape[axis] = len res) This portion of ma.apply along axis should be backported to numpy.apply along axis.","bug"
|
|
"6013","2583","Data can be truncated in creation of an array from a mixed sequence of numeric values and strings Trac 1990).
|
|
Original ticket http projects.scipy.org numpy ticket 1990 on 2011 12 05 by thouis, assigned to unknown.
|
|
When an array is created from a sequence of numeric boolean, int, float, complex) and string bytes, str, unicode) values, the resulting array type is string, but only the string values are being used to choose the string length, leading to truncation of data.
|
|
Code import numpy as np for numericval in [True, 1234, 123.4, complex 1, 3)] for stringconversion in [str, unicode, bytes] print repr np.array [numericval, stringconversion abc )])) Result array [ Tru , abc ], dtype= |S3 ) array [u Tru , u abc ], dtype= <U3 ) array [ Tru , abc ], dtype= |S3 ) array [ 123 , abc ], dtype= |S3 ) array [u 123 , u abc ], dtype= <U3 ) array [ 123 , abc ], dtype= |S3 ) array [ 123 , abc ], dtype= |S3 ) array [u 123 , u abc ], dtype= <U3 ) array [ 123 , abc ], dtype= |S3 ) array [ 1 , abc ], dtype= |S3 ) array [u 1 , u abc ], dtype= <U3 ) array [ 1 , abc ], dtype= |S3 )","bug"
|
|
"6013","2584","NumPy for Python26 to Python32 on Win 7 64 Trac 1991).
|
|
Original ticket http projects.scipy.org numpy ticket 1991 on 2011 12 07 by trac user Duman, assigned to unknown.
|
|
This may be a silly question, if so please forgive me.
|
|
I was recently turned on to Python by a geek friend, and I m attempting to load NumPy on my Win 7 64 system.
|
|
However,there does not appear to be a Python26, or higher, distribution of NumPy for Win 7 64.
|
|
The current NumPy fails to install on my system.
|
|
Am I missing something?","enhancement"
|
|
"6013","2585","data leakage with S0 in record dtype Trac 1992).
|
|
Original ticket http projects.scipy.org numpy ticket 1992 on 2011 12 12 by trac user hpaulj, assigned to unknown.
|
|
As of numpy 1.5.1 and earlier) >>> x = np.zeros 1, S,i ) >>> x[ f0 ]= A >>> x array [ , 65)], dtype=[ f0 , |S0 ), f1 , <i4 )]) In contrast, if the assigned dtype is S1,i , x is then A ,0) So not only is there ambiguity as to what a S dtype produces as discussed in other tickets), but assignment to that record leaks over to the adjacent field.
|
|
I first encountered this in an introduction to numpy for Ironpython, but it occurs in all versions that I ve tested.","bug"
|
|
"6013","2586","extend logical and, logical or etc.
|
|
to tuples as input Trac 1993).
|
|
Original ticket http projects.scipy.org numpy ticket 1993 on 2011 12 13 by trac user andreash, assigned to unknown.
|
|
I think it would be a great enhancement if the logical X methods would also work when giving them a tuple of arrays as input, so that logical and A, B, C)) would yield A B C","enhancement"
|
|
"6013","2587","Empty indices in empty arrays raise IndexError Trac 1994).
|
|
Original ticket http projects.scipy.org numpy ticket 1994 on 2011 12 13 by trac user jordigh, assigned to unknown.
|
|
Let x be an array and consider the following expression x[ , idx] This should be valid as long as all elements of idx are nonnegative integers less than x.shape[1], right?
|
|
However, this is not true if, say, x.shape == 5, 0) and idx = [].
|
|
The statement all elements of idx are nonnegative integers less than 0 is still true, however this causes an error.
|
|
Note that the following does not error out, however x = np.array []) x[ [] ] So this looks like a bug to me.
|
|
I have had independent confirmation in IRC of this behaviour in 1.6.1 and 1.5.0.
|
|
The reason why I would like this fixed is that I don t want to have to special case my indexing code for indexing arrays with empty dimensions with an empty index array.
|
|
I could work around this, but I would rather not.","bug"
|
|
"6013","2588","DOC numpy.correlate ) docstring is wrong Trac 1995).
|
|
Original ticket http projects.scipy.org numpy ticket 1995 on 2011 12 13 by trac user O.C., assigned to unknown.
|
|
The docstring of numpy.correlate ) indicates z[k] = sum n a[n] conj v[n k]) This is wrong.
|
|
According to my tests, the real behavior is in fact z[k] = sum n a[n] conj v[n k]) or equivalently z[k] = sum n a[k n] conj v[n])","bug"
|
|
"6013","2589","sin 180) != 0 Trac 1996).
|
|
Original ticket http projects.scipy.org numpy ticket 1996 on 2011 12 14 by trac user ilshat, assigned to unknown.
|
|
>>> import math >>> math.sin math.pi) 1.2246467991473532e 16 Must calculate more precisely!","bug"
|
|
"6013","2590","asarray may truncate strings if the input has numbers Trac 1997).
|
|
Original ticket http projects.scipy.org numpy ticket 1997 on 2011 12 15 by trac user dalleyg, assigned to unknown.
|
|
In the following example, numpy.asarray chooses the dtype |S1 , causing the integer 10 s string form to be truncated to 1 >>> numpy.asarray [ 1 , 10]) array [ 1 , 1 ], dtype= |S1 ) >>> numpy.
|
|
version 1.6.1","bug"
|
|
"6013","2591","loadtxt should give line numbers of problems Trac 1998).
|
|
Original ticket http projects.scipy.org numpy ticket 1998 on 2011 12 15 by samtygier, assigned to unknown.
|
|
currently input like 1 2 3 4 5 6 78 will cause loadtxt to give the following error Traceback most recent call last) File .
|
|
np lt.py , line 5, in <module> numpy.loadtxt sys.argv[1]) File usr lib python2.7 site packages numpy lib npyio.py , line 804, in loadtxt X = np.array X, dtype) ValueError setting an array element with a sequence.
|
|
This would be far more useful, if it contained some information about where the problem was.
|
|
This is similar to the bug 1810 , but for the loadtxt function.
|
|
I attach a patch that gives the following message ValueError Wrong number of columns at line 3 it also provides a test that the input raises the correct error.
|
|
Inconveniently it is the same error type as before, so the test is not very useful.
|
|
However I thing ValueError is probably the most appropriate in this case.
|
|
Maybe there is a way to check the exception message.","bug"
|
|
"6013","2592","fromiter swallows exceptions Trac 1999).
|
|
Original ticket http projects.scipy.org numpy ticket 1999 on 2011 12 16 by trac user ged, assigned to unknown.
|
|
numpy.fromiter with a count argument hides any exception that might be raised by the iterator as ValueError iterator too short This is quite annoying because it makes for a totally ununderstandable unrelated exception in many cases.
|
|
This was already reported on the mailing list by xc3x85smund Hjulstad a few months ago but AFAIK nobody ever answered him.
|
|
This bug is present in numpy 1.6.1 but I think it has been there for quite a while.
|
|
See attached test case.","bug"
|
|
"6013","2593","numpy.add.reduceat no longer works on degenerate arrays Trac 2000).
|
|
Original ticket http projects.scipy.org numpy ticket 2000 on 2011 12 16 by trac user dalleyg, assigned to unknown.
|
|
The following works on numpy 1.4.1 python c from numpy import ; print version ; print add.reduceat zeros 0,)), []) 1.4.1 [] but does not work on 1.6.1 python c from numpy import ; print version ; print add.reduceat zeros 0,)), []) 1.6.1 Traceback most recent call last) File <string> , line 1, in <module> ValueError cannot remove a zero sized axis from an iterator It would be helpful if it could be made to work again e.g.
|
|
it should return the input array if both the a and indices input arrays are empty).","bug"
|
|
"6013","2594","Seg fault when using r incorrectly Trac 2001).
|
|
Original ticket http projects.scipy.org numpy ticket 2001 on 2011 12 16 by trac user claumann, assigned to unknown.
|
|
Howdy Just discovered a segfault when calling r on a list of arrays of record type in a recent snapshot.
|
|
See minimal session below.
|
|
Best, C python Enthought Python Distribution www.enthought.com Version 7.1 1 64 bit) Python 2.7.2 |EPD 7.1 1 64 bit)| default, Jul 3 2011, 15 56 02) [GCC 4.0.1 Apple Inc. build 5493)] on darwin Type packages , demo or enthought for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 2.0.0.dev 26aa3cf >>> dat1 = np.zeros 5, dtype=[ a , <f8 )]) >>> dat2 = np.zeros 5, dtype=[ a , <f8 )]) >>> np.r [dat1,dat2] array [ 0.0,), 0.0,), 0.0,), 0.0,), 0.0,), 0.0,), 0.0,), 0.0,), 0.0,), 0.0,)], dtype=[ a , <f8 )]) >>> np.r [[dat1,dat2]] Segmentation fault 11 uname a Darwin blackbear.local 11.2.0 Darwin Kernel Version 11.2.0 Tue Aug 9 20 54 00 PDT 2011; root xnu 1699.24.8 1 RELEASE X86 64 x86 64","bug"
|
|
"6013","2595","doc for concatenate with ndarray argument Trac 2002).
|
|
Original ticket http projects.scipy.org numpy ticket 2002 on 2011 12 16 by trac user hpaulj, assigned to unknown.
|
|
The doc string for concatenate does not explain what happens when the argument is a single ndarray, as opposed to a tuple of arrays.
|
|
kron ), for example, uses concatenate like this to consolidate axes.
|
|
It appears that concatenate a,axis=n) is equivalent to concatenate [x for x in a], axis=n) That is, it splits the array on the 1st axis, and joins the subarrays on their n th axis.
|
|
The net effect is to consolidate the 1st and n 1 axes of a.
|
|
The same issue applies to the docs for hstack and vstack.","bug"
|
|
"6013","2596","matlib parse fails on MacOS10.7 but works on Ubuntu 10.04 Trac 2003).
|
|
Original ticket http projects.scipy.org numpy ticket 2003 on 2011 12 22 by trac user hajons, assigned to unknown.
|
|
I have attached a script that parses a matlab file and stores data in a file.
|
|
It works well on Ubuntu 10.04 with python 2.6.5, but fails on MacOs 10.7 both on python 2.6.1 and 2.7.1 MacOS version of python).
|
|
I don t have any other problems with python on MacOS.
|
|
Scipy version 0.11 and numpy version 2.0.0.
|
|
The error message I get on MacOS is this seldmac6145 Proximates 23056164 python parse proximity.py realitymining.mat Done loading matlab data.
|
|
Traceback most recent call last) File parse proximity.py , line 79, in <module> get events matlab obj) File parse proximity.py , line 18, in get events subject mac = int subject object.my mac[0][0][0], 16) AttributeError numpy.void object has no attribute my mac","bug"
|
|
"6013","2597","NA incompatible with astype method.
|
|
Trac 2004).
|
|
Original ticket http projects.scipy.org numpy ticket 2004 on 2011 12 23 by charris, assigned to unknown.
|
|
To wit In [13] x = ones 5, maskna=1) In [14] x[2] = NA In [15] x.astype double) ValueError Traceback most recent call last) home charris <ipython input 15 c95438174230> in <module> ) > 1 x.astype double) ValueError Cannot assign NA to an array which does not support NAs I suspect the easiest thing to do here is propagate the MASKNA flag.","bug"
|
|
"6013","2598","Overwrite of alpha values when casting a string slice to int Trac 2005).
|
|
Original ticket http projects.scipy.org numpy ticket 2005 on 2011 12 23 by trac user markhmoulton ..., assigned to unknown.
|
|
When trying to cast a slice of string values to int, where one of the cell values is alpha, the alpha value is mistakenly converted into an integer.
|
|
Instead, it should return a ValueError exception.
|
|
>>> B = np.zeros 5,3)).astype S10 ) >>> B[ ,0] = range 5) >>> B[0,0] = StudentID >>> B array [[ StudentID , 0.0 , 0.0 ], [ 1 , 0.0 , 0.0 ], [ 2 , 0.0 , 0.0 ], [ 3 , 0.0 , 0.0 ], [ 4 , 0.0 , 0.0 ]], dtype= |S10 ) Casting the whole array to int throws an error, as it should.
|
|
>>> B.astype int) Traceback most recent call last) File <pyshell 152> , line 1, in <module> B.astype int) ValueError invalid literal for int ) with base 10 StudentID However, casting a slice of the same array to int actually changes StudentID into an integer 11 in this case).
|
|
>>> B[ ,0].astype int) array [11, 1, 2, 3, 4]) >>> This behavior makes it hard to catch important ValueError exceptions, besides mangling my data.
|
|
My versions are Python 2.7.2 and Numpy 1.6.0, as distributed by Enthought EPD 7.1 1 64 bit)), on a Mac Pro.","bug"
|
|
"6013","2599","Why does recarray assignment fail cannot call setfield on an object array ?
|
|
Trac 2006).
|
|
Original ticket http projects.scipy.org numpy ticket 2006 on 2011 12 23 by trac user kevin000, assigned to unknown.
|
|
Using a recarray with member assignment fails if any column uses the object data type.
|
|
I don t understand why this fails, especially since the dictionary style assignment works.
|
|
Brief example showing what recarray assignments succeed or fail import numpy as np recarray with integer datatypes dt = np.dtype [ foo , i8 ), bar , i8 )]) r = np.zeros 1,3), dtype=dt).view np.recarray) r[ foo ] = np.array [1, 2, 3]) OK r.foo = np.array [1, 2, 3]) OK recarray with an object datatype dt = np.dtype [ foo , i8 ), bar , O8 )]) r = np.zeros 1,3), dtype=dt).view np.recarray) r[ foo ] = np.array [1, 2, 3]) OK r.foo = np.array [1, 2, 3]) RuntimeError For the relevant C code search for cannot call setfield on an object array in this file https github.com numpy numpy blob master numpy core src multiarray methods.c Thanks!
|
|
Kevin","bug"
|
|
"6013","2600","Correct handling of empty arrays Trac 2007).
|
|
Original ticket http projects.scipy.org numpy ticket 2007 on 2011 12 28 by rgommers, assigned to unknown.
|
|
This is a single ticket to describe issues related to empty arrays.
|
|
Triggered by 2587 and discussion on ML starting on 25 Dec 2011 titled Indexing empty dimensions with empty arrays .
|
|
Empty arrays are not handled well in NumPy.
|
|
Operations like indexing should have well defined behavior for empty arrays, and all functions should do something sensible.
|
|
A summary of how all this is supposed to work should be added to the User Guide.
|
|
Issue indexing with empty index array doesn t work on empty arrays with shape other than 0,) >>> x = np.ones 5,0)) >>> x[idx] works array [], shape= 0, 0), dtype=float64) >>> x[ , idx] doesn t work Traceback most recent call last) File <ipython input 27 7038691cb565> , line 1, in <module> x[ , idx] doesn t work IndexError invalid index Issue calculating a determinant of an empty array results in a segfault np.linalg.det np.zeros 0,0))) Parameter 4 to routine DGETRF was incorrect <segfault> All functions should do something sensible for all empty input.
|
|
Example >>> np.meshgrid [], []) ok array [], shape= 0, 0), dtype=float64), array [], shape= 0, 0), dtype=float64)) >>> np.meshgrid [], np.ones 3,0))) not ok ... ValueError Patches to improve handling of empty array are very welcome.
|
|
New specific issues can be added to this ticket.","bug"
|
|
"6013","2601","Make putmask use copyto under the covers.
|
|
Trac 2008).
|
|
Original ticket http projects.scipy.org numpy ticket 2008 on 2011 12 31 by charris, assigned to unknown.
|
|
Summary says it.","enhancement"
|
|
"6013","2602","Forward operations of numpy.ndarray incompatible with numpy.float Trac 2009).
|
|
Original ticket http projects.scipy.org numpy ticket 2009 on 2012 01 04 by trac user mrclary, assigned to unknown.
|
|
Forward operations ) involving a numpy.ndarray of dtype=object and numpy.float give a TypeError.
|
|
Leftward operations do not give a TypeError.
|
|
[[BR]] numpy.array [1.
|
|
], dtype=object) numpy.float 1) OK numpy.array [1.
|
|
], dtype=object) numpy.float16 1) Not OK numpy.array [1.
|
|
], dtype=object) numpy.float32 1) Not OK numpy.array [1.
|
|
], dtype=object) numpy.float64 1) Not OK numpy.array [1.
|
|
], dtype=object) numpy.float128 1) Not OK numpy.float16 1) numpy.array [1.
|
|
], dtype=object) OK numpy.float32 1) numpy.array [1.
|
|
], dtype=object) OK numpy.float64 1) numpy.array [1.
|
|
], dtype=object) OK numpy.float128 1) numpy.array [1.
|
|
], dtype=object) OK Numpy 1.6.1 Python 2.7.2 Both via MacPorts Mac OS X 10.7.2","bug"
|
|
"6013","2603","dtype ) construction fails when metadata = None Trac 2010).
|
|
Original ticket http projects.scipy.org numpy ticket 2010 on 2012 01 04 by trac user tilkau, assigned to unknown.
|
|
>>> np.dtype H , metadata={ foo 1}) dtype uint16 ) >>> np.dtype H , metadata=None) TypeError argument 4 must be dict, not None IMO metadata=None makes sense it means don t update the metadata .
|
|
Note that this would produce a different result than passing {} that results in mydtype.metadata returning dict proxy {}) rather than None.
|
|
I m writing a higher level dtype support system constructor, and currently, due to this bug, I m using everywhere this wrapper function instead of np.dtype def dtype dt, metadata=None, kwargs) DType wrapper that avoids crashery when metadata=None if metadata return np.dtype dt,metadata = metadata, kwargs) return np.dtype dt, kwargs) The offending line seems to be around line 237 of core src multiarray descriptor.c","bug"
|
|
"6013","2604","nested iters does not accept length zero nest and no documentation) Trac 2011).
|
|
Original ticket http projects.scipy.org numpy ticket 2011 on 2012 01 05 by trac user jsalvatier, assigned to unknown.
|
|
While using the nested iters function, I ve noticed that it does not accept length zero nestings.
|
|
For example, the following fails nested iters [ones 3),ones 3)], [[], [0]]) with ValueError If op axes or itershape is not NULL in theiterator constructor, oa ndim must be greater than zero This makes a certain amount of sense to me, but I think having the iterator with the empty axes have a single iteration would be more useful.
|
|
For example, if you are using nested iters to ally a function along a specific set of axes, you ll otherwise have to special case the case where those axes take up the whole array which is my use case).
|
|
This is not much of a hassle for me, but I thought other people might like to know.
|
|
Also, I could not find any nested iters documentation.","bug"
|
|
"6013","2605","Creating an array from a sequence of arrays of datetime64 s gives error Trac 2012).
|
|
Original ticket http projects.scipy.org numpy ticket 2012 on 2012 01 05 by trac user jsalvatier, assigned to unknown.
|
|
This works In [43] a = array 1.5, complex64) In [44] array a,a)) Out[44] array [ 1.5 0.j, 1.5 0.j], dtype=complex64) but this does not In [46] b = array datetime 2009, 1, 1, 0, 0), datetime64) In [47] array b,b)) ValueError Traceback most recent call last) C \Users\jsalvatier\workspace\varscale\src\varscale\timevars.py in <module> ) > 1 2 3 4 5 ValueError error setting an array element with a sequence","bug"
|
|
"6013","2606","f2py is not proper translated by 2to3 in linux 64 on python3.2 Trac 2013).
|
|
Original ticket http projects.scipy.org numpy ticket 2013 on 2012 01 09 by trac user thelen ben ..., assigned to pearu.
|
|
I m running on Ubuntu 11.10 64 and have installed python3.2 and numpy 1.6.1 by python3.2 setup.py install All seems to have installed properly and is fully working.
|
|
but if I run f2py2e it will produce various errors caused by 2to3 translation.
|
|
See appended file I guessed this is caused by numpy for linux only supporting python version 2.x Any updates for numpy release for linux python 3.x?
|
|
Thanks,","enhancement"
|
|
"6013","2607","m68k xe2x80x9clong doublexe2x80x9d format isnxe2x80x99t recognised Trac 2014).
|
|
Original ticket http projects.scipy.org numpy ticket 2014 on 2012 01 11 by sandrotosi, assigned to unknown.
|
|
Hello, a fellow Debian Developer has rebuilt numpy on m68k architecture and reported http; bugs.debian.org 655388) a failure in detecting the long double format [ ] compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr include python2.6 d c gcc configtest.c configtest.c 4 5 warning function declaration isn t a prototype [ Wstrict prototypes] success!
|
|
removing configtest.c configtest.o C compiler gcc pthread fno strict aliasing g Wall Wstrict prototypes fPIC compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core include I usr include python2.6 d c gcc configtest.c removing configtest.c configtest.o Traceback most recent call last) File setup.py , line 211, in <module> setup package ) File setup.py , line 204, in setup package configuration=configuration ) File tmp buildd python numpy 1.5.1 numpy distutils core.py , line 186, in setup return old setup new attr) File usr lib python2.6 distutils core.py , line 152, in setup dist.run commands ) File usr lib python2.6 distutils dist.py , line 975, in run commands self.run command cmd) File usr lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File tmp buildd python numpy 1.5.1 numpy distutils command build.py , line 37, in run old build.run self) File usr lib python2.6 distutils command build.py , line 135, in run self.run command cmd name) File usr lib python2.6 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File tmp buildd python numpy 1.5.1 numpy distutils command build src.py , line 152, in run self.build sources ) File tmp buildd python numpy 1.5.1 numpy distutils command build src.py , line 169, in build sources self.build extension sources ext) File tmp buildd python numpy 1.5.1 numpy distutils command build src.py , line 328, in build extension sources sources = self.generate sources sources, ext) File tmp buildd python numpy 1.5.1 numpy distutils command build src.py , line 385, in generate sources source = func extension, build dir) File numpy core setup.py , line 434, in generate config h rep = check long double representation config cmd) File numpy core setup common.py , line 136, in check long double representation type = long double representation pyod object)) File numpy core setup common.py , line 268, in long double representation raise ValueError Unrecognized format s) saw) ValueError Unrecognized format [ 000 , 000 , 000 , 000 , 001 , 043 , 105 , 147 , 211 , 253 , 315 , 357 , 300 , 031 , 000 , 000 , 353 , 171 , 242 , 240 , 000 , 000 , 000 , 000 , 376 , 334 , 272 , 230 , 166 , 124 , 062 , 020 ]) [58723 refs] dh auto build python2.6 dbg setup.py build force returned exit code 1 the reported traceback is for 1.5.1 but it applies also for 1.6.1.
|
|
Regards, Sandro","bug"
|
|
"6013","2608","Misleading warning Line got X columns instead of Y) Trac 2015).
|
|
Original ticket http projects.scipy.org numpy ticket 2015 on 2012 01 12 by trac user wes waugh, assigned to unknown.
|
|
In ... numpy lib npyio.py 1438 I get this warning pretty frequently when using `usecols`.
|
|
For example, `usecols = [1,15,36,110]` but the input data will have only 108 columns.
|
|
So, I get an error message like Got 104 columns instead of 4 .
|
|
What the heck does that mean?
|
|
Neither 4 nor 104 are directly helpful.
|
|
What I need is the number of columns on a line 108) and the column I m trying to use that isn t present 110).
|
|
Maybe the IndexError on line 1396 could raise a warning like Line X Trying to use column Y, but only Z columns are present I know this is an older version of numpy I ll try upgrading.
|
|
Thanks!","enhancement"
|
|
"6013","2609","build fails due to error in distutils mingw32ccompiler.py Trac 2016).
|
|
Original ticket http projects.scipy.org numpy ticket 2016 on 2012 01 12 by trac user acdlab, assigned to cournape.
|
|
python 2.7.2 c compiler gcc 4.5.2 f compiler gfortran 4.5.2 OS Windows 7 When building various packages which use numpy, distutils\version.py raises ValueError invalid version number 4.
|
|
This can be fixed by editing Python27 Lib site packages numpy distutils mingw32ccompiler.py and replacing the two instances of 4. with 4.0.0 they are located on lines 107 and 127 then packages will build correctly","bug"
|
|
"6013","2610","On 32 bit, uintp and uint32 do not have the same hash Trac 2017).
|
|
Original ticket http projects.scipy.org numpy ticket 2017 on 2012 01 12 by trac user twiecki, assigned to unknown.
|
|
>>> np.uintp numpy.uint32 >>> np.dtype np.uintp) == np.dtype np.uint32) True >>> hash np.dtype np.uintp)) == hash np.dtype np.uint32)) False This creates a problem for using dtypes as dict keys.
|
|
Also, it seems that this bug is not present on 64 bit, where uintp and uint64 have the same hash.","bug"
|
|
"6013","2611","Bug in distutils compiling on Windows fails if Python is installed in a path with spaces.
|
|
Trac 2018).
|
|
Original ticket http projects.scipy.org numpy ticket 2018 on 2012 01 13 by trac user almar, assigned to cournape.
|
|
I found a bug in numpy.distutils, which is causing problems with compiling Cython code on Windows.
|
|
I have reproduced this on Python 2.6 and Python 3.2 32 bit).
|
|
I first thought this issue had to do with the normal distutils package bugs.python.org issue13765) The problem occurs with the native msvc compiler.
|
|
Using gcc MinGW) works fine.
|
|
The problem is that the command to link the libraries does not put double quotes around paths that have spaces in them.
|
|
Unfortunately, the path where I have Python installed has spaces in it c program files python26 ).
|
|
Small example of part of a link command LIBPATH C \Program Files x86)\python32\libs.
|
|
Note that the include dirs DO have double quotes around them.
|
|
I traced the problem down to the custom CCompiler spawn ) function which replaces the distutils spawn function.
|
|
If I understand correctly, the idea is to use numpy s quote args ) in gen lib options ) to put quotes around given library dirs, and to not check for the need of quotes in spawn.
|
|
It seems, however, as if one forgot that the libary dirs returned by distutils gen lib options ) can also contain spaces.
|
|
And they do if you install Python in program files .
|
|
Anyway, I propose to fix the problem by placing the following line at the very start of the CCompiler spawn ) function cmd = quote args cmd) Additionally, I propose to change quote args to also check whether the LAST character is a quote.
|
|
This seems a bit more robust.
|
|
Consider the two cases LIBPATH C \Program Files x86)\python32\libs LIBPATH C \Program Files x86)\python32\libs ===== Below follows a minimal Cython example and traceback ===== ===== test .pyx def foo ) print hello ) ===== setup.py import os, sys from Cython.Distutils import build ext from distutils.core import setup from distutils.extension import Extension from numpy.distutils.misc util import get numpy include dirs Ugly hack so I can run setup.py in my IDE sys.argv = [ setup.py , build ext , inplace ] Init include dirs include dirs = [ . ]
|
|
include dirs.extend get numpy include dirs )) Creat Extensions ext modules = [ Extension test , [ test .pyx ], include dirs=include dirs, ), ] Compile setup cmdclass = { build ext build ext}, ext modules = ext modules, ) print Successfully compiled cython file test ) ===== output when running setup.py running build ext No module named msvccompiler in numpy.distutils; trying from distutils cythoning test .pyx to test .c building test extension C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\cl.exe c nologo Ox MD W3 GS DNDEBUG I. I C \Program Files x86)\python32\lib\site packages\numpy\core\include I C \Program Files x86)\python32\include I C \Program Files x86)\python32\PC Tctest .c Fobuild\temp.win32 3.2\Release\test .obj Found executable C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\cl.exe C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\link.exe DLL nologo INCREMENTAL NO LIBPATH C \Program Files x86)\python32\libs LIBPATH C \Program Files x86)\python32\PCbuild EXPORT PyInit test build\temp.win32 3.2\Release\test .obj OUT C \almar\projects\py\cmu1394\test .pyd IMPLIB build\temp.win32 3.2\Release\test .lib MANIFESTFILE build\temp.win32 3.2\Release\test .pyd.manifest Found executable C \Program Files x86)\Microsoft Visual Studio 9.0\VC\BIN\link.exe LINK fatal error LNK1181 cannot open input file Files.obj","bug"
|
|
"6013","2612","concatenate does not understand dict.values ).
|
|
Original report below edit by seberg).
|
|
I think there might still be an issue here.
|
|
It does raise an error now, but `dict.values )` does have a `len`, so if this makes more problems in other areas as well, NumPy should maybe use iteration to fetch such values, or just convert to a tuple.
|
|
Original ticket http projects.scipy.org numpy ticket 2019 on 2012 01 15 by trac user takluyver, assigned to unknown.
|
|
Steps to reproduce In [1] import numpy as np In [2] np.
|
|
version Out[2] 1.6.1 In [3] d = {1 np.ones 5), 2 np.zeros 5)} In [4] np.concatenate d.values )) Segmentation fault gdb output This GDB was configured as i686 linux gnu .
|
|
For bug reporting instructions, please see <http bugs.launchpad.net gdb linaro >... Reading symbols from home thomas Code virtualenvs pymc3 bin python3... no debugging symbols found)...done.
|
|
gdb) run c import numpy as np;d = {1 np.ones 5), 2 np.zeros 5)};np.concatenate d.values )) Starting program home thomas Code virtualenvs pymc3 bin python3 c import numpy as np;d = {1 np.ones 5), 2 np.zeros 5)};np.concatenate d.values )) [Thread debugging using libthread db enabled] Program received signal SIGSEGV, Segmentation fault.
|
|
PyArray ConvertToCommonType op=0x8aa0584, retn=0xbfffee4c) at numpy core src multiarray convert datatype.c 1515 1515 if !PyArray CheckAnyScalar otmp)) { gdb) c Continuing.
|
|
Program terminated with signal SIGSEGV, Segmentation fault.
|
|
The program no longer exists.
|
|
This is with Python 3.2.2 on 32 bit Linux.
|
|
Converting it to a list works fine In [3] np.concatenate list d.values ))) Out[3] array [ 1., 1., 1., 1., 1., 0., 0., 0., 0., 0.])","bug"
|
|
"6013","2613","Reduce doesn t call inner loop for scalars.
|
|
Trac 2020).
|
|
Original ticket http projects.scipy.org numpy ticket 2020 on 2012 01 15 by charris, assigned to unknown.
|
|
This isn t a problem for current ufuncs, but suppose there was a binary ufunc addrounded a, b) that added the rounded values of a and b.
|
|
Then addrounded.reduce 1.5) should be 1, but currently it will just return the element 1.5 without calling the inner loop that would round the result.
|
|
On the other hand, it would work fine with addrounded.reduce [1.5])","bug"
|
|
"6013","2614","segfault on searchsorted Trac 2021).
|
|
Original ticket http projects.scipy.org numpy ticket 2021 on 2012 01 17 by trac user adamklein, assigned to unknown.
|
|
This results in a segfault In [10] x = np.array [1,2,3], datetime64[D] ) In [11] x.searchsorted 2)","bug"
|
|
"6013","2615","Allow user extensions of numpy to run without the GIL but occasionally grab it Trac 2022).
|
|
Original ticket http projects.scipy.org numpy ticket 2022 on 2012 01 18 by girving, assigned to unknown.
|
|
Currently numpy supports two execution modes for the execution of extension code with the Python API and GIL) and entirely without.
|
|
However, a reasonable middle ground is code that runs without accessing the Python API most of the time, but occasionally needs to grab access, for example to set a Python exception.
|
|
While in the current version of numpy it is possible to set the no api flag and explicitly grab the GIL if access is needed, the resulting behavior is broken since numpy does not check for any resulting exceptions.
|
|
The easiest fix for this problem is simply to check for exceptions regardless if the user code claims to need the Python API, as in this patch https github.com girving numpy commit 730b05a892371d6f18d9317e5ae6dc306c0211b0 However, such a modification would conflict with any future modifications of numpy to parallelize array operations over multiple threads.
|
|
Therefore, I propose a pair of numpy api functions which acquire and release the GIL safely from inside numpy loop code.
|
|
The semantics of these functions would be the same as PyGILState Ensure Release, but would work correctly even if called from a non Python thread created by numpy for parallelization purposes.","enhancement"
|
|
"6013","2616","float128 does not convert correctly from strings Trac 2023).
|
|
Original ticket http projects.scipy.org numpy ticket 2023 on 2012 01 18 by trac user parejkoj, assigned to unknown.
|
|
Converting strings to float128 dtype f16 ) does not produce the desired precision, but rather converts to float64 first.
|
|
I m using OS X 10.6.8, numpy 1.6.1, python 2.6.7, compiled under Fink.
|
|
Example a=np.float128 0.12345678901234567890 ) b=np.float64 0.12345678901234567890 ) repr a) repr b) a==b produces 0.12345678901234567737 0.12345678901234568 True When it should produce 0.12345678901234567890 0.12345678901234568 False I couldn t find another bug report about this, but this email thread describes the same problem http comments.gmane.org gmane.comp.python.numeric.general 46380 Whereas this bug report from two years ago seems to show correct behavior http projects.scipy.org numpy ticket 1311","bug"
|
|
"6013","2617","NULL bytes removed from end of string in NumPy array Trac 2024).
|
|
Original ticket http projects.scipy.org numpy ticket 2024 on 2012 01 19 by trac user edcjones, assigned to unknown.
|
|
I use Debian stable on a PC.
|
|
The distribution uses Python 2.6.6 and NumPy 1.4.1.
|
|
In Python and I assume in NumPy, strings are not NULL terminated.
|
|
The following program strips the zero bytess from the end of a string !
|
|
usr bin env python import numpy DT = numpy.dtype [ hashcode , numpy.str , 16)]) badstring = 4 chr 0) ABCDEFGH 4 chr 0) print repr badstring) arr = numpy.array badstring,), dtype=DT) print repr arr) s = str arr[ hashcode ]) print len s), repr s) The output is \x00\x00\x00\x00ABCDEFGH\x00\x00\x00\x00 array \x00\x00\x00\x00ABCDEFGH ,), dtype=[ hashcode , |S16 )]) 12 \x00\x00\x00\x00ABCDEFGH It appears that the string is being treated like a C string somewhere in NumPy.","bug"
|
|
"6013","2618","NA causes problems for the real imag views.
|
|
Trac 2025).
|
|
Original ticket http projects.scipy.org numpy ticket 2025 on 2012 01 21 by charris, assigned to unknown.
|
|
In [5] a = np.array [np.NA], maskna=1) In [6] a.real Out[6] array [ NA], dtype=float64) In [7] a.imag Out[7] array [ 0.])
|
|
Imaginary part should show NA.
|
|
In [8] a = np.array [np.NA], dtype=np.complex128, maskna=1) In [9] a.real ValueError Traceback most recent call last) home charris <ipython input 9 f93368bb6a37> in <module> ) > 1 a.real ValueError Cannot construct a view of data together with the NPY ARRAY MASKNA flag, the NA mask must be added later In [10] a.imag ValueError Traceback most recent call last) home charris <ipython input 10 e8e5f986ad96> in <module> ) > 1 a.imag ValueError Cannot construct a view of data together with the NPY ARRAY MASKNA flag, the NA mask must be added later","bug"
|
|
"6013","2619","einsum iterator bug Trac 2027).
|
|
Original ticket http projects.scipy.org numpy ticket 2027 on 2012 01 21 by mwiebe, assigned to unknown.
|
|
This was reported by royf by reopening ticket 2478.
|
|
I m moving it to a new ticket, as it s a new bug, even though it s very similar in nature to the one reopened.
|
|
Problem not solved in 1.6.1, or a very similar problem remains import numpy; print numpy.version.full version); a = numpy.arange 1, 3); b = numpy.arange 1, 5).reshape 2, 2); c = numpy.arange 1, 9).reshape 4, 2); numpy.einsum x,yx,zx >xzy , a, b, c) 1.6.1 array [[[ 1, 3], [ 3, 9], [ 1, 3], [ 3, 9]], [[ 8, 16], [16, 32], [ 8, 16], [16, 32]]]) The correct result is numpy.array [[[a[x] b[y, x] c[z, x] for y in range 2)] for z in range 4)] for x in range 2)]) array [[[ 1, 3], [ 3, 9], [ 5, 15], [ 7, 21]], [[ 8, 16], [16, 32], [24, 48], [32, 64]]])","bug"
|
|
"6013","2620","dtype detection doesn t check for unsigned long long Trac 2028).
|
|
Original ticket http projects.scipy.org numpy ticket 2028 on 2012 01 25 by mwiebe, assigned to unknown.
|
|
This was reported on the mailing list here http comments.gmane.org gmane.comp.python.numeric.general 47557 The fix needs to go in the function array find python scalar type in common.c, here as of this bug report https github.com numpy numpy blob master numpy core src multiarray common.c L18","bug"
|
|
"6013","2621","Should not allow views of NA masked arrays which change the array size Trac 2029).
|
|
Original ticket http projects.scipy.org numpy ticket 2029 on 2012 01 25 by mwiebe, assigned to unknown.
|
|
Chuck posted an example Looks like views of masked arrays have other problems In [13] a = ones 3, int16, maskna=1) In [14] a.view int8) Out[14] array [1, 0, 1, NA, 1, NA], dtype=int8) This looks like a serious bug to me, to avoid memory corruption issues it should raise an exception.","bug"
|
|
"6013","2622","random Power Law distribution a global approach Trac 2030).
|
|
Original ticket http projects.scipy.org numpy ticket 2030 on 2012 01 25 by trac user tristan, assigned to unknown.
|
|
Hi, It s my first post in Numpy so if I m not clear or if you need more informations, let me know.
|
|
I search for a Numpy function which permit to draw a random distribution from a power law N x) = x alpha.
|
|
I m surprised to not find an easy way to make it with Numpy.
|
|
Of course there are numpy.random.power only for alpha = a 1 > 1) and numpy.random.pareto only for alpha = a 1 < 1) respectively in [http docs.scipy.org doc numpy reference generated numpy.random.power.html] and [http docs.scipy.org doc numpy reference generated numpy.random.pareto.html].
|
|
In the Pareto distribution, there is no way at least as explain in the documentation) to change the shape m .
|
|
Finally, there is no solution to have a power law with alpha = 1!
|
|
So my proposition is quiet easy.
|
|
I propose to add a new numpy.random function called for example genpower for GENeral POWER law) that permit us to draw a random distribution from a power law where alpha is a real.
|
|
In this aim, I wrote a simple function based on the IDL randomP.pro [http astro.uni tuebingen.de software idl astrolib math randomp.html]).
|
|
import numpy as np def genpower alpha, min, max, N=1000, seed=None) Random Power Law distribution generator PDF x,alpha) = x alpha Input arg alpha Power Law index min Minimal value for X max Maximal value for X arg N Number of value to draw default N=1000) sedd Value for the seed generator default None) Convert min, max a, and alpha to float min = float min) max = float max) alpha = float alpha) Fixe seed if seed != None np.random.seed seed) Get uniform vector rand = np.random.uniform 0.0, 1.0, N) Test alpha != 1 if alpha != 1.0 General case pow = alpha 1.0 X = min pow rand max pow min pow) ) 1.0 pow) Normalisation integral = 1.0 pow max pow min pow) else Particular case alpha = 1 X = min max min) rand Normalisation integral = np.log max min) endIF Return return X, integral return X endDEF if name == main alpha = 1.5 min = 3. max = 1e3 N = 1e5 X = genpower alpha, min, max, N=N) Show distribution import pylab as py py.figure ) ax = py.subplot 2, 1, 1) n, b, p = py.hist X, 100, histtype= step , color= k ) bin = np.array [ b[i] b[i 1]) 2. for i in range 0,len b) 1) ]) Get centered bin value powerlaw = bin alpha np.sum n) np.sum bin alpha) py.plot bin, powerlaw) ax = py.subplot 2, 1, 2) n, b, p = py.hist X, 100, histtype= step , color= k ) bin = np.array [ b[i] b[i 1]) 2. for i in range 0,len b) 1) ]) Get centered bin value powerlaw = bin alpha np.sum n) np.sum bin alpha) py.loglog bin, powerlaw) py.show ) endMAIN I the attached a file randomPL.eps , which shows some distribution for different value of alpha.
|
|
I hope to help Numpy community.
|
|
Tristan","enhancement"
|
|
"6013","2623","numpy.cov behaviour change when called with 2 arrays cross covariance) Trac 2031).
|
|
Original ticket http projects.scipy.org numpy ticket 2031 on 2012 01 26 by trac user pierreh, assigned to unknown.
|
|
Following up a discussion on Numpy discussion ML, I propose a behaviour change in numpy.cov function, when called with 2 arrays.
|
|
cov should only compute the cross covariance of the two arrays Today s behaviour description cov X,Y) is just an equivalent of cov concatenate X,Y)).
|
|
The output can be written in terms of blocks 1 N X X.T X Y.T) Y X.T Y Y.T) Therefore, 3 blocks out of 4 are often useless, since diagonal blocks are simply cov X) and cov Y) while the off diagonal blocks are transpose of each other only one them is needed.
|
|
This brings two issues today s behavior is somehow useless since cov concatenate X,Y)) is equivalent it consumes 3 times as many resources as what is needed.
|
|
Proposed behaviour the output of cov X,Y) should only be one of the cross covariance block 1 N X Y.T) This would put numpy in line with R s behaviour http stat.ethz.ch R manual R patched library stats html cor.html) Also, cov X,X) would become equivalent to cov X) Implementation skeleton def cov X, Y=None) if Y is None Y = X else assert Y.shape == X.shape or something like that [...jumping to the end of the existing code...] if not rowvar return dot X.T, Y.conj )) fact).squeeze ) else return dot X, Y.T.conj )) fact).squeeze )","enhancement"
|
|
"6013","2624","cross doesn t broadcast correctly Trac 2032).
|
|
Original ticket http projects.scipy.org numpy ticket 2032 on 2012 01 27 by girving, assigned to unknown.
|
|
cross doesn t broadcast correctly in all cases.
|
|
Here s an example >>> cross random.randn 2,1,3),random.randn 5,3)) Traceback most recent call last) File <stdin> , line 1, in <module> File opt local Library Frameworks Python.framework Versions 2.6 lib python2.6 site packages numpy core numeric.py , line 1377, in cross x = a[1] b[2] a[2] b[1] ValueError operands could not be broadcast together with shapes 1,2) 5)","bug"
|
|
"6013","2625","a 2.0 has incorrect type Trac 2033).
|
|
Original ticket http projects.scipy.org numpy ticket 2033 on 2012 01 30 by trac user kasal, assigned to teoliphant.
|
|
Hello, I observe an incorrect evaluation of the power operator for the exponent = 2.0 print NumPy , np.
|
|
version a = np.array [2, 4], dtype=np.int16) for e in np.arange 5.0) s = a {} .format e) print { 9} {} .format s, repr eval s))) !NumPy 1.6.0 a 0.0 array [ 1., 1.])
|
|
a 1.0 array [ 2., 4.])
|
|
a 2.0 array [ 4, 16], dtype=int16) a 3.0 array [ 8., 64.])
|
|
a 4.0 array [ 16., 256.])
|
|
In all other cases, the output dtype is float, because of te float exponent.
|
|
But a 2.0 seems to be incorrectly reduced to a a .","bug"
|
|
"6013","2626","Implement sincos ) Trac 2034).
|
|
Original ticket http projects.scipy.org numpy ticket 2034 on 2012 01 30 by trac user nschloe, assigned to unknown.
|
|
Hi, NumPy has capabilites for all kinds of trigonometric functions, but still missing out one important bit.
|
|
In most every case I know of, both the sine and cosine of an angle need to be computed e.g.
|
|
for the representation of a circle).
|
|
Numerically, sine and cosine can be evaluated at the same time, thus saving a great deal of complexity.
|
|
This numerical method is implemented in http linux.die.net man 3 sincos, for example.
|
|
Cheers, Nico","enhancement"
|
|
"6013","2627","Broadcasting with masked array doesn t work with divide after tile instruction) Trac 2035).
|
|
Original ticket http projects.scipy.org numpy ticket 2035 on 2012 02 02 by trac user elephantterrible, assigned to unknown.
|
|
Broadcasting in a division does not work after a tile instruction.
|
|
See attached script to reproduce.
|
|
If the tile instruction is commented out, the division works.
|
|
Division also works if numpy.divide ) is used, or if there s no masked value.
|
|
Python gives the following error Traceback most recent call last) File ma tiling issue.py , line 18, in <module> a 100. raises ValueError invalid return array shape File usr local lib python2.7 dist packages numpy ma core.py , line 3654, in div return divide self, other) File usr local lib python2.7 dist packages numpy ma core.py , line 1078, in call m |= ma ValueError non broadcastable output operand with shape 3,3,3) doesn t match the broadcast shape 1,3,3,3) The error occurs in all numpy versions 1.5.1, 1.6.1, 2.0.0.dev) with python 2.7.1 The original bug discussion is on numpy discuss http www.mail archive.com numpy discussion scipy.org msg35146.html","bug"
|
|
"6013","2628","Leak when mixing OpenCV GetMat and NumPy asarray Trac 2036).
|
|
Original ticket http projects.scipy.org numpy ticket 2036 on 2012 02 02 by trac user r2d3, assigned to unknown.
|
|
The following code leaks My profiling tool Instruments on OSX) that OpenCV cvmat array struct allocated the memory.
|
|
But I suspect asarray to keep a reference on it in some way.
|
|
OSX 10.6 Python 2.6 Numpy 1.6.1 OpenCV 2.3.1 The leaking sample code import cv import numpy as N while True img = cv.CreateImage 800,600), 8, 1); numpy Img = N.asarray cv.GetMat img))","bug"
|
|
"6013","2629","genfromtxt with dtype different behavior in 32bit or 64bit numpy Trac 2037).
|
|
Original ticket http projects.scipy.org numpy ticket 2037 on 2012 02 03 by trac user risotto, assigned to unknown.
|
|
I ve got installed on one machine debian linux 6) numpy in 32bit, with virtual environment, and on the other machine is numpy in 64bit installed via virtual environment.
|
|
There s an issue, if in your file is an float value, but want to read it as integer.
|
|
The example will be more apparent.
|
|
On 32bit >>> import sys >>> sys.maxint 2147483647 >>> from StringIO import StringIO >>> import numpy as np >>> s = StringIO 333 12.5 22 ) >>> data = np.genfromtxt s, dtype=[ x , float), y , int), z , int)]) >>> data array 333.0, 12, 22), dtype=[ x , <f8 ), y , <i4 ), z , <i4 )]) >>> s = StringIO 333 12.5 22 ) >>> data = np.genfromtxt s, dtype=[ x , float), y , long), z , int)]) >>> data array 333.0, 1L, 22), dtype=[ x , <f8 ), y , <i8 ), z , <i4 )]) On 64bit >>> import sys >>> sys.maxint 9223372036854775807 >>> from StringIO import StringIO >>> import numpy as np >>> s = StringIO 333 12.5 22 ) >>> data = np.genfromtxt s) >>> data array [ 333. , 12.5, 22. ])
|
|
>>> s = StringIO 333 12.5 22 ) >>> data = np.genfromtxt s, dtype=[ x , float), y , int), z , int)]) >>> data array 333.0, 1, 22), dtype=[ x , <f8 ), y , <i8 ), z , <i8 )]) >>> s = StringIO 333 12.5 22 ) >>> data = np.genfromtxt s, dtype=[ x , float), y , long), z , int)]) >>> data array 333.0, 12L, 22), dtype=[ x , <f8 ), y , <i8 ), z , <i8 )])","bug"
|
|
"6013","2630","numpy.core.arrayprint.IntegerFormatter uses wrong exception catching code Trac 2038).
|
|
Original ticket http projects.scipy.org numpy ticket 2038 on 2012 02 03 by trac user fijal, assigned to unknown.
|
|
code goes like this except TypeError, NotImplementedError if reduce data) fails, this instance will not be called, just instantiated in formatdict.
|
|
pass except ValueError this occurs when everything is NA pass should be except TypeError, NotImplementedError) ... instead","bug"
|
|
"6013","2631","Numpy wrongly assumes that any non i386 OS X Darwin) system is PowerPC Trac 2039).
|
|
Original ticket http projects.scipy.org numpy ticket 2039 on 2012 02 04 by trac user jdemeyer, assigned to cournape.
|
|
Apparently Numpy forgot about the existence of x86 64 systems Python 2.7.2 default, Feb 3 2012, 15 29 15) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> from distutils.util import get platform >>> get platform ) macosx 10.6 x86 64 The bug is in `src numpy distutils system info.py` if sys.platform== darwin and not os.environ.get ATLAS ,None) args = [] link args = [] if get platform )[ 4 ] == i386 intel = 1 else intel = 0 if os.path.exists System Library Frameworks Accelerate.framework ) if intel args.extend [ msse3 ]) else args.extend [ faltivec ]) link args.extend [ Wl, framework , Wl,Accelerate ])","bug"
|
|
"6013","2632","f2py does not treat quad precision real complex Trac 2040).
|
|
Original ticket http projects.scipy.org numpy ticket 2040 on 2012 02 07 by inducer, assigned to unknown.
|
|
As is visible from this code https github.com numpy numpy blob master numpy f2py capi maps.py L35 f2py incorrectly maps quad precision real complex down to their real valued counterparts, leading to crashes and or incorrect data in user applications.
|
|
Compiler support for quad in icc and gcc both C and Fortran side) is decent nowadays, so there s little reason to do this, IMO.
|
|
If this has a chance of getting in, I d try to provide a patch.","bug"
|
|
"6013","2633","support multiple comment strings in loadtxt ) Trac 2041).
|
|
Original ticket http projects.scipy.org numpy ticket 2041 on 2012 02 08 by trac user andreash, assigned to unknown.
|
|
I would like to see the following enhancement in !NumPy In the numpy.loadtxt method, support passing multiple comment strings, perhaps as a list.
|
|
In reality, people change dataformats quite often, and too many people don t care too much about the users of their data at least in my field of science).
|
|
So when processing large datasets, I have to deal with files which either have ` ` or ` ` as comment string.
|
|
If !NumPy supported something like the following, I d be happy loadtxt myfile.txt , comments=[ , ])","enhancement"
|
|
"6013","2634","complex log breaks log conj x)) = conj log x)) invariant Trac 2042).
|
|
Original ticket http projects.scipy.org numpy ticket 2042 on 2012 02 09 by trac user radford, assigned to unknown.
|
|
The following should both be true, but the second fails.
|
|
numpy.log 1.0 0.0j) == numpy.pi 1j numpy.log 1.0 0.0j) == numpy.pi 1j Background The c99 clog ) function guarantees clog conj x)) == conj clog x)) and numpy attempts to conform to conform to the C99 standard, but doesn t in the above case.","bug"
|
|
"6013","2635","Can t index with a zero length bool array Trac 2043).
|
|
Original ticket http projects.scipy.org numpy ticket 2043 on 2012 02 09 by trac user radford, assigned to unknown.
|
|
You can index an array with a boolean array for all lengths except zero.
|
|
Specifically this works for i>0, but not when it s i==0.
|
|
i = 0; np.arange 3 i).reshape 3,i)[ ,np.bool [True] i)] Attached is a proposed patch to fix the problem.","bug"
|
|
"6013","2636","Trying to delete ndarray.flat from an instance segfaults Trac 2044).
|
|
Original ticket http projects.scipy.org numpy ticket 2044 on 2012 02 10 by alex, assigned to unknown.
|
|
Tested at the latest master of numpy from github e5dd45b0 c56f 4e03 8b53 e3fda8cc68bc)alex alex gaynor laptop .virtualenvs e5dd45b0 c56f 4e03 8b53 e3fda8cc68bc python Python 2.7.2 default, Oct 4 2011, 20 06 09) [GCC 4.6.1] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.arange 2) array [0, 1]) >>> numpy.arange 2).flat <numpy.flatiter object at 0x2ea3800> >>> del numpy.arange 2).flat Segmentation fault","bug"
|
|
"6013","2637","Trying to delete an item from a flatiter segfaults Trac 2045).
|
|
Original ticket http projects.scipy.org numpy ticket 2045 on 2012 02 10 by alex, assigned to unknown.
|
|
Tested at the latest master of the github repo e268cfc7 ff3f 4d97 a02c 27481c8a8186)alex alex gaynor laptop .virtualenvs e268cfc7 ff3f 4d97 a02c 27481c8a8186 python Python 2.7.2 default, Oct 4 2011, 20 06 09) [GCC 4.6.1] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.arange 2) array [0, 1]) >>> numpy.arange 2).flat <numpy.flatiter object at 0x2e49910> >>> del numpy.arange 2).flat[1] Segmentation fault","bug"
|
|
"6013","2638","Deleting ndarray.real on an instance segfaults Trac 2046).
|
|
Original ticket http projects.scipy.org numpy ticket 2046 on 2012 02 10 by alex, assigned to charris.
|
|
Tested on the latest numpy master from github e268cfc7 ff3f 4d97 a02c 27481c8a8186)alex alex gaynor laptop .virtualenvs e268cfc7 ff3f 4d97 a02c 27481c8a8186 python Python 2.7.2 default, Oct 4 2011, 20 06 09) [GCC 4.6.1] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> del numpy.arange 5).real Segmentation fault","bug"
|
|
"6013","2639","f2py needs to be updated to use non deprecated macros functions Trac 2047).
|
|
Original ticket http projects.scipy.org numpy ticket 2047 on 2012 02 11 by charris, assigned to pearu.
|
|
This is a work in progress.
|
|
There is a working branch at [https github.com charris numpy tree f2py deprecated fixups]","bug"
|
|
"6013","2640","memory overflow when using numpy load in a loop Trac 2048).
|
|
Original ticket http projects.scipy.org numpy ticket 2048 on 2012 02 11 by trac user eldada, assigned to unknown.
|
|
Looping over npz files load causes memory overflow depending on the file list length).
|
|
None of the following seems to help[[BR]] 1.
|
|
Deleting the variable which stores the data in the file.
|
|
[[BR]] 2.
|
|
Using mmap.
|
|
[[BR]] 3. calling gc.collect ) garbage collection).
|
|
[[BR]] [[BR]] The following code should reproduce the phenomenon [[BR]] import numpy as np[[BR]] generate a file for the demo[[BR]] X = np.random.randn 1000,1000)[[BR]] np.savez tmp.npz ,X=X)[[BR]] [[BR]] here come the overflow [[BR]] for i in xrange 1000000) [[BR]] data = np.load tmp.npz )[[BR]] data.close ) avoid the too many files are open error[[BR]] [[BR]] in my real application the loop is over a list of files and the overflow exceeds 24GB of RAM!
|
|
[[BR]] please note that this was tried on ubuntu 11.10, and for both numpy v 1.5.1 as well as 1.6.0","bug"
|
|
"6013","2641","Building 64 bit using VS2008 EE Trac 2049).
|
|
Original ticket http projects.scipy.org numpy ticket 2049 on 2012 02 12 by trac user jusic, assigned to pv.
|
|
Hi, I just managed to compile a 64 bit version of latest numpy git master on Windows 7 Prof. 64 bit and VS2008 Express Edition PSDK 6.1.
|
|
I would like to provide some input for Microsoft Toolchain Support wiki page http projects.scipy.org numpy wiki MicrosoftToolchainSupport), since it s not too specific on how this is done.
|
|
I installed the following packages Windows SDK for Windows Server 2008 and .NET Framework 3.5 [[BR]] download link http www.microsoft.com download en details.aspx?id=24826 [[BR]] file 6.0.6001.18000.367 KRMSDK EN.iso Microsoft Visual Studio 2008 Express Edition [[BR]] download link http www.microsoft.com download en details.aspx?id=3092 [[BR]] file VS2008ExpressENUX1397868.iso Python 2.7.2 amd64) distutils all that is needed.
|
|
C \Python27\ is in PATH .
|
|
To make amd64 compilers work, the batch file C \Program Files x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat needs to be fixed on lines 19 and 20 to if not exist dp0bin\vcvars64.bat goto missing call dp0bin\vcvars64.bat Then, to build numpy install package, open command prompt and cd to numpy git directory, and run > C \Program Files x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat amd64 > python setup.py config > python setup.py build > python setup.py bdist wininst The package seems to work even if I didn t have a Fortran compiler.
|
|
I had MinGW installed but not in PATH .
|
|
Thanks.","enhancement"
|
|
"6013","2642","setup.py Git revision doesn t work on Windows Trac 2050).
|
|
Original ticket http projects.scipy.org numpy ticket 2050 on 2012 02 12 by trac user jusic, assigned to rgommers.
|
|
Building as described in 2641, I ran into a problem that the installation package version ends up being 2.0.0dev Unknown .
|
|
This is because the current setup.py doesn t find PATH correctly on Windows for some reason.
|
|
Please see the attached patch, which works on my setup.
|
|
git is perfectly runnable from command line as is.","bug"
|
|
"6013","2643","Use the shortest repr for np.float32 float64 Trac 2051).
|
|
Original ticket http projects.scipy.org numpy ticket 2051 on 2012 02 12 by trac user beaujolais, assigned to unknown.
|
|
Python 2.7 and 3.1 now always use the shortest decimal representation for numbers that are not exactly representable by a binary float see [http bugs.python.org issue1580]).
|
|
This causes some very surprising inconsistencies with NumPy s own repr >>> x = 0.1 >>> x 0.1 >>> a = np.array [x]) >>> a array [ 0.1]) >>> a[0] 0.10000000000000001 >>> float a[0]) 0.1 >>> type a[0]) <class numpy.float64 > For the sake of consistency I suggest adopting the same repr in NumPy.","enhancement"
|
|
"6013","2644","numpy record dtypes segfault Trac 2052).
|
|
Original ticket http projects.scipy.org numpy ticket 2052 on 2012 02 13 by trac user fijal, assigned to unknown.
|
|
An example session >>> from numpy import >>> a = zeros 2, dtype=[ x , int), y , float)]) >>> a[ x ] array [0, 0]) >>> a[0] 0, 0.0) >>> del a[0][ x ] Segmentation fault","bug"
|
|
"6013","2645","void.tolist segfault Trac 2053).
|
|
Original ticket http projects.scipy.org numpy ticket 2053 on 2012 02 13 by trac user fijal, assigned to unknown.
|
|
Example >>> from numpy import void >>> void 5).tolist )[0] = a Segmentation fault","bug"
|
|
"6013","2646","f2py s handling of strided arrays.
|
|
Trac 2054).
|
|
Original ticket http projects.scipy.org numpy ticket 2054 on 2012 02 14 by trac user pch, assigned to pearu.
|
|
There was some discussion some time ago on numpy s mailing list to include strided array handling in f2py.
|
|
I badly need this feature, so here I file the request for enhancement.
|
|
Given a Fortran routine, subroutine multiply me a) integer 8, intent inout) a , ) a = a 2 end subroutine multiply me The following Python code should give > a = np.ones 2,4), np.int64) > multiply me a[ 2, 2]) > print a [[2,1,2,1] [1,1,1,1]]","enhancement"
|
|
"6013","2647","Fix can cast to deal with struct dtypes Trac 2055).
|
|
Original ticket http projects.scipy.org numpy ticket 2055 on 2012 02 15 by mwiebe, assigned to unknown.
|
|
Matthew Brett noticed that can cast has some weird behaviors when calling it with struct dtypes.
|
|
The email thread starts here http mail.scipy.org pipermail numpy discussion 2012 February 060388.html He had a suggestion for beginning to approach the problem How about, at a first pass, returning True for conversion of void types only if input dtype == output dtype, then adding more sophisticated rules later?","bug"
|
|
"6013","2648","String to complex type conversion fails in Python3 Trac 2056).
|
|
Original ticket http projects.scipy.org numpy ticket 2056 on 2012 02 17 by charris, assigned to unknown.
|
|
String to complex type conversion fails in Python3 >>> a = np.ones 2, S ) >>> a.astype D ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError complex ) argument must be a string or a number","bug"
|
|
"6013","2649","Invalid output shape from np.dot with matrixes Trac 2057).
|
|
Original ticket http projects.scipy.org numpy ticket 2057 on 2012 02 17 by pv, assigned to unknown.
|
|
http permalink.gmane.org gmane.comp.python.scientific.user 31095 Consider this import numpy as np x = np.arange 5) I = np.asmatrix np.identity 5)) print np.dot I, x).shape > 1, 5) while ` 5, 1)` would be what is expected based on usual linear algebra rules.
|
|
It might also be a mistake to return a matrix from mixed matrix array computations.","bug"
|
|
"6013","2650","Strange behavior of binary ufunc with scipy sparse matrices Trac 2058).
|
|
Original ticket http projects.scipy.org numpy ticket 2058 on 2012 02 17 by pv, assigned to unknown.
|
|
The ufuncs behave strangely when fed Scipy sparse matrices import numpy as np import scipy.sparse as sp A = sp.csc matrix np.identity 5)) B = np.asmatrix np.ones 5,5))) print np.multiply A,A) <5x5 sparse matrix of type <class numpy.float64 > with 5 stored elements in Compressed Sparse Column format> print np.multiply B,A) NotImplemented print np.multiply A,B).dtype object There are a couple things amiss here, the reason for which probably resides in Numpy returning NotImplemented instead of raising ValueError same as 2091 ) it seems that in the last case, `csc matrix.multiply` is not called.
|
|
See also http projects.scipy.org scipy ticket 1598","bug"
|
|
"6013","2651","Numpy float plus times poly1d gives array Trac 2059).
|
|
Original ticket http projects.scipy.org numpy ticket 2059 on 2012 02 20 by trac user ezander, assigned to unknown.
|
|
Multiplication or addition of a poly1d object with a numpy float from the left, gives an ndarray object, instead of an poly1d.
|
|
See the following code In [38] a=np.float64 3.0) In [39] a np.poly1d [1,1,1]) Out[39] array [ 4., 4., 4.])
|
|
In [40] np.poly1d [1,1,1]) a Out[40] poly1d [ 1., 1., 4.])","bug"
|
|
"6013","2652","Blasdot tests with out argument failing Trac 2060).
|
|
Original ticket http projects.scipy.org numpy ticket 2060 on 2012 02 20 by rgommers, assigned to unknown.
|
|
On Python 2.5 seeing these failures, works on 3.1 ====================================================================== ERROR test blasdot.test dot 3args Traceback most recent call last) File C \Python25\lib\site packages\nose 0.11.3 py2.5.egg\nose\case.py , line 186, in runTest self.test self.arg) File C \Python25\Lib\site packages\numpy\core\tests\test blasdot.py , line 51, in test dot 3args np.dot f,v,r) TypeError function takes exactly 2 arguments 3 given) ====================================================================== ERROR test blasdot.test dot 3args errors Traceback most recent call last) File C \Python25\lib\site packages\nose 0.11.3 py2.5.egg\nose\case.py , line 186, in runTest self.test self.arg) File C \Python25\Lib\site packages\numpy\core\tests\test blasdot.py , line 69, in test dot 3args errors assert raises ValueError, np.dot, f, v, r) File C \Python25\Lib\site packages\numpy\testing\utils.py , line 1053, in assert raises return nose.tools.assert raises args, kwargs) File C \Python25\lib\unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) TypeError function takes exactly 2 arguments 3 given)","bug"
|
|
"6013","2653","difference between behavior of assignment division and future division not documented Trac 2061).
|
|
Original ticket http projects.scipy.org numpy ticket 2061 on 2012 02 21 by trac user AlanFrankel, assigned to pv.
|
|
I just ran into the same issue described in bug 1734 assignment division not compatible with future division).
|
|
Namely, I expected to see that the results of array a = scalar s would depend on whether a from future import division statement was included in the code.
|
|
Eventually, I figured out this had no effect; the only way I could get true division, which I wanted, was to use array a = array a scalar s I see that bug 1734 has been marked won t fix .
|
|
Whether or not I agree with that assessment, I definitely think that this behavior should be in user accessible documentation not just in the bug database).
|
|
The Mathematical Functions page http docs.scipy.org doc numpy reference routines.math.html ) doesn t have any reference to the assignment operators at all =, =, =, =), which in my opinion is a glaring omission, especially since it s hard to search for these strings via a punctuation stripping search engine.
|
|
The assignment operators should be listed or there should be a link to them), and the fact that = gives different results from should be stated.","bug"
|
|
"6013","2654","Numpy array direct indexing of complex numbers causes crash Trac 2062).
|
|
Original ticket http projects.scipy.org numpy ticket 2062 on 2012 02 23 by trac user nschutten, assigned to unknown.
|
|
On my computer with Python 2.6 32b), I get the following from numpy import a=array [1j,2j]) a[0 1 1] this gives output array [0.
|
|
1,j]) If I then try the direct index e.g.
|
|
a[1] , python crashes.
|
|
I have tried different 2.6 versions of Python as well as the latest Numpy Scipy, but the issue remains).
|
|
Any thoughts?
|
|
Regards, Niels","bug"
|
|
"6013","2655","unique ) does not seem to return correct index if array larger than 16 Trac 2063).
|
|
Original ticket http projects.scipy.org numpy ticket 2063 on 2012 02 23 by trac user lolowizard, assigned to unknown.
|
|
Unique seem to return the index of the last occurence rather than first occurence if len array)>16 To reproduce v=[0,0,0,0,0,1,1,1,1,1,1,2,2,2,2,2,2] w=[0,0,0,0,0,1,1,1,1,1,1,2,2,2,2] Only difference between v and w is the two extra values at the end.
|
|
Then unique v,return index=True) unique w,return index=True) Look at the results for both commands array [0, 1, 2]), array [ 0, 10, 15])) array [0, 1, 2]), array [0, 5, 9])) The two commands should return the same index!","bug"
|
|
"6013","2656","numpy.digitize uses a linear search when it should be using binary search Trac 2064).
|
|
Original ticket http projects.scipy.org numpy ticket 2064 on 2012 02 25 by trac user jpeel, assigned to unknown.
|
|
In numpy lib src compiled base.c, numpy.digitize s code uses linear searches rather than binary searches as it should.
|
|
The difference in speed is quite enormous when using very many bins.
|
|
Quite frankly, it should probably use searchsorted after checking that the bins are monotonic.
|
|
This was opened as a result of a question on stackoverflow [http stackoverflow.com questions 9444409 why is numpy much slower than matlab on a digitize example 9445500 9445500] You can see an example of the speed issue there.","bug"
|
|
"6013","2657","Poor ndarray.take performance on Fortran order arrays Trac 2065).
|
|
Original ticket http projects.scipy.org numpy ticket 2065 on 2012 02 26 by wesm, assigned to unknown.
|
|
3000x slowdown observed In [25] arr = np.random.randn 350000, 5) In [26] timeit arr.take np.arange 5), axis=0) 100000 loops, best of 3 2.86 us per loop In [27] arr = np.random.randn 350000, 5).copy F ) In [28] timeit arr.take np.arange 5), axis=0) 100 loops, best of 3 9.03 ms per loop","bug"
|
|
"6013","2658","np.searchsorted segfaults when input is a recarray Trac 2066).
|
|
Original ticket http projects.scipy.org numpy ticket 2066 on 2012 02 26 by trac user josh.ayers, assigned to unknown.
|
|
The following code segfaults.
|
|
I m running NumPy 1.6.1 on Windows 7. import numpy as np dtype = np.format parser [ i4 , i4 ], [], []) a = np.recarray 2, ), dtype) np.searchsorted a, 1) The gdb output is below gdb) run c import crash; crash.run ) Starting program c \python27\python.exe c import crash; crash.run ) [New Thread 3520.0x9cc] Program received signal SIGSEGV, Segmentation fault.
|
|
0x673982ac in PyArray SearchSorted ) from c \python27\lib\site packages\numpy\core\multiarray.pyd gdb) bt 0 0x673982ac in PyArray SearchSorted ) from c \python27\lib\site packages\numpy\core\multiarray.pyd 1 0x673986e8 in array searchsorted ) from c \python27\lib\site packages\numpy\core\multiarray.pyd 2 0x1e0887c7 in python27!PyCFunction Call ) from c \python27\python27.dll 3 0x1e0bf781 in python27!PyEval GetFuncDesc ) from c \python27\python27.dll 4 0x0262abe8 in ?? )
|
|
5 0x1e1d57f8 in python27!PySuper Type ) from c \python27\python27.dll 6 0x00000002 in ?? )
|
|
7 0x01cedf30 in ?? )
|
|
Backtrace stopped previous frame inner to this frame corrupt stack?)
|
|
gdb)","bug"
|
|
"6013","2659","when x is an N D array, list x) produces surprising results Trac 2067).
|
|
Original ticket http projects.scipy.org numpy ticket 2067 on 2012 02 26 by trac user pfeldman, assigned to unknown.
|
|
x= array [[1,2],[3,4]]) list x) [array [1, 2]), array [3, 4])] list x.ravel )) [1, 2, 3, 4] The behavior of list x) is not what any reasonable person would want, and does not make sense conceptually either.
|
|
A Numpy 2 D array is not a nesting of 1 D arrays).
|
|
list x.ravel )) works but is cumbersome.","bug"
|
|
"6013","2660","Named dtype array Difference between a[0][ name ] and a[ name ][0]?
|
|
Trac 2068).
|
|
Original ticket http projects.scipy.org numpy ticket 2068 on 2012 02 27 by trac user nschloe, assigned to unknown.
|
|
Hi!
|
|
I came across the following oddity in numpy which may or may not be a bug import numpy as np dt = np.dtype [ tuple , int, 2))]) a = np.zeros 3, dt) type a[ tuple ][0]) ndarray type a[0][ tuple ]) ndarray a[ tuple ][0] = 1,2) ok a[0][ tuple ] = 1,2) ValueError shape mismatch on array construction I would have expected that both of the options below work.
|
|
Opinions?","bug"
|
|
"6013","2661","boolean indexing with no matches should leave shape same Trac 2069).
|
|
Original ticket http projects.scipy.org numpy ticket 2069 on 2012 03 02 by trac user shaunc, assigned to unknown.
|
|
>>> a = array [ [ 1, 1 ] ] ) >>> a[ a == 0 ].shape 0, ) This is also puzzling as >>> aa == 0 array [[False, False]], dtype=bool) >>> b = [ [ False, False ] ] >>> a [ b ] >>> aa[ bb ] array [[1, 1], [1, 1]])","bug"
|
|
"6013","2662","Arithmetic operations with objects that implement array don t make use of their rmul or r ) Trac 2070).
|
|
Original ticket http projects.scipy.org numpy ticket 2070 on 2012 03 03 by trac user phippo, assigned to unknown.
|
|
In order to not subclass `ndarray` or other classes that are based on `ndarray`, like `quantites.Quantity`, I prefere to work with ` array ` methods.
|
|
But there seems to be a bug when it comes to right sided arithmetic operatons.
|
|
I attached a file that demonstrates the problem.","bug"
|
|
"6013","2663","Python 2.4 compatibility broken Trac 2071).
|
|
Original ticket http projects.scipy.org numpy ticket 2071 on 2012 03 03 by rgommers, assigned to mwiebe.
|
|
`itertools.combinations` doesn t even exist in Python 2.5, so that should certainly be fixed.
|
|
====================================================================== ERROR test dtype.TestDtypeAttributeDeletion.test dtype non writable attributes deletion Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages nose 1.1.2 py2.4.egg nose case.py , line 197, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test dtype.py , line 412, in test dtype non writable attributes deletion assert raises AttributeError, delattr, dt, s) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy testing utils.py , line 1053, in assert raises return nose.tools.assert raises args, kwargs) File Library Frameworks Python.framework Versions 2.4 lib python2.4 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) TypeError attribute subdtype of numpy.dtype objects is not writable ====================================================================== ERROR test maskna.test array maskna view dtype Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages nose 1.1.2 py2.4.egg nose case.py , line 197, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test maskna.py , line 467, in test array maskna view dtype for x in itertools.combinations tcs, 2) AttributeError module object has no attribute combinations ====================================================================== ERROR test nditer.test iter writable attribute deletion Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages nose 1.1.2 py2.4.egg nose case.py , line 197, in runTest self.test self.arg) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test nditer.py , line 2564, in test iter writable attribute deletion assert raises AttributeError, delattr, it, s) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy testing utils.py , line 1053, in assert raises return nose.tools.assert raises args, kwargs) File Library Frameworks Python.framework Versions 2.4 lib python2.4 unittest.py , line 320, in failUnlessRaises callableObj args, kwargs) TypeError attribute value of numpy.nditer objects is not writable","bug"
|
|
"6013","2664","Label NA masked arrays and datetime as experimental Trac 2072).
|
|
Original ticket http projects.scipy.org numpy ticket 2072 on 2012 03 03 by rgommers, assigned to pv.
|
|
In order to signal to users that the implementation of masked arrays and datetime can still change, we had decided to label both as experimental.
|
|
This should be done in several obvious places in the documentation, and in the release notes.","bug"
|
|
"6013","2665","stride tricks with float16 Trac 2073).
|
|
Original ticket http projects.scipy.org numpy ticket 2073 on 2012 03 05 by trac user npinto, assigned to unknown.
|
|
It looks like stride tricks don t work with float16 python bug.py develop ) trying with float32 trying with float16 Traceback most recent call last) File bug.py , line 18, in <module> arr out = as strided arr in, shape=new shape, strides=new strides) File usr lib64 python2.6 site packages numpy lib stride tricks.py , line 28, in as strided return np.asarray DummyArray interface, base=x)) File usr lib64 python2.6 site packages numpy core numeric.py , line 235, in asarray return array a, dtype, copy=False, order=order) ValueError unsupported typestring","bug"
|
|
"6013","2666","random.shuffle fails if row in multidimensional array is None Trac 2074).
|
|
Original ticket http projects.scipy.org numpy ticket 2074 on 2012 03 05 by trac user tillahoffmann, assigned to unknown.
|
|
The random.shuffle fails if an array is multidimensional and one of the rows is None from numpy.random import shuffle a = [1,2,3,None] shuffle a) Works fine b = [ 1,1), 2,2), 3,3),None] shuffle b) Fails with the following error message Traceback most recent call last) File <stdin> , line 1, in <module> File mtrand.pyx , line 4283, in mtrand.RandomState.shuffle numpy random mtrand mtrand.c 18131) TypeError NoneType object is unsliceable","bug"
|
|
"6013","2667","Writing to a transpose results in strange error Trac 2075).
|
|
Original ticket http projects.scipy.org numpy ticket 2075 on 2012 03 07 by trac user lxop, assigned to unknown.
|
|
When I try to write to the transpose of an array, I get `AttributeError attribute T of numpy.ndarray objects is not writable` which seems fine, but the write has actually taken place.
|
|
In particular, this occurs with in place operations = etc).
|
|
Test code a = np.array [1,2,3]).reshape 1,3)) b = np.array [4,5,6]).reshape 3,1)) a.T = b The result is the above error is raised and the operation actually goes ahead anyway.
|
|
I would expect only one or the other.","bug"
|
|
"6013","2668","Bus error for F order ndarray creation on SPARC Trac 2076).
|
|
Original ticket http projects.scipy.org numpy ticket 2076 on 2012 03 08 by matthew brett, assigned to unknown.
|
|
This test, using numpy 1.7.0.dev 7c07089 import numpy as np from StringIO import StringIO as BytesIO from numpy.testing import assert array equal def test 2d buf ) dtt = np.complex64 arr = np.arange 10, dtype=dtt) 2D array arr2 = np.reshape arr, 2, 5)) Fortran write followed by C or F) read caused bus error data str = arr2.tostring F ) data back = np.ndarray arr2.shape, arr2.dtype, buffer=data str, order= F ) assert array equal arr2, data back) causes a bus error on Debian squeezy on SPARC64.
|
|
This is only so for complex types as `dtt`, and if the array is > 1D.
|
|
Debug log attached.","bug"
|
|
"6013","2669","Incorrect finfo from PPC longdouble Trac 2077).
|
|
Original ticket http projects.scipy.org numpy ticket 2077 on 2012 03 10 by matthew brett, assigned to unknown.
|
|
In [1] import numpy as np In [2] print np.finfo np.longdouble) Users mb312 dev trees numpy numpy core machar.py 125 RuntimeWarning overflow encountered in add a = a a Users mb312 dev trees numpy numpy core machar.py 127 RuntimeWarning invalid value encountered in subtract temp1 = temp a Users mb312 dev trees numpy numpy core machar.py 136 RuntimeWarning invalid value encountered in subtract itemp = int conv temp a) Users mb312 dev trees numpy numpy core machar.py 160 RuntimeWarning overflow encountered in add a = a a Users mb312 dev trees numpy numpy core machar.py 162 RuntimeWarning invalid value encountered in subtract temp1 = temp a Users mb312 dev trees numpy numpy core machar.py 169 RuntimeWarning invalid value encountered in subtract if any temp a != zero) Machine parameters for float128 precision= 75 resolution= 1e 75 machep= 4 eps= 1.3817869701e 76 negep = 4 epsneg= 1.3817869701e 76 minexp= 1 tiny= 1.08420217274e 19 maxexp= 1 max= 9.22337203471e 18 nexp = 1 min= max In [3] print np.finfo np.longdouble).nmant 1 In [4] np.
|
|
version Out[4] 1.7.0.dev aae5b0a I think the correct answers for nexp, nmant are 11, 106 respectively; see https developer.apple.com library mac documentation Darwin Reference Manpages man3 float.3.html np devel)[mb312 jerry ] uname a Darwin jerry.bic.berkeley.edu 8.11.0 Darwin Kernel Version 8.11.0 Wed Oct 10 18 26 00 PDT 2007; root xnu 792.24.17 1 RELEASE PPC Power Macintosh powerpc np devel)[mb312 jerry ] gcc v Using built in specs.
|
|
Target powerpc apple darwin8 Configured with var tmp gcc gcc 5370 2 src configure disable checking enable werror prefix= usr mandir= share man enable languages=c,objc,c ,obj c program transform name= [cg][ . ]
|
|
s 4.0 with gxx include dir= include c 4.0.0 with slibdir= usr lib build=powerpc apple darwin8 host=powerpc apple darwin8 target=powerpc apple darwin8 Thread model posix gcc version 4.0.1 Apple Computer, Inc. build 5370)","bug"
|
|
"6013","2670","min max mean of empty arrays Trac 2078).
|
|
Original ticket http projects.scipy.org numpy ticket 2078 on 2012 03 11 by trac user muellner, assigned to unknown.
|
|
[]In NumPy release 1.5.1, the minimum maximum mean of empty arrays is handled in a sensible way, namely by returning an empty array >>> numpy.min numpy.zeros 0,2)), axis=1) array [], dtype=float64) Likewise for max and mean.
|
|
This behavior is broken in NumPy 1.7.0.dev 56f66bb, and the line above raises a ValueError.
|
|
Is this a bug or a feature?
|
|
Personally, I prefer the old behavior, as it allows the easy handling of special cases zero sized arrays) without the need for extra checks.","bug"
|
|
"6013","2671","calling multiple times numpy.ctypeslib.as array pointer, shape) always uses the same shape Trac 2079).
|
|
Original ticket http projects.scipy.org numpy ticket 2079 on 2012 03 13 by trac user pieleric, assigned to unknown.
|
|
The latest version of numpy supports numpy.ctypeslib.as array ) for a ctypes.POINTER.
|
|
However, when using numpy.ctypeslib.as array ) on a pointer multiple times in a row with different shapes, it s always the shape of the first call which is used.
|
|
Example import ctypes import numpy >>> buffer = ctypes.pointer ctypes.c uint16 100) )) >>> p = ctypes.cast buffer, ctypes.POINTER ctypes.c uint16)) >>> a = numpy.ctypeslib.as array p, shape= 100,)) >>> a.shape 100,) >>> b = numpy.ctypeslib.as array p, shape= 10,10)) >>> b.shape 100,) We would expect b.shape == 10,10) .
|
|
The bug is in as array ) and prep pointer ).
|
|
They are too much a copy paste of the case of an ndarray only create the array interface if it is not yet there.
|
|
For an ndarray this works fine as the array interface is saved on the type.
|
|
For a pointer, it should be updated every time.
|
|
So, I propose something like this def prep pointer pointer obj, shape) contents = pointer obj.contents dtype = dtype type contents)) inter = { version 3, typestr dtype.str, data ct.addressof contents), False), shape shape} pointer obj.
|
|
array interface = inter def as array obj, shape=None) if hasattr obj, contents ) prep pointer obj, shape) else tp = type obj) try tp.
|
|
array interface except AttributeError prep array tp) return array obj, copy=False) BTW, the description of as array ) should be updated to mention the shape parameter instead of the size parameter.","bug"
|
|
"6013","2672","Issue with memory mapping large files Trac 2080).
|
|
Original ticket http projects.scipy.org numpy ticket 2080 on 2012 03 13 by trac user for usenet, assigned to unknown.
|
|
I ve been using Numpy to read, organize and process data files.
|
|
The size of these files are quite variable, but they have a fixed size header, after which, the data payload can range anywhere from a few MB to a few GB in size.
|
|
I ve been able to read and process smaller files just fine, but when I tried to read a 12 GB file, I get the following error File Library Python 2.6 site packages numpy 1.6.1 py2.6 macosx 10.6 universal.egg numpy core memmap.py , line 237, in new mm = mmap.mmap fid.fileno ), bytes, access=acc, offset=start) ValueError mmap length is greater than file size The error occurs for Numpy 1.6.1 on Mac OS X 10.6.8 64 bit with Python 2.6, and on openSUSE 11.4 64 bit with python 2.7.
|
|
The snippet of code that gives the above error is self.rawDataHeader.seek 0, 0) self.rawData4Recon = np.memmap self.rawDataHeader, mode = r , dtype = np.int16, offset = self.rawDataOffset ), shape = self.rawDataShape )) Please let me know if there s anything I can do to help track down or debug this error.","bug"
|
|
"6013","2673","Unicode string is truncated when python uses 2 byte unicode.
|
|
Trac 2081).
|
|
Original ticket http projects.scipy.org numpy ticket 2081 on 2012 03 14 by trac user taldcroft, assigned to unknown.
|
|
Using numpy devel with Python 3.2 gives an unexpected result trying to initialize an array from numpy string objects.
|
|
This is a regression from previous behavior devel prior to about two weeks ago and 1.6.1).
|
|
devel >>> d = np.array [ a ]) >>> np.asarray [d[0]]) array [ ], dtype= <U0 ) In numpy 1.6.1 >>> d = np.array [ a ]) >>> np.asarray [d[0]]) array [ a ], dtype= <U1 ) If one uses a longer string like abcd instead of a then the result is ab , so it looks like the actual output length is 2 from the desired length.
|
|
Perhaps this is related to https github.com numpy numpy commit 91f87e1f613630ff0ad9864017f059afcd6e57f1","bug"
|
|
"6013","2674","Updating the shrink keyword functionality in masked values Trac 2082).
|
|
Original ticket http projects.scipy.org numpy ticket 2082 on 2012 03 15 by trac user gsever, assigned to pierregm.
|
|
Considering this simple case x = np.array [1, 1.1, 2, 1.1, 3]) ma.masked values x, 1.5) masked array data = [ 1.
|
|
1.1 2.
|
|
1.1 3.
|
|
], mask = False, fill value=1.5) ma.masked values x, 1.5, shrink=False) masked array data = [ 1.
|
|
1.1 2.
|
|
1.1 3.
|
|
], mask = False, fill value=1.5) Setting the shrink keyword to False should create a mask with False filled array provided the condition is not met.
|
|
The improved shrink keyword handling should yield the following masked array.
|
|
ma.masked values x, 1.5, shrink=False) masked array data = [1.0 1.1 2.0 1.1 3.0], mask = [False False False False False], fill value = 1.5) The documentation about masked values is at http docs.scipy.org doc numpy reference generated numpy.ma.masked values.html","enhancement"
|
|
"6013","2675","numpy installation problem Trac 2083).
|
|
Original ticket http projects.scipy.org numpy ticket 2083 on 2012 03 16 by trac user joujisan, assigned to unknown.
|
|
Numpy version 1.6.1 OS Windows 7 x64 and python 3.2 Error Message Python 2.6 is required, which was not found in the registry Solution proposed at http tech.valgog.com 2010 01 after installing 64 bit windows 7 at.html The problem is the pywin32 is a 32 bit installer, it will look for the Python reg entries in the Wow6432Node, the following .reg file will add the entries, ensure the paths are correct for your install.
|
|
[HKEY LOCAL MACHINE\SOFTWARE\Wow6432Node\Python\PythonCore\3.2\InstallPath] = C \\Python32\\ [HKEY LOCAL MACHINE\SOFTWARE\Wow6432Node\Python\PythonCore\3.2\InstallPath\InstallGroup] = Python 3.2 [HKEY LOCAL MACHINE\SOFTWARE\Wow6432Node\Python\PythonCore\3.2\Modules] [HKEY LOCAL MACHINE\SOFTWARE\Wow6432Node\Python\PythonCore\3.2\PythonPath] = C \\Python32\\Lib;C \\Python32\\DLLs Is it correct?","bug"
|
|
"6013","2676","Issue with concatenating structured arrays segmentation fault) Trac 2084).
|
|
Original ticket http projects.scipy.org numpy ticket 2084 on 2012 03 19 by astrofrog, assigned to unknown.
|
|
The following demonstrates the issue even though Numpy knows how to concatenate the fields of a structured array, it crashes if one tries to concatenate the structured arrays themselves In [1] import numpy as np In [2] d1 = np.array zip [ a , b , c ]), dtype=[ b , |S1 )]) In [3] d2 = np.array zip [ aa , bb , cc ]), dtype=[ b , |S2 )]) In [4] np.hstack [d1[ b ],d2[ b ]]) Out[4] array [ a , b , c , aa , bb , cc ], dtype= |S2 ) In [5] np.hstack [d1, d2]) TypeError Traceback most recent call last) Users tom <ipython input 5 bd5cc420043d> in <module> ) > 1 np.hstack [d1, d2]) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core shape base.pyc in hstack tup) 268 269 > 270 return nx.concatenate map atleast 1d,tup),1) 271 TypeError invalid type promotion A similar issue occurs with floating point values of different endian ness In [1] import numpy as np In [2] d1 = np.array zip [1,2,3]), dtype=[ a , <f4 )]) In [3] d2 = np.array zip [1,2,3]), dtype=[ a , >f4 )]) In [4] np.hstack [d1[ a ],d2[ a ]]) Out[4] array [ 1., 2., 3., 1., 2., 3.
|
|
], dtype=float32) In [5] np.hstack [d1, d2]) TypeError Traceback most recent call last) Users tom <ipython input 5 bd5cc420043d> in <module> ) > 1 np.hstack [d1, d2]) Users tom Library Python 2.7 lib python site packages numpy core shape base.pyc in hstack tup) 271 As a special case, dimension 0 of 1 dimensional arrays is horizontal 272 if arrs[0].ndim == 1 > 273 return nx.concatenate arrs, 0) 274 else 275 return nx.concatenate arrs, 1) TypeError invalid type promotion In some cases, this can even cause segmentation faults, though I have yet to find a way to reproduce this consistently.","bug"
|
|
"6013","2677","Testsuite fails with Python 2.7.3rc1 and 3.2.3rc1 Debian) Trac 2085).
|
|
Original ticket http projects.scipy.org numpy ticket 2085 on 2012 03 19 by sandrotosi, assigned to unknown.
|
|
Hello, running numpy test suite with python versions 2.7.3rc1 and 3.2.3rc1 as available in Debian), returns this errors 2.7.3rc1 Running unit tests for numpy NumPy version 1.6.1 NumPy is installed in build buildd python numpy 1.6.1 6 i386 lYkcLV python numpy 1.6.1 debian tmp usr lib python2.7 dist packages numpy Python version 2.7.3rc1 default, Mar 10 2012, 00 01 06) [GCC 4.6.3] nose version 1.1.2 .........................Debug memory block at address p=0xaa13400 API m 8 bytes originally requested The 3 pad bytes at p 3 are FORBIDDENBYTE, as expected.
|
|
The 4 pad bytes at tail=0xaa13408 are FORBIDDENBYTE, as expected.
|
|
The block was made by call 1026630 to debug malloc realloc.
|
|
Data at p a3 03 00 00 00 00 00 00 Fatal Python error bad ID Allocated using API m , verified using API o 3.2.3rc1 Running unit tests for numpy NumPy version 1.6.1 NumPy is installed in build buildd python numpy 1.6.1 6 i386 lYkcLV python numpy 1.6.1 debian tmp usr lib python3 dist packages numpy Python version 3.2.3rc1 default, Mar 9 2012, 20 32 02) [GCC 4.6.3] nose version 1.1.2 usr lib python3 dist packages nose core.py 247 ResourceWarning unclosed file < io.TextIOWrapper name= usr lib python3 dist packages nose usage.txt mode= r encoding= ANSI X3.4 1968 > os.path.dirname file ), usage.txt ), r ).read ) .........................Debug memory block at address p=0xa625310 API m 8 bytes originally requested The 3 pad bytes at p 3 are FORBIDDENBYTE, as expected.
|
|
The 4 pad bytes at tail=0xa625318 are FORBIDDENBYTE, as expected.
|
|
The block was made by call 1475913 to debug malloc realloc.
|
|
Data at p a3 03 00 00 00 00 00 00 Fatal Python error bad ID Allocated using API m , verified using API o Aborted It might be something related to newer CPython versions, given testsuite executed on 2.6.7 works fine, but I m asking you first, just to be sure.
|
|
Regards, Sandro","bug"
|
|
"6013","2678","dtype constructor ignores byte order for datetime, timedelta Trac 2086).
|
|
Original ticket http projects.scipy.org numpy ticket 2086 on 2012 03 19 by charris, assigned to unknown.
|
|
On Intel little endian) In [1] dtype >M8[s] ) Out[1] dtype <M8[s] ) In [2] dtype >m8[s] ) Out[2] dtype <m8[s] )","bug"
|
|
"6013","2679","In place math functions like exp A, out=A) giving wrong results with integer arguments Trac 2087).
|
|
Original ticket http projects.scipy.org numpy ticket 2087 on 2012 03 20 by trac user chris, assigned to unknown.
|
|
Given an input matrix A with dtype int64 , a mathematical function like exp incorrectly keeps the type of A as an integer type producing a wrong result if used as an in place operation numpy.exp A, out=A) The other version utilising a new result matrix works correctly A = numpy.exp A) Tested on MacPort Mac OS X 10.6.8), Python 2.7.2 and Python 3.2.2.
|
|
Example A = numpy.array [[1, 0],[0,1]]) A.dtype >dtype int64 ) numpy.exp A) >array [[ 2.71828183, 1.
|
|
], > [ 1. , 2.71828183]]) correct result A = numpy.array [[1, 0],[0,1]]) numpy.exp A, out=A) >array [[2, 1], > [1, 2]]) wrong result A.dtype > dtype int64 ) A == numpy.array [[2, 1],[1,2]]) > array [[ True, True], > [ True, True]], dtype=bool)","bug"
|
|
"6013","2680","cov ) fails for complex input Trac 2088).
|
|
Original ticket http projects.scipy.org numpy ticket 2088 on 2012 03 21 by trac user shoyer, assigned to unknown.
|
|
The cov ) function in lib function base.py casts all input to float and hence fails on complex inputs.
|
|
Note that the code already includes the appropriate .conj ) to handle complex inputs properly [1], so it only fails on complex inputs because of the type casting.
|
|
This should be an easy fix.
|
|
[1] http en.wikipedia.org wiki Covariance matrix Complex random vectors","bug"
|
|
"6013","2681","Wrong method resolution for `multiply` other arithmetic ufuncs Trac 2089).
|
|
Original ticket http projects.scipy.org numpy ticket 2089 on 2012 03 22 by pv, assigned to unknown.
|
|
`multiply` and the other arithmetic ufuncs fall back to `PyNumber ` methods before trying `obj.<ufunc name>` methods.
|
|
This is problematic for sparse matrices and matrices import numpy as np import scipy.sparse as sp A = sp.rand 20,20,density=0.1) B = sp.rand 20,20,density=0.1) print np.multiply A,B) A.dot B)).todense ).max ) out 0.0 So, `multiply` actually calls the matrix multiply.
|
|
The reason seems to be that in `numpy core code generators generate umath.py` for these ufuncs the resolution ends with TD O, f= PyNumber Multiply ), which ends up calling ` mul `.
|
|
The fix could be to change this to TD O, f= NPy ElementwiseMultiply ), which would try to call `multiply` before falling back to `PyNumber `.","bug"
|
|
"6013","2682","`npy PyFile Dup` should check file object subclass Trac 2090).
|
|
Original ticket http projects.scipy.org numpy ticket 2090 on 2012 03 22 by pv, assigned to unknown.
|
|
Consider this on Python 3 import gzip import numpy as np f = gzip.open foo.gz , wb ) f.write b 1 2 3 ) f.close ) f = gzip.open foo.gz , rb ) print np.fromfile f, sep= )) > [ 1.]
|
|
On Python 2 this fails with an IOError.
|
|
The problem here is that `fromfile` seems to read the compressed stream.
|
|
This is probably a bug in Python 3 `PyObject AsFileDescriptor` when called on a `GzipFile` object does not fail although there is no OS level file handle corresponding to the uncompressed stream!).
|
|
As a workaround, `npy PyFile Dup` should probably check that the object passed in inherits from a suitable subclass in the io object hierarchy.","bug"
|
|
"6013","2683","Problem on OSX 10.7 with lapack lite functions and multiprocessing Trac 2091).
|
|
Original ticket http projects.scipy.org numpy ticket 2091 on 2012 03 26 by trac user dougal, assigned to pv.
|
|
The following code segfaults for me on OSX 10.7.3. from future import print function import numpy as np import multiprocessing as mp import scipy.linalg def f a) print about to call ) these all cause crashes sign, x = np.linalg.slogdet a) x = np.linalg.det a) x = np.linalg.inv a).sum ) these are all fine x = scipy.linalg.expm3 a).sum ) x = np.dot a, a.T).sum ) print result , x) return x def call proc a) print \ncalling with multiprocessing ) p = mp.Process target=f, args= a,)) p.start ) p.join ) if name == main import sys n = int sys.argv[1]) if len sys.argv) > 1 else 50 a = np.random.normal 0, 2, n, n)) f a) call proc a) This code causes a segfault trying to access e.g.
|
|
`0x0000000000000108`); when I do a core dump, the backtrace is 0 0x00007fff8832c324 in dispatch group async f ) 1 0x00007fff8b2fed3e in dgetrf ) 2 0x000000010ac0c26a in initlapack lite ) 3 0x000000010a5a8d77 in PyEval EvalFrameEx ) 4 0x000000010a5abdf7 in PyEval EvalCode ) 5 0x000000010a5a8e0a in PyEval EvalFrameEx ) 6 0x000000010a5abcd8 in PyEval EvalCodeEx ) 7 0x000000010a549abf in PyClassMethod New ) 8 0x000000010a528d32 in PyObject Call ) 9 0x000000010a5a95ec in PyEval EvalFrameEx ) 10 0x000000010a5abdf7 in PyEval EvalCode ) 11 0x000000010a5a8e0a in PyEval EvalFrameEx ) 12 0x000000010a5abdf7 in PyEval EvalCode ) 13 0x000000010a5a8e0a in PyEval EvalFrameEx ) 14 0x000000010a5abcd8 in PyEval EvalCodeEx ) 15 0x000000010a549abf in PyClassMethod New ) 16 0x000000010a528d32 in PyObject Call ) 17 0x000000010a5376e9 in PyInstance New ) 18 0x000000010a528d32 in PyObject Call ) 19 0x000000010a573484 in PyObject SlotCompare ) 20 0x000000010a56db7a in PyType Modified ) 21 0x000000010a528d32 in PyObject Call ) 22 0x000000010a5a8f63 in PyEval EvalFrameEx ) 23 0x000000010a5abdf7 in PyEval EvalCode ) 24 0x000000010a5a8e0a in PyEval EvalFrameEx ) 25 0x000000010a5abcd8 in PyEval EvalCodeEx ) 26 0x000000010a5abe6c in PyEval EvalCode ) 27 0x000000010a5a8e0a in PyEval EvalFrameEx ) 28 0x000000010a5abcd8 in PyEval EvalCodeEx ) 29 0x000000010a5abd4d in PyEval EvalCode ) 30 0x000000010a5c308f in Py CompileString ) 31 0x000000010a5c314f in PyRun FileExFlags ) 32 0x000000010a5c42a2 in PyRun SimpleFileExFlags ) 33 0x000000010a5d42af in Py Main ) 34 0x000000010a519e88 in ?? )
|
|
So it seems like the Apple Grand Central Dispatch stuff isn t playing nice with multiprocessing in this case.
|
|
If it s helpful, the full OSX problem report is [https gist.github.com 2209271 here].
|
|
This happens when `f )` calls any of the `numpy.linalg` functions there, but not for the matrix multiplication or the `scipy.linalg.expm3` call which doesn t call anything from lapack lite; if I use one of the other `scipy.linalg` functions that do call e.g.
|
|
`solve`, then it also segfaults).
|
|
If I comment out the call to `f` from the main process, it runs fine.
|
|
This happens for `n >= 33`; for `n <= 32`, it s fine.
|
|
Note that 32 32=1024, which seems like a reasonable cutoff point for when Accelerate would start parallelizing.
|
|
Of course, it doesn t matter if the original `f` call is on the same matrix it s being pickled, after all) as long as it s also big.
|
|
Calls with one big, one small matrix or two small matrices defining small as n x n for n <= 32 ) run fine.
|
|
It can also be a different function, e.g.
|
|
first `np.linalg.inv` and then call out to another process for `np.linalg.slogdet`.
|
|
This happens for me on my OSX 10.7.3 desktop with python 2.6.7, numpy 1.5.1 python 2.7.1, numpy 1.5.1, scipy 0.10.0 python 3.2.2, numpy 1.6.1, scipy 0.10.1 The 2.6 and 2.7 are I think the default system installs; I installed the 3.2 versions manually from the source tarballs.
|
|
All of those numpys are linked to the system Accelerate framework otool L `python3.2 c from numpy.core import dotblas; print dotblas.
|
|
file ) ` Library Frameworks Python.framework Versions 3.2 lib python3.2 site packages numpy core dotblas.so System Library Frameworks Accelerate.framework Versions A Accelerate compatibility version 1.0.0, current version 4.0.0) usr lib libSystem.B.dylib compatibility version 1.0.0, current version 125.2.1) I get the same behavior on another Mac with a similar setup.
|
|
But all of the options for `f` work on other machines running OSX 10.6.8 with Python 2.6.1 and numpy 1.2.1 linked to Accelerate 4 and vecLib 268 except that it doesn t have scipy or slogdet) Debian 6 with Python 3.2.2, numpy 1.6.1, and scipy 0.10.1 linked to the system ATLAS Ubuntu 11.04 with Python 2.7.1, numpy 1.5.1 and scipy 0.8.0 linked to system ATLAS So, I m not 100 sure this counts as a numpy bug, rather than one in multiprocessing or Accelerate; if you think I should just report it to one of them, let me know.
|
|
Any workarounds would also be appreciated, as this is currently making my life harder. )","bug"
|
|
"6013","2684","numpy.ma.average casts complex to float under certain conditions Trac 2092).
|
|
Original ticket http projects.scipy.org numpy ticket 2092 on 2012 03 27 by trac user wucknitz, assigned to pierregm.
|
|
Hi, when calling numpy.ma.average with a complex array, with weights and an argument for axis , the complex number is reduced to real.
|
|
See attached test script and the output here numpy version 1.3.0 with mask, no axis and with axis 2 3j) 2.0 without mask, no axis and with axis 2 2j) 2 2j) Version 1.6.1 still has the problem, even though under python 2.7) it at leasts complains about discarding an imaginary part.
|
|
It is easy to track the bug down in numpy ma extras.py, where the type is sometimes cast to float.
|
|
For me it does the trick to add the following lines use type complex if necessary if np.iscomplexobj a) type= complex else type= float And then use type instead of float in all lines with n = ... that now use float explicitly.
|
|
There may be better ways to fix the problem.
|
|
Cheers, Olaf","bug"
|
|
"6013","2685","Casting rule change since 1.6.1 Trac 2093).
|
|
Original ticket http projects.scipy.org numpy ticket 2093 on 2012 03 28 by matthew brett, assigned to unknown.
|
|
This, in numpy 1.6.1 In [2] np.zeros 2,), dtype=np.uint8) np.int16 257)).dtype Out[2] dtype int16 ) changed to this in current trunk In [2] np.zeros 2,), dtype=np.uint8) np.int16 257)).dtype Out[2] dtype uint16 ) which is different still in previous versions of numpy e.g.
|
|
1.4.1) In [2] np.zeros 2,), dtype=np.uint8) np.int16 257)).dtype Out[2] dtype uint8 ) see below for more versions) My impression had been that the plan was to avoid changes in the casting rules if possible.
|
|
I assume this change was not intentional?
|
|
Testing with this script import numpy as np Adata = np.zeros 2,), dtype=np.uint8) Bdata = np.zeros 2,), dtype=np.int16) Bzero = np.int16 0) Bbig = np.int16 256) print np.
|
|
version print Array add , Adata Bdata).dtype print Scalar 0 add , Adata Bzero).dtype print Scalar 256 add , Adata Bbig).dtype Gives the following output 1.4.1 Array add int16 Scalar 0 add uint8 Scalar 256 add uint8 1.5.1 Array add int16 Scalar 0 add uint8 Scalar 256 add uint8 1.6.1 Array add int16 Scalar 0 add uint8 Scalar 256 add int16 1.7.0.dev aae5b0a Array add int16 Scalar 0 add uint8 Scalar 256 add uint16","bug"
|
|
"6013","2686","np.array equal does not works for string arrays Trac 2095).
|
|
Original ticket http projects.scipy.org numpy ticket 2095 on 2012 03 30 by trac user shima shima, assigned to unknown.
|
|
if dtype of arrays are string, such as |S10 , np.array equal does not work.
|
|
This function is implemented as follows numpy core numeric.py, line 1966 return bool logical and.reduce equal a1,a2).ravel ))) However, np.equal is not implemented for strings.
|
|
This issue can be avoided by replacing this line by return bool logical and.reduce a1 == a2).ravel ))) The same issue occurs in array equiv at line 2008.","bug"
|
|
"6013","2687","ndarray.dot rejects out argument Trac 2096).
|
|
Original ticket http projects.scipy.org numpy ticket 2096 on 2012 04 02 by trac user JLull, assigned to unknown.
|
|
Testing with python 2.7.2, Pywin32 build 217, and numpy 1.6.1 win32 superpack python2.7.exe, any call to ndarray.dot ) with an out argument fails >>> import numpy >>> numpy.version.version 1.6.1 >>> a=numpy.array 1,2), 3,4))) >>> b=numpy.array 5,6)) >>> c=numpy.array 7,8)) >>> a.dot b) array [17, 39]) >>> a.dot b,c) Traceback most recent call last) File <interactive input> , line 1, in <module> TypeError function takes exactly 1 argument 2 given) >>> a.dot b,out=c) Traceback most recent call last) File <interactive input> , line 1, in <module> TypeError dot ) takes no keyword arguments Similar calls to numpy.dot ) succeed >>> numpy.dot a,b,c) array [17, 39]) >>> numpy.dot a,b,out=c) array [17, 39]) Similar calls to matrix.dot ) with an out argument also fail not shown)","bug"
|
|
"6013","2688","searchsorted sorter=) test error Trac 2097).
|
|
Original ticket http projects.scipy.org numpy ticket 2097 on 2012 04 04 by rgommers, assigned to charris.
|
|
The new sorter keyword for searchsorted added in [313fe460] http github.com numpy numpy commit 313fe460) has a problem ====================================================================== ERROR test searchsorted with sorter test multiarray.TestMethods) Traceback most recent call last) File Users rgommers Code numpy numpy core tests test multiarray.py , line 763, in test searchsorted with sorter assert equal b.searchsorted k), a.searchsorted k, sorter=s)) ValueError sorter elements must be non negative I checked that s doesn t contain non negative values; something goes wrong in the error checking.","bug"
|
|
"6013","2689","Documentation of numpy.histogram2d ) doesn t state x and y need the same size Trac 2098).
|
|
Original ticket http projects.scipy.org numpy ticket 2098 on 2012 04 05 by trac user Sven Marnach, assigned to pv.
|
|
The documentation describes the parameters as x array like, shape N,) y array like, shape M,) This gives the impression that `x` and `y` may have different shapes and left me confused as to what this function actually does.
|
|
If I pass two arrays of different sizes, I get the even more confusing error message ValueError setting an array element with a sequence.
|
|
The documentation should probably state that the values from `x` and `y` are paired, so in contrast to an 1D histogram, the order of the entries in `x` and `y` matter, at least relative to each other.","bug"
|
|
"6013","2690","Online NumPy documentation is missing parts.
|
|
Trac 2099).
|
|
Original ticket http projects.scipy.org numpy ticket 2099 on 2012 04 05 by trac user ewtoombs, assigned to unknown.
|
|
Documentation is missing in the current default version http docs.scipy.org doc numpy reference routines.array manipulation.html that is present in an older version http docs.scipy.org doc numpy 1.6.0 reference routines.array manipulation.html I don t know how that could possibly have happened, but this is pretty serious and it should be fixed.","bug"
|
|
"6013","2692","Documentation misleading example in `numpy.linalg.solve` .
|
|
The linalg.solve ) documentation gives a misleading way to check that its result is correct.
|
|
Quoting from the doc .
|
|
Check that the solution is correct .
|
|
Although this works with the above example, the test won t work in general due to rounding errors.
|
|
As this is certainly one of the most used functions in numerical computing, likely used by untrained users, it should be stated more correctly Check that the solution is correct .","bug"
|
|
"6013","2695","KeyError dev gitwash patching in sphinx doc build.
|
|
I tried building the numpy html docs locally for c5ccca92c5f136833ad85614feb2aa4f5bd8b7c3 which fails like this .
|
|
The numpy tests for that version run without error.
|
|
Here s the full output I get for `cd docs; make html` https gist.github.com 3937452 Is this a numpy issue or a problem with my sphinx version setup?","bug"
|
|
"6013","2700","BUG AttributeError incompatible shape for a non contiguous array.
|
|
The following is an error when running the matplotlib test suite.
|
|
The bug is traced back to this commit https github.com numpy numpy commit c48156dfdc408f0a1e59ef54ac490cccbd6b8d73 Traceback most recent call last) File home Patrick.Marsh .local lib python2.7 site packages nose 1.2.1 py2.7.egg nose case.py , line 197, in runTest self.test self.arg) File home Patrick.Marsh .local lib python2.7 site packages matplotlib testing decorators.py , line 39, in failer result = f args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib testing decorators.py , line 145, in do test figure.savefig actual fname) File home Patrick.Marsh .local lib python2.7 site packages matplotlib figure.py , line 1363, in savefig self.canvas.print figure args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib backend bases.py , line 2118, in print figure kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib backend bases.py , line 1894, in print svg return svg.print svg args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib backends backend svg.py , line 1103, in print svg return self.
|
|
print svg filename, svgwriter, fh to close, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib backends backend svg.py , line 1139, in print svg self.figure.draw renderer) File home Patrick.Marsh .local lib python2.7 site packages matplotlib artist.py , line 54, in draw wrapper draw artist, renderer, args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib figure.py , line 999, in draw func args) File home Patrick.Marsh .local lib python2.7 site packages matplotlib artist.py , line 54, in draw wrapper draw artist, renderer, args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib axes.py , line 2086, in draw a.draw renderer) File home Patrick.Marsh .local lib python2.7 site packages matplotlib artist.py , line 54, in draw wrapper draw artist, renderer, args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib collections.py , line 699, in draw return Collection.draw self, renderer) File home Patrick.Marsh .local lib python2.7 site packages matplotlib artist.py , line 54, in draw wrapper draw artist, renderer, args, kwargs) File home Patrick.Marsh .local lib python2.7 site packages matplotlib collections.py , line 248, in draw transform, transOffset, offsets, paths = self.
|
|
prepare points ) File home Patrick.Marsh .local lib python2.7 site packages matplotlib collections.py , line 222, in prepare points offsets.shape = 1, 2) Make it Nx2 AttributeError incompatible shape for a non contiguous array","bug"
|
|
"6013","2704","Boolean indexing edge case.
|
|
This looks like a bug to me .
|
|
NumPy 1.6.1, have not checked git master yet)","bug"
|
|
"6013","2705","wrong output of inplace operators x = y, x = y, ... for nonstandard strides.
|
|
The following script works for numpy 1.4.1 but not for numpy > 1.6.
|
|
See also the discussion on http thread.gmane.org gmane.comp.python.numeric.general 51379 .
|
|
NumPy 1.4.1 output .
|
|
NumPy 1.8.0.dev 9597b1f output .","bug"
|
|
"6013","2711","DOC create documentation guide.
|
|
On http docs.scipy.org doc there s a link Write, review and proof the documentation to http docs.scipy.org numpy .
|
|
I suggest to use the file in the doc root https github.com numpy numpy tree master doc to create a nice and RST based documentation guideline.","bug"
|
|
"6013","2713","Indexing crash.
|
|
Following snippet import numpy as np a = np.array [[0.]])
|
|
a[np.array [[True]]).squeeze )] is crashing Python for me.
|
|
System win32, numpy 1.6.2 mkl build from Christoph.","bug"
|
|
"6013","2714","FloatingPointError with logaddexp [ inf], [ inf]).
|
|
Get this with NumPy 1.6.2 32 bit) on Windows 7 .
|
|
The correct result .
|
|
is returned, but I think this is a common enough use case that .
|
|
should account for it without raising any warnings exceptions.","bug"
|
|
"6013","2715","npyio.loadtxt is bytes casting text file input, even with str dtype specified..
|
|
Environment Python Version 3.3 python.org release) on OS X Mountain Lion numpy Cloned from git master When calling `numpy.loadtxt` on file containing strings as follows .
|
|
Where NYSE dates.txt is simply a list of dates could be anything really) .
|
|
Output is .
|
|
As you can see, all the strings have been bytes casted, then stringified through `conv`, as you would get the same result from `str str 12 31 2020 ).encode latin1 ))`, per `conv` `compat.asbytes`.
|
|
After looking at the code, it appears that all strings are bytes casted with `asbytes ...)` pretty much throughout, as for example in `split line ...)`, so this must mean every routine in the module is broken.","bug"
|
|
"6013","2718","Unexpected linalg.matrix rank behavior.
|
|
I had a couple of unexpected results using matrix rank.
|
|
I m on 1.6.1.
|
|
One was that removing a row from a matrix increased the rank, another was that adding a row to a matrix increased the rank by 2.
|
|
Also, these are not really reproducible and seem to be related to slicing.
|
|
I get something like .
|
|
I m not sure where to go from here.
|
|
The matrix in this case has shape 80, 92) and three non zero entries per row.
|
|
Is was that a known issue that has been fixed in later version?
|
|
Should I try master?
|
|
Any help would be appreciated.
|
|
Thanks, Andy","bug"
|
|
"6013","2719","Can t build on Mac OS X 10.7.5 with Xcode 4.5.2.
|
|
I m following http www.scipy.org Installing SciPy Mac OS X I installed http r.research.att.com tools gcc 42 5666.3 darwin11.pkg as specified then ran export CC=clang export CXX=clang python setup.py build I get two obvious issue.
|
|
First there s the don t know how to compile Fortran code message.
|
|
Second, there s the error failing to find usr bin .. llvm gcc 4.2 bin powerpc apple darwin11 llvm gcc 4.2 which is weird because I m specifying clang, not llvm gcc, and because I don t want a ppc build.
|
|
Running from numpy source directory.
|
|
non existing path in numpy distutils site.cfg F2PY Version 2 numpy core setup common.py 86 MismatchCAPIWarning API mismatch detected, the C API version numbers have to be updated.
|
|
Current C api version is 8, with checksum f4362353e2d72f889fda0128aa015037, but recorded checksum for C API version 8 in codegen dir cversions.txt is 17321775fc884de0b1eda478cd61c74b.
|
|
If functions were added in the C API, you have to update C API VERSION in numpy core setup common.pyc.
|
|
MismatchCAPIWarning) blas opt info FOUND extra link args = [ Wl, framework , Wl,Accelerate ] define macros = [ NO ATLAS INFO , 3)] extra compile args = [ msse3 , I System Library Frameworks vecLib.framework Headers ] lapack opt info FOUND extra link args = [ Wl, framework , Wl,Accelerate ] define macros = [ NO ATLAS INFO , 3)] extra compile args = [ msse3 ] running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building py modules sources creating build creating build src.macosx 10.3 fat 2.7 creating build src.macosx 10.3 fat 2.7 numpy creating build src.macosx 10.3 fat 2.7 numpy distutils building library npymath sources customize Gnu95FCompiler Could not locate executable gfortran Could not locate executable f95 customize NAGFCompiler customize AbsoftFCompiler Could not locate executable f90 Could not locate executable f77 customize IBMFCompiler Could not locate executable xlf90 Could not locate executable xlf customize IntelFCompiler Could not locate executable ifort Could not locate executable ifc customize GnuFCompiler Could not locate executable g77 customize G95FCompiler Could not locate executable g95 customize PGroupFCompiler Could not locate executable pgfortran don t know how to compile Fortran code on platform posix C compiler clang fno strict aliasing fno common dynamic arch ppc arch i386 g O2 DNDEBUG g O3 compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I Library Frameworks Python.framework Versions 2.7 include python2.7 c clang configtest.c llvm gcc 4.2 error trying to exec usr bin .. llvm gcc 4.2 bin powerpc apple darwin11 llvm gcc 4.2 execvp No such file or directory clang error assembler via gcc) command failed with exit code 255 use v to see invocation) llvm gcc 4.2 error trying to exec usr bin .. llvm gcc 4.2 bin powerpc apple darwin11 llvm gcc 4.2 execvp No such file or directory clang error assembler via gcc) command failed with exit code 255 use v to see invocation) failure.
|
|
removing configtest.c configtest.o Traceback most recent call last) File setup.py , line 214, in <module> setup package ) File setup.py , line 207, in setup package configuration=configuration ) File Users jwatt src numpy numpy distutils core.py , line 186, in setup return old setup new attr) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File Users jwatt src numpy numpy distutils command build.py , line 37, in run old build.run self) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils command build.py , line 127, in run self.run command cmd name) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File Library Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File Users jwatt src numpy numpy distutils command build src.py , line 152, in run self.build sources ) File Users jwatt src numpy numpy distutils command build src.py , line 163, in build sources self.build library sources libname info) File Users jwatt src numpy numpy distutils command build src.py , line 298, in build library sources sources = self.generate sources sources, lib name, build info)) File Users jwatt src numpy numpy distutils command build src.py , line 385, in generate sources source = func extension, build dir) File numpy core setup.py , line 648, in get mathlib info raise RuntimeError Broken toolchain cannot link a simple C program ) RuntimeError Broken toolchain cannot link a simple C program","bug"
|
|
"6013","2720","Installer doesn t work on Mac OS X 10.7.5.
|
|
I installed python v2.7.3 from python.org on my Mac and then tried to install the latest NumPy from http sourceforge.net projects numpy files NumPy Unfortunately the installer refuses to let me proceed, telling me numpy 1.6.2 can t be installed on this disk.
|
|
numpy requires python.org Python 2.6 to install.
|
|
The default install location for python.org python is usr local bin python, so initially I thought that s the problem, but if I symlink usr bin python to usr local bin python I still get the same error message.
|
|
Is the installer really hardcoded to only accept 2.6 as an install target?
|
|
If so, can it be fixed to accept later versions too?","bug"
|
|
"6013","2724","random choice should support multidimensional arrays.
|
|
random.choice enhancement request random.choice should support multidimensional arrays by taking an axis number.
|
|
With this enhancement, axis=None would choose from a flattened array, while an integer argument would chose from the subarrays along that axis.
|
|
Apologies if this is a duplicate.
|
|
My previous submission seems to have vanished.","enhancement"
|
|
"6013","2728","Impossible to use ACML with system installed ATLAS.
|
|
As far as I can tell it s impossible to configure `site.cfg` to link to ACML CBLAS) with a system installed ATLAS, using maintenance 1.6.x.
|
|
I haven t tried with newer versions).
|
|
For example, `site.cfg` would have me believe that I should be able to link to ACML with .
|
|
But running `python setup.py build` instead links to the system installed ATLAS .
|
|
This is easily seen as well by running `get info blas opt )` .
|
|
How can I convince the setup process to not ignore my site.cfg file?","bug"
|
|
"6013","2729","1.7.x invalid value encountered in absolute.
|
|
In Python 2.7 I am getting .
|
|
and in python 3.2 I am also getting `ResourceWarning unclosed file` besides the above) .","bug"
|
|
"6013","2731","1.7.x in Python 3.2 debug PyArray DTypeFromObjectHelper Assertion ... failed.
|
|
.","bug"
|
|
"6013","2732","counterintuitive argument handling in cov and corrcoef.
|
|
The behavior of cov and corrcoef seems somewhat counterintuitive.
|
|
As far as I can tell, if the user passes an array like y in in addition to x), x and y are simply concatenated before computing the covariance or correlation matrix.
|
|
It s not clear what the purpose of this is, since the user could presumably have just concatenated the two arrays before calling cov ).
|
|
Moreover, it seems confusing given the behavior of standard correlation routines in R and Matlab, where the expected result if you pass in two 2D arrays x and y is an M x N matrix where M and N are the number of columns in x and y, respectively).
|
|
Situations where one needs to correlate two matrices column wise or a matrix and a vector) are quite common; perhaps it would make sense to follow the same convention in numpy?","enhancement"
|
|
"6013","2738","Python 3.2 Segmentation fault in numpy testing print coercion tables.py.
|
|
.
|
|
The problem is here .
|
|
Full stacktrace .","bug"
|
|
"6013","2743","Support Python 3.3.
|
|
From Peter Cock Having looked at the README.txt and INSTALL.txt files on the branch, I see no mention of which Python 3.x versions are supported https github.com numpy numpy blob maintenance 1.7.x README.txt https github.com numpy numpy blob maintenance 1.7.x INSTALL.txt Is NumPy 1.7 intended to support Python 3.3?
|
|
My impression from this thread is probably yes http mail.scipy.org pipermail numpy discussion 2012 July 063483.html ... http mail.scipy.org pipermail numpy discussion 2012 August 063597.html If so, then under Windows 32 bit at least) where the Python.org provided Python 3.3 is compiled with MSCV 2010 there are some problems see https github.com numpy numpy pull 2726","bug"
|
|
"6013","2750","test unicode fails on Python 3.3.
|
|
Test unicode fails with numpy 1.7.0rc1.dev win py3.3 3a52aa06c5), built with Visual Studio 2010 and Intel MKL 11.
|
|
All other tests pass.
|
|
.","bug"
|
|
"6013","2752","set printoptions with threshold as nan bug.
|
|
According to the numpy docs, if you want to disable summary printing for ndarray, you call `set printoptions threshold= nan )`, but this results in the following error when doing the print .
|
|
Line 240 of arrayprint.py should be changed to .","bug"
|
|
"6013","2753","Eulerxe2x80x93Mascheroni constant should be accessable.
|
|
Being already present in the numpy core libraries as NPY EULER, this constant is not so easily accessible as pi or e, as one can not access it directly from vanilla numpy, but presumably) from custom C extensions only.","bug"
|
|
"6013","2755","in1d needs an asarray call?.
|
|
Consider this .","bug"
|
|
"6013","2756","array reshape crashes when size in one dimension is greater than 2 31.
|
|
I am having a problem with reshape crashing > python Python 2.6.4 r264 75706, Jan 16 2010, 21 11 47) [GCC 4.3.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.version.version 1.6.2 >>> npData = numpy.ones [701,701,7899],dtype=numpy.dtype b )) >>> npDataSubset = npData[[slice 0,700),slice 0,700),slice 0,5000)]] >>> npDataOutput = npDataSubset.reshape [700 700 5000],order= F ) Segmentation fault If I change the 5000 to a 4000 , everything is fine.
|
|
I m not running out of memory my system had 48 GB of memory and nothing else is using a significant portion of this memory.
|
|
Note 700x700x4000 = 1,960,000,000 < 2 31 and 700x700x5000 = 2450000000 > 2 31.
|
|
I suspect somewhere in the underlying code there is a signed 32 bit integer being used for an index pointer offset this is running on a 64 bit machine).
|
|
I did some searching of the archives and didn t find a match for this problem and posted to the numpy discussion group.
|
|
Francesc Alted suggested I submit a ticket.
|
|
Thank you for any and all help!","bug"
|
|
"6013","2757","numpy ma tests test core.py Heisenbug at line 2731.
|
|
This is the odd invalid value encountered in power bug at line 2731 of test core.py that shows up in the Travis tests now and then.
|
|
It can be made reproducible by modifying test varstd specialcases as follows.
|
|
def test varstd specialcases self) Test a special case for var nout = np.ones 1, dtype=float) mout = ones 1, dtype=float) This is either a bug in the test, or a bug in the ma var std method code.","bug"
|
|
"6013","2759","ndarray var method doesn t cope gracefully with empty arrays..
|
|
In [3] arange 0).var ) home charris .local lib python2.7 site packages numpy core methods.py 72 RuntimeWarning invalid value encountered in true divide out=arrmean, casting= unsafe , subok=False) home charris .local lib python2.7 site packages numpy core methods.py 96 RuntimeWarning invalid value encountered in double scalars ret = ret float rcount) Out[3] nan In [4] arange 0).std ) Out[4] nan I also wonder if nan is the proper return for this case.","bug"
|
|
"6013","2761","build breaks with new virtualenv 1.8.3 on py3.
|
|
.
|
|
I ve opened a coressponding issue in venv pypa virtualenv 359.
|
|
If the numpy team could lend a hand in debugging this, that d be great.","bug"
|
|
"6013","2762","f2py python2 3 exec ) incompatibility.
|
|
.","bug"
|
|
"6013","2763","f2py with associate < > end associate .
|
|
f2py break with functions subroutines containing associate construct.
|
|
my hack .","bug"
|
|
"6013","2764","np.random.choice without replacement or weights is less efficient than random.sample?.
|
|
I was interested in choosing k elements out of n without replacement or weighted probabilities, and with k<<n .
|
|
I ve tested using the following ipython notebook [source] https gist.github.com 4134617) [view] http nbviewer.ipython.org 4134617 ).
|
|
The outcome was that using `random.sample xrange n),k)` was 50 fold faster than using `np.random.permutation n)[ k]`, which is the way it is [currently coded] https github.com numpy numpy blob master numpy random mtrand mtrand.pyx).","enhancement"
|
|
"6013","2768","Travisbot is running extraneous test files, causing irrelevant failures..
|
|
There is a 1.6.2 test module, not present in 1.7.x, that is left hanging about and raising errors.
|
|
njsmith Is there an easy way to ensure that the environment starts up without an installed version of numpy?","bug"
|
|
"6013","2770","A C contiguous array can now have strides[1] == 0.
|
|
Since https github.com numpy numpy commit c48156dfdc408f0a1e59ef54ac490cccbd6b8d73 a C contiguous array can now have strides[1] == 0.
|
|
This is a change in behavior since numpy 1.7 and it induces a bunch of bugs in scikit learn.
|
|
The question is should this be considered as a bug, or should the scikit learn adapt to this change in behavior.","bug"
|
|
"6013","2771","Missing argument to internal makenames list in PyArray FieldNames.
|
|
In numpy core internal.py, a function is defined as def makenames list adict, align) However, in numpy core src multiarray scalarapi.c, line 504, it is called by tup = PyObject CallMethod numpy internal, makenames list , O , fields); The inputs mismatch which makes the function PyArray FieldNames unusable.","bug"
|
|
"6013","2775","Crash for some 0 d output arrays in dot.
|
|
There seems to be some problem with 0 sized output in dot.
|
|
However I do not get this on numpy 1.6.2. so I am not sure what is going on.
|
|
.
|
|
The error seems a bit random, but my guess is that maybe invalid memory is being written to, so maybe that is to be expected.
|
|
The reason I found this, was that I tried changing the strides of 0 sized arrays such that they are zero after the zero dimension), which however also changed memory allocation for 0 sized arrays, as in this case the full size array would be allocated for `z` in current numpy if `z` was not given, which I would consider a bug, even if the memory should never be used.
|
|
This is probably the reason why the error does not occur if the output is not given.","bug"
|
|
"6013","2776","ndarray should derive from collections.abc.Sequence?.
|
|
juliantaylor raised this in a pandas issue.
|
|
The example from the ticket .
|
|
This occurs on 3.3 with 1.7.0rc1.dev 3a52aa0, and on 3.2 with 1.6.2.
|
|
2.7 is unaffected of course.
|
|
The relavent code from cpython Lib random.py 297 .
|
|
I couldn t grep another location in the stdlib with a similar test, but lib2to3 did show an assumed equivalence .
|
|
in 2.7 .
|
|
but in 3.3 3.2 .","bug"
|
|
"6013","2778","numpydoc nostesting modules for adherance to numpydoc docstring standards?.
|
|
I ve always thought it would be cool to automatically to automatically validate at least in a minimal way) my docstrings during build testing.
|
|
Because of issues like [this] http www.voidspace.org.uk python weblog arch d7 2009 05 16.shtml e1090), its not going to be perfect, but whatever.)
|
|
Well, I went ahead and did it.
|
|
I m wondering if there s any interest in the numpy community for adding this feature into the numpydoc codebase.
|
|
With code like this in the test suite .
|
|
you get behavior like .
|
|
The mechanism is just a class factory that builds classes with names like Test XXX so that they get discovered by nose.
|
|
Perhaps it could be done more cleverly with a nose extension.
|
|
If there s any interest in this Perhaps it s outside the scope of numpy numpydoc) I will write it up for a PR.","bug"
|
|
"6013","2781","Is change in behavior of ``ndindex`` intentional?.
|
|
It looks like the behavior for ``ndindex`` has changed since 1.6.2.
|
|
For example, the following used to work .
|
|
but now they fail .
|
|
so I just wanted to check that these changes were intentional?
|
|
I can easily update my code if needed).","bug"
|
|
"6013","2785","np.unique TypeError requested sort not available for type.
|
|
Upon recently upgrading from 1.6.1 to 1.6.2, I discovered that the well intentioned 74b9f5ee broke my usage of `np.unqiue` .
|
|
Can I work around this in some way, without creating my own copy of `np.unique`?
|
|
Also why can t mergesort work for this type?","bug"
|
|
"6013","2790","Polyfit fails for some complex values..
|
|
The following will fail with the current numpy .
|
|
The problem is [line 557 of lib polynomial.py] https github.com numpy numpy blob af22fc43921e2e7f77e909a1acf77011f682ff7b numpy lib polynomial.py L577) .
|
|
The two terms create a zero scale factor here, making `lhs` infinite or nan, sending `lstsq` into never never land.
|
|
There is probably another bug in `lstsq` in that it should catch the invalid input and raise an error before charging ahead and consuming all the computer resources!)
|
|
It should probably be something like this or maybe using `norm`) .
|
|
With this modification, `polyfit` should work nicely with complex valued polynomials assuming analyticity).
|
|
I will play with this for a bit to see if there are any other problems, then submit a PR.","bug"
|
|
"6013","2799","Make unique behave like sort when the argument is a list.
|
|
The sort and unique functions do not handle lists in a consistent manner.
|
|
.
|
|
but .","bug"
|
|
"6013","2801","Deprecate NPY CHAR.
|
|
This broke ABI compatibility in 1.6.x.
|
|
Should be deprecated and removed, see http projects.scipy.org numpy ticket 2228","bug"
|
|
"6013","2802","Raise a warning when assigning to a temporary fix described, patch needed).
|
|
It would be nice if numpy raised a warning when people assigned into temporary arrays, since it s rather easy to end up with these accidentally and not realize what s going on, so people keep barking their shins on this.
|
|
The original report below is one example; another would be .
|
|
which will usually modify `a`... but sometimes, depending on `a` s memory layout, it may be a no op instead.
|
|
Quite sneaky.
|
|
The way to do this would be to write a function like `PyArray DefinitelyTemporary arr)` that returned true iff all entries in `arr` s `.base` chain fulfilled the properties is an ndarray with reference count exactly 1 Then in the Python entry points for numpy assignments `np.copyto`, ` setitem `, ufunc `out=` parameters, etc.
|
|
have some code like .
|
|
NB I say Python entry points above, because any numpy C API entry points can legitimately take borrowed references to non temporary arrays that will have reference count 1.
|
|
So we need to be careful that these checks are done before we hit code that could be reached by C API users.)
|
|
Original report I know that I should be using slice notation instead of ``np.copyto``, but I also feel like case 1 should throw an error.
|
|
.","bug"
|
|
"6013","2803","triu does not return boolean array when the input is boolean.
|
|
When tril is given a boolean array as an input it returns a boolean array as it supposed to.
|
|
However, triu returns an int array.
|
|
.","bug"
|
|
"6013","2805","Make np.lib.recfunctions more visible.
|
|
Currently a number of very useful functions live in `np.lib.recfunctions`.
|
|
The reason for this is apparently historical, as these routines originally came from matplotlib.
|
|
However, not only is this not an entirely obvious place for these functions, this module doesn t even appear in the documentation.
|
|
For this reason, one can find numerous questions on Stack Overflow and mailing lists looking for basic functionality that really ought to be easily found.
|
|
This has come up a [few] http mail.scipy.org pipermail numpy discussion 2011 July 057398.html) [times] http mail.scipy.org pipermail numpy discussion 2010 February 048438.html) in the past but I would argue that the current state of things can still use improvement.
|
|
If I understand correctly the last discussion ended with the conclusion that the `recfunctions` namespace should be imported into the root `numpy` namespace, although this apparently never happened.
|
|
At very least `np.lib.recfunctions` should be included in the documentation.
|
|
Additionally, these routines should really be in a more visible point in the namespace although as pointed out [previously] http mail.scipy.org pipermail numpy discussion 2011 July 057465.html) this isn t entirely trivial due to circular import issues).","enhancement"
|
|
"6013","2806","save load and tofile fromfile fail silently for large arrays on Mac os X.
|
|
It seems that saving large numpy arrays to disk using numpy.save fails silently for very large matrices, example .
|
|
Note zeros in the end of the matrix that do not exist in a I m running numpy version 1.6.2 .
|
|
on python 2.7.3 on Mac os X 10.8.2 if that helps","bug"
|
|
"6013","2807","Docstring for numpy.load ) incorrectly describes file like object requirements.. numpy.load ) says that the `file` parameter must support seek ) and read ).
|
|
However, in the event that pickle.load ) is invoked on the file i.e.
|
|
via the cload ) function), the file must also support readline ).
|
|
According to the pickle documentation and based on my actual experience), the pickle.load ) requirements on the file like object are The file like object must have two methods, a read ) method that takes an integer argument, and a readline ) method that requires no arguments.
|
|
So, it seems that the numpy.load ) docstring should be updated to indicate this somehow.","bug"
|
|
"6013","2810","Deprecate the use of non integer non boolean indexes for arrays.
|
|
Currently, we have .
|
|
Silently truncating floating point numbers like this is both risky and inconsistent with how we handle indexes which are arrays.
|
|
Step 1 as soon as possible) make the indexing code issue a deprecation warning whenever we cast a non integer non boolean scalar.
|
|
Step 2 a relase or two later) replace this warning with an error.","enhancement"
|
|
"6013","2811","When indexing numpy arrays, treat lists and array input the same way.
|
|
Right now, if you use a Python list object to index an ndarray, then it is silently hard cast to integer see also gh 2810) .
|
|
The behaviour for floating point input is confusing, dangerous, and inconsistent with how ndarray input is handled `a[np.array [1.5, 2.5])]` is an error.
|
|
The behaviour for boolean input is even more confusing and dangerous.
|
|
Srsly wtf?
|
|
Obviously this gives entirely different results than `a[np.array [False, False, False])]`.
|
|
We want to make it so that if `L` is a list object, then `a[L]` and `a[np.array L)]` produce the same result.
|
|
But, the boolean handling makes this a bit trickier to fix than gh 2810, though, because there we re just adding an error; here we re converting one non error into a different non error, so there s more danger of silently breaking user code.
|
|
Plan Step 1 as soon as possible) when we get a list as input to indexing, run the dtype auto detection code, and if it returns a non integer dtype, then produce a FutureWarning not a DeprecationWarning).
|
|
Step 2 a release or two later) when we get a list as input to indexing, just convert it to an ndarray using the equivalent of `np.asarray`, and then use the ndarray input path.
|
|
EXCEPT, if the resulting ndarray has a boolean type, produce an error Step 3 a release or two later) Remove the error on boolean lists, so that they too act like ndarray input.","bug"
|
|
"6013","2813","Deprecate non integer axes.
|
|
I remembered this because of the discussion about deprecating non inter non ` index `) indices.
|
|
I think similar should be done for the axes argument to numpy functions.
|
|
Right now most all of the C backed functions silently allow a non integer axis argument not as bad as indexing that even allows strings...), while most python functions at some point do something like `arr.shape[axes]` that fails.
|
|
Its a smaller thing then the indexing, but I think it should be handled just the same as indexing with scalars.","enhancement"
|
|
"6013","2814","numpy.ma.mean over an axis doesn t work.
|
|
Simple example with one missing value .
|
|
It gives the following error and traceback .
|
|
If I understand the code in [numpy ma core.py] https github.com numpy numpy blob maintenance 1.6.x numpy ma core.py) correctly, the problem is that the variable `result` from [line 1069] https github.com numpy numpy blob maintenance 1.6.x numpy ma core.py L1069) has `dtype=object`, and subsequently the call to `umath.isfinite )` fails because it doesn t support this type [line 1073] https github.com numpy numpy blob maintenance 1.6.x numpy ma core.py L1073).
|
|
I m using Fedora 17 x86 64) with numpy 1.6.2 1.fc17.x86 64 python 2.7.3 7.2.fc17.x86 64","bug"
|
|
"6013","2824","overlapping src dest memory corruption with 1.6.2 on certain libc.
|
|
closed the 2819 PR and moved this to an issue.
|
|
324 or equiv solution needs beckporting to 1.6.2","bug"
|
|
"6013","2826","Fix the RandomState bug.
|
|
This was introduced in 2727.","bug"
|
|
"6013","2830","Drop support for Python 2.4, and possibly Python 2.5.
|
|
As discussed here http mail.scipy.org pipermail numpy discussion 2012 December 064714.html And here http mail.scipy.org pipermail numpy discussion 2012 December 064725.html Todo Document this in the release notes.
|
|
grep for PY VERSION HEX checks in the source and clean up no longer necessary workarounds ...?","bug"
|
|
"6013","2832","Get rid of spurious lines in Travis tests output.
|
|
Now the test results at Travis contain the lines .
|
|
For example .
|
|
This seems to be caused by 2812.","bug"
|
|
"6013","2843","Segfault ufunc object.c 1750.
|
|
Hi repeatedly getting a segfault in umath.
|
|
Its an internal test tool that performs some analysis on PCM voice streams using scipy.
|
|
I don t really know how to debug the issue but here is the information I collected from the last segfault.
|
|
I can provide more info if necessary, just need to be pointed to some instructions.
|
|
Cheers, Hank OS Linux localhost.localhost 2.6.18 308.16.1.el5 1 SMP Tue Oct 2 22 01 37 EDT 2012 i686 i686 i386 GNU Linux CentOS release 5.8 Final) Python Python2.7 Numpy NAME numpy.version THIS FILE IS GENERATED FROM NUMPY SETUP.PY FILE usr local lib python2.7 site packages numpy version.py DATA full version = 1.6.2 git revision = ca07bce202ae26b6f0a73870eb2ef0b88e0210c5 release = True short version = 1.6.2 version = 1.6.2 SciPy Help on module scipy.version in scipy NAME scipy.version THIS FILE IS GENERATED FROM SCIPY SETUP.PY FILE usr local lib python2.7 site packages scipy version.py DATA full version = 0.11.0 git revision = 827679b6f0149abb2c0e121749887012946b99c3 release = True short version = 0.11.0 version = 0.11.0 Matplotlib version = 1.2.0 version numpy = 1.4 .","bug"
|
|
"6013","2845","histogramdd fails on data points that are close to rightmost bin edge.
|
|
My numpy version is 1.6.1 6 on ubuntu 12.04 64 bit with python 2.7.3.
|
|
I want to create histograms with the function histogramdd.
|
|
When I have data points that are very close to the rightmost bin edge, they will land in the wrong bin, that is, in the second to right bin, instead of the rightmost bin!
|
|
Here is some code to reproduce this behavior .
|
|
The last bin is empty and all data points with value 2 are in the second bin.
|
|
Note that the last bin edge is very close to 2 but not exactly 2!
|
|
This doesn t happen using the 1d function .
|
|
The problem seems to be in line 347 of function base.py in my version) .
|
|
...together with the following code line 356ff), where all the data points that fall on the rightmost edge are shifted one bin to the left .
|
|
What happens is that the data points that are 2 fall into the 3rd bin, because they are smaller then the last edge 2.0000001.
|
|
The digitize function would put all points that were exactly 2.0000001 into an outlier bin!
|
|
Then, in order to correct for the edge points that were classified as outliers, these are shifted one bin to the left.
|
|
However, to find them the data points and edges are rounded to decimal precision, which has not be done in the first digitize operation.
|
|
So this time all the 2s are thought to be on the edge and shifted to the 2nd bin, although they were already in the correct bin!
|
|
I recommend to change line 347 to round the edges already here to decimal precision .
|
|
I hope I could make clear what the problem is.
|
|
For now I found the error and can design my bin edges to be far enough away or exactly at the maximum of the data.
|
|
But it would be nice if this could be fixed in future releases.","bug"
|
|
"6013","2846","Regression numpy.ndindex ) fail in NumPy 1.7 branch and in master.
|
|
In NumPy 1.7, I have this error >>> numpy.ndindex ) Traceback most recent call last) File <stdin> , line 1, in <module> File usr local lib python2.7 dist packages numpy lib index tricks.py , line 536, in init x = as strided nx.zeros 1), shape=shape, strides= nx.zeros like shape)) File usr local lib python2.7 dist packages numpy lib stride tricks.py , line 28, in as strided return np.asarray DummyArray interface, base=x)) File usr local lib python2.7 dist packages numpy core numeric.py , line 324, in asarray return array a, dtype, copy=False, order=order) ValueError array interface shape must be at least size 1 In NumPy 1.6.1, I have this >>> numpy.ndindex ) <numpy.lib.index tricks.ndindex object at 0x7f9015364750> >>>numpy.ndindex ).next ) ) This make this Theano test fail nosetests s pdb pdb failure theano tensor signal tests test downsample.py TestDownsampleFactorMax.test max pool 2d 2D","bug"
|
|
"6013","2848","f2py fortran parsing fails in case commented line present in parameter list.
|
|
A simple piece of valid fortran code, using commented lines in between parameters in a subroutine parameter list currently lets f2py produce bogus results.
|
|
See for example this code .
|
|
this gfortran statements compiles the code without a single warning .
|
|
When producing the signature file the following warning is shown .
|
|
In the signature file subroutine name and associated parameter list remain undefined and a bogus subroutine named unknown subroutine without parameters is defined in stead .
|
|
after removing the 2 comments the test runs just fine .
|
|
The f2py version used for testing was numpy f2py 1.6.2 1.fc17.x86 64 python version was python 2.7.3 7.2.fc17.x86 64 gfortran version was gcc gfortran 4.7.2 2.fc17.x86 64 executed 20 Dec 2012 on Fedora 17 64 bit)","bug"
|
|
"6013","2855","fftfreq dosent work with 64bit ints.
|
|
On a 64bit system the following dosen t work fftfreq 2L) Casting to python int first works fftfreq int 2L))","bug"
|
|
"6013","2860","numpy.test ) fails for 1.8.0.dev a8c641f on OS X 10.8.2.
|
|
On OS X 10.8.2 Core i7), numpy.test ) for numpy 1.8.0.dev a8c641f fails the following test .
|
|
Here s my compiler info .
|
|
.","bug"
|
|
"6013","2862","Update paver for Python 3.3.
|
|
I want to release for Python 3.3 on windows as well.
|
|
And things seems to work except that Paver isn t ready yet .
|
|
So we need to fix it.","bug"
|
|
"6013","2865","Segfault from using probably invalid dtype.
|
|
On a 64 bit machine this leads to a segfault.
|
|
In [1] a = np.array [1], dtype= O ).astype O , [ name , Q )])) In [2] a[ name ] The type on the right should probably be illegal.","bug"
|
|
"6013","2867","FFT2 issue.
|
|
I have run into a problem with fft2.
|
|
I have noticed difficulties getting the symmetry that should exist after a fft2 with a size M,N) image and indexes starting at 0, F u,v) = F M 1 u, N 1 v) Rather than F 0,0) being equal to F M 1, N 1), F 1,1) is equal to F M 1, N 1), and the entire first row column does not line up with the rest.
|
|
I used a 4x4 image as an extreme case so that I could print out the results of the fft2 !
|
|
[ 2ExtremeCase] https f.cloud.github.com assets 3146323 35018 03175d78 517b 11e2 99c1 836c857d6e4b.jpg) it s tiny, and I used the python image library to load it and then numpy.asarray ) to make it into an array for the fft2 ).
|
|
I saved the code into a py file, and each time I want to run it, I drag the image onto the py file.
|
|
Here are it s contents import sys import numpy from PIL import Image imgOrig = Image.open sys.argv[1]) temp = numpy.asarray imgOrig ) fft = numpy.fft.fft2 temp) print fft) input Press Enter to quit ) Finally, here is the output.
|
|
!
|
|
[Result] https f.cloud.github.com assets 3146323 35021 542cebc8 517c 11e2 9bba a571b7c6949d.jpg) And, I moved the rows over manually in GIMP since it was hard to see the asymmetry symmetry when it wrapped !
|
|
[Result2] https f.cloud.github.com assets 3146323 35023 da52c2e0 517c 11e2 8e50 6d8a3d547b3d.jpg)","bug"
|
|
"6013","2870","Verify and possibly fix test failures in Cython and Bottleneck.
|
|
As cgohlke reported Looks good so far.
|
|
I tested numpy MKL 1.7.0rc1.win amd64 py2.7 with some packages that were compiled with numpy 1.6.x <http www.lfd.uci.edu gohlke pythonlibs tests 20121228 win amd64 py2.7 numpy 1.7.0rc1 >.
|
|
There are a few additional test failures in [bottleneck] http pypi.python.org pypi Bottleneck) and Cython, but they don t look serious.
|
|
The rc works well on Python 3.3 too <http www.lfd.uci.edu gohlke pythonlibs tests 20121229 win amd64 py3.3 >.
|
|
We should try to reproduce the failures and see if this is a bug in NumPy and fix it.","bug"
|
|
"6013","2871","Add unique row, for finding unique rows in a 2D array.
|
|
In matlab the function unique works also for selecting the unique rows.
|
|
That should be useful if you add this possibility also in numpy.
|
|
For example > a = np.array [ [1,2][2,3][1,2] ] ) >np.unique a , row=True ) for example >[[1,2] > [2,3]] see matlab docs http www.mathworks.com help matlab ref unique.html Now in numpy for performing this operation we must use some unreadable code like > np.unique1d c.view [ ,c.dtype)] c.shape[1])).view c.dtype).reshape 1,c.shape[1])","enhancement"
|
|
"6013","2877","uint64 causing errors when used to address byte position in a large binary file.
|
|
I am working with large binary files several gigabytes each), and I need to be able to address specific bytes in those files.
|
|
So I keep addresses of data within these files using 64 bit integers, specifically numpy types int64 and uint64.
|
|
However, I noticed that the unsigned version is giving an error >>> ids1 = np.fromfile TR 45 3 ch 0 cluster 1.txt , dtype = np.uint64, sep = ) >>> ids1[ 1] 4999998 >>> tes4530.goto ids1[ 1]) Traceback most recent call last) File <pyshell 55> , line 1, in <module> tes4530.goto ids1[ 1]) File xxx\tes utility.py , line 32, in goto self.f.seek newpos 2 self.rec len) OverflowError Python int too large to convert to C long If record addresses are stored using the signed version numpy.int64), no such error occurs.
|
|
tes4530 above is an instance of a class I wrote to help accessing these files.
|
|
Relevant code is reproduced below class TESstream1 Objects of this class allow loading and searching data from a a binary file containing single channel data def init self, fname, rec len = 256, outtype = np.float32) rec len number of data points per record outtype number type used in the outputs self.records = os.path.getsize fname) 2 rec len) self.f = open fname, rb ) self.rec len = rec len self.fname = fname self.outtype = outtype self.x = 2 15 offset to compensate for data being unsigned def goto self, newpos) go to the beginning of the record specified by newpos) self.f.seek newpos 2 self.rec len) def read self, Nrec = 1) returns the next Nrec records d = np.fromfile self.f, dtype = np.uint16, count = Nrec self.rec len) if self.outtype != np.uint16 d = np.array d, dtype = self.outtype) return np.reshape d, Nrec, self.rec len)) self.x In particular, errors begin to arise when the record address is 4194304 or higher this translates to byte address 2 31 256 data points per record, 2 bytes per data point).
|
|
Using the signed version numpy.int64) does not raise any errors, and it appears that the correct records are being addressed access does not appear to reset to the beginning of file or read random bits, for example).
|
|
I am using Python 2.7.2 64 bit, numpy numpy 1.6.1 64 bit from Christoph Gohlke s repository), running on Windows 7 x64.","bug"
|
|
"6013","2878","Decide whether the value of scalars not just the type) should affect result type of mixed scalar array operations.
|
|
Consider an operation between an array and a scalar .
|
|
Our job is to pick `result.dtype`.
|
|
In numpy before 1.6, if `arr` and `scalar` had the same basic kind of dtype e.g.
|
|
both some kind of float), then we simply used `arr.dtype` as the result dtype so in particular, if `arr.dtype` was `float32`, and `scalar.dtype` as `float64`, then the result was `float32`.
|
|
Perhaps a surprising special case the first time you see it, but the rule is easy to remember, and provides important usability since `scalar` here is usually a python literal, and python literals always have a fixed size that shouldn t be causing upcasting all the time).
|
|
In numpy 1.6, this was changed so that the value of `scalar` was also taken into account.
|
|
So far as I can tell, the rule is now, if the `scalar` cannot be exactly represented using `arr.dtype`, then we go ahead and perform an upcast to a type that can hold both the value of `scalar` and is also consistent with `arr.dtype`.
|
|
So for example it s entirely possible to have `int8 int32` give you an `int16`.
|
|
Examples .
|
|
This change is documented a bit obscurely) here http docs.scipy.org doc numpy reference generated numpy.result type.html It s also been discussed on the list http thread.gmane.org gmane.comp.python.numeric.general 52044 http mail.scipy.org pipermail numpy discussion 2011 September 058563.html http thread.gmane.org gmane.comp.python.numeric.general 47986 This change has caused a fair amount of confusion, and AFAICT everyone who s expressed an opinion has disliked it.
|
|
But, in the discussions so far it s been hard to separate out from all the other changes made in 1.6 or even tell which changes we re talking about), so no consensus has been reached.
|
|
So, we need to decide what to do about this.
|
|
Some options 1.
|
|
Keep the current behaviour as is, `128 int8 > int16` 2.
|
|
Revert to 1.5 and earlier behaviour, `128 int8 > int8` 3.
|
|
Revert to 1.5 and earlier behaviour when it comes to selecting an output dtype, but then use `safe` casting rules when coercing the input to this dtype.
|
|
Right now IIUC we use the exact same rules for `scalar arr` and `np.add scalar, arr, dtype=selected dtype)`, i.e., we treat the selected dtype as if it were explicitly requested and do an `unsafe` forced cast.
|
|
The suggestion here is to treat the implicitly selected output dtype as if it were implicitly selected, and use more conservative casting.)
|
|
Result `128 int8 > error`, and you can use the full `np.add ...)` form if you really meant this).
|
|
Right now I like option 3 the best but am open to discussion.
|
|
CC delallea, matthew brett","bug"
|
|
"6013","2880","Get financial functions out of main namespace.
|
|
We have these financial functions that are exported in the main `numpy` namespace, and are really out of place http docs.scipy.org doc numpy reference routines.financial.html Also They have terrible names even if they weren t in the main namespace, but especially so given that they are) There are multiple documented, but unimplemented, functions here.
|
|
WTF.
|
|
Let s move them to numpy.financial currently they re also in `numpy.lib.financial`, but the `.lib.` part is just noise IMO) with better names and deprecate the current names.","enhancement"
|
|
"6013","2881","If Python warnings module is used to catch RuntimeWarning, numpy.log raises TypeError on issubclass ).
|
|
If the Python `warnings` module is used to catch and handle `RuntimeWarning` as an error, `numpy.log 0.
|
|
)` will raise a `TypeError` on `issubclass )` not called by the user).
|
|
For example Python prompt) >>> import numpy >>> import warnings >>> numpy.log 0.)
|
|
main 1 RuntimeWarning divide by zero encountered in log inf >>> warnings.simplefilter error , RuntimeWarning ) >>> try ... numpy.log 0.)
|
|
...except RuntimeWarning ... pass ... Traceback most recent call last) File <stdin> , line 2, in <module> TypeError issubclass ) arg 2 must be a class or tuple of classes","bug"
|
|
"6013","2885","median calls mean, apparently for no good reason.
|
|
Complexity of algorithm is wrong.. numpy.median is defined in numpy libs function base.py.
|
|
At the very last step mean is called.
|
|
I think this can be replaced by something i.e.
|
|
len 2 ) that I m guessing will be O 1) rather than O n).
|
|
Use mean in odd and even case to coerce data type and check, use out array.
|
|
return mean sorted[indexer], axis=axis, out=out) I would almost consider this a bug since the complexity of the algorithm is completely wrong.","bug"
|
|
"6013","2889","Backport reminder.
|
|
Backport 2886 and 2887 into 1.7.","bug"
|
|
"6013","2890","np.ma.argsort uses default fill value not largest representable value.
|
|
example for bug with wrong sorting http projects.scipy.org scipy ticket 1805 np.ma.sort uses fill value=np.ma.minimum fill value aam) which uses the largest representable value fix should be to use this also in argsort","bug"
|
|
"6013","2892","Regression in ufunc.reduceat with zero sized index array.
|
|
The test code .
|
|
works in numpy 1.4, but raises an error in 1.8.
|
|
I ve got a patch for this, will make a pull request for it.","bug"
|
|
"6013","2894","np.linalg.svd fails to perform decomposition on vector.
|
|
Code to reproduce A = np.random.normal size=10) decomp = np.linalg.svd There are several reasons why not allowing an SVD on a vector are a very bad idea The SVD is properly mathematically defined for any real valued matrix with m >= 1 columns and n >= 1 rows.
|
|
As such, it is a bit silly to not provide this functionality pinv calls SVD and a pseudoinverse is also completely mathematically valid for a vector or a scalar!).
|
|
The real reason In many greedy algorithms such as forward subset selection, orthogonal matching pursuit, etc.
|
|
), you d have to modify the code on the first step which is very inelegant) in which you would be computing the pseudoinverse of an n by 1 matrix i.e., n dimensional vector).
|
|
Of course it s easy to code the corresponding mathematical operator w dot products and what not, but you shouldn t have to, and in fact you don t in R. In short, I think this should be fixed.
|
|
Adam","bug"
|
|
"6013","2895","np.ndindex ) is now failing.
|
|
On master and maintenance 1.7.x the following code is now failing np.ndindex ) This has been traced to the code added in 444 ctors.c 2082 2086) and 2815 ctors.c 2083 2087) which does what seems to be an unnecessary but explicit check to invalidate array scalars in the array interface.
|
|
This can be fixed by removing that check.","bug"
|
|
"6013","2896","assigning with mask of float32 array elements fails for int argument.
|
|
Hello, The following code raises TypeError array cannot be safely cast to required type a = np.zeros 1, dtype=np.int) b = np.zeros 1, dtype=np.float32) mask = np.array [True], dtype=np.bool) b[mask] = a[ ] Notice, that if b is of type np.float64, this code passes a = np.zeros 1, dtype=np.int) b = np.zeros 1, dtype=np.float64) mask = np.array [True], dtype=np.bool) b[mask] = a[ ] print ok )","bug"
|
|
"6013","2898","Stop mixing up integer and floating point overflow.
|
|
The integer arithmetic code in `scalarmathmodule.c.src` has overflow detection logic in it.
|
|
That s great.
|
|
When it detects an overflow, the way it signals that is by calling `npy set floatstatus overflow )`.
|
|
That s not so great.
|
|
Two problems Integer rollover is really not the same thing as floating point overflow.
|
|
I m much more worried about integer overflow for one thing, since it doesn t nicely go to `inf`, it just scrambles your data in some impossible to notice way.)
|
|
Setting and checking the float status bits is REALLY EXPENSIVE.
|
|
Right now we set and check the float status bits on every scalar integer operation, and this makes those operations slower by a factor of two .
|
|
So yeah, we shouldn t do that.
|
|
Proposed solution Add a new arithmetic error condition, let s call it rollover , which is specifically for integer overflow, and make the current overflow just be used for floats.
|
|
So this is a new argument to np.seterr, etc.)
|
|
Test and signal this condition using a new thread local storage key, rather than the floating point status bits.
|
|
This should be much cheaper, though you know, double check this.)
|
|
Python exports an undocumented C API for thread local storage .
|
|
For our purposes, we d do something like .","bug"
|
|
"6013","2899","The epic dtype cleanup plan.
|
|
[this is very drafty notes to self that I m hitting save on now because I was dumb and typed it into a webform where it might get lost, and because I m lazy and don t want to pull it out to stick in some file on my hard drive where I ll lose it.
|
|
So no guarantees any of this actually makes sense or anything, but since we ve been running into a number of nasty aspects of dtypes recently I wanted to write down a bunch of ideas together so we can try and organize them somewhat.
|
|
Feel free to comment even at this early date.
|
|
Also, don t tell anyone, but part of the secret goal here is enabling the NA dtype work too though not all of this is a prerequisite for that by any means).]
|
|
End goal dtypes basically function like Python objects WRT to subclassing etc.
|
|
; isinstance issubclass work in a useful way; parametric dtypes are not horrible; user defined dtypes are on closer to equal footing with built in dtypes Pass dtypes to ufunc inner loops Make dtypes immutable and make sure that we don t actually allocate new ones on ufunc return or when passing cast data to inner loops, but instead just incref and go.
|
|
immutability is a prerequisite for this.)
|
|
Move all the special case operations for void string datetime dtypes out of the core ufunc code and into ufunc loops.
|
|
Merge ufunc.so and multiarray.so, the split just ends up making us contort code to let them interface with each other Make overflow etc.
|
|
error flag clearing setting checking methods on dtypes.
|
|
Use this to split the integer and float error flag stuff 2898 is to make this possible, but leaves the problem of arranging for the correct functions to be called on the correct loops.
|
|
this is the clean way to actually call the correct pieces at the correct time, while also as a bonus allowing custom dtypes to do their own error signaling, or not, as they prefer.)
|
|
This is an example of a place where having to export this interface between multiarray.so, where dtypes live, and ufunc.so, where ufuncs live, would be pointlessly cumbersome.)
|
|
Make dtype the class) inherit from `type` and define ` instancecheck ` and ` subclasscheck ` methods, so `isinstance foo, dtype instance)` can do something useful.
|
|
This will completely change the dtype struct s memory layout though, `type` is a huge object, and all of dtype s fields will get shifted down in memory.)
|
|
Give all the dtype functions `self` arguments.
|
|
Turn them into the equivalent of cpdef methods.
|
|
not sure how we make this work follow MRO when doing dtype operations?
|
|
or in dtype.
|
|
new , copy parent fn ptrs to undefined child fn ptrs and ha ha no multiple inheritance is not supported?)
|
|
Make it possible to add new dtype methods without breaking the memory representation in the future right now they have to go in the middle of the dtype struct...) Make it possible to define new dtypes in Python.
|
|
A set of C level dtype methods that just call python level methods) Re arrange all the weird stuff in the current dtype memory representation so the parts that are specific to a single dtype go into that subclass e.g.
|
|
the field names, the datetime unit, etc., should not be part of the core dtype object) ...figure out how to do this without completely messing up existing code.
|
|
Really only code that actually messes with dtypes should be affected like, code that defines new dtypes).
|
|
Binary compatibility much harder than source compatibility can we tell people to recompile?
|
|
Possibility make sure we can distinguish a pointer to old dtype object from a pointer to new dtype object should be easy, all PyObject s are compatible enough to let you get their type), and at all the public entry points do `if !
|
|
arg = CheckOrConvertThisDtypeThing arg)) return 1;`, and that function sets up a new style dtype in place of the old style one, while raising a DeprecationWarning.
|
|
This might not be so horrible even, though we would want to make sure we did all the binary compatibility breaking changes at once.
|
|
This still won t help for user code that reaches into dtype objects by hand, though, as opposed to just creating them.
|
|
It s possible that all user dtypes do this, so eh.)
|
|
Rationalize the dtype constructor arguments.
|
|
Ideally move the horrible build a dtype from strings and tuples and sealing wax stuff to Python code.
|
|
Note there is an argument for getting all the dtype struct changes out of the way before exposing them to ufunc inner loops, since that will open the door to more user code that depends on the contents of dtype structs.
|
|
Of course if the problem is already as bad as it could possibly be, then this doesn t matter ).
|
|
And letting ufuncs get at dtypes is probably the single biggest win on this list, so it would be nice to prioritize it.","enhancement"
|
|
"6013","2902","AttributeError when comparing inequally sized ndarrays with complicated dtype.
|
|
When comparing two ndarrays with complicated dtypes, and the sizes of the arrays are not the same, the comparison fails with an `AttributeError`.
|
|
Below is some example code.
|
|
issue 790 appears similar, but was closed in 2006.
|
|
The problem I experience is not identical, because a==a succeeds ; it fails if and only if the sizes are not the same.
|
|
It also fails with a slightly different error message it doesn t say anything about a `bool`).
|
|
In [35] zeros 4) == zeros 4) Out[35] array [ True, True, True, True], dtype=bool) In [36] zeros 4) == zeros 5) Out[36] False In [37] dtp = dtype [ A , int8, 2), B , int8, 3)]) In [38] zeros 4, dtp) == zeros 4, dtp) Out[38] array [ True, True, True, True], dtype=bool) In [39] zeros 4, dtp) == zeros 5, dtp) AttributeError Traceback most recent call last) <ipython input 39 cb9b1c180c1c> in <module> ) > 1 zeros 4, dtp) == zeros 5, dtp) AttributeError logical and I m using numpy 1.6.2, git revision ca07bce202ae26b6f0a73870eb2ef0b88e0210c5.","bug"
|
|
"6013","2905","Add a big scary warning to oldnumeric and numarray compat modules.
|
|
As per the thread starting here http mail.scipy.org pipermail numpy discussion 2013 January 065111.html","bug"
|
|
"6013","2906","docstring typo.
|
|
https github.com numpy numpy blob master numpy core code generators ufunc docstrings.py L1102 exp calculate x p.","bug"
|
|
"6013","291","Changes in PyArray FromAny between 1.5.x and 1.6.x.
|
|
In Numpy 1.5.x, we have .
|
|
In 1.6, we get the following, .
|
|
This seems to be do to the changes in PyArray\ FromAny introduced in http github.com mwhansen numpy commit 2635398db3f26529ce2aaea4028a8118844f3c48 .
|
|
In particular, \ array\ find\ type used to be used to query our \ \ array\ interface\ \ attribute, and it no longer seems to work.
|
|
It should be reproducible with the following minimal example .","bug"
|
|
"6013","2912","numpy.testing.assert allclose reports wrong mismatch .
|
|
numpy.testing.assert allclose [1,2,3],[1,2.01,3]) returns mismatch 100.0 numpy.testing.assert almost equal [1,2,3],[1,2.01,3]) correctly returns mismatch 33","bug"
|
|
"6013","2913","Switch to separate compilation by default.
|
|
Let s make 1.8 use separate compilation by default, with an eye towards eventually phasing out the one big blob style of compilation though for now we will certainly continue to support both).
|
|
Todo Make the code that checks for NPY SEPARATE COMPILATION instead check for NPY ALL IN ONE COMPILATION or something like that), and flip the default Update the .travis.yml so that it runs some tests with this new variable set instead of the current NPY SEPARATE COMPILATION tests Make sure to document the new variable somewhere so in case anyone runs into problems caused by this, they have some hope of reverting to the old behaviour without having to work it out from first principles.","bug"
|
|
"6013","2915","numpy 1.7.0rc1 breaks inline fortran code in sage.
|
|
The following commit ad9c2f47b491fbf19c48e9c2dc7fd56ddca61685 breaks sage s ability to compile inline fortran <pre> sage t long devel sage main sage misc inline fortran.py File home work fbissey sandbox sage 5.6.rc0 devel sage main sage misc inline fortran.py , line 29 sage fortran example) Exception raised Traceback most recent call last) File home work fbissey sandbox sage 5.6.rc0 local bin ncadoctest.py , line 1231, in run one test self.run one example test, example, filename, compileflags) File home work fbissey sandbox sage 5.6.rc0 local bin sagedoctest.py , line 38, in run one example OrigDocTestRunner.run one example self, test, example, filename, compileflags) File home work fbissey sandbox sage 5.6.rc0 local bin ncadoctest.py , line 1172, in run one example compileflags, 1) in test.globs File <doctest main .example 1[4]> , line 1, in <module> fortran example) line 29 sage fortran example) File home work fbissey sandbox sage 5.6.rc0 local lib python site packages sage misc inline fortran.py , line 21, in call return self.eval args, kwds) File home work fbissey sandbox sage 5.6.rc0 local lib python site packages sage misc inline fortran.py , line 84, in eval f2py.compile x, name, extra args = extra args, source fn=fortran file) File home work fbissey sandbox sage 5.6.rc0 local lib python site packages numpy f2py init .py , line 40, in compile s,o = exec command c) File home work fbissey sandbox sage 5.6.rc0 local lib python site packages numpy distutils exec command.py , line 197, in exec command if with python and 0 or sys.stdout.fileno )== 1) AttributeError SpoofOut instance has no attribute fileno < pre>","bug"
|
|
"6013","2916","Numpy installation cannot find GnuFCompiler executable.
|
|
While configuring numpy, I see customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 customize IntelFCompiler It seems that the installer tries to find a GnuFCompiler executable by looking for the names g77 and f77 and then gives.
|
|
However, on my system and I think on many others) the executable is called gfortran , which acts as both a f77 and a f90 95 2003 compiler.
|
|
Adding gfortran to the list of potential executable names would solve this.","bug"
|
|
"6013","2917","`numpy.lib.
|
|
iotools.StringConverter.upgrade` should have a return value.
|
|
According to the documentation, routine `upgrade` should return the value of the conversion, but it does not.","bug"
|
|
"6013","2918","Make numpydoc Python 3 compatible.
|
|
Numpydoc in doc sphinxext) does not support Python 3.","bug"
|
|
"6013","2921","Backport 2897.
|
|
I need to get useful commits from 2897 and backport them.","bug"
|
|
"6013","2922","Cannot load large data >2GB archived with savez on 64bit system.
|
|
When large arrays are saved in npz format with `savez` function, they can not be loaded back with `load` function.
|
|
This should reproduce the problem .
|
|
On my system numpy 1.8.0.dev 27690e3, python 2.7.3, ubuntu 12.04 64bit) this produces the error .
|
|
This bug might be related to bug in cStringIO used by zipfile) that has file size limitation of 2 31 1 http bugs.python.org issue11564","bug"
|
|
"6013","2925","In the numpydoc the class attributes not included.
|
|
I would like to use Sphinx with numpydoc, but the Attribute not included in the documentation, what can be the problem?
|
|
.","bug"
|
|
"6013","2926","faster array equal.
|
|
I think `array equal` [here] https github.com numpy numpy blob master numpy core numeric.py)) can be at least twice as fast on average if instead of checking all elements of both arrays for equality `a1==a2`) and then using `all )`, the code will check one by one and return `False` when an element pair is not equal .
|
|
Of course a pure python version is not as fast as the current Numpy version, but I ve implemented this with Cython and got good results when comparing arrays with 1000 elements and a single randomly chosen element that is not equal it runs twice as fast as the current Numpy function.
|
|
The problem with the Cython code is that it expects a certain `dtype`, and I don t know how to make it generic for all types.
|
|
Also, I didn t find where is the relevant `.pyx` file.
|
|
So I m posting this as an issue instead of a PR.
|
|
So here is my Cython code .","bug"
|
|
"6013","2927","Memory leak in np.array a ) .
|
|
Code that reproduces it .
|
|
This is in both 1.7 and master, and I ve got a patch to fix it.","bug"
|
|
"6013","2929","genfromtxt docstring does not explain the arguments loose , missing and skiprows ..
|
|
The Parameters section of the docstring for `genfromtxt` does not include the arguments `loose`, `missing` and `skiprows`.","bug"
|
|
"6013","2930","unsupported operand type s) for numpy.ndarray and numpy.float64 .
|
|
Please see post at stackoverflow http stackoverflow.com questions 14408122 unsupported operand types for numpy ndarray and numpy float64 comment20050672 14408122","bug"
|
|
"6013","2934","Bento build broken by commit d2122ea1.
|
|
charris in commit d2122ea1f, why are you including the header starting with an underscore?
|
|
Removing the underscore fixes the build issue.
|
|
[29 51] c numpy core src umath umathmodule onefile.c > build numpy core src umath umathmodule onefile.c.6.o In file included from home rgommers Code numpy numpy core src private npy config.h 5 0, from home rgommers Code numpy build numpy core src umath loops.c 8, from .. numpy core src umath umathmodule onefile.c 1 home rgommers Code numpy build numpy core include numpy numpyconfig.h 2 2 error error this header should not be included directly, always include numpyconfig.h instead File home rgommers Code Bento bentomakerlib bentomaker.py , line 486, in noexc main main argv) File home rgommers Code Bento bentomakerlib bentomaker.py , line 251, in main return wrapped main global context, popts, run node, top node, build node) File home rgommers Code Bento bentomakerlib bentomaker.py , line 325, in wrapped main return main global context, cached package, popts, run node, top node, build node) File home rgommers Code Bento bentomakerlib bentomaker.py , line 402, in main run cmd global context, cached package, cmd name, cmd argv, run node, top node, build node) File home rgommers Code Bento bentomakerlib bentomaker.py , line 469, in run cmd run with dependencies global context, cmd name, cmd argv, run node, top node, running package) File home rgommers Code Bento bento commands wrapper utils.py , line 18, in run with dependencies resolve and run command global context, cmd name, cmd argv, run node, package) File home rgommers Code Bento bento commands wrapper utils.py , line 32, in resolve and run command run command in context context, cmd, pre hooks, post hooks) File home rgommers Code Bento bento commands wrapper utils.py , line 68, in run command in context cmd func context) File home rgommers Code Bento bento commands build.py , line 55, in run ctx.compile ) File home rgommers Code Bento bento backends waf backend.py , line 363, in compile self.waf context.compile ) File home rgommers Code waf waflib Build.py , line 382, in compile raise Errors.BuildError self.producer.error) bentomaker Error bentomaker crashed uncaught exception <class waflib.Errors.BuildError > Build failed > task in umath failed exit status 1) {task 156165292 c umathmodule onefile.c > umathmodule onefile.c.6.o} [ usr bin gcc , fPIC , I home rgommers Code numpy build numpy core src umath , I home rgommers Code numpy numpy core src umath , I home rgommers Code numpy build numpy core src private , I home rgommers Code numpy numpy core src private , I home rgommers Code numpy build numpy core , I home rgommers Code numpy numpy core , I home rgommers Code numpy build numpy core include , I home rgommers Code numpy numpy core include , I home rgommers Code numpy build numpy core include numpy , I home rgommers Code numpy numpy core include numpy , I usr include python2.7 , .. numpy core src umath umathmodule onefile.c , c , o , numpy core src umath umathmodule onefile.c.6.o ]).","bug"
|
|
"6013","2935","repr float32 0.0)) lacks decimal point.
|
|
The behavior of `float.
|
|
repr ` and `float32.
|
|
repr ` disagree at ` 0.0`.
|
|
The former gives a decimal point, just as `print f , 0.0f)`, but the latter lacks the point.
|
|
.
|
|
System Info os=archlinux python=2.7.3 4 numpy=1.7.0 1 LANG=en US.UTF 8","bug"
|
|
"6013","2939","2738 fix breaks scipy test decomp.py in 1.7.0rc1.
|
|
edit the issue is actually hash randomization and a bug in scipy, see below the fix for gh 2738 commit e208de6adfdd38aabe0b5a87ba22f80e04a014ad) in numpy 1.7.0rc1 breaks a scipy 0.11 and current git head test with python3.3.
|
|
This was previously reported to scipy in http mail.scipy.org pipermail scipy dev 2012 September 017995.html but I did not find a reaction.
|
|
Bisecting numpy lead to above commit.
|
|
The crash can be reproduced on ubuntu 13.04 with python3.3.
|
|
It occurs with the debug and regular python3.3 variant.
|
|
It manifests itself with random segfaults, glibc memory corruption aborts and hangs when you execute scipy linalg tests test decomp.py reverting the commit from the rc1 tag fixes the crash.
|
|
numpy git head does not seem affected.","bug"
|
|
"6013","294","Typestring h2 works in 1.6, fails in master.
|
|
Problem which breaks scipy) http mail.scipy.org pipermail numpy discussion 2012 June 062605.html Analysis http mail.scipy.org pipermail numpy discussion 2012 June 062606.html Todo Re enable parsing of type strings like h2 before the 1.7 release [DONE] Give a deprecation warning for type strings like h100 .
|
|
Possibly also for type strings like h2 this isn t obvious.
|
|
[DONE] Disable at least h100 style type strings for 1.9, or whenever we decide to finalize the deprecations that started in 1.7.","bug"
|
|
"6013","2944","ENH allow linspace to take 1D vectors as start and end.
|
|
I have a line segment in N dimensional space, defined by its begin and endpoints both 1D vectors).
|
|
For example in 3D start = array [0,1,2]) end = array [1,3,6]) I would like to be able segment the line in k segments, for example 4, or 5 including the endpoints.
|
|
The result would then be array [[ 0. , 0.25, 0.5 , 0.75, 1.
|
|
], [ 1. , 1.5 , 2. , 2.5 , 3.
|
|
], [ 2. , 3. , 4. , 5. , 6. ]])
|
|
I extended and broke) linspace to not only accept scalar, but also 1D vector start and endpoints def linspace start,stop,num=50,endpoint=True,retstep=False) spannedRange = stop start numberOfSteps = num 1 if endpoint else 0) step = array spannedRange numberOfSteps,ndmin=2).T coefficientColumn = numpy.arange num) samples = numpy.kron coefficientColumn,step) array start,ndmin=2).T if retstep return samples,step) else return samples Would it be possible without breaking linspace former behaviour) to incorporate this functionality into NumPy?","bug"
|
|
"6013","2946","Incorrect documentation of mgrid on scipy.org.
|
|
on the following website http www.scipy.org NumPy for Matlab Users) the comparison to Matlab indicates that the two lines mgrid[0 9.,0 6.]
|
|
and meshgrid r [0 9.
|
|
],r [0 6.])
|
|
are equivalent, but the results are transposed.
|
|
The meshgrid result is the one that matches Matlab.","bug"
|
|
"6013","2948","Data type problem between numpy and scipy in function label ) in scipy.ndimage\measurements.py.
|
|
The following error occurs with installed versions of Python 2.7.3 Python 2.7 numpy 1.6.2 Python 2.7 scipy 0.11.0 File C \Python27\lib\site packages\scipy\ndimage\measurements.py , line 157, in label max label = nd image.label input, structure, output) RuntimeError data type not supported The error does NOT occur when I install Python 2.7 numpy 1.5.1 instead of Python 2.7 numpy 1.6.2 !
|
|
Good luck, thanks a lot!
|
|
Mathias !
|
|
[error] https f.cloud.github.com assets 3370697 93891 a88ab7f6 6638 11e2 982c e96012442ff2.jpg)","bug"
|
|
"6013","2949","numpy.reshape return wrong strides information in master.
|
|
On master in commit 2b4c9a66c8485973291076a6a1d3a3e9a3, I have this >>> import numpy >>> numpy.asfortranarray numpy.random.rand 3,6,1,1)).reshape 3,6,1).strides 8, 24, 140247363096948) With numpy 1.6.1, I have this output 8,24,144) I expect the numpy 1.6.1 give the right outputs.
|
|
This cause problem in Theano tests as we do not support not aligned memory.
|
|
In master, the ndarray isn t aligned due to the bad strides.
|
|
We can work around in theano by checking the output of reshape, but maybe this can affect other people.
|
|
I think there was some discussion on strides for dimensions of shape 1, but I m not sure and I probably didn t followed them closely.
|
|
Do someone remember that?","bug"
|
|
"6013","2951","numpy.int64 is not instance of int.
|
|
As reported in [xlwt] https github.com python excel xlwt issues 15) Here is an examination of numpy behaviour Python 2.7.3, numpy 1.6.2) .
|
|
Looks like numpy has done the work to make its int32 and float64 recognisable by other software but not int64.","bug"
|
|
"6013","2952","Unused function `simple capsule dtor`.
|
|
When compiling matplotlib with clang under py2.7, I get a compiler warning and I thought I d report the issue.
|
|
It stems from the agg backend source including the `npy 3kcompat.h` file and the warning is .","bug"
|
|
"6013","2955","bitwise and with uint64.
|
|
The bit wise `` `` operator works fine with all Numpy integer types, except ``uint64`` .
|
|
Even Python long integers can be used, so is there a reason why ``np.uint64`` cannot be used, or is it a bug?","bug"
|
|
"6013","2956","Spurious C contiguity issues on git master.
|
|
This Cython code fails on Numpy git master 4600b2fe1d) .
|
|
called with .
|
|
ascontiguousarray should always return C contiguous arrays with correct flags, but that doesn t seem to be the case currently.
|
|
Doesn t fail on 1.7.0rc1 and 1.7.x branch.
|
|
Observed in Scipy tests with Numpy master http projects.scipy.org scipy ticket 1832 http projects.scipy.org scipy ticket 1830","bug"
|
|
"6013","2958","Bug when making a structured array view a transposed array.
|
|
A view of a transpose does not work properly.
|
|
A transpose is itself a kind of view, so this is a view of a view of an array.
|
|
I see nothing in the documentation that says that views of views should not work.
|
|
Nor does creating a view of a view raise any error or warning.
|
|
It just returns a strange answer leaving the user to track down the cause.
|
|
import numpy as np Define two arrays, one directly and one as a transpose view >>> my array 1 = np.array [[1,2],[3,4],[5,6],[7,8]], dtype=float) >>> my array 2 = np.array [[1,3,5,7],[2,4,6,8]], dtype=float).T These two arrays seem to be the same... >>> print my array 1) [[ 1.
|
|
2.]
|
|
[ 3.
|
|
4.]
|
|
[ 5.
|
|
6.]
|
|
[ 7.
|
|
8.]]
|
|
>>> print my array 2) [[ 1.
|
|
2.]
|
|
[ 3.
|
|
4.]
|
|
[ 5.
|
|
6.]
|
|
[ 7.
|
|
8.]]
|
|
But if I view them as structured arrays, they re different!
|
|
>>> print my array 1.view dtype=[ ,np.float), ,np.float)])) [[ 1.0, 2.0)] [ 3.0, 4.0)] [ 5.0, 6.0)] [ 7.0, 8.0)]] >>> print my array 2.view dtype=[ ,np.float), ,np.float)])) [[ 1.0, 3.0) 2.0, 4.0)] [ 5.0, 7.0) 6.0, 8.0)]] I m using NumPy 1.6.1.
|
|
Is the latest version different?)
|
|
Any numpy expert will immediately recognize a workaround Use x.T.copy ) instead of x.T before constructing the view.
|
|
Sure, OK, that s what I ll do from now on ... but still, this behavior should be fixed, and or documented, and or raise a warning or error flag.
|
|
Thanks in advance!","bug"
|
|
"6013","2959","multiarray failed to import.
|
|
I have installed Python 2.7.3 on a redhat 5.5 system.
|
|
Did an installation of numpy 1.6.2 and I get the following python >>>import numeric Traceback most recent call last) File <stdin> , line 1, in <module> File opt oss python Python 2.7.3 lib python2.7 site packages numpy core numeric.py , line 28, in <module> import umath ImportError numpy.core.multiarray failed to import I can import numpy without errors.","bug"
|
|
"6013","2961","qr decomposition on non square matrices broken.
|
|
Typing numpy.linalg.qr array [[1],[1]])) into the interpreter yield a tuple q = array [[ 0.70711], [ 0.70711]]) r = array [[ 1.41421]]) Using octave etc gives the correct values of q = array [[ 0.70711, 0.70711],[ 0.70711, 0.70711]]) r = array [[ 1.41421],[0]]) The dimensions of the components of the QR decomposition of a rectangular non square) matrix are defined in https en.wikipedia.org wiki QR decomposition Numpy does not seem to honor this definition.","bug"
|
|
"6013","2969","1.7 should include this memory leak fix.
|
|
While investigating a memory leak in scikit learn https github.com scikit learn scikit learn issues 1663), I discovered that the memory leak was caused by a bug in ``numpy``.
|
|
The fix to the memory leak was committed to ``numpy`` in September 2012 https github.com numpy numpy commit 80b3a3401382cb3f14c5b76dd90d9f932f50ad15), but isn t in the maintenance 1.7.x branch.
|
|
Are there plans to backport it to 1.7 and make a new release?
|
|
Thanks,","bug"
|
|
"6013","2971","Polynomials should automatically scale their degree when their upper term is set to 0.
|
|
For example from numpy.polynomial import Polynomial as P >>> p Polynomial [ 1., 2., 3., 4., 5., 6., 7., 8., 9.
|
|
], [ 1., 1.
|
|
], [ 1., 1.])
|
|
>>> p. coef [ 1] = 0 >>> p Polynomial [ 1., 2., 3., 4., 5., 6., 7., 8., 0.
|
|
], [ 1., 1.
|
|
], [ 1., 1.])
|
|
>>> p. coef [ 1] 0.0 >>> p. degree ) 8 Should be 7, and the list should get rid of the upper 0, so p. coef [ 1] should be 8.
|
|
It would maybe worth writing a setitem function for Polynomials that would take into account this kind of side effect?
|
|
V.","bug"
|
|
"6013","2972","Accessing fields for a masked structured array fails with ValueError.
|
|
I m trying to create a mask for a structured array.
|
|
When I try to access a field from the array or simply represent it, numpy throws a `ValueError field names A not found`, as illustrated below.
|
|
.
|
|
I tested it with the stable version `1.6.2` and the bleeding edge version obtained directly from `git`.
|
|
It also fails in earlier stable versions tested with 1.6.2).
|
|
Note if I instead create directly a masked array with `numpy.ma.empty`, I get a different error .","bug"
|
|
"6013","2977","backport 2976.
|
|
2976 is a trivial bug fix.","bug"
|
|
"6013","2978","nansum bool array) is broken in numpy 1.7 regression).
|
|
This worked fine in 1.6.2 and earlier versions of numpy.
|
|
There is of course an easy workaround to check the type of the data first and then use sum instead of nansum in case it is boolean, but this is tedious.
|
|
FWIW, it still) works fine for integer columns.
|
|
.","bug"
|
|
"6013","2980","DOC Use tp basicsize, not NPY SIZEOF PYARRAYOBJECT, for allocation.
|
|
As charris mentioned in 2975, we should mention the `NPY SIZEOF PYARRAYOBJECT` > `type >tp basicsize` change in `doc release 1.8.0 notes.rst`.
|
|
Does this look okay?
|
|
I ll create a pull request when we have agreed upon language.
|
|
.","bug"
|
|
"6013","2981","f2py issue when returning a single tuple.
|
|
From http projects.scipy.org scipy ticket 1187, comment of Pearu f2py supports user defined functions that have multiple return values.
|
|
These return values are expected to be returned as a single tuple object.
|
|
So, the issue here is not a result of f2py bug but due to this f2py feature that is unintentionally triggered by the given example.
|
|
As a possible fix, we could have the f2py code to treat the returned tuples as ordinary sequences when the expected number of returned values is exactly 1.
|
|
On the other hand, in the case of f2py misuse returning 2 values when 1 is expected, for instance) the values will be silently converted to an array and may result even a more complicated issues to be analyzed.
|
|
I think the best fix is to document this f2py feature in the corresponding codes, that is, never return tuple as a single return value, and perhaps improve the error message giving a hint why the crash might have been occurred.","bug"
|
|
"6013","2986","shape parameter for np.random function should be called `shape`.
|
|
Everything in numpy uses the convention that arrays `shape`s are called `shape`... except `np.random`.
|
|
This just confused me yet again.
|
|
Is there any good reason we can t switch to `shape=` as the canonical name for this argument?
|
|
Obviously we would continue to support `size=` as an alias for some long time, but at least I would stop stubbing my toe on this...","bug"
|
|
"6013","2987","random integer function names are less helpful than they could be.
|
|
np.random contains two functions to generate random integers `randint` and `random integers`.
|
|
Anyone know what the difference is without checking the docs...?
|
|
`randint` treats its range as half open [low, high) like other ranges in python; `random integers` treats it as closed like [low, high].)
|
|
The Python stdlib is not much better, with `randrange` and `randint`, but at least `randrange` makes its semantics obvious from the name it has the same semantics as the standard `range` function), and `randint` is a wart that predates `randrange` and survives for backwards compatibility.
|
|
For extra points, the mapping is numpy `randint` = stdlib `randrange` numpy `random integers` = stdlib `randint` Can we at least add a proper `np.random.randrange` function?
|
|
Extra points if we add support for the `step` argument at the same time; stdlib `randrange` supports this.)
|
|
That s the key one we need to follow Python convention.
|
|
And maybe we should deprecate our `randint` at the same time?","enhancement"
|
|
"6013","2989","argsort always uses int64 as a result type.
|
|
argsort always produces array of type int64, which is a waste of space, as any value in resulting array cannot be bigger than total number of elements in original array, which is known beforehead.
|
|
In following example it would be enough to produce array of int8 .","bug"
|
|
"6013","2990","easy install2.7 numpy broken on systems without previous version of numpy?).
|
|
My system has no scipy or numpy.
|
|
Let me know if you need anything else.
|
|
Thanks!
|
|
.
|
|
.","bug"
|
|
"6013","2991","bug in np.ndarray.
|
|
iadd when adding the transpose.
|
|
I want to symmetrize a matrix a by adding its transpose, but there is a strange bug when using the = operator.
|
|
see below for a test case >>> import numpy as np >>> a=np.vstack [np.arange 95)] 95) >>> c=np.vstack [np.arange 95)] 95) >>> b = c c.T >>> a = a.T at this point both a and b should be symmetric, however a is not!
|
|
>>> print a a.T array [[ 0, 0, 0, ..., 0, 0, 0], [ 0, 0, 0, ..., 1, 1, 1], [ 0, 0, 0, ..., 2, 2, 2], ..., [ 0, 1, 2, ..., 0, 0, 0], [ 0, 1, 2, ..., 0, 0, 0], [ 0, 1, 2, ..., 0, 0, 0]]) >>> print np.where b b.T !=0) array [], dtype=int64), array [], dtype=int64)) This only appears to happen on larger arrays at least 91,91) This happens for me on my mac as well as a linux box running python 2.7.3 with numpy 1.6.1 I also upgraded my mac to numpy 1.7.0 with pip install upgrade numpy and the problem persists.","bug"
|
|
"6013","2993","Outdated bento build instructions.
|
|
It is said in BENTO BUILD.txt that waf should be downloaded with svn.
|
|
That is strange, as the waf project seems to have a git repository, not svn.","bug"
|
|
"6013","2995","Multiarray Double Free or Unmap Pointer error for huge datasets.
|
|
Hello NumPy team, I seem to have encountered some unusual behavior when running NumPy SciPy on massive datasets.
|
|
Please see below for the error.
|
|
Also, this ticket bug report has been cross posted on stackoverflow.com http stackoverflow.com questions 14906962 python double free error for huge datasets) and the SciPy developers list http projects.scipy.org scipy ticket 1846).
|
|
It was upon the suggestion of an SOer that I decided to post here and on the SciPy dev list.
|
|
I am running Linux x86 64 bit OpenSuSE 11.4, NumPy version 1.5.1, SciPy version 0.9.0, Python 2.7.
|
|
I have a very simple script in Python, but for some reason I get the following error when running a large amount of data .
|
|
I am used to these errors coming up in C or C , when one tries to free memory that has already been freed.
|
|
However, by my understanding of Python and especially the way I ve written the code), I really don t understand why this should happen.
|
|
Here is the code .
|
|
One may ask, why declare a P and a Q numpy array?
|
|
I simply do that to reflect the actual conditions as this code is simply a segment of what I actually do, where I need a P matrix and declare it beforehand).
|
|
I have access to a 192GB machine, and so I tested this out on a very large SciPy sparse matrix 2.2 million by 2.2 million, but very sparse, that s not the issue).
|
|
The main memory is taken up by the Q, P, and mat matrices, as they are all 2.2 million by 2000 dense matrices size = 2.2 million, numlabels = 2000).
|
|
The peak memory goes up to 131GB, which comfortably fits in memory.
|
|
While the mat matrix is being computed, I get the glibc error, and my process automatically goes into the sleep S) state, without deallocating the 131GB it has taken up.
|
|
Given the bizarre for Python) error I am not explicitly deallocating anything), and the fact that this works nicely for smaller matrix sizes around 1.5 million by 2000), I am really not sure where to start to debug this.
|
|
As a starting point, I have set ulimit s unlimited before running, but to no avail.
|
|
Any help or insight into numpy s behavior with really large amounts of data would be welcome.
|
|
Note that this is NOT an out of memory error I have 196GB, and my process reaches around 131GB and stays there for some time before giving the error above.
|
|
As per suggestions, I ran Python with GDB.
|
|
Interestingly, on one GDB run I forgot to set the stack size limit to unlimited , and got the following output .
|
|
When I set the stack size limit to unlimited, I get the following .
|
|
This makes me believe the basic issue is with the numpy multiarray core module line 4 in the first output and line 18 in the second).
|
|
I will bring it up as a bug report in both numpy and scipy just in case.
|
|
Has anyone seen this before?","bug"
|
|
"6013","2997","export multiarray api to numpy.core.multiarray.
|
|
As recently [discussed on the mailing list] http mail.scipy.org pipermail numpy discussion 2013 February 065516.html), NumPy s API is unnecessarily tricky to link from a non C language because the `numpy.core.multiarray.
|
|
ARRAY API` lookup table of API pointers `PyArray API` in C) has a meaning that can only be determined by parsing the ` multiarray api.h` header file.
|
|
A simple improvement would be to export an ` ARRAY API NAMES` variable or similar) as well in Python, a simple list of strings corresponding to the symbols `PyArray GetNDArrayCVersion` etc.)
|
|
pointed to by the `ARRAY API` `PyArray API` array.
|
|
This should hopefully be a straightforward modification to `generate numpy api.py` and related files.
|
|
Or even just export the variables in the `numpy core code generators numpy api.py` file.","bug"
|
|
"6013","2998","need exported PyArray API) functions for Array API data structure access.
|
|
As recently [discussed on the mailing list] http mail.scipy.org pipermail numpy discussion 2013 February 065516.html), the NumPy API s array data structure information is unnecessarily hard to access from non C languages because the accessor functions `PyArray NDIM` etcetera are defined either as macros or as inline functions, neither of which are exported for dynamic linking.
|
|
Please consider exporting non inline) versions of these functions adding them to `PyArray API` for dynamic linking).
|
|
e.g.
|
|
you could call them `PyArray ndim` etcetera.
|
|
This is very similar to how, in the Python API, important macros like `Py XDECREF` have equivalent exported functions `Py DecRef` [ for runtime dynamic embedding ] http docs.python.org dev c api refcounting.html).)","enhancement"
|
|
"6013","2999","Assume we can use sys.stdout.fileno ) commit breaks weave.inline run under nosetests.
|
|
2766 seems to have broken our testsuite.
|
|
If I interpret it correctly, it seems to be a combination of weave using `numpy.distutils.exec command` and nose redirecting stdout to a `StringIO` instance.
|
|
This is similar to SAGE s problem reported in 2915 but they appear to be using their own replacement for sys.stdout so they can add the missing fileno attribute.
|
|
There is apparently a discussion going on for a while to fix issues like this at the nosetest side https github.com nose devs nose issues 6 However, it looks like it could be fixed on numpy s side with a relatively unintrusive `hasattr sys.stdout, fileno)`, doesn t it?
|
|
Here s the full stack trace of one of my crashes Traceback most recent call last) File usr lib python2.7 unittest case.py , line 332, in run testMethod ) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages nose case.py , line 197, in runTest self.test self.arg) File home jenkins workspace Linux Test OS linux PythonVersion python2.7 packages newest brian tests testinterface test directcontrol.py , line 253, in test poissoninput net.run 500 ms) File home jenkins workspace Linux Test OS linux PythonVersion python2.7 packages newest brian network.py , line 565, in run self.update ) File home jenkins workspace Linux Test OS linux PythonVersion python2.7 packages newest brian network.py , line 509, in update f ) File home jenkins workspace Linux Test OS linux PythonVersion python2.7 packages newest brian neurongroup.py , line 486, in update spikes = self.
|
|
threshold self) get spikes File home jenkins workspace Linux Test OS linux PythonVersion python2.7 packages newest brian threshold.py , line 179, in call extra compile args=self.
|
|
extra compile args) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages scipy weave inline tools.py , line 355, in inline kw) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages scipy weave inline tools.py , line 482, in compile function verbose=verbose, kw) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages scipy weave ext tools.py , line 367, in compile verbose = verbose, kw) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages scipy weave build tools.py , line 272, in build extension setup name = module name, ext modules = [ext],verbose=verb) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils core.py , line 186, in setup return old setup new attr) File usr lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File usr lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File usr lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils command build ext.py , line 232, in run self.build extensions ) File usr lib python2.7 distutils command build ext.py , line 448, in build extensions self.build extension ext) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils command build ext.py , line 345, in build extension kws) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils ccompiler.py , line 203, in CCompiler compile self.
|
|
compile obj, src, ext, cc args, extra postargs, pp opts) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils unixccompiler.py , line 35, in UnixCCompiler compile extra postargs, display = display) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils ccompiler.py , line 57, in CCompiler spawn s,o = exec command cmd) File home jenkins .jenkins virtual envs python2.7 newest local lib python2.7 site packages numpy distutils exec command.py , line 197, in exec command if with python and 0 or sys.stdout.fileno )== 1) AttributeError StringIO instance has no attribute fileno","bug"
|
|
"6013","3000","numpy.int32 != numpy.int32.
|
|
I have a function that converts all numpy variables to python ones before insertion into a mongo database.
|
|
.
|
|
But every once in a while a strange np.int32 slips through!
|
|
.
|
|
It took me forever to figure this out and I see another was bitten by this http projects.scipy.org numpy ticket 1246 Maybe isinstance behavior should be fixed but also the repr should represent the dtype.char.","bug"
|
|
"6013","3001","BUG take on object array sometimes?)
|
|
leaves self in inconsistent state when IndexError is raised.
|
|
Hello, pydata pandas 2892 is apparently due to the following, with Python 3.3 and latest dev numpy on 32 bit ubuntu) .
|
|
No segfault occurs if `astype` is omitted above.
|
|
I m not familiar with the numpy code, so I m not sure if my description of the issue in the subject line is 100 correct, but it seems to be the gist of it from testing.
|
|
Thanks in advance for any help with this!","bug"
|
|
"6013","3003","pickle of record array elements results in corruption and my cause segfault.
|
|
It seems that pickle of record scalars does not work properly and results in corruption.
|
|
That may even cause segfaults if a field happens to be an object.
|
|
Note that the first field seems to be ok.
|
|
Note, all tested on mac os X 64 bit) Here is a sample script that raises the problem .
|
|
Running in a console this results in .
|
|
Note the garbage in the second field of the unpickled object.
|
|
A similar script with objects .
|
|
results in a segfault .
|
|
My bet only first field gets unpickled and the rest is left uninitialized","bug"
|
|
"6013","3004","matrix.
|
|
add vs. matrix.
|
|
mul .
|
|
.
|
|
In other words, `matrix.
|
|
add ` gives a `matrix`, but `matrix.
|
|
mul ` returns `NotImplemented` or it just isn t implemented), and so falls back to `sympy.Matrix.
|
|
rmul `.
|
|
I don t particularly care which behavior is used, but it should be consistant.
|
|
Note that this is new in numpy 1.7.0.
|
|
Previously, both of the above would return a SymPy `Matrix`.","bug"
|
|
"6013","3008","NPY NO DEPRECATED API check is backwards.
|
|
There s an annoying warning on all of my files for basically every project that uses numpy .
|
|
Looking at the check in ``numpy core include numpy ndarraytypes.h`` .
|
|
The user is supposed to set ``NPY NO DEPRECATED API`` to the API version that they wish to check their codebase against for deprecated code, and the version `` defines`` are just integers.
|
|
So setting it to ``7`` should check that you are clean against numpy ``1.7``.
|
|
Also, the ``NPY API VERSION`` should be greater than the version you re trying to check against.
|
|
Lengthy explanation, skip this if you re good at boolean algebra...
|
|
Looking at the check, you don t want to include ``npy deprecated api.h``, so you want the inside of the check to be ``true`` so the ``not`` evaluates to ``false`` and the `` if`` is skipped.
|
|
The first check ``NPY NO DEPRECATED API`` is ``True``, the second check ``NPY API VERSION <= NPY NO DEPRECATED API`` is ``8 <=7``, which is ``false``.
|
|
So the inside evaluates to ``false``, we take the ``not`` of that and include the file.
|
|
Changing the check to this fixes it .","bug"
|
|
"6013","3011","Einsum does not work for multiple scalar inputs.
|
|
The following command raises an error .
|
|
although it should return 6 i.e., 3 2).","bug"
|
|
"6013","3013","Why no cotangent function for radians?.
|
|
I saw that there is a cotdg function to compute the cotangent of an angle in degrees.
|
|
Why its counterpart in radians is not present?","bug"
|
|
"6013","3015","array equal fails on string dtypes.
|
|
If I pass two ndarrays with string dtype to `array equal`, it blows up .
|
|
Expected result `True`","bug"
|
|
"6013","3016","Contains method is not consistent for subarrays.
|
|
The ` contains ` method is written to be used for a single array element.
|
|
However for example in list of list, ` contains ` does a check more equivalent to subarrays.
|
|
`in` must return a single boolean.
|
|
After some discussion on the list, there are three main possibilities 1.
|
|
The first item must be an element.
|
|
That means that an array `a` for `a in b` will normally be a simple error.
|
|
As nathaniel mentioned on the list).
|
|
2.
|
|
Do a list of list like comparison.
|
|
I.e.
|
|
`in` operates on the first dimension.
|
|
3.
|
|
Do some kind of subarray matching there are many different versions of this allowing different things) Point 2. seems wrong, since arrays are not list of lists.
|
|
Point 3. has some merit, it can go as far as allowing things similar to strings ` a in cat `, however there are some problems with the details.
|
|
Point 1. is the simplest and safest solution.
|
|
One problem with 2. is that for object arrays it can be not quite clear how to interpret for example a tuple list.
|
|
At this time there was not much discussion yet though), it seems that the best solution is to just raise an error i.e.
|
|
solution 1.).
|
|
Finding subarrays is better suited for a dedicated function.","enhancement"
|
|
"6013","3017","Inconsistency in indexing 0 d arrays with Ellipsis.
|
|
Attempting to set the element of a 0 d array by indexing it with ` ...,)` causes an IndexError, but similar operations with Ellipsis work correctly .
|
|
I assume this should be allowed?
|
|
It came up in some code along the following lines .
|
|
which fails to work when c is zero dimensional.","bug"
|
|
"6013","3018","Structured array allows non integer size but shouldn t. The creation of a dtype where any size is not a strictly positive integer should raise an error ValueError or TypeError).
|
|
In numpy 1.8.0.dev b8bfcd0 or 1.6.2, it gets instead silently truncated when an object with this .
|
|
The same for `ones` or `empty`.
|
|
A different case, that probably boils down to the same problem .
|
|
That probably shouldn t say too big, but invalid.
|
|
When I pass a complex number, numpy does complain about invalid shape .
|
|
I guess it should raise the same ValueError invalid shape for any size that is not a strictly positive integer.
|
|
Tested with the latest build, version 1.8.0.dev b8bfcd0, git revision b8bfcd02a2f246a9c23675e1650c3d316d733306.","bug"
|
|
"6013","3019","loadtxt is slow?.
|
|
Based on [this question] http stackoverflow.com q 15096269 1301710) on Stackoverflow I tested some ways to parse a file .
|
|
So `loadtxt` is very slow compared to a simple loop!
|
|
Note that this was also noted in [this answer] http stackoverflow.com a 15097561 1301710) to the same question where `genfromtxt` is faster than `loadtxt`!).
|
|
I m using 1.7.
|
|
Was something changed in `loadtxt` between 1.6 and 1.7 or what am I doing wrong?","bug"
|
|
"6013","3021","has key should be changed to key in .","bug"
|
|
"6013","3022","Update except syntax..
|
|
Replace except Exception, msg by except Exception as msg","bug"
|
|
"6013","3030","nanargmin broken when using axis kwarg.
|
|
The following snippet was broken by commit 8362e081291 import numpy as np np.nanargmin np.array [[ nan , nan ], [1, 2]], dtype=np.float), axis=1) Before the change this would yield [ 9223372036854775808 0] Now this raises a ValueError .
|
|
This is very likely a re incarnation of bug 643 in some form, but I m opening separately since this was working up until a month ago.","bug"
|
|
"6013","3031","2to3 basestring.
|
|
Python 2 isinstance a, basestring) has no analog in python 3.","bug"
|
|
"6013","3032","2to3 buffer vs memoryview.
|
|
memoryview hasn t been backported to python 2.6, only 2.7.","bug"
|
|
"6013","3033","2to3 idioms fail on python 2.
|
|
I think these are fixable failures look fixable.
|
|
The code in general looks better after the fixes.","bug"
|
|
"6013","3034","2to3 itertools fails on python 2.7.
|
|
These look fixable.
|
|
But do we really want to use itertools.izip instead of zip?
|
|
Only for very large files, I think.","bug"
|
|
"6013","3035","2to3 long will fail on python 2.7.
|
|
Python 3 only has int.
|
|
That is a problem.","bug"
|
|
"6013","3037","Linear least squares with MKL is suddenly very sensitive to input after version upgrade.
|
|
First, I am not 100 sure this is due to NumPy, it could well be Intel has changed something in their implementation.
|
|
Anyway, it will be good to have it documented somewhere on the web, and maybe there is something that can be done about it?
|
|
I am running an application that needs to find the unknown amplitude and phase of a sine function with known frequency.
|
|
I discovered that the linalg.lstsq function was suddenly misbehaving.
|
|
What I had done was upgrading OS to Scientific Linux 6.3), recompiling the latest numpy 1.7.0) with MKL as I had done previously, but now with the latest MKL that follows with version 2013.1.117 of the Intel suite).
|
|
I made the following test script .
|
|
I have tested it on numpy compiled the normal way no options, no Intel environmental variables, just python setup.py build ) and compiled with Intel compilers and MKL as described for 64 bit on http software.intel.com en us articles numpyscipy with intel mkl and the following site.cfg [mkl] library dirs = home applications intel composer xe 2013.1.117 mkl lib intel64 include dirs = home applications intel composer xe 2013.1.117 mkl include mkl libs = mkl def,mkl intel lp64,mkl sequential,mkl core lapack libs = The above script output is like this with standard numpy compile as expected) 100 [ 4.86957337e 02 5.34110751e 03 2.09610861e 00] 9.33021001596 200 [ 4.86997334e 02 5.34109942e 03 2.08323202e 00] 15.8074498248 400 [ 4.87007719e 02 5.34111513e 03 2.12535154e 00] 32.2812787113 800 [ 4.86997302e 02 5.34109373e 03 2.09251290e 00] 72.6769634985 1600 [ 4.86997470e 02 5.34109944e 03 2.10325586e 00] 133.946312709 3200 [ 4.86998548e 02 5.34109880e 03 2.11239048e 00] 272.671579315 6400 [ 4.87007892e 02 5.34110036e 03 2.11281644e 00] 534.042146499 12800 [ 4.87002423e 02 5.34110530e 03 2.10787976e 00] 1072.48157652 25600 [ 4.86999417e 02 5.34109930e 03 2.11220371e 00] 2129.87663949 51200 [ 4.86995656e 02 5.34109902e 03 2.11009978e 00] 4261.57387148 102400 [ 4.87000230e 02 5.34109903e 03 2.11173973e 00] 8519.57051824 204800 [ 4.87002204e 02 5.34110040e 03 2.11092556e 00] 17083.6947101 409600 [ 4.86999883e 02 5.34109949e 03 2.11209901e 00] 34067.4717676 819200 [ 4.86999624e 02 5.34109942e 03 2.11141873e 00] 68311.5049075 1638400 [ 4.86999957e 02 5.34109994e 03 2.11185559e 00] 136809.17059 And like this with the MKL compile 100 [ 4.86977088e 02 5.34117438e 03 2.11153548e 00] 7.54282461483 200 [ 4.86967406e 02 5.34112226e 03 2.12100667e 00] 16.4320682438 400 [ 4.87024012e 02 5.34110389e 03 2.10866724e 00] 30.2492304951 800 [ 5.01498162e 04 1.57589440e 06 1.68693811e 08] 1.13214757643e 19 1600 [ 5.57560725e 04 1.47778117e 06 3.37186430e 08] 9.07121758628e 19 3200 [ 8.11002072e 04 1.41033602e 06 6.74166645e 08] 7.26280220711e 20 6400 [ 1.38426677e 05 1.36980407e 06 1.34812497e 09] 5.81261897473e 21 12800 [ 2.56237521e 05 1.34750103e 06 2.69603292e 09] 4.65102881282e 22 25600 [ 4.93384385e 05 1.33564315e 06 5.39179644e 09] 3.72112675765e 23 51200 [ 9.68441305e 05 1.32954258e 06 1.07834510e 10] 2.97709382163e 24 102400 [ 1.91890860e 06 1.32650199e 06 2.15665725e 10] 2.3817106729e 25 204800 [ 3.82003630e 06 1.32495338e 06 4.31329078e 10] 1.90539095992e 26 409600 [ 7.62238107e 06 1.32423851e 06 8.62655378e 10] 1.52432030925e 27 819200 [ 1.52271691e 07 1.32380692e 06 1.72531373e 11] 1.21946739889e 28 1638400 [ 3.04366929e 07 1.32358614e 06 3.45062231e 11] 9.75573777678e 28 Notice the drastic change between 400 and 800 time steps ...
|
|
It is scary that it suddenly goes from working to blowing up.
|
|
I know that the values in b will be closer and closer for each increase in number of time steps and that the problem above could probably be mitigated by taking a sample of point in b instead of the whole array, but it has allways worked to push all data through before.
|
|
I do not have results for the previous MKL installation, but I never experienced anything like this as far as I can remember.
|
|
On Windows with Numpy 1.6.1 from http www.lfd.uci.edu gohlke pythonlibs I see very similar results to the good results above, and it seems like that is also using MKL.
|
|
Feel free to tell me to take this up with Intel support.
|
|
I feel better for now knowing that this issue is at least googlable for the next person that hits this bug","bug"
|
|
"6013","3038","Old numeric, `sys.maxint`, ` .
|
|
nonzero`.
|
|
These only show up in oldnumeric when 2to3 with the 2to3 renames and nonzero fixers.
|
|
The replacement for `sys.maxint` in python3 is `sys.maxsize`, but the two aren t the same, the latter the size of `Py ssize t`.
|
|
Likewise the the replacement for ` nonzero` is ` bool `, but anyone using old numeric is likely to have made the change.
|
|
Because oldnumeric will go away, I suggest we leave those attributes unchanged and advertize that oldnumeric isn t supported in python 3.","enhancement"
|
|
"6013","3040","Add `from import division` to every python file..
|
|
This should be harmless and will provide a place to put the `print function` and `absolute import` options in the future.
|
|
The division change will also provide more testing of this change as it will be present in python 2 which is more heavily used at this point.
|
|
This is addressed by gh 3026","enhancement"
|
|
"6013","3041","2to3 run `apply` fixer.","bug"
|
|
"6013","3042","2to3 run `basestring` fixer.","bug"
|
|
"6013","3043","2to3 run `buffer` fixer.","bug"
|
|
"6013","3044","2to3 run `callable` fixer.","bug"
|
|
"6013","3046","Inconsistent numpy.angle behavior.
|
|
Following up on a Stackoverflow question of mine http stackoverflow.com questions 15137003 strange complex phase displayed I am bringing this here.
|
|
The issue is that angle called on a zero matrix causes inconsistent values of the phase of a complex number.
|
|
I understand the underlying float implementation is the cause, but I think it is important to understand e.g.
|
|
Matlab does not have this problem, and instead does the more sensible thing and returns 0.
|
|
For completeness, the numpy script .
|
|
And the corresponding Matlab code .
|
|
I agree the phase of a zero isn t strictly defined and GNU Octave... yuck, handles it even worse than numpy), but when showing these types of images, this is a hindrance and having to code around this seems artificial and adds an extra line of code ; )).
|
|
If it has to be undefined, make it at least numerically handy undefined.","bug"
|
|
"6013","3048","2to3 run `exec` fixer..","bug"
|
|
"6013","3050","2to3 run `dict` fixer..
|
|
This works, but the changes are extensive, we may want to take a closer look here.","bug"
|
|
"6013","3051","2to3 run `execfile` fixer.","bug"
|
|
"6013","3052","2to3 run `exitfunc` fixer..","bug"
|
|
"6013","3053","2to3 run `filter` fixer.","bug"
|
|
"6013","3054","2to3 run `funcattrs` fixer.","bug"
|
|
"6013","3057","distutils.misc util.get shared lib extension is python ext=False) returns d.so with python dbg.
|
|
distutils.misc util.get shared lib extension is python ext=False) returns d.so with python dbg with numpy 1.7.0 on debian unstable and numpy 1.8.0.dev 0934653 ubuntu 12.10 is this really intentional?
|
|
I would have though is python ext=False gets me the the regular library extension which is just .so on linux.
|
|
If not the documentation should make this clear.","bug"
|
|
"6013","3058","2to3 run `funcattrs` fixer.","bug"
|
|
"6013","3060","2to3 run `getcwdu` fixer.","bug"
|
|
"6013","3061","2to3 run `has key` fixer.","bug"
|
|
"6013","3062","2to3 run `idioms` fixer.
|
|
This has some problems but it does make better looking code in a number of places.
|
|
Worth fixing up the result.","bug"
|
|
"6013","3063","2to3 run `input` fixer.","bug"
|
|
"6013","3064","2to3 run `intern` fixer.","bug"
|
|
"6013","3065","2to3 run `isinstance` fixer.","bug"
|
|
"6013","3066","2to3 run `long` fixer.
|
|
Known to be hairy, needs work.","bug"
|
|
"6013","3067","2to3 run `long` fixer.
|
|
Known to need work.
|
|
Python 3 doesn t have long.","bug"
|
|
"6013","3068","2to3 run `map` fixer.","bug"
|
|
"6013","3069","2to3 run `metatclass` fixer.","bug"
|
|
"6013","3070","2to3 run `methodattrs` fixer.","bug"
|
|
"6013","3071","2to3 run `ne` fixer.","bug"
|
|
"6013","3072","2to3 run `next` fixer.","bug"
|
|
"6013","3073","2to3 run `nonzero` fixer.","bug"
|
|
"6013","3074","2to3 run `numliteral` fixer.
|
|
This refers to the L suffix for python 2 long integers.
|
|
Will be tricky to fix.","bug"
|
|
"6013","3075","2to3 run `operator` fixer.","bug"
|
|
"6013","3076","2to3 run `paren` fixer.","bug"
|
|
"6013","3077","2to3 run `raise` fixer.","bug"
|
|
"6013","3078","2to3 run `print` fixer.","bug"
|
|
"6013","3079","2to3 run `raw input` fixer.","bug"
|
|
"6013","3080","2to3 run `reduce` fixer.","bug"
|
|
"6013","3081","2to3 run `reduce` fixer.","bug"
|
|
"6013","3082","2to3 run `renames` fixer.
|
|
.
|
|
These aren t the same things.
|
|
In python 2 it is the size of the python small integer, in python 3 there is no small integer and it refers to Py ssize t. This only affects oldnumeric, we should probably leave as is and raise an error if run in python 3.","bug"
|
|
"6013","3083","2to3 run `repr` fixer.
|
|
replace the backtic notation `a` by `repr a)`.","bug"
|
|
"6013","3084","2to3 run `setliteral` fixer after 2.6 support is dropped.. Set literals are in not available in Python 2.6 and Numpy will likely continue support for a number of years.
|
|
The set literals are nice, but not needed, the current syntax in Numpy is forward portable.
|
|
So mark this as a task but attach it to no milestone.","enhancement"
|
|
"6013","3085","2to3 run `standarderror` fixer.","bug"
|
|
"6013","3086","2to3 run `sys exc` fixer.","bug"
|
|
"6013","3087","2to3 run `throw` fixer.","bug"
|
|
"6013","3088","2to3 run `tuple params` fixer.","bug"
|
|
"6013","3089","2to3 run `unicode` fixer.
|
|
This is probably going to be major work.","bug"
|
|
"6013","3090","2to3 run `urllib` fixer.","bug"
|
|
"6013","3091","2to3 run `ws comma` fixer.","bug"
|
|
"6013","3092","2to3 run `xrange` fixer.. Replaces `xrange` by `range` and replaces `range ...)` by `list range ...))` where needed.","bug"
|
|
"6013","3093","2to3 run `xreadlines` fixer..
|
|
This should probably be like xrange, we should use xreadlines every where until we are ready to dispense with 2to3.","bug"
|
|
"6013","3094","2to3 run `zip` fixer.
|
|
zip is now an iterator.
|
|
This is simiar to the filter and map changers.","bug"
|
|
"6013","3095","2to3 run range fixer.. Before doing this, replace range by xrange in for loops before running the range fixer to get the other range uses.
|
|
When the xrange fixer is run later, things will come right.","bug"
|
|
"6013","3096","2to3 replace readlines by xreadlines in for loops.
|
|
The strategy here is to protect readlines against the `readlines` fixer.","bug"
|
|
"6013","311","npyio.py > recfromcsv seems to have a bug that ignores any dtypes provided.
|
|
I found this due to [someone s problem loading large integers with recfromcsv] http stackoverflow.com questions 11064908 getting numpy overflow despite declaring dtype int64 11065583 comment14481167 11065583).
|
|
After playing with it, I think it is simply ignoring all the dtype information provided.
|
|
[Line 1859 in npyio.py, part of recfromcsv )] https github.com numpy numpy blob master numpy lib npyio.py L1859) tries to either bring in the dtypes passed or set None.
|
|
However, it actually brings in a kwarg that I couldn t find anywhere else called update which looks suspiciously like a copied mistype from earlier in the line.
|
|
If I change update > dtype , then the types get picked up as expected.
|
|
More explicitly, in recfromcsv, change `kwargs.update dtype=kwargs.get update , None),` to this `kwargs.update dtype=kwargs.get dtype , None),` I found this in the github repo although I may have been looking at the wrong branch) as well as the 1.6.2 release that I am using.
|
|
apologies for double posting this.
|
|
I [posted in trac] http projects.scipy.org numpy ticket 2168) first and then heard that maybe that won t be looked at.
|
|
I haven t found a way to delete or close that ticket.)","bug"
|
|
"6013","3110","Bug in np.matrix indexing.
|
|
This .
|
|
Occurs only with `np.matrix`, `np.ndarray` indexing works OK. Indexing with ` 1` instead of `np.array 1)` also works OK.
|
|
This bug has probably been around forever, occurs also on Numpy 1.2.0...","bug"
|
|
"6013","3111","ma.MaskedArray.mean ) ma.mean ).
|
|
When these functions are run on arrays which are completely masked, they return a MaskedConstant type.
|
|
What s the rationale behind this?
|
|
Why not just return a nan float, like a normal ndarray does?
|
|
>>> import numpy as np >>> x = np.array []) >>> x.mean ) usr lib python2.7 site packages numpy core methods.py 57 RuntimeWarning invalid value encountered in double scalars ret = ret float rcount) nan >>> x msk = np.ma.array [5,5,5,5], mask=[1,1,1,1]) >>> x msk.mean ) masked","bug"
|
|
"6013","3112","AttributeError vectorize object has no attribute name in Python 2.
|
|
I was trying to decorate `np.vectorize` but if one tries to use functions like `update wrapper` from Python functools it fails because in Python 2 vectorize objects have no ` name ` attribute.
|
|
Surprisingly, this is not an issue in Python 3 because in this version it does have ` name `.","bug"
|
|
"6013","3113","Disable unused 2to3 fixers.
|
|
After applying 2to3 fixers permanently to the source, we should disable them from running during install, so that the need for them doesn t creep back in.
|
|
IPython s setup.py has an example https github.com ipython ipython blob master setup.py L268","bug"
|
|
"6013","3116","Memory leak using division and trace ) with Intel MKL.
|
|
I have discovered that the following code produces a memory leak with NumPy 1.7.0 and MKL 10.3.6.
|
|
On Stackoverflow I was told that this is likely a NumPy bug.
|
|
With my outdated setup of NumPy 1.4.1 and ATLAS it works fine.
|
|
import numpy as np t = np.random.rand 10,10) while True t = t np.trace t) http stackoverflow.com questions 15191391 how to avoid this four line memory leak with numpymkl","bug"
|
|
"6013","3118","Sum of uint64 and int is a float risk of data precision loss).
|
|
Please consider the following import numpy as np a=np.uint64 1) a 1 2.0 Adding an uint64 to an int gives a float.
|
|
This is particularly problematic in indexing.
|
|
When you have an index i that is a numpy int type and you do something like a[i 1] you are in trouble.","bug"
|
|
"6013","3119","Polyfit is returning coefficients from low to high instead of high to low.. Docs say that np.polyfit returns Polynomial coefficients, highest power first.
|
|
If y was 2 D, the coefficients for k th data set are in p[ ,k].
|
|
However, at least for the linear case, it is actually returning coefficients lowest power first.
|
|
Thus, for a fit y=mx b, polyfit returns b,m) and not m,b) as the docs indicate.
|
|
y=np.linspace 0,10,10) x=y plt.plot x,y) m,b=polyfit x,y,1) Returns highest power first print m is , m, b is , b Output gives m is 0.0 b is 1.0 See that it s reversed?","bug"
|
|
"6013","312","numpy.load documentation regarding gzip compression does not match implementation.
|
|
The documentation of numpy.load indicates If the filename extension is ``.gz``, the file is first decompressed.
|
|
Contrary to numpy.loadtxt, however, numpy.load does not seem to check the extension, but rather whether the file object is an instance of gzip.GzipFile see https github.com numpy numpy blob master numpy lib npyio.py L360) The following example seems to be correct according to the documentation, but fails in practice .
|
|
The easy solution is to update the documentation to reflect this behavior, though I believe it would be preferable to correct the implementation instead I can prepare a patch if requested).","bug"
|
|
"6013","3125","np.finfo np.clongdouble).precision wrong on powerpc.
|
|
Hello, I m forwarding here the Debian bug http bugs.debian.org 702169 .
|
|
Thanks, Sandro","bug"
|
|
"6013","3126","Inconsistent array assignment for structured arrays.
|
|
When assigning arrays inside structured arrays without a left sided `[ ]`, only the first element gets actually assigned.
|
|
This is shown in the following code snippet, where a structured array is initialized with zeros, and then filled by assignment of an array, and the assignment of a structured array.
|
|
The assignment of the array only assigns the first value, where the assignment of the structured array assigns all values .
|
|
The expected result is that all elements of `x[0][ field ]` are 1.
|
|
However, `x[0][ field ]` recognizes some part of the assigned array, as it will complain about broadcasting problems for wrong sizes .
|
|
As mentioned in the beginning, adding `[ ]` on the left side In [6]) will correctly set all elements to 1.
|
|
If the shown assignment is not allowed, I would have expected that an exception gets raised?","bug"
|
|
"6013","3129","Backport gh 365.
|
|
Please backport gh 365 to the 1.7.x branch.
|
|
It s a one liner and trivially correct.","bug"
|
|
"6013","3137","Adopt Mersenne Twister 64bit?.
|
|
The current implementation of the RNG seems to be MT19937 32.
|
|
Since 64 bit machines are common nowadays, I am suggesting adding or upgrading to MT19937 64.
|
|
Thoughts?","bug"
|
|
"6013","3138","Unsafe type conversion in numpy.bincount when passing list instead of array.
|
|
`np.bincount` is [supposed] http docs.scipy.org doc numpy reference generated numpy.bincount.html) to raise a `TypeError` if the input array like contains elements of float or complex type.
|
|
However, this is not the observed behavior when the input is an array like in this case, a list), rather than an array.
|
|
.","bug"
|
|
"6013","3139","Numpy scalar vs Python scalar comparison slow.
|
|
Numpy scalar comparisons can be very slow .
|
|
This slowdown seems to be related to type conversions .","bug"
|
|
"6013","314","NumPy fails to compile in Ubuntu 11.10.
|
|
The latest NumPy version 36681eb2c4b731d5f8e639607aff41404aa8fc45) fails to compile in Ubuntu 11.10 with the error .
|
|
The fix is either to apply the following patch as is done in the official Debian Ubuntu NumPy deb package) .
|
|
Or simply .
|
|
As such, it should be at least documented in the INSTALL.txt.","bug"
|
|
"6013","3140","np.tile modifies the mask of the inputted masked array.
|
|
Here is an example of the problem .
|
|
The issue is that the call to `np.tile` changed `A. mask` from `array [False, False], dtype=bool)` to `array [[False, False]], dtype=bool)`, which made the ` repr )` function return the error.
|
|
Digging deeper, I found that the issue is with the call ` nx.array A,copy=False,subok=True,ndmin=d)` in the [tile function] https github.com numpy numpy blob master numpy lib shape base.py L829).
|
|
Before this call, `A` has the correct ` mask` and after this call, `A` s ` mask` changes.
|
|
Does anyone have any ideas how to fix this?
|
|
This is on Fedora rawhide with python 2.7.3 34.fc19.x86 64 numpy 1.7.0 1.fc19.x86 64","bug"
|
|
"6013","3142","Einsum gives wrong results sometimes.
|
|
I have encountered a very weird behaviour with einsum.
|
|
I try to compute something like R A R , where denotes a kind of matrix multiplication .
|
|
However, for particular shapes of R and A, the results are extremely bad.
|
|
I compare two einsum results First, I compute in two einsum calls as R A) R .
|
|
Second, I compute the whole result in one einsum call.
|
|
However, the results are significantly different for some shapes.
|
|
My test .
|
|
Output .
|
|
So, for D={5,6,7,9}, allclose returns False and the error norm is HUGE.
|
|
It doesn t seem like just some small numerical inaccuracy because the error norm is so large.
|
|
I don t know which one is correct Y1 or Y2) but at least either one is wrong in my opinion.
|
|
I ran the same test several times, and each time same values of D fail.
|
|
If I change the shapes somehow, the failing values of D might change too, but I usually have several failing values.
|
|
I m running the latest version from github commit bd7104cef4) under Python 3.2.3.
|
|
With NumPy 1.6.1 under Python 2.7.3 the test crashes and Python exits printing Floating point exception .
|
|
This seems so weird to me that I wonder if I m just doing something stupid..","bug"
|
|
"6013","3143","Memmap cannot use existing file handles..
|
|
Currently if one does .
|
|
an exception is raised stating that .
|
|
However, numpy.memmap is more than happy to work with files.
|
|
This restriction should be lifted if at all possible as it would be nice to be able to seek to a point in an open file and have numpy memmap this.","enhancement"
|
|
"6013","3144","Use macro access to flatiter objects in mtrand.pyx.
|
|
Cython 0.18 s `numpy.pxd` removes direct member access to `flatiter` and other objects in favor of the macros.
|
|
`mtrand.pyx` still uses direct member access.
|
|
E.g.
|
|
https github.com numpy numpy blob a530098e816e1164efdb356ac2280ba25bfb1c7f numpy random mtrand mtrand.pyx L178","bug"
|
|
"6013","3145","ma.empty like copies reference to mask instead of making a new mask.
|
|
ma.empty like on a masked array creates a new masked array with the mask pointing to the same mask as the source array, so if the target array mask is changed, the mask of the source array changes as well.
|
|
Code sample <code> from numpy import \ create some random data a = random.random 5,5)) \ make this a masked array a = ma.masked greater a, 0.3) \ use empty like to create a new array b = ma.empty like a) \ change b, in particular do something that sets the mask to False b[ , ] = 2.0 \ check whether at least one element in a is still masked print a.mask.any ) < code> You will see that this destroys the mask of a.
|
|
Note that this issue is similar to https github.com numpy numpy issues 2572, except that this exists even for numpy.ma.empty like, which, being from the numpy.ma module, should in principle know how to handle masked arrays.","bug"
|
|
"6013","3146","test build.py gives NameError global name libfile is not defined.
|
|
There is an undefined variable name in test build.py.
|
|
In numpy 1.7.0 running python c import numpy; numpy.test ) gives the following error ERROR test lapack test build.TestF77Mismatch) Traceback most recent call last) File var local mol pythonenv flynerves lib python2.6 site packages numpy testing decorators.py , line 146, in skipper func return f args, kwargs) File var local mol pythonenv flynerves lib python2.6 site packages numpy linalg tests test build.py , line 47, in test lapack asbytes nested [ libg2c , libgfortran ])) File var local mol pythonenv flynerves lib python2.6 site packages numpy linalg tests test build.py , line 30, in grep dependencies stdout = self.get dependencies file) File var local mol pythonenv flynerves lib python2.6 site packages numpy linalg tests test build.py , line 25, in get dependencies raise RuntimeError Failed to check dependencies for s libfile) NameError global name libfile is not defined Ran 4817 tests in 18.223s FAILED KNOWNFAIL=5, SKIP=2, errors=1)","bug"
|
|
"6013","3147","constructing bytearray or bytes from numpy array of length 1. tested with numpy 1.6.2, python 3.2.3.
|
|
Win7, 64 bit.
|
|
When creates a bytes object from a numpy array of length 1, the result is a bytes string with the length of the value of the single element, not a single byte equal to the single element.
|
|
This is inconsistent with creating a bytes string from a list.
|
|
Since an array of length 1 is still an array and still iterable, the result should be a byte string matching the contents of the array.
|
|
Example \>\>\> bytes np.array [3], np.uint8)) b \x00\x00\x00 \>\>\> bytes [3]) b \x03 I would expect both to give the second result.
|
|
The same behavior is observed with bytearray.","bug"
|
|
"6013","315","Build with NPY SEPARATE COMPILATION=1 fails.
|
|
Building with .
|
|
fails after recent changes works still in 651ef74c4ebe7d) .
|
|
Time to scrap the one file build?","bug"
|
|
"6013","3150","Probable bug in core src multiarray dtype transfer.c.
|
|
strided to strided one to n with finish contains what looks like a bug.
|
|
NpyAuxData subdata = d >data, data finish src = data finish src; data finish src gets set to itself.
|
|
I guess it s supposed to be data finish src = d >data finish src;","bug"
|
|
"6013","3151","OSError with numpy.fromfile ) on python3 32 bit) after seeking on a big file.
|
|
Code !
|
|
usr bin env python3 import os import sys assert sys.maxsize == 0x7fffffff path = <some big file over 2 GiB)> filesize = os.path.getsize path) assert filesize >= sys.maxsize f = open path, rb ) f.seek sys.maxsize) numpy.fromfile f, numpy.uint8, 1) f.seek filesize 1) numpy.fromfile f, numpy.uint8, 1) Either OSError [Errno 22] Invalid argument or OSError first argument must be an open file happens.
|
|
Only Python 3 32 bit) seems to be affected.
|
|
Tested on Windows 7 and Ubuntu 12.10.","bug"
|
|
"6013","3154","Nditer bug for multiple reduction axes.
|
|
Changing in `op axes` to `[ 0,1,2), 0, 1, 1)]` in `test iter buffered reduce reuse` of https github.com seberg numpy blob da86ee9792c0d742df5d781a0dd5bc8bb53888ee numpy core tests test nditer.py makes the test fail.
|
|
However, the double reduction should work fine as far as I can see.
|
|
This seems not related to buffer reuse, since skipping the buffer reuse completely still has the test failing.
|
|
So there seems to be some bug with multiple reductions at least in combination to non standard strides.
|
|
I am not sure if it is, but it does seem like it may be a big issues.","bug"
|
|
"6013","3155","random sample for type float32.
|
|
Hi Numpy I need a numpy.float32 array with a distribution between [0...1).
|
|
I can not find a way to generate this array using the existing numpy.random tools as converting from the default double to float causes the distribution to change to [0..1].
|
|
You can see the problem in the code below.
|
|
Can you suggest a better approach or should an optional argument for dtype be added to the existing methods?
|
|
Thanks David import numpy sz = 10000000 rnd f32 = numpy.zeros sz,numpy.float32) while True rnd f64 = numpy.random.rand sz) Creates doubles!
|
|
rnd f32[ ] = rnd 64 cast to float32 if rnd f32 >= 1.0).any ) print FooBar!
|
|
print Rnd64 Max s max rnd f64) print Rnd32 Max s max rnd f32) break","bug"
|
|
"6013","3156","MaskedArray.filled does not preserve array order.
|
|
Title says it all I have an F contiguous masked array which I want for scikit learn purposes) but as soon as I fill the masked values I get a C contiguous version.
|
|
This is because ndarray.copy defaults to C order and is [used by MaskedArray.filled] https github.com numpy numpy blob master numpy ma core.py L3449).
|
|
I believe fixing this simply requires changing that line to `result = self.
|
|
data.copy A )` and similarly for line 3444.
|
|
As far as I can tell that s all of it.
|
|
apologies if there is some reason for this behavior that I don t understand.
|
|
If there is, I would be glad to hear it)","bug"
|
|
"6013","3157","git master numpy setup is broken on mac 10.8.3 and linux.
|
|
I did a ``python setup.py install`` and I get tons of warnings and two errors.
|
|
.","bug"
|
|
"6013","3158","List of fixes that may be backportable.
|
|
I thought there should be a bit of a list of things that might be worth back porting, so this is a start.
|
|
It is the output from `git log master maintenance 1.7.x no merges pretty=oneline abbrev commit | grep BUG` I guess grepping BUG does not cover everything though), minus those that already had a commit with the same commit message in the 1.7.x branch on march 19th.
|
|
I started removing some by hand, but not many yet as for many I simply don t know if they effect 1.7. at all.
|
|
Please feel free to edit to remove unimportant or too complex ones and marking those that probably should be back ported have started marking a few that I thought were good with > , though not sure how much needs back porting).
|
|
.","bug"
|
|
"6013","3159","New dtype behavior in numpy 1.7.0.
|
|
Is this a bug note the different dtypes of the output)?
|
|
In [1] np.array [1, 2, 3], dtype=str) Out[1] array [ 1 , 2 , 3 ], dtype= |S1 ) In [2] np.array [1, 2, 3]).astype str) Out[2] array [ 1 , 2 , 3 ], dtype= |S24 ) This behavior is new to numpy 1.7.0 and breaks some unit tests in one of my packages).","bug"
|
|
"6013","316","NPY NO EXPORT should be a private macro.
|
|
One should also handle the extern vs no extern inside this macro to avoid the ifdef NPY SEPARATE COMPILATION dance for every variable definition.","bug"
|
|
"6013","3160","SandboxViolation easyinstalling 1.7.0 on Mac OS X 10.8.3.
|
|
I didn t see this error reported, apologies if it is a duplicate my current numpy is 1.6.2. .","bug"
|
|
"6013","3161","Single element array passed to `np.float64` returns as numpy scalar, not array.
|
|
If a single element array is passed to `np.float64 )` in the following way, it is returned not as an array but as a numpy scalar.
|
|
I have not observed this behavior when converting in this way via other dtypes, though my tests are not exhaustive.
|
|
Code to reproduce below .
|
|
This was observed on both NumPy 1.6.1 and 1.6.2.
|
|
I realize a workaround is to use the `.astype )` method, and that would probably be more correct.
|
|
However, this functionality is exposed and I do not believe it is expected behavior.","bug"
|
|
"6013","3164","Why does array priority not take precedence over array ?.
|
|
If I run the following dummy example .
|
|
I get .
|
|
which in my view is incorrect `` array `` should not get called because `` array priority `` is higher than that for ``np.array [1])``).
|
|
If I remove the `` array `` method, I get the expected result .
|
|
as expected.
|
|
This suggests to me that the presence of `` array `` takes precedence over the `` array priority ``.
|
|
Is there a reason for this, or is it an oversight?
|
|
Most importantly, is there a way to get around this, and ensure that if `` array `` is present, `` rmul `` still gets called?","bug"
|
|
"6013","3165","ERROR test exec command.test exec command stderr.
|
|
Numpy1.7.1rc1 fails to run one test on Python 3.x for Windows msvc MKL builds) .
|
|
The test passes when opening the TemporaryFile in ASCII instead of binary mode .","bug"
|
|
"6013","3171","ma.polyfit ) returns nans when there are masked nans in the input.
|
|
This used to return almost) [1., 1.]
|
|
but returns [nan, nan] in numpy 1.7. import numpy as np x = np.ma.array [np.nan, 1., 2.
|
|
], mask=[True, False, False]) y = np.ma.array [np.nan, 2., 3.
|
|
], mask=[True, False, False]) np.ma.polyfit x, y, 1) Changing the last line of ma.ployfit ) this way seems to work for me return np.polyfit x.filled ), y.filled ), deg, rcond, full, w, cov) But it does not make much sense to me, to be honest.","bug"
|
|
"6013","3172","2to3 run `import` fixer.","bug"
|
|
"6013","3173","Segmentation fault when size argument to numpy.random.multinomial has type numpy.uint32.. Steps to reproduce Python 2.7.3 default, Aug 1 2012, 05 14 39) [GCC 4.6.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.random.multinomial 1,[0.5,0.5],np.uint32 1)) array [0, 1, 1]) >>> np.uint32 1) 1 >>> np.random.multinomial 1,[0.5,0.5],np.uint32 1)) array [0, 1, 0]) >>> np.random.multinomial 1,[0.5,0.5],np.uint32 1)) array [0, 1, 0]) >>> np.random.multinomial 1,[0.5,0.5],np.uint32 1)) glibc detected python free ) invalid next size fast) 0x0000000001bbfb90 It also seems that numpy.random.multinomial behaves incorrectly when the `size` argument is of type `numpy.uint32` in the above example, each vector is of length 3 whereas the expected behaviour is an array of length 2.
|
|
Some details about the platform >>> numpy.
|
|
version 1.6.1 lsb release a No LSB modules are available.
|
|
Distributor ID Ubuntu Description Ubuntu 12.04.1 LTS Release 12.04 Codename precise uname ar Linux mlui 3.2.0 38 generic 61 Ubuntu SMP Tue Feb 19 12 18 21 UTC 2013 x86 64 x86 64 x86 64 GNU Linux","bug"
|
|
"6013","3174","`int64` range documentation has missing negative sign.. From http docs.scipy.org doc numpy user basics.types.html .
|
|
That should presumably read .","bug"
|
|
"6013","3175","segmentation fault with numpy.array ) from bytearray.
|
|
I experience segmentation faults on certain systems when constructing a numpy array from a bytearray.
|
|
The following code shows the problem .
|
|
I am getting segfaults on these systems Python 2.6.6, Numpy 1.6.2, 32bit Python 2.6.6, Numpy 1.7.0, 32bit Python 2.6.6, Numpy 1.6.2, 64bit Python 2.6.6, Numpy 1.7.0, 64bit In contrast, the above example is running on this system Python 2.7.3, Numpy 1.6.2, 64bit So it seems the problem is connected to Numpy s interaction with Python 2.6?
|
|
Using `np.frombuffer` instead of `np.array` works on all systems .","bug"
|
|
"6013","3176","Dtype descr inconsistency with invisible fields.
|
|
Declaring a dtype with the dict name , format , offset format could result in invisible fields.
|
|
The same dtype however is not reproduced by its own descr that uses the list of tuples format.
|
|
The invisibility does matter especially at comparisons sorting.
|
|
In the following case e.g.
|
|
B and C should behave the same way .
|
|
Also, on save and load into .npy file, the invisible field becomes visible.","bug"
|
|
"6013","3177","order of items returned by boolean indexing and nonzero.
|
|
What is the order of elements returned by boolean indexing?
|
|
In other words, what are the order of indices returned by nonzero?
|
|
Are they always sorted in the 1 D case?)
|
|
The order should be mentioned in the docs.
|
|
If .
|
|
then the result of A[flag] always looks to be [5,6,7] even if A is created with order= F .
|
|
I guess the order is actually inherited from flatiter that says Iteration is done in C contiguous style, with the last index varying the fastest.
|
|
but this should be confirmed.","bug"
|
|
"6013","3180","2to3 run `imports` fixer.","bug"
|
|
"6013","3181","2to3 run `imports2` fixer.","bug"
|
|
"6013","3183","allclose and assert allclose use different defaults.
|
|
numpy.testing.assert allclose, at least from the documentation, seems to be equivalent to asserting allclose.
|
|
However, in practice the two are not equivalent, because they use different default arguments.
|
|
Specifically, allclose uses default atol=1e 08, while assert allclose uses default atol=0.","bug"
|
|
"6013","3184","np.genfromtxt hates text files.
|
|
OK, the lead is a bit extreme.
|
|
However, in python3 both genfromtxt and savetxt, and auxiliary functions expect bytes strings, generators returning bytes strings, and files opened in binary mode.
|
|
That seems a bit extreme for functions advertised to deal with text and text files.
|
|
Among other things, the file encoding can t be specified.
|
|
I think what happened here is that turning everything into bytes looked like the easy way to deal with unicode, but it may have ended up papering over problems.","bug"
|
|
"6013","3189","DEP in 1.8. nditer oa ndim == 0 for no op axes..
|
|
The use of `oa ndim == 0` to signal that no `op axes` `itershape` is passed into `AdvancedNew` is deprecated starting with Numpy 1.8. by gh 3104.
|
|
Finalizing this requires editing `numpy core src multiarray nditer constr.c` to remove .","bug"
|
|
"6013","319","Test NumPy 1.7 against as many packages as possible.
|
|
This is important to find things that break either at compile time or run time) and clearly document how to upgrade to NumPy 1.7 in the release notes.
|
|
Packages to test at least feel free to edit, add more packages) scipy, matplotlib, pandas, h5py Also try to simply update the package in Debian Ubuntu and see if it works well with other packages.","bug"
|
|
"6013","3190","What is an invalid value in a NumPy Array?.
|
|
I apologize for the vagueness of this bug, but I do not know exactly how to summarize it, nor even if it really is a bug.
|
|
I have encountered arrays on which any arithmetic operation e.g.
|
|
`arr 1`) generates `RuntimeWarning invalid value encountered in multiply` and does not return any data!
|
|
Neither `NaN`s nor `Inf`s reproduce this behaviour, because arithmetic operations on such values are well defined `NaN` or `Inf` respectively).
|
|
I assumed this was the result of using uninitialized arrays `ma.empty )`) however, even after generating thousands of large 1000) uninitialized arrays with `ma.empty` and performing `arr 1` on them, I have yet to reproduce the behaviour.
|
|
So, the invalid value appears to be something very particular, and I have no idea where it comes from.
|
|
The reason this is important is because if these errant values exist in a masked array, the error and failure to return data also occurs even if the invalid value is masked .
|
|
Can anyone clarify this?
|
|
Many thanks.
|
|
Here is a pickle of an offending array .
|
|
And here is a minimal failing example that loads that data .","bug"
|
|
"6013","3192","f2py crashes with UnboundLocalError exception.
|
|
This is a copy http projects.scipy.org numpy ticket 1932 since that one seems no longer in use and no one responded to it in almost 2 years, even though I posted a patch that solves the problem....
|
|
I have been trying to use the f2py routinely to build an interface around the ECMWF BUFR library.
|
|
See http code.google.com p pybufr ecmwf for details.
|
|
This works fine for the python 2.x versions.
|
|
However, I also tried to add the possibility to use python3 and here I encountered some trouble.
|
|
My python code converts properly to python3 using to 2to3 tool, but when I try to use the python3 version of f2py I encounter this error somewhere deep inside the f2py internals Post processing stage 2)...
|
|
Saving signatures to file f2py build signatures.pyf Traceback most recent call last) File .
|
|
run f2py tool.py , line 11, in <module> main ) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 563, in main run main sys.argv[1 ]) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 342, in run main postlist=callcrackfortran files,options) File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 279, in callcrackfortran pyf=crackfortran.crack2fortran postlist) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2693, in crack2fortran pyf=crack2fortrangen block) \n File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2505, in crack2fortrangen body=crack2fortrangen blockbody?,tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2505, in crack2fortrangen body=crack2fortrangen blockbody?,tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2465, in crack2fortrangen ret=ret crack2fortrangen g,tab) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2506, in crack2fortrangen vars=vars2fortran block,blockvars?,al,tab tabchar) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2655, in vars2fortran lst = true intent list vars[a]) File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2557, in true intent list if c UnboundLocalError?
|
|
local variable c referenced before assignment ERROR build of python wrapper failed the signatures file could not be found Traceback most recent call last) File .
|
|
build interface.py , line 1803, in <module> BI.build ) File .
|
|
build interface.py , line 543, in build self.generate python wrapper source dir) File .
|
|
build interface.py , line 1667, in generate python wrapper raise InterfaceBuildError?
|
|
main.InterfaceBuildError?
|
|
I any additional information is needed I would be happy to provide it.","bug"
|
|
"6013","3193","numpy.distutils.misc util.get shared lib extension ) incompatible with Python >=3.2.4 and >=3.3.1.
|
|
Meaning of `SO` variable has been reverted in Python >=3.2.4 and >=3.3.1 from .cpython 32.so .cpython 33.so to .so .
|
|
`SO` variable also has been deprecated in favor of `SHLIB SUFFIX` variable and is removed in Python 3.4.)
|
|
NumPy should use new `EXT SUFFIX` variable, which has value .cpython 32.so .cpython 33.so .cpython 34.so .
|
|
See http bugs.python.org issue16754 and http mail.python.org pipermail python dev 2013 March 124817.html.
|
|
Error in test suite of NumPy .
|
|
Patch .","bug"
|
|
"6013","3194","Document NPY RELAXED STRIDES CHECKING.
|
|
Also the other minor changes 0 d, and 1 element array corner cases) should be mentioned, and the user guide get a real explanation, how contiguous is defined.
|
|
The small change for the aligned flag needs mention as well.","bug"
|
|
"6013","3195","segmentation fault with numpy.place.
|
|
Hi, On my install at least see below), I append to find an easy way to make a segmentation fault .
|
|
I understand this is not be the right way to set some elements of an object array to an empty list sadly), but still it should probably not crash python os mac10.6 python 2.7.3 with ipython 0.13.1 numpy 1.6.2","bug"
|
|
"6013","320","Faster column stack?.
|
|
I was experimenting with a faster version of ``column stack``, and came up with this .
|
|
It seems to be about 50 faster is there a drawback to doing it this way, or is this approach better?","enhancement"
|
|
"6013","3204","argsort segmentation fault .
|
|
Sorting a large number of int64 numbers causes a segmentation fault for numpy 1.7.0 import numpy from numpy import random N=5000000000 id1=random.random integers 2 31 1,size=N) id2=random.random integers 2 31 1,size=N) id=id1 id2 << 32) index=numpy.argsort id)","bug"
|
|
"6013","3207","numpy.fromfile cannot read from zipfile .
|
|
Trying to use fromfile ) on a file opened within a zip archive with zipfile.ZipFile ) always yields an error.
|
|
import numpy as np import zipfile z=zipfile.ZipFile archive.zip ) f=z.open file.txt ) np.fromfile f, count=1) IOError first argument must be an open file This is with numpy version 1.6.1.","bug"
|
|
"6013","3209","Possible memory leak in np.diagonal ).
|
|
On my computer this leaks memory on numpy 1.7.0 but not 1.6.2 In [1] a = np.ones 500,500)) In [2] for i in range 10000000) a.diagonal ) Functions that call a.diagonal, such as np.diag ) and np.trace ), will also leak.","bug"
|
|
"6013","3211","testing, will close.. testing","bug"
|
|
"6013","3212","Test ticket.
|
|
checking if non committers can tag issues","bug"
|
|
"6013","3213","Assigning a fill value in masked array fails silently for wrong type.
|
|
I try to assign a `fill value` to a masked array.
|
|
If it s the right type e.g.
|
|
an `int` in an array of type `int`), it works, but if I try to assign a `fill value` of the wrong dtype e.g.
|
|
a `string` in an array of type `int`, it fails silently.
|
|
I don t know if this is a bug or a feature, but I had expected a `ValueError` or something similar if the value cannot be broadcasted to the required type.
|
|
.","bug"
|
|
"6013","3214","Incorrect warning rasied by logaddexp2.
|
|
Issue np.logaddexp2 raises a Warning RuntimeWarning invalid value encountered in logaddexp2 when provided with two values of ` np.inf`.
|
|
While the correct value is returned Inf), this is inconsistent with other logarithms of 0 that return Inf with no warnings.
|
|
Reproduce You can simply execute .
|
|
even though those two calculations are identical.
|
|
I am not sure how or why logaddexp2 is raising any different warnings than log2, since as far as I can tell it s just a wrapper for internal addition for log2.
|
|
Obviously you could manually supress this error, but it s inconsistent.
|
|
Can somebody with a better understanding of how these errors are raised make the quick fix?","bug"
|
|
"6013","3215","datetime.datetime to np.datetime64 conversion in astype.
|
|
This used to be possible in numpy 1.6 but isn t in 1.7, a bug or a feature?
|
|
dts = [datetime.datetime 2013,4,1) i datetime.timedelta days=1) for i in range 10)] npDts = np.array dts) npDts.astype datetime64) TypeError Traceback most recent call last) <ipython input 6 c682ed3993b0> in <module> ) 1 dts = [datetime.datetime 2013,4,1) i datetime.timedelta days=1) for i in range 10)] 2 npDts = np.array dts) > 3 npDts.astype datetime64) TypeError Cannot cast datetime.datetime object from metadata [us] to [D] according to the rule same kind","bug"
|
|
"6013","3217","Merge generalized linear algebra routines into numpy.linalg.
|
|
The features in gh 2954 should be brought into numpy.linalg.
|
|
A suggested plan for action 1.
|
|
Combine the lite libraries.
|
|
2.
|
|
Merge in gh 2954, without exposing the features in a public API.
|
|
3.
|
|
Change gh 3220) the implementation of routines in numpy.linalg to use umath linalg while preserving full backward compatibility, but allow stacked input.
|
|
4.
|
|
Add the new routines from ` gufunc linalg.py` into numpy.linalg keep eye on Scipy function signatures, if applicable) 5.
|
|
Figure out how to deal with error handling in numpy.linalg and implement any changes.
|
|
6.
|
|
Figure out how to allow numpy.linalg routines compute single precision in single precision rather than using double precision LAPACK) 7.
|
|
Extend umath linalg to cover all functions if possible?)
|
|
in numpy.linalg, and drop the old lapack lite wrapper.
|
|
Open questions A.
|
|
Error handling.
|
|
Raising LinAlgErrors is not necessarily the right thing to do especially in the stacked case, so this should be made user controllable.
|
|
B.
|
|
The numpy.linalg routines compute results for single precision by using double precision internally.
|
|
This should perhaps be made tunable?
|
|
C. Which of the generalized ufunc arguments should be exposed in the numpy.linalg API?
|
|
D. The umath linalg error handling goes through by mis using the invalid FP flag.
|
|
Should ufunc error handling be extended here?
|
|
scipy.special would also have an use case) Possible solutions to questions A.
|
|
Add keyword argument `raise errors` to all functions `inv`, `solve`, `eig`, ...) that used to raise a LinAlgError on LAPACK errors.
|
|
Perhaps also add global state, e.g.
|
|
`np.linalg.seterr` for tuning the default value however, global state here may have unexpected effects in 3rd party libraries.
|
|
B.
|
|
Changing the default can cause changes in results, and may make some algorithms not converge?
|
|
C. Dtype, perhaps, and maybe also axis control arguments?","enhancement"
|
|
"6013","3218","issubdtype fails with string representations.
|
|
>>> np.issubdtype np.float64,np.float32) this is ok False >>> np.issubdtype float64 , float32 ) but this is not ok True According to the docs, the arguments of issubdtype can be dtype or string representing a typecode , but it fails in the second case.","bug"
|
|
"6013","3219","MaskedArray.dot not considering mask?.
|
|
`masked.dot other)` is equivalent to `numpy.dot masked, other)`, rather than `numpy.ma.dot masked, other)`.
|
|
.","bug"
|
|
"6013","3222","dtype keyword doesn t work as expected for generalized ufuncs.
|
|
With the generalized ufuncs in gh 2954 gh 3220 .","bug"
|
|
"6013","3228","make sure there aren t any deprecation warnings that incorrectly say stuff will happen in 1.8.
|
|
There are a number of deprecations that we expected would be finalized in 1.8, but that won t be, e.g.
|
|
gh 596, gh 456, gh 294.
|
|
We ve bumped them to reconsider for 1.9, but we should run a grep over the source to make sure that we haven t left any behind any messages that say Warning blah balh will happen in numpy 1.8 , and if there are any we should replace them with something more vague.","bug"
|
|
"6013","3233","2to3 run `itertools` fixer.","bug"
|
|
"6013","3234","2to3 run `itertools imports` fixer.","bug"
|
|
"6013","3240","2to3 run `types` fixer.","bug"
|
|
"6013","3246","Add imports from future builtins.
|
|
There is a `future builtins` module available since Python 2.6 that supplies iterator versions of map, filter, and zip.
|
|
These should be imported in all python files to provide additional testing and help keep the code compatible between Python versions.","bug"
|
|
"6013","3247","2to3 remove tools py3tool.py after all the fixers are run..","bug"
|
|
"6013","3251","Return value of asarray does not behave like ndarray in arithmetic operations.
|
|
The documentation of asarray ..) claims it does not pass through array subclasses, unlike asanyarray.
|
|
However, passing a sparse matrix to asarray ..) returns a value which still behaves in some ways expected of sparse matrices but surprising for arrays.
|
|
In [38] import numpy as np In [39] import scipy.sparse as sp In [40] a = np.asarray sp.csr matrix [1])) In [41] type a) Out[41] numpy.ndarray In [42] type 2 a) Out[42] scipy.sparse.csr.csr matrix In [43] 2 a).ravel ) AttributeError Traceback most recent call last) <ipython input 43 fc4ac7b3ba92> in <module> ) > 1 2 a).ravel ) C \Python27\lib\site packages\scipy\sparse\base.pyc in getattr self, attr) 391 return self.getnnz ) 392 else > 393 raise AttributeError attr not found ) 394 395 def transpose self) AttributeError ravel not found In [44] 2 a) NotImplementedError Traceback most recent call last) <ipython input 44 c1aa2665906d> in <module> ) > 1 2 a) C \Python27\lib\site packages\scipy\sparse\compressed.pyc in radd self, other) 187 188 def radd self,other) > 189 return self.
|
|
add other) 190 191 def sub self,other) C \Python27\lib\site packages\scipy\sparse\compressed.pyc in add self, other) 173 return self.copy ) 174 else Now we would add this scalar to every element.
|
|
> 175 raise NotImplementedError adding a nonzero scalar to a 176 sparse matrix is not supported ) 177 elif isspmatrix other) NotImplementedError adding a nonzero scalar to a sparse matrix is not supported","bug"
|
|
"6013","3253","Cannot get a view from a composite array with object data members.
|
|
With NumPy 1.7.1 on Win64 with Python 2.7 I see a new issue that was not a problem on a previous version e.g., 1.6.2).
|
|
Consider a composite data type, where a view is obtained from .
|
|
cool note back in 1.4.1 although a subset could be made, the order was the same as the original).
|
|
However, if one of types used in the composite data type is |O8 , I can no longer obtain a view .
|
|
> Traceback most recent call last) > File <interactive input> , line 1, in <module> > File C \Python27\lib\site packages\numpy\core\ internal.py , line 296, in index fields > view = ary.view dtype=view dtype) > TypeError Cannot change data type for object array.","bug"
|
|
"6013","3254","numpy.select affected by numpy.choose regression..
|
|
There is a bug in numpy.select ), originating from an unsolved regression bug in numpy.choose ), which limits the number of items from which to select to 32 NPY MAXARGS).
|
|
This behaviour is not documented in the numpy documentation for numpy.select.
|
|
This bug affects a project I m working on and unfortunately is going to be tricky to work around due to a constraint on how I can use numpy.
|
|
I can only execute a single expression; numpy.select allowed an elegant and obvious solution).
|
|
A code sample that shows the bug, with expected and actual output, and without explicit use of numpy.choose Code .
|
|
Expected output .
|
|
Actual output .
|
|
A related mailing list thread about the regression in numpy.choose from June 2011. http www.mail archive.com numpy discussion scipy.org msg31915.html A related bugfix pull request which is apparently partially incomplete not integrated.
|
|
https github.com numpy numpy pull 226 Other related links discussion http mail.scipy.org pipermail numpy discussion 2012 August 063646.html http stackoverflow.com questions 11570157 numpy choose 32 choice limitation http comments.gmane.org gmane.comp.python.scientific.user 28521 http www.mail archive.com numpy discussion scipy.org msg31915.html","bug"
|
|
"6013","3255","ImportError Undefined symbol PyGILState Release.
|
|
When compiling and using numpy with a single threaded python, an import numpy will lead to an ImportError saying that PyGILState Release is an undefined symbol.
|
|
In e.g.
|
|
numpy random mtrand mtrand.c, the python C API calls which require threading are enclosed by a ifdef WITH THREAD endif .
|
|
This is not the case in numpy core src multiarray multiarraymodule.c and numpy core src multiarray nditer api.c.
|
|
I would suggest to enclose the PyGILState calls in these two files, too.","bug"
|
|
"6013","3256","recarray.
|
|
getitem with field list gives TypeError Cannot change data type for object array.
|
|
when dtype contains object.
|
|
This code works for me on numpy 1.6.2 but on 1.7.1 I get an exception, because in the new version numpy.core.
|
|
internal.
|
|
index fields calls ary.view ) and this doesn t work on object arrays.
|
|
.
|
|
Example on 1.7.1 .","bug"
|
|
"6013","3258","appranent inconsistency in numpy.unicode on narrow builds follow up to dated Trac 525).
|
|
I m assigning a numpy.unicode to a slice of array of numpy.unicode and see the following In [85] l = np.array [u blah ,u blah ,u blah ]) In [87] type l[0]) Out[87] numpy.unicode In [88] v = np.unicode bubu ) In [89] l[0] = v In [90] l Out[90] array [u bubu , u blah , u blah ], dtype= <U4 ) In [91] l[1 3] = v <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< In [92] l Out[92] array [u bubu , u \uf500\udc62\uf500\udc62\u8d00\udc00 , u \uf500\udc62\uf500\udc62\u8d00\udc00 ], dtype= <U4 ) <<<<<<<<<<<< I consider this a problem In [93] {k v for k, v in sysconfig.get config vars ).items ) if unicode in k.lower )} Out[93] { PY UNICODE TYPE unsigned short , Py UNICODE SIZE 2, Py USING UNICODE 1, UNICODE OBJS Objects unicodeobject.o Objects unicodectype.o } In [94] np.
|
|
version Out[94] 1.7.1 uname a Linux ... 2.6.32 279.19.1.el6.x86 64 1 SMP Wed Dec 19 07 05 20 UTC 2012 x86 64 x86 64 x86 64 GNU Linux On a wide unicode build I can do the same with no problem In [81] l = np.array [u blah ,u blah ,u blah ]) In [82] type l[0]) Out[82] numpy.unicode In [83] v = np.unicode bubu ) In [84] l[0] = v In [85] l Out[85] array [u bubu , u blah , u blah ], dtype= <U4 ) In [86] l[1 3] = v <<<<<<<<<<<<<<<<<<<<<<<<<<<<< In [87] l Out[87] array [u bubu , u bubu , u bubu ], dtype= <U4 ) <<<<<<<<<<<<<<<<<<<<<<<<<<<< this works fine In [88] {k v for k, v in sysconfig.get config vars ).items ) if unicode in k.lower )} Out[88] { PY UNICODE TYPE 0, Py UNICODE SIZE 4, Py USING UNICODE 1, UNICODE OBJS Objects unicodeobject.o Objects unicodectype.o } In [89] np.
|
|
version Out[89] 1.7.1 my27py)pavel pavel ThinkPad T520 uname a Linux pavel ThinkPad T520 3.2.0 40 generic 64 Ubuntu SMP Mon Mar 25 21 22 10 UTC 2013 x86 64 x86 64 x86 64 GNU Linux","bug"
|
|
"6013","3259","More bugs in numpy.select numpy.choose.
|
|
In issue 3254, I noted this part of the traceback .
|
|
I decided to test this error message, and found the following extra bugs, separate to issue 3254.
|
|
BUG 1.
|
|
The error message is wrong and precisely but incorrectly describes numpy s behaviour.
|
|
BUG 2.
|
|
Numpy does not handle size 32 inclusive arrays, but rather, size 30 inclusive.
|
|
Both 31 and 32 element arrays generate errors.
|
|
this may indicate an off by 2 array handling error somewhere BUG 3.
|
|
Zero and 1 element arrays do not generate errors, but according to the error message) should.
|
|
My feeling is that this is a feature rather than a bug.
|
|
Not sure about correct result to return for case zero.
|
|
these two cases also seem like an off by 2 array handling error BUG 4.
|
|
I assume) there are no test cases that check for these bugs and detect them automatically or they would be known already.
|
|
Since there are a few bugs in relation to choose select array size handling, maybe test cases can be added to stop this from happening again in future?
|
|
I hope the bug demo code below is useful for this.
|
|
The array related bugs may have been fixed already by patch https github.com numpy numpy issues 226 See https github.com numpy numpy issues 3254 issuecomment 16499519 But this would need to be checked after the merge conflicts are sorted.
|
|
==== Test code ==== .
|
|
==== Expected Output ==== .
|
|
==== Actual Output ==== .
|
|
N.B.
|
|
cases 0, 1, 31, 32.
|
|
Complete traceback .
|
|
Related issues https github.com numpy numpy issues 3254 https github.com numpy numpy pull 226 http mail.scipy.org pipermail numpy discussion 2011 July 057831.html Thanks for your time in reading this issue report.","bug"
|
|
"6013","3260","linalg.svd hangs on matix containing inf windows only) numpy 1.6.2).
|
|
Example import numpy as np c= np.array [[1,2,3],[1,float inf ),3],[1,2,3]]) print np.linalg.svd c) Output Linux) array [[ 0., 1., 0.
|
|
], [ 1., 0., 0.
|
|
], [ 0., 0., 1.
|
|
]]), array [ nan, nan, nan]), array [[ 0., 1., 0.
|
|
], [ 1., 0., 0.
|
|
], [ 0., 0., 1.]]))
|
|
On windows it just hangs does not throw the SVD did not converge exception).
|
|
If you replace inf by nan , SVD did not converge is thrown on both plattforms.","bug"
|
|
"6013","3263","Possible regression numpy.random.shuffle clobbers mask of a MaskedArray.
|
|
Title sums it up.
|
|
If I use random.shuffle to shuffle the rows of a masked array, the mask gets messed up.
|
|
Sometimes it disappears entirely, sometimes a few masks remain.
|
|
It appears that this appeared somewhere between 1.6.2 seems to be work fine) and 1.7.0 bug appears).
|
|
Sorry I don t have better information than that, I just noticed the bug when switching between versions.
|
|
.
|
|
Note that some masked values are now exposed but not all.
|
|
Maybe the masked rows were not moved by shuffle.
|
|
In 1.6.2 it works as expected.","bug"
|
|
"6013","3267","Doc in meshgrid is wrong.
|
|
Hi all, The docs at http docs.scipy.org doc numpy reference generated numpy.meshgrid.html present the wrong order for nx and ny in the loop for the example.
|
|
It says .
|
|
But it should be .
|
|
As i is the index for the number of lines, and j is the index for the number of rows.","bug"
|
|
"6013","3273","1.8 dev fails badly for Python built with 16 bit unicode..
|
|
Gave this a try with ActiveState python 2.7 with the following test result .","bug"
|
|
"6013","3274","np.concatenate with axis=None.
|
|
I ran into the same issue as reported in https github.com numpy numpy issues 442 using numpy 1.7.1.
|
|
Can the fix for this issue be backported to 1.7.x?
|
|
Sorry if this was already discussed somewhere.","bug"
|
|
"6013","3276","Segmentation fault using ma.masked all with Python 3.3.
|
|
Using `ma.masked all )` with a unicode string on linux 64 with python 3.3 and numpy 1.7.0, there is a segmentation fault shown below.
|
|
This is not present in python 3.2 or python 2.7. .","bug"
|
|
"6013","3278","New issues in 1.7.x.
|
|
While backporting a fix I noticed the following 1) Unreproducible segfault while running python 3.3 tests.
|
|
2) Two unclosed resource warnings in python 3.3 tests.
|
|
I opened this issue just to make a note of these.","bug"
|
|
"6013","3279","np.insert ) writes over second argument.
|
|
This is a regression since 1.7.x.
|
|
The following code on current master .
|
|
On 1.7.x, one gets `[1 1 1]` as expected.","bug"
|
|
"6013","3282","Segmentation fault when passing None to numpy.result type.
|
|
Tested on 1) OSX 10.8.3, python 2.7.4, numpy 1.7.1 2) Ubuntu 12.04 x64, python 2.7.3, numpy 1.7.1 >>> import numpy >>> numpy.result type numpy.int32, None) fish Job 1, python terminated by signal SIGSEGV Address boundary error) >>> import numpy >>> numpy.result type None) fish Job 1, python terminated by signal SIGSEGV Address boundary error) If I start Python in 32 bit mode with ``arch i386 python``, the error is different >>> import numpy >>> numpy.result type None) fish Job 1, arch i386 python terminated by signal SIGBUS Misaligned address error)","bug"
|
|
"6013","3283","stats.f oneway needs floats.
|
|
Tested on Mac OS X 10.8.3 scipy 0.11.0 stats.f oneway does not handle integer data properly >>> import numpy as np >>> from scipy import stats >>> y0 = np.array [655, 788, 734, 721, 679, 699], dtype=np.uint16) >>> y1 = np.array [789, 772, 786, 686, 732, 774], dtype=np.uint16) >>> print stats.f oneway y0, y1) yields f = 0.009, p = nan but repeating with floats >>> y0 = np.array [655, 788, 734, 721, 679, 699], dtype=float) >>> y1 = np.array [789, 772, 786, 686, 732, 774], dtype=float) >>> print stats.f oneway y0, y1) yields the correct results f = 3.052, p = 0.111 Suggested fix change the first line of f oneway from >>> args = map np.asarray, args) to >>> args = [np.asarray arg, dtype=float) for arg in args]","bug"
|
|
"6013","3284","can t download numpy.
|
|
Hi, Downloading links mainly i need this one http numpy.scipy.org ) seems to do not work anymore.
|
|
I would be gratefull if you will make it working again.
|
|
Greets.
|
|
Duz","bug"
|
|
"6013","3285","vectorize call error when otype attribute is set after vectorize object creation.
|
|
if a vectorize object is created without otype argument, an attempt to set it later causes an error .","bug"
|
|
"6013","3286","cannot field slice structured array or recarray) containing objects.
|
|
Usually one can select a subset of fields from a structured array using `rarray[list of fields]`.
|
|
If the structured array contains `object`s dtype `O`) even if an object field isn t being selected an error is thrown in numpy 1.7.1).
|
|
.
|
|
This field slice process functions by [creating a view] https github.com numpy numpy blob master numpy core internal.py L300) with a modified dtype; for this example, the new dtype is `{ names [ a , b ], formats [dtype float32 ), dtype float32 )], offsets [0, 4], itemsize 74}`.
|
|
It seems the checks in [`array descr set`] https github.com numpy numpy blob master numpy core src multiarray getset.c L451) are too conservative, discarding alternative views of any array including objects.
|
|
While it is understandable that this check ensures against arbitrary modification of memory, it makes structured arrays containing objects very unwieldy.
|
|
Surely this validation can be made more precise, or else disabled for particular safe operations.
|
|
Related to [TRAC 674 closed)] http projects.scipy.org numpy ticket 674) and 2346.","bug"
|
|
"6013","3287","array division failure.
|
|
FloatingPointError underflow encountered in multiply when executing a division of two 3 element arrays.
|
|
I can easily reproduce the error, but I can t isolate it for when the offending code is run in isolation, even with the same numeric values, no exception is raised.
|
|
python 2.7.1 numpy 1.6.x, 1.7.1 windows XP pro 32 the code from numpy.ma.core import array, sqrt from numpy.core.numeric import NaN import numpy SCV = array [0.20, 0.80, 0.00]) Scattering coefficient of leaves for visible radiation PAR) print SCV.
|
|
class , SCV , SCV SQV = sqrt 1.
|
|
SCV) 6.14 GVL) [0.894427191 0.4472135955 1.0] print SQV.
|
|
class , SQV , SQV reflection coefficients for horizontal and spherical leaf angle distribution REFH = 1.
|
|
SQV) 1.
|
|
SQV) 6.19) print SCV.
|
|
class , REFH , REFH print numpy.version.version expected output When run in isolation, the correct) output is <class numpy.ma.core.MaskedArray > SCV [ 0.2 0.8 0. ]
|
|
<class numpy.ma.core.MaskedArray > SQV [0.8944271909999159 0.44721359549995787 1.0] <class numpy.ma.core.MaskedArray > REFH [0.05572809000084123 0.3819660112501052 0.0] 1.7.1 However, when run in place, an exception is raised.
|
|
The numeric values of SCV are identical.
|
|
The module s imports are identical.
|
|
But it is executed within a method of an object, that function s well apart from this.
|
|
the exception <class numpy.ma.core.MaskedArray > SCV [ 0.2 0.8 0. ]
|
|
<class numpy.ma.core.MaskedArray > SQV [0.8944271909999159 0.44721359549995787 1.0] ...
|
|
File D \Projects\ForSpace\Model\src\light\plantItem.py , line 183, in reflectionAndExtinction REFH = 1.
|
|
SQV) 1.
|
|
SQV) 6.19) File C \Program Files\Python\EPD 7.3 2\lib\site packages\numpy\ma\core.py , line 3656, in div return divide self, other) File C \Program Files\Python\EPD 7.3 2\lib\site packages\numpy\ma\core.py , line 1084, in call m |= filled domain da, db), True) File C \Program Files\Python\EPD 7.3 2\lib\site packages\numpy\ma\core.py , line 777, in call return umath.absolute a) self.tolerance >= umath.absolute b) FloatingPointError underflow encountered in multiply","bug"
|
|
"6013","3288","Broken link http scipy.org NumPy) in README.
|
|
There is a broken link http scipy.org NumPy) in the README file.","bug"
|
|
"6013","329","genfromtxt can fail when using None as a comments keyword argument.
|
|
genfromtxt has a bug where using None as a comments keyword argument causes the text string None to be treated as a comment.
|
|
This causes exceptions when loading large datasets if the string None occurs by chance.
|
|
eg >>> genfromtxt [ test1\ttestNonethrestofthedata ],dtype=None, comments=None, delimiter= \t ) array [ test1 , test ], dtype= |S5 ) The problem is due to a conversion to str using asbytes on Lines 1295 1296 of numpy numpy lib npyio.py Py3 data conversions to bytes, for convenience comments = asbytes comments) This should be wrapped in a conditional if comments != None comments = asbytes comments)","bug"
|
|
"6013","3290","Timezones and Datetime64.
|
|
A new version of datetime64 dtype was added to numpy recently, and as of 1.7, is considered experimental.
|
|
As of 1.7, datetime64 attempts to handle timezones by Assuming all datetime64 objects are in UTC Applying timezone offsets when parsing ISO 8601 strings Applying the Locale timezone offset when the ISO string does not specify a TZ.
|
|
Applying the Locale timezone offset when printing, etc.
|
|
note that no timezone handling is done when converting to from the stdlib datetime object) However, this behavior makes it very difficult to use datetime64 in a timezone agnostic way, or if you need to work with a timezone that is not the system Locale one.
|
|
Complete and proper timezone management could be a very useful, but the current system is only useful in a few small cases, and makes common use cases very difficult.
|
|
Good timezone handling is a large and difficult problem it is unclear whether numpy wants to take it on.
|
|
There has been much discussion on the numpy mailing list, and a new datetime64 NEP is progress, but in the meantime, I propose we make the following changes to make datetime64 far more useful and less prone to error Assume that datetime64 is timezone naive , as the standard library datetime object does it is up to the user to do any time zone offset conversions.
|
|
Do not apply a timezone offset when converting to an ISO string Never use the Locale settings for anything, without the user explicitly asking for it.
|
|
raise an exception when parsing an ISO 8601 string that has a timezone offset other than 0 Z)","bug"
|
|
"6013","3291","linalg.inv and linalg.solve fail with the 0 by 0 matrix.
|
|
In several codes I maintain it is handy to be able to invert the zero by zero matrix.
|
|
The inverse of this matrix should again be the zero by zero matrix since it is the only linear operator on a 0 dimensional vector space).
|
|
Both of the following lines currently fail with numpy .
|
|
The benefit of allowing the inversion of the zero by zero matrix is that unnecessary if else blocks can be avoided which often leads to more generic code with better readability.
|
|
If you need an example to see that this is a useful feature, I d be glad to post it here.","bug"
|
|
"6013","3292","Changing the function name ptp to something more meaningful.
|
|
The peak to peak function, `ptp )`, is a simple and handy function to use, but it confuses the PEP20 suggestions a bit.
|
|
See this discussion for reference http d.pr i snXG Based on the discussion, it may be a good idea to make the name more relevant?
|
|
Some suggestions include `span )` or `valuerange )`.","enhancement"
|
|
"6013","3295","numpy.ma default fill value for strings not appropriate for S1 and S2.
|
|
For string arrays of type ``S1`` or ``S2``, the default fill value in Numpy masked arrays ``N A``) is not appropriate .
|
|
this causes issues for codes that assume the fill value is present in the filled array and indicated masked values.
|
|
It s not clear to me what the fill value should be, but I d just thought I d raise the question.
|
|
cc taldcroft","bug"
|
|
"6013","3299","test xerbla segfaults in linalg tests.
|
|
Running .
|
|
results in last line) .
|
|
and throws me right out of python so no test results are displayed.
|
|
It does not seem to fail on travis, though.
|
|
I m running Ubuntu 13.04, perhaps it s an incompatibility with the shipped LAPACK?
|
|
Commit 8eebee82b88f26f40c75422f5f655dea9f86f4b6 introduced the test.","bug"
|
|
"6013","3300","086721ce8d760 build fails with python 3.3.1 and clang .
|
|
numpy core src umath test rational.c.src 1264 9 error non void function PyInit test rational should return a value [ Wreturn type] return; numpy core src umath test rational.c.src 1269 9 error non void function PyInit test rational should return a value [ Wreturn type] return; 5 warnings and 2 errors generated.
|
|
The compile completed after changing return; to return NULL;","bug"
|
|
"6013","3302","memmap does not work with TemporaryFile in python3.
|
|
affects python 3.3.1, 3.2 and numpy git head and older TemporaryFile ).name returns a integer in python3 leading to .","bug"
|
|
"6013","3311","bug in deepcopy ) of rank zero arrays.
|
|
I cite Chris Baker s message from numpy discussion, since I encountered the exact same bug and I could not find it reporter here Hi folks, In [264] np.
|
|
version Out[264] 1.7.0 I just noticed that deep copying a rank zero array yields a scalar probably not what we want.
|
|
In [242] a1 = np.array 3) In [243] type a1), a1 Out[243] numpy.ndarray, array 3)) In [244] a2 = copy.deepcopy a1) In [245] type a2), a2 Out[245] numpy.int32, 3) regular copy.copy ) seems to work fine In [246] a3 = copy.copy a1) In [247] type a3), a3 Out[247] numpy.ndarray, array 3)) Higher rank arrays seem to work fine In [253] a1 = np.array 3,4)) In [254] type a1), a1 Out[254] numpy.ndarray, array [3, 4])) In [255] a2 = copy.deepcopy a1) In [256] type a2), a2 Out[256] numpy.ndarray, array [3, 4])) Array scalars seem to work fine as well In [257] s1 = np.float32 3) In [258] s2 = copy.deepcopy s1) In [261] type s1), s1 Out[261] numpy.float32, 3.0) In [262] type s2), s2 Out[262] numpy.float32, 3.0) There are other ways to copy arrays, but in this case, I had a dict with a bunch of arrays in it, and needed a deepcopy of the dict.
|
|
I was surprised to find that my rank 0 array got turned into a scalar.","bug"
|
|
"6013","3312","bug in count nonzero.
|
|
I am using the numpy function count nonzero in one of my programs Python 2.7.3, Numpy 1.6.1; the standards for Ubuntu 12.04 LTS) but I ran in to some problems with the count nonzero function crashing when colleagues running the 64 bit windows version of python 2.7.3 and numpy 1.7.1 taken from http www.lfd.uci.edu gohlke pythonlibs numpy) ran the program.
|
|
This problem also occurs on OS X according to a fellow SO user, link to the SO question is http stackoverflow.com q 16418332 1093485.
|
|
Summary, doing the following steps will recreate this bug 1.
|
|
Install Python 2.7.3.amd64.msi 2.
|
|
Install numpy MKL 1.7.1rc1.win amd64 py2.7.msi 3.
|
|
Run the following code import numpy as np data = np.zeros 2500,dtype= float32, 25000,2)float32 ) The following 3 lines are not necessary but asking for the amount of non zeros in an array filled with zeros feels silly data[0][0] = 1 data[0][1][0] = 2 data[0][1][1] = 3 time elements = np.count nonzero data) 4.
|
|
Watch python crash Kind Regards, Bas Jansen Leiden University Medical Center","bug"
|
|
"6013","3314","BUG any and all give unexpected results for 0 dimensional arrays.. With NumPy 1.7.1 and Python 2.7.3) .
|
|
I expected to get `False` instead of `0.0` in both cases.
|
|
This appears to be a regression from NumPy 1.6.
|
|
Originally reported in http stackoverflow.com q 16426547 270986","bug"
|
|
"6013","3315","savetxt bug with structured arrays.
|
|
Savetxt raises a TypeError exception when saving a structured array float argument required, not numpy.void , using numpy 1.7.1.
|
|
Minimum example follows import numpy as np a=np.zeros 10,2),dtype=[ label1 , float ), label2 , float )]) np.savetxt test.dat ,a)","bug"
|
|
"6013","3319","numpy.distutils.mingw32ccompiler.py can t find pythonXX.dll if not in sys.prefix.
|
|
using numpy 1.6.1, numpy.distutils.mingw32ccompiler.py if pythonXX.dll XX is python major and minor revs, e.g., 27) is not in sys.prefix, find python dll ) fails.
|
|
Recommend enlarging search for pythonXX.dll to include the entire path from ==== def find python dll ) maj, min, micro = [int i) for i in sys.version info[ 3]] dllname = python d d.dll maj, min) path = os.path.join sys.prefix, dllname) assert os.path.isfile path) return path to == def find python dll ) maj, min, micro = [int i) for i in sys.version info[ 3]] dllname = python d d.dll maj, min) for p in sys.path path = os.path.join p, dllname) if os.path.isfile path) return path raise ValueError python d d.dll not found in sys.path maj, min))","bug"
|
|
"6013","3320","ndarray norm member function.
|
|
I assume that there s a reason this doesn t exist, but it would be convenient to use.
|
|
For example you would be able to have X.norm 1) without caring whether X is an ndarray or matrix or a scipy sparse matrix, or maybe even a more abstract linear operator.
|
|
I think that .diagonal ), .transpose ), .dot ), and other functions already work like this.
|
|
Currently numpy.linalg.norm X, 1) only works with things that asarray ) understands.","enhancement"
|
|
"6013","3321","numpy.ind1 with masked arrays dtype=boolean).
|
|
The code works fine .
|
|
But if I use boolean dtype in the code, the result is different all elements becames masked) .
|
|
numpy version is 1.6.1","bug"
|
|
"6013","3326","numpy trace leaks memory.
|
|
The following simple code leads to a memory leak.
|
|
Calculating the trace directly in a for loop solves this issue.
|
|
import numpy as np A = np.zeros 1000,1000)) A[1,1] = 1.0 B = np.ones 1000,1000)) C = np.zeros 5000) for t in range 5000) C[t] = C[t] np.trace np.dot A, B))","bug"
|
|
"6013","3331","Can t get numpy to work at any level.
|
|
Hello, This isn t really a bug but I don t know where to get help I ve been searching Google for hours trying to find an answer).
|
|
I m getting the error ImportError Error importing numpy you should not try to import numpy from its source directory; please exit the numpy source tree, and relaunch your python intepreter from there.
|
|
when I try to use numpy with Eclipse on Windows.
|
|
I know I have something set up wrong but I can t find it.
|
|
If you could direct me to some more detailed instructions I would be most appreciative.
|
|
Thanks!
|
|
Joe Piccoli joe13676 comcast.net","bug"
|
|
"6013","3332","fliplr documentation incorrectly states 2 d not required.
|
|
[`fliplr` documentation] http docs.scipy.org doc numpy reference generated numpy.fliplr.html) incorrectly states 2 d not required, but that is not true.
|
|
I think it is a typo, since `fliplr` was probably copied from `flipud` which really doesn t require a 2 d array.
|
|
>>> import numpy as np >>> a = np.array [1,2,3,4]) >>> np.fliplr a) Traceback most recent call last) File <stdin> , line 1, in <module> File C \Python27\lib\site packages\numpy\lib\twodim base.py , line 61, in fliplr raise ValueError Input must be >= 2 d. ) ValueError Input must be >= 2 d.","bug"
|
|
"6013","3334","numpy.distutils 1.7.0 breaks with easy install.
|
|
Platform win32, python 2.6 binary installer from python.org, numpy binary installer of versions given, mingw32 compiler specified in distutils.cfg) easy install ffnet works fine with numpy 1.6.2; with numpy 1.7.0, it raises a SandboxViolation > Wrote C API module pikaia to file build\src.win32 2.6\ffnet\fortran\ pikaiamodule.c > Fortran 77 wrappers are saved to build\src.win32 2.6\ffnet\fortran\ pikaia f2pywrappers.f > error Setup script exited with error SandboxViolation open C \\Python26\\lib s\\libmsvcr90.def , w ) {} I know distribution setup tools are a little flavour of the month, but I haven t seen a call that numpy.distutils is incompatible with easy install now ffnet still recommends easy install).","bug"
|
|
"6013","3336","ma.corrcoef bug.
|
|
The documentation states that the masked version is just like the regular version of corrcoef, just with handling of missing data.
|
|
But how it handles that missing data isn t obvious, and I m getting unexpected results.
|
|
import numpy as np x=np.array [1,2,3,np.nan,5]) y=np.array [2,2.5,3.1,3,5]) xm=np.ma.masked array x,np.isnan x)) ym=np.ma.masked array y,np.isnan y)) m= xm.mask|ym.mask) a=np.ma.corrcoef xm,ym,ddof=0)[0,1] b=np.corrcoef x[m],y[m],ddof=0)[0,1] print a,b) 0.74153834037860156, 0.98871778717146885) What exactly is the masking doing here, and why?
|
|
A naive approach to handling missing data in this case for me at least) would be to discard datapoints where either x or y is nan result b ).
|
|
Yet the result of ma.corrcoef a ) is quite far from that.","bug"
|
|
"6013","3338","Generalized ufunc inner1d problems with 1 element inputs.
|
|
In these new linear algebra routines as generalized ufuncs, I think the following is a bug .
|
|
The result should be .
|
|
as far as I understand.
|
|
Also, the following is weird .
|
|
I m not sure whether it should return an error because the last axes have different sizes or whether the answer should be .
|
|
In any case, these results seem bugs to me.
|
|
Any help?
|
|
Thanks!","bug"
|
|
"6013","3343","eigh crashes python for illegal input on win32.
|
|
Using Python 2.7.2 default, Jun 12 2011, 15 08 59) [MSC v.1500 32 bit Intel)] on win32) and numpy 1.7.1 as [packaged by python x,y)] https code.google.com p pythonxy wiki Downloads Plugin updates))), the following snippet .
|
|
results in a python crash with the following output .
|
|
The expected result is an exception, with either the same text or a more descriptive error cannot calculate eigenvalues of a 0x0 matrix ).
|
|
On Python 2.7.3 default, Aug 1 2012, 05 14 39, [GCC 4.6.3] on linux2) and numpy 1.6.1, the same snippet results in .","bug"
|
|
"6013","3348","Access violation in descriptor from pep3118 format.
|
|
We managed to trigger an access violation in NumPy, coming from ` descriptor from pep3118 format` in `numpy core src multiarray buffer.c`.
|
|
The relevant lines are .
|
|
Here `buf` refers to a buffer that was `malloc`ed earlier in the function; by the time we get to this line, `buf` has already been `free`d, so we get an access violation when building the exception message.","bug"
|
|
"6013","3351","Possible conversion bug with record array.
|
|
I got a weird output from the following script .
|
|
Output [[0.0, 1.875, 0.0, 0.0], [0.0, 0.0, 0.0, 0.0], [0.0, 0.0, 0.0, 1.875], [0.0, 0.0, 0.0, 0.0]],) Tested on numpy 1.7.1, confirmed on 1.6.0.
|
|
As noted by Chirs Barker on the mailing list, U is a rank zero array, not a 4X4 array.","bug"
|
|
"6013","3352","bug in mtrand.
|
|
In mtrand distributions.c line 312, int should be long.","bug"
|
|
"6013","3353","Incorrect argsort answer with small float?.
|
|
<code>numpy.array [0.99999994, 0.99999994, 0.99999994, 0.99999714]).argsort )< code> with answer <code>array [3, 0, 1, 2])< code>","bug"
|
|
"6013","3354","numpy.core.umath.multiply TypeError, while executing numpy.var.
|
|
Numpy 1.7.1 numpy.core.umath.multiply break execution of numpy.var or numpy.std).
|
|
Where .
|
|
Expected output .
|
|
I see .
|
|
It s not very clear to calculate variance std) for one value scalar), but I was expected the similar behaviour as in numpy.mean or older version of numpy.var in numpy 1.6.1).","bug"
|
|
"6013","3355","Incoherence in meshgrid documentation.
|
|
In the meshgrid documentation is written Returns X, Y ndarray For vectors `x`, `y` with lengths ``Nx=len x)`` and ``Ny=len y)``, return `X`, `Y` where `X` and `Y` are `` Ny, Nx)`` shaped arrays with the elements of `x` and y repeated to fill the matrix along the first dimension for `x`, the second for `y`.
|
|
This makes the function the same of the matlab meshgrid, but in the formalism of Python the first dimension is row and the second colums, so would not be more logical to produce as output an Nx,Ny) matrix?
|
|
It would be nice to introduce a switch option in meshgrid or add a Meshgrid function that will do this.","bug"
|
|
"6013","3357","Problem with dot a,b,out=a).
|
|
Here is a short example showing the problem when out = a or out =b) This works as expected .
|
|
This doesn t work as expected wrong result) .","bug"
|
|
"6013","3359","test ufunc.py should test exceptions thrown from ufuncs.
|
|
`test ufunc.py` runs a few tests against the rational dtype that I wrote, but doesn t seem to run any of the exception related tests.
|
|
In particular, if in my code I leave the `ACQUIRE GIL` define commented out see `test rational.c.src`), I get a segfault in the `test numpy cast` test here https github.com girving rational blob 69c7c8ec2fe49461e87501503a9407ab0f47ae0b test rational.py L174 I have only tested this against 1.7.1, so it s possible the bug is fixed in more recent versions.
|
|
However, it should be tested regardless of whether it s fixed.","bug"
|
|
"6013","3363","test operand flags failure on 32 bit.
|
|
with current git head .
|
|
it works on 64 bit.","bug"
|
|
"6013","3364","severe performance regression when slicing on memory map.
|
|
in numpy 1.7 since commit 731cf3aaaa95c714361f4062e6929f5a324586cc slicing in memory maps is extremely slow due to the may share memory call in array finalize .
|
|
numpy 1.7 with the overlap check .
|
|
when just removing the check .","bug"
|
|
"6013","3367","BUG ma Several functions in the ma module return None when given a scalar..
|
|
Here s an example, using `ma.atleast 1d` In [4] np.
|
|
version Out[4] 1.8.0.dev 31a5501 When given a scalar, the regular version of `atleast 1d` returns an array In [5] x = np.atleast 1d 1.0) In [6] x Out[6] array [ 1.])
|
|
`ma.atleast 1d` returns `None` In [7] x = np.ma.atleast 1d 1.0) In [8] print x None `ma.atleast 1d` and several other functions in the `ma` module are defined using the class ` fromnxfunction` defined in numpy ma extras.py).
|
|
In the ` call ` method of this class, in the case where `len args) == 1`, the code handles the cases where the single argument is an ndarray, a tuple or list.
|
|
If the single argument is not one of these, the code falls through to the end without returning anything.
|
|
The non masked versions of these accept a scalar, so the masked version should too atleast 1d atleast 2d atleast 3d diagflat","bug"
|
|
"6013","3368","BUG ma Inconsistent return type from the count method..
|
|
Example In [6] a = ma.masked array [1, 2, 3]) In [7] b = ma.masked array [1, 2, 3], mask=False) In [8] a Out[8] masked array data = [1 2 3], mask = False, fill value = 999999) In [9] b Out[9] masked array data = [1 2 3], mask = [False False False], fill value = 999999) Notice the difference `a.mask` is the scalar value `False`, while `b.mask` is an array In [10] a.mask Out[10] False In [11] b.mask Out[11] array [False, False, False], dtype=bool) That s not the problem.
|
|
They both represent the same data a 1 D array with the values [1, 2, 3], with none of the values masked.
|
|
Because these represent the same data, I would expect the `count` method for these two arrays to return exactly the same thing.
|
|
However, In [12] a.count axis=0) Returns an integer Out[12] 3 In [13] b.count axis=0) Returns a scalar array Out[13] array 3) While the different return values act similarly in many cases, it is a nuisance and a bug, in my opinion) that the return types are not the same.","bug"
|
|
"6013","3370","test mode raw test linalg.TestQR) fails with Accelerate.
|
|
On Mac OS X 10.8.3 with the current numpy master, the test fails as such .
|
|
Commenting out lines in the test shows that all of the `old assert almost equal` calls eventually fail, but I suppose the outputs are not that informative so I am not pasting them.
|
|
I assume this is, again, an Accelerate issue.","bug"
|
|
"6013","3375","numpy scalar array like == performance horror.
|
|
When I run this .
|
|
I get this .
|
|
Is there a way to tell `numpy`, no, don t worry about it, just call ` rmul ` on the whole thing, instead of picking it apart?
|
|
In my specific case, `MyThing` is an array like object that lives on a GPU, and while it s possible and not necessarily incorrect) to pick the array apart in this way, it s unexpected and has really terrible performance.
|
|
sorry about the many edits)","bug"
|
|
"6013","3377","Heisenbug MemoryError.
|
|
This shows up now an then on the Travisbot after recent commits.
|
|
.","bug"
|
|
"6013","338","Indexing columns of an empty numpy array could return an empty numpy array.
|
|
I have a function that returns a numpy array of key value pairs or an empty numpy array and wish to access the keys values using slice notation with columns.
|
|
My issue is that when I have an empty numpy array, column indexing throws exceptions.
|
|
.
|
|
For empty numpy arrays, indexing columns breaks .
|
|
Also, returning all the columns has a similar behavior for an empty array.
|
|
.
|
|
Indexing any column of an empty numpy array could return an empty numpy array instead of throwing an ``Invalid index`` exception, and likewise in the case of returning all the columns.","bug"
|
|
"6013","3381","Inefficient memory and cpu usage by numpy ndarray operations.
|
|
When working with large 2D arrays, numpy runs out of memory very fast, probably because the memory is not used efficiently.
|
|
For example the following commands produce a MemoryError medians = np.ma.median arr, axis=1) arr medians However, I can work around this problem by serializing the my operations with a for loop.
|
|
Moreover, only one CPU is used by numpy, although it should use all by default.
|
|
So I had to code this all by hand as follows def get medians spec) median procs = [] pool = multiprocessing.Pool multiprocessing.cpu count )) channels = np.shape spec)[0] for i in range channels) median procs.append pool.apply async np.ma.median, spec[i, ],))) pool.close ) medians = np.ones channels) for i in range channels) medians[i] = median procs[i].get ) return medians Numpy should do this by itself.
|
|
The same is true for multiplications and divisions spec = spec medians[ , np.newaxis]","enhancement"
|
|
"6013","3382","np.round 4.5) != round 4.5).
|
|
Is there any reason why numpy.round 4.5) does not return 5.0 but 4.0?
|
|
pythons round 4.5) returns 5.0, as expected.","bug"
|
|
"6013","3384","Memory leak in dot in 1.8.0.dev 780d829.
|
|
I think I ve found a memory leak in 1.8.0.dev 780d829.
|
|
I m using Python 3.3.
|
|
The leak isn t present in numpy 1.7.1 .","bug"
|
|
"6013","3386","genfromtxt with names=True doesn t skip comments.
|
|
If numpy.genfromtxt ) is called with `names=True`, as is the case with numpy.recfromcsv ), then the comment lines are ignored.
|
|
This is due to the following code block .
|
|
Is this the intended behavior?
|
|
It seems like comments should be read and ignored regardless if `names is True`.","bug"
|
|
"6013","3388","Datetime timezone handling broken in 1.7.x.
|
|
For a description, see http thread.gmane.org gmane.comp.python.numeric.general 53805 Something needs to be done about this for 1.8.0","bug"
|
|
"6013","3389","document npv change.
|
|
Just so it isn t forgotten, the npv change should be mentioned in the release notes.","bug"
|
|
"6013","3390","transpose ) does not accept keyword argument `axis`.
|
|
.","bug"
|
|
"6013","3391","Documentation Missing negative sign in integer limits.
|
|
Looking at http docs.scipy.org doc numpy user basics.types.html The lower limit for int64 is missing the appropriate minus sign.
|
|
The description for int64 should read Integer 9223372036854775808 to 9223372036854775807)","bug"
|
|
"6013","3392","np.array ) very slow when args contain None or list of lists.
|
|
Discussed on StackOverflow here http stackoverflow.com questions 16819261 why is numpy array is sometimes very slow Patch files to fix this are here https github.com m d w numpy ctor patch.git Details I m using the numpy.array ) function to create numpy.float64 ndarrays from lists.
|
|
I noticed that this is very slow when either the list contains None or a list of lists is provided.
|
|
Below are some examples with times.
|
|
There are obvious workarounds but why is this so slow?
|
|
Examples for list of None Very slow to call array ) with list of None In [3] timeit numpy.array [None] 100000, dtype=numpy.float64) 1 loops, best of 3 240 ms per loop Problem doesn t exist with array of zeroes In [4] timeit numpy.array [0.0] 100000, dtype=numpy.float64) 100 loops, best of 3 9.94 ms per loop Also fast if we use dtype=object and convert to float64 In [5] timeit numpy.array [None] 100000, dtype=numpy.object).astype numpy.float64) 100 loops, best of 3 4.92 ms per loop Also fast if we use fromiter ) insead of array ) In [6] timeit numpy.fromiter [None] 100000, dtype=numpy.float64) 100 loops, best of 3 3.29 ms per loop Examples for list of lists Very slow to create column matrix In [7] timeit numpy.array [[0.0]] 100000, dtype=numpy.float64) 1 loops, best of 3 353 ms per loop No problem to create column vector and reshape In [8] timeit numpy.array [0.0] 100000, dtype=numpy.float64).reshape 1,1)) 100 loops, best of 3 10 ms per loop Can use itertools to flatten input lists In [9] timeit numpy.fromiter itertools.chain.from iterable [[0.0]] 100000),dtype=numpy.float64).reshape 1,1)) 100 loops, best of 3 9.65 ms per loop After patching was 240 ms, best alternate version was 3.29 In [5] timeit numpy.array [None] 100000) 100 loops, best of 3 7.49 ms per loop After patching was 353 ms, best alternate version was 9.65 In [6] timeit numpy.array [[0.0]] 100000) 10 loops, best of 3 23.7 ms per loop","bug"
|
|
"6013","3395","Heisenbug reference count test mismatch.
|
|
I suspect this is a bug in the test.
|
|
.","bug"
|
|
"6013","3398","String dtype inference conversion broken recently.
|
|
In NumPy 1.8. there is a regression probably caused by gh 3270 for string creation from objects types.
|
|
I get bad results here on master for `np.array 12345678901234567890, dtype=object).astype str)` which gives `array 12345678 , dtype= |S8 )`.
|
|
Plus I just get a segfault for `np.array 123456789012345678901234567890, dtype=str)`.
|
|
I still think that we should maybe go through `PyArray GetArrayParamsFromObject` or doing something similar and infer the dtype from the actual number object length which means adding object string length inference there), instead of this conversion size table.","bug"
|
|
"6013","3399","Test failure and two errors.
|
|
This is on Ubuntu 12.04.
|
|
Python 3.3.1 was built from source.
|
|
Numpy is from the master branch, commit e9e490a54.
|
|
The failure was discussed on the mailing list on June 1, and can be fixed by using appropriate precision specifications in some of the array comparison assertions in the function `test dot array order )` in the file `numpy core tests test blasdot.py`.
|
|
.","bug"
|
|
"6013","3405","f2py support for MinGW w64.
|
|
I d like to build 64 bit Python extensions on Windows using the MinGW w64 compiler, which provides gfortran.
|
|
I ve [explored this question] http stackoverflow.com questions 16929544 building 64 bit python extensions with f2py on windows), and it seems that much of f2py somewhat supports it with [a small hack] http scientificcomputingco.blogspot.com.au 2013 02 f2py on 64bit windows python27.html).
|
|
The hack fix is to pass and comment out `NotImplementedError Only MS compiler supported with gfortran on win64 )`.
|
|
I can then successfully compile an extension from `cmd.exe` using .
|
|
The resulting foo.pyd file works with the same version of Python ... success!
|
|
However, this is not yet feature rich, as I cannot seem to statically link in the GCC gfortran components, as the resulting pyd file depends on 64 bit versions of libgfortran .dll.
|
|
I d like to package this Python extension to others without supplying these GCC DLLs.
|
|
I ve tried using either ` static` or ` static libgfortran` flags to one of either the ` f90flags=` argument, or the environment variables `FFLAGS` and `CFLAGS`.
|
|
These don t seem to work for f2py, and I m not sure why, even though I can confirm that the same gfortran compiler can statically build executables that do not depend on the GCC DLLs.","enhancement"
|
|
"6013","3406","regression in 1.7.x series within ENH missingdata .. ) incorrect np.asarray s result shape when fed ctypes data.
|
|
.
|
|
Initially detected with PyMVPA s unittest http github.com PyMVPA PyMVPA blob HEAD mvpa2 tests test procrust.py testing functionality of https github.com PyMVPA PyMVPA blob HEAD mvpa2 support lapack svd.py and crashing with .
|
|
without any obvious reason, in differing locations depending on numpy s treeish the issue was bisect to lie somewhere within 321b8bf..25a64c8 where intermediate commits didn t build some still might... forgot details) Problem was troubleshooted with Sebastian AKA seberg on IRC) many kudos to him, and as a resolution use of .frombuffer instead of .asarray was suggested and seems to resolve the crash in my case .","bug"
|
|
"6013","3408","SVD instructions Example code has a mistake.
|
|
Wasn t sure to leave this here or scipy.
|
|
This url http docs.scipy.org doc numpy reference generated numpy.linalg.svd.html Reconstruction based on a full SVD reads like this >>> U, s, V = np.linalg.svd a, full matrices=True) >>> U.shape, V.shape, s.shape 9, 6), 6, 6), 6,)) However, if full matrices=True, then the shape of U is 9,9).
|
|
From the parameters section full matrices bool, optional If True default), u and v have the shapes M, M) and N, N), respectively.
|
|
Otherwise, the shapes are M, K) and K, N), respectively, where K = min M, N).
|
|
I was working through this and simultaneously trying to understand SVD, and this threw me for a loop when I could not execute this u, s, v = np.linalg.svd a) np.dot u, np.dot s,v)) This gives an error because the matrix dimensions do not align.","bug"
|
|
"6013","3409","Float to int data type coercion fails with opaque AttributeError when some magic number is exceeded.
|
|
On vanilla 1.7.1 from Ubuntu Raring sources <pre> [GCC 4.7.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.7.1 >>> X = numpy.zeros 1,dtype=int) >>> numpy.iinfo int).max 9223372036854775807 >>> X[0] = 9223372036854775807 >>> X[0] = 9223372036854775808 Traceback most recent call last) File <stdin> , line 1, in <module> OverflowError Python int too large to convert to C long >>> so far so good ... >>> X[0] = numpy.array [9223372036854775807 512],dtype=float) >>> X[0] = numpy.array [9223372036854775807 511],dtype=float) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError setting an array element with a sequence.
|
|
>>> WTF?!
|
|
< pre> The magic number is the system s maximum int64 value minus 512.
|
|
Weird.","bug"
|
|
"6013","3413","einsum crashes with default memory layout.
|
|
Using numpy version 1.7.1 the following code crashes <pre> import numpy as np A = np.ones 10000) B = np.array [A, A, A]).T T = np.ones 3, 3)) OK np.einsum iq,ik,pk >ipq , B, B, T, order= C ) Segmentation fault np.einsum iq,ik,pk >ipq , B, B, T) < pre>","bug"
|
|
"6013","3423","ENH Add class methods as callback function subroutines from fortran in f2py.
|
|
I would like to use a class method as callback function subroutine from Fortran when using Fortran code with the help of f2py.
|
|
Therefore, it is necessary to somehow store the `self` pointer in Fortran, because it is necessary as first argument when calling a class method.
|
|
How can this be accomplished?
|
|
If I use static class methods, so I do not need `self` everything works) Here is an example which compiles but fails during runtime classCallback.py .
|
|
classCallback.f90 .
|
|
The f2py command using cygwin) .
|
|
This is the runtime error .
|
|
Obviously self is not a function subroutine pointer but this was the only way to get it compiled or pass the Python type check.
|
|
E.g.
|
|
if I use `INTEGER 4)` for the pointer `self` it also fails .","bug"
|
|
"6013","3424","test kind.TestKind.test all fails on sparc.
|
|
current git head 75cdf3d82e test kind.TestKind.test all fails on sparc64 debian unstable, linux 2.6.32 5, big endian, 16 byte long double, no unaligned loads) .","bug"
|
|
"6013","3425","f2py fails for functions returning character len=len input string)).
|
|
I ve written simple toupper ) and tolower ) functions in Fortran that start like this .
|
|
When wrapping these with f2py, I get an error looking like this > tmp tmpmrZ5Ra src.linux x86 64 2.7 RMCfunctionsmodule.c In function xe2x80x98f2py rout RMCfunctions mod functions toupperxe2x80x99 > tmp tmpmrZ5Ra src.linux x86 64 2.7 RMCfunctionsmodule.c 926 23 error xe2x80x98str Dimsxe2x80x99 undeclared first use in this function) > .
|
|
I guess the `len str)` needs to be replaced by `slen str)`.
|
|
This problem occurs with both f2py 2.7 and f2py 3.3 It does not depend on the name of the variable `str` I tried a couple of other names.","bug"
|
|
"6013","3426","intent in) breaks f2py h for python 3 UnboundLocalError in crackfortran.py).
|
|
.
|
|
Compiling with `f2py m foo breaksf2py.f90` works fine, but `f2py m foo h foo.pyf breaksf2py.f90` gives the following error for python3 f2py tested with 3.2 and 3.3) > Reading fortran codes... > Reading file breaksf2py.f90 format free) > Post processing... > Block foo > Block breaksf2py > Post processing stage 2)... > Saving signatures to file .
|
|
foo.pyf > Traceback most recent call last) > File usr bin f2py 3.2 , line 24, in <module> > main ) > File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 590, in main > run main sys.argv[1 ]) > File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 361, in run main > postlist=callcrackfortran files,options) > File usr lib64 python3.2 site packages numpy f2py f2py2e.py , line 286, in callcrackfortran > pyf=crackfortran.crack2fortran postlist) > File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2749, in crack2fortran > pyf=crack2fortrangen block) \n > File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2517, in crack2fortrangen > ret=ret crack2fortrangen g,tab,as interface=as interface) > File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2561, in crack2fortrangen > vars=vars2fortran block,block[ vars ],argsl,tab tabchar, as interface=as interface) > File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2711, in vars2fortran > lst = true intent list vars[a]) > File usr lib64 python3.2 site packages numpy f2py crackfortran.py , line 2612, in true intent list > if c > UnboundLocalError local variable c referenced before assignment Python2 f2py can run `f2py 2.7 m foo h foo.pyf breaksf2py.f90` fine, and python3 f2py can compile the resulting pyf file with `f2py 3.3 m foo c foo.pyf breaksf2py.f90`.","bug"
|
|
"6013","3427","PyArray Transpose does not anticipate subclasses.
|
|
Transposing seems to be well defined for subclasses.
|
|
However, the machinery does not anticipate the subclass doing anything fancy it could in principle change the shape, ignoring that, it could force a copy).
|
|
It is simply assumed that the subclass will not do anything about shape or data pointer as far as I can see.
|
|
Probably anything it tries to do about shape will simply get overwritten in any case.","bug"
|
|
"6013","3428","Regression in 1.7.1 ctor.c setArrayFromSequence ).
|
|
setArrayFromSequence ) returns error code 0 on error >>> arr = np.array [[1,2,3],[4,5,6]]) >>> arr[ ,1] = [] This blows up in 1.6.2, not in 1.7.1 >>> arr[0,1] 2 >>> arr Kaboom!
|
|
Here is the patch diff ur numpy 1.7.1 orig numpy core src multiarray ctors.c numpy 1.7.1 numpy core src multiarray ctors.c numpy 1.7.1 orig numpy core src multiarray ctors.c 2013 04 07 06 04 05.000000000 0100 numpy 1.7.1 numpy core src multiarray ctors.c 2013 06 12 15 25 59.423961489 0100 383,7 383,7 int dim, npy intp offset) { Py ssize t i, slen; int res = 0; int res = 1; This code is to ensure that the sequence access below will diff ur numpy 1.7.1 orig numpy core tests test regression.py numpy 1.7.1 numpy core tests test regression.py numpy 1.7.1 orig numpy core tests test regression.py 2013 04 07 06 04 05.000000000 0100 numpy 1.7.1 numpy core tests test regression.py 2013 06 12 14 42 54.469727026 0100 1816,7 1816,14 order= F ) assert array equal arr2, data back) def test setArrayFromSequence fails self) 1.6.2 passes this test, 1.7.1 fails arr = np.array [[1,2,3],[4,5,6]]) try arr[ ,1] = [] except ValueError pass str arr) if name == main","bug"
|
|
"6013","3430","Unexpected return type of numpy.mean ).
|
|
The docstring for `numpy.mean` has this example that demonstrates the use of the `dtype` argument >>> a = np.zeros 2, 512 512), dtype=np.float32) >>> a[0, ] = 1.0 >>> a[1, ] = 0.1 >>> m32 = np.mean a) >>> m32 0.546875 >>> m64 = np.mean a, dtype=np.float64) >>> m64 0.55000000074505806 What is strange is that in the first case, even though the accumulator was 32 bit, the return value is 64 bit >>> type m32) <class numpy.float64 > That s unexpected.
|
|
Why do the work in 32 bit, and then return a 64 bit result?","bug"
|
|
"6013","3433","No documentation for np.save format.
|
|
The documentation page has a broken link http docs.scipy.org doc numpy dev reference generated numpy.save.html?highlight=save numpy.save It says .
|
|
but format is not a link.
|
|
It looks like ``format.py`` doesn t get installed as ``np.format``, hence why I think the link is broken.","bug"
|
|
"6013","3434","Linear Equations in NumPy.
|
|
I solve a system of linear equations and get the wrong result.
|
|
Is it my error or Numpy?
|
|
>>> a matrix [[ 1, 2, 3, 4, 5, 6, 7, 8], [ 8, 3, 5, 6, 9, 13, 19, 47], [ 10, 20, 30, 40, 50, 60, 70, 80], [ 11, 12, 13, 14, 15, 16, 17, 18], [ 21, 22, 23, 24, 25, 26, 27, 28], [ 31, 32, 43, 54, 65, 26, 47, 68], [ 11, 223, 43, 54, 65, 36, 87, 28], [ 31, 52, 73, 4, 95, 66, 47, 38]]) >>> f matrix [[12], [13], [24], [56], [68], [36], [96], [45]]) >>> X = a.I f >>> X matrix [[ 2.84378223e 15], [ 3.72400812e 15], [ 2.18442758e 16], [ 5.29098869e 15], [ 1.21532679e 16], [ 1.35353493e 16], [ 1.77236028e 16], [ 2.02048222e 15]]) >>> a X matrix [[ 16.
|
|
], [ 0.
|
|
], [ 160.
|
|
], [ 16.
|
|
], [ 88.
|
|
], [ 416.
|
|
], [ 296.
|
|
], [ 32.]])","bug"
|
|
"6013","3437","numpy scalar types behave differently from python types with len , getitem , and custom operators.
|
|
I get inconsistent behavior when a user defined class implements ` len `, ` getitem `, and binary operators.
|
|
.
|
|
The output of this code is the following .
|
|
In this example, I would expect to be able to make the following assertions .","bug"
|
|
"6013","3442","Regression float64 and Python long int multiplication fails.
|
|
Multiplying float64 and Python long integers with large values fails on current Numpy master.
|
|
This used to work and works on 1.7.1).
|
|
On master at 65fe95ccc73 .
|
|
Spotted in Scipy.","bug"
|
|
"6013","3443","Current master branch does not compile on cygwin fa5bc1c0e6f96f50dc4f20195ab6b1180726cb23).
|
|
The current master ID fa5bc1c0e6f96f50dc4f20195ab6b1180726cb23) does not compile on cygwin.
|
|
.
|
|
It is the same error message as in Issue 2537.
|
|
Version 1.7.1 does compile.","bug"
|
|
"6013","3446","possible performance regression from 1.6.2 > 1.7.0 np.any ) and np.all ) are unexpectedly slow over large arrays.
|
|
When `np.all` encounters a zero element it should return `False` immediately without testing any further elements.
|
|
Therefore, the time taken by `np.all` should not increase with increasing array size, provided that the first element in the array is always zero.
|
|
The same should be true for `np.any` if the first element in the array is nonzero.
|
|
Test script .
|
|
Results .
|
|
Bug was initially reported in relation to this [SO question] http stackoverflow.com questions 17128116 why is numpy any so slow over large arrays).","bug"
|
|
"6013","3454","np 1.7.1 savez load does not work.
|
|
If I try to follow the example in the documents, I get a gigantic traceback with an assertion error .","bug"
|
|
"6013","3457","BUG close ) method of NpzFile returned by np.load triggers ResourceWarning.
|
|
EDIT I had missed a key point in the docstring of np.load, so my original description of the issue was wrong.
|
|
When `np.load` is given an npz file, it returns an instance of `NpzFile`.
|
|
To close the file held by the `NpzFile` instance, the `close` method must be called.
|
|
However, this results in a `ResourceWarning`.
|
|
For example python3 Wdefault Python 3.3.2 default, Jun 14 2013, 12 12 22) [GCC 4.6.3] on linux Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> xy = np.load xy.npz ) >>> xy.close ) home warren local py332 lib python3.3 site packages numpy lib npyio.py 219 ResourceWarning unclosed file < io.BufferedReader name= xy.npz > self.zip.close ) And for the record >>> np.
|
|
version 1.8.0.dev 6dd20ff This ResourceWarning shows up in the numpy and scipy test suites.","bug"
|
|
"6013","3458","array ass boolean subscript ) writes non existent data to array.
|
|
Numpy 1.7.1 has tightened up the behaviour of .
|
|
compared with 1.6.2 in the case when the length of mask is different from the length of the array of values to assign.
|
|
However there are a couple of corner cases that appear wrong, namely zero length arrays and unit length arrays.
|
|
For example .
|
|
In the first case .
|
|
in mapping.c is writing the 0th element of an empty array to the matching slot which seems wrong.
|
|
It seems to me that raising a ValueError in all cases where the lengths miss match is desirable and probably intended) not least as it is identical to Pythons extended slicing rules.
|
|
Here is a patch that fixes it, and adds tests, in 1.7.1 and works for 1.8) .","bug"
|
|
"6013","3459","np.polynomial.legendre.leggauss 1).
|
|
This gives an error with my numpy version 1.7.1, but maybe it should return ` array [0]), array [2.0]))` instead?","bug"
|
|
"6013","346","Backport PR 345 to 1.6.x.
|
|
1.7.x was cut from master on July 18th, 2012 with PR already applied","bug"
|
|
"6013","3462","DeprecationWarning for = with python3.
|
|
Hello, I m not sure if this behavior is intentional.
|
|
Basically .
|
|
Is being thrown, even when there is no casting, example .
|
|
There is a possibility for float output, but the warning is still thrown if nothing actually casted.","bug"
|
|
"6013","3464","savez fails when tmp files are size limited..
|
|
The size limit causes problems because savez goes through savez, which creates a temporary file, then zips it.
|
|
The size limit arises in recent Linux distros Ubuntu, Fedora, others that have started mounting tmpfs on tmp.
|
|
The tmpfs file system is essentially a ram disk and uses memory, which limits the size available.
|
|
The advantage is spped and that SSD life is extended by not doing large numbers of small writes.
|
|
On Linux is it probably better to use var tmp when writing big temporary files, but probably the easiest solution would be to put the temporary file in the user s directory.
|
|
It gets deleted in any case.","bug"
|
|
"6013","3467","BUG np.polynomial.
|
|
.fromroots unexpected result with domain.
|
|
I need to construct monic polynomials with prescribed roots in an interval ..
|
|
According to the [documentation] http docs.scipy.org doc numpy reference generated numpy.polynomial.chebyshev.Chebyshev.fromroots.html numpy.polynomial.chebyshev.Chebyshev.fromroots) I can provide the roots and the domain in the .
|
|
function in order to obtain a numerically stable representation e.g.
|
|
in the Chebyshev basis) of the monic polynomial in the domain .
|
|
by shifting and scaling the domain to .).
|
|
However, the resulting polynomials are not monic anymore when a domain is specified.
|
|
Here is a minimal example with the polynomial .
|
|
on the domain .
|
|
.
|
|
This results in .
|
|
I expect that also .. .
|
|
does not seem to handle the domain correctly but it can be fixed by scaling the polynomial .
|
|
I m not sure whether this is actually a bug or it is intended to behave this way but the documentation is ambiguous.
|
|
One way or the other, I think it should be addressed.
|
|
;)","bug"
|
|
"6013","3469","Slicing corner case using a length 0 mask as a fancy index raises IndexError..
|
|
The script below Python 2.7, NumPy 1.7.1) fails for me with an IndexError .
|
|
Is this intended behaviour?
|
|
I was expecting `data[mask]` to give an array of shape ` 0, 3)`.
|
|
.","bug"
|
|
"6013","347","assert array almost equal nulp tests only pass by chance.
|
|
There are a ton of tests in numpy testing tests test utils.py that use np.random without setting a seed.
|
|
So first of all, that s a bug just by itself, because it means that the actual tests that are being run depends randomly on what other tests have been run as I discovered today while adding a totally different test that touch np.random and suddenly the tests in test utils started failing.
|
|
So all these tests should be using an explicit random seed.
|
|
Second, the fact that these tests are failing randomly as soon as the input is perturbed suggests that either `assert array almost equal nulp` or its tests are broken.
|
|
And that we should be testing a lot more cases, since obviously the existing tests have not really been working it s really easy to do `for i in xrange 100) do random test )`!).
|
|
Since I m trying to get this other patch submitted, for now I ll just add a loop to `TestArrayAlmostEqualNulp.test simple` to force it to fail consistently, and knownfail it.
|
|
But someone who knows about this module should really take a look...","bug"
|
|
"6013","3473","Save large array with numpy.savez compressed).
|
|
longer version http stackoverflow.com questions 17298129 python pickle size limit Save data with .
|
|
Load data again with .
|
|
results in error message http pastebin.com ms3xhsaV OS MacOSX 10.8.4 64bit) 4GB RAM and Linux CentOS 6.4 64bit) 250GB RAM Python 2.7.4 Mac) 2.73 Linux) Numpy 1.8.0.dev 20224ea Mac) 1.7.0 Linux)","bug"
|
|
"6013","3474","np.copy masked array) has unexpected behavior.
|
|
....where unexpected means I didn t expect it.
|
|
This may just be a documentation fix, or maybe I will be scolded for not realizing how `np.ma` works, but I don t think this behavior is appropriate .
|
|
I would at least expect a warning, but would prefer to see, in order of preference `np.copy x)` returns a masked array since x is a masked array) `np.copy x)` warns that it is unmasking the array, then returns the same array as above `np.copy x)` raises an exception `np.copy x)` returns `[1,1,nan,1,1]` this is probably not ideal)","bug"
|
|
"6013","3475","Implement GSVD in numpy.linalg.
|
|
numpy.linalg gives an svd function to calculate the singular value decomposition of a matrix, but not a gsvd for a generalized SVD of a matrix pair.
|
|
This would be very handy to have on hand for exploratory data analysis.","bug"
|
|
"6013","3477","genfromtxt very slow and use 10x more memory than needs.
|
|
Hello, I tried to load a 380 MB .xsf file which is basically 3D array 400x800x100 = 32 milions of of floats using numpy.genfromtxt ).
|
|
It took > 1 minute and it consumed 3.6 GB of memory the I killed it).
|
|
I don t know how this function is working but the resulting ndarray for type double 32bit) should has size just 128 MB.
|
|
Thus even if it would read the whole file into memory, and then parse it to ndarray word by word it should not consume more than 128 380 =508 MB of memory.
|
|
also parsing the same data in fortran, C or GDL is severaltimes faster.","bug"
|
|
"6013","3479","np.ma should complain that it can t compute a .mean on a object array .
|
|
The following code causes a division by zero exception .
|
|
Strangely, if `x` doesn t have any `None`s but `mx` is created with the same elements masked, the `mean )` call doesn t throw an exception.
|
|
This issue is important to address when analyzing data in which some measurements are invalid `None` in this case).","bug"
|
|
"6013","3480","numpy.random.binomial ) gives error when n = 0.
|
|
The code import numpy numpy.random.binomial 0, 0.5) Gives `ValueError n <= 0` This is correct according to the documentation of the function, but it is counter intuitive, since the binomial distribution is well defined for n=0.
|
|
random.binomial 0, anything) should always return 0.","bug"
|
|
"6013","3481","Addition and multiplication by a number on string numpy arrays raise exceptions.
|
|
In my opinion the following code should work following Python s regular operation on arrays instead of raising exceptions import numpy as np a = np.array [ a , b , c ]) print a a TypeError unsupported operand type s) for numpy.ndarray and numpy.ndarray print 2 a TypeError unsupported operand type s) for int and numpy.ndarray","enhancement"
|
|
"6013","3484","Division broken when dtype specified.
|
|
In a head build using python 3 .","bug"
|
|
"6013","3485","Fast Fourier Transform based on FFTW library.
|
|
Hello, Guys ) I found intresting python module, it names anfft https code.google.com p anfft It provides fast Fourier Transform based on FFTW library, that works in several times more faster then numpy.fft.fft.
|
|
!
|
|
This very simple to use import anfft result = anfft.fft data) It would be very nice to see such function in Numpy ) .
|
|
More functions based on FFT became more faster when works on anfft.fft ).
|
|
I will really happy if Numpy will have fft function based on FFTW library )","enhancement"
|
|
"6013","3487","Documentation bug for PyArray GETPTR1.
|
|
The signature for PyArray GETPTR1 is given as void PyArray GETPTR1 PyObject obj, \<npy intp\> i) in http docs.scipy.org doc numpy reference c api.array.html However, it should be void PyArray GETPTR1 PyArrayObject obj, npy intp i) Similarly for PyArray GETPTR2, PyArray GETPTR3, PyArray GETPTR4.","bug"
|
|
"6013","3488","`column stack )` error in documentation.
|
|
in `numy lib shape base.py`, around line 277 This function is equivalent to ``np.vstack tup).T``.
|
|
If I m not mistaken, this was true for the old behaviour.
|
|
Currently, inputs with >= 2 dimensions are not transposed which is good!)
|
|
and therefore it is not equivalent anymore.
|
|
Here is an example, the commented line gives an error .","bug"
|
|
"6013","3492","bad hash for int64 objects in 32bit systems.
|
|
There s a problem with hashing numpy.int64 objects on 32 bit systems.
|
|
It doesn t follow the basic pythonic rule a == b > hash a) == hash b) In [53] a = 356487790608L In [54] b = numpy.int64 a) In [55] a == b Out[55] True In [56] hash a) == hash b) Out[56] False Needless to say this leads to wierd things that cause errors, such as In [57] set [a,b]) Out[57] set [356487790608, 356487790608L]) Everything seems to works fine on my 64bit system.","bug"
|
|
"6013","3493","bug on rank method.
|
|
The following example illustrates the bug the array is a bit long but it needs to be ate least 32 lines long) array [[1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0], [1, 0, 1, 0]]) >>> a=np.matrix 1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0;1,0,1,0 ) >>> np.linalg.matrix rank a) 2 Clearly, the answer should be 1.
|
|
Thanks in advance.
|
|
I.","bug"
|
|
"6013","3494","np.insert with negative axis scalar.
|
|
Just a quick npte, as was reported on the mailing list, np.insert is broken with negative axis when insert obj is a scalar.
|
|
This is becuase of np.rollaxis use with axis 1 and a regression in both 1.7 and 1.8.","bug"
|
|
"6013","3498","non integer deprecation warning in as strided.
|
|
.
|
|
throws an deprecation warning even though no float was used in the user code .","bug"
|
|
"6013","3499","numpy.prod do not warn when integer overflow arise.
|
|
In [9] np.seterr all= raise ) Out[9] { divide warn , invalid warn , over warn , under ignore } In [10] np.prod [i for i in range 1,55)]) Out[10] 7927461244078915584","bug"
|
|
"6013","3502","Don t try to) handle array priority in ufunc object.c badly).
|
|
This issue is a continuation of the disussion with seberg in PR 3501 In my opinion, having ` array priority ` logic in `umath` is just broken.
|
|
It should be ripped out.
|
|
A tell tale sign of this is that the logic is activated only for two argument ufuncs with one output.
|
|
What s so special about them?
|
|
Why should they behave differently from three argument ufuncs?
|
|
Also, as seberg pointed out, why should a ufunc ever return `NotImplemented`?
|
|
See here https github.com numpy numpy blob master numpy core src umath ufunc object.c L2362 https github.com numpy numpy blob master numpy core src umath ufunc object.c L1971","bug"
|
|
"6013","3503","array multiply != the multiply ufunc.
|
|
This issue is a continuation of the disussion with seberg in 3501, along with 3502 Lots of places such as [1] https github.com numpy numpy blob master numpy core src multiarray multiarraymodule.c L950), [2] https github.com numpy numpy blob master numpy core src multiarray scalartypes.c.src L255)) use `PyArray Type >tp as number >nb multiply` to call the `multiply` ufunc etc.).
|
|
Because of ` array priority `, the semantics of `nb multiply` are no longer the same as the ufunc.
|
|
Specifically, `nb multiply` is required to to return `NotImplemented` in cases of priority inversion, whereas the ufunc has a job to do and just needs to do it, without asking questions about priorities.","bug"
|
|
"6013","3504","linspace behaves unexpectedly when start argument is an integer array scalar.. start and stop arguments of linspace ) are supposed to be scalar, but it seems like linspace also accepts array scalars.
|
|
However it fails due to truncation in some cases.
|
|
3 examples .
|
|
The first example points out what s going wrong, but the other ones are really tricky because the user doesn t get any warning.
|
|
A simple fix would be to add the following 2 lines at the beginning of the linspace ) function .","bug"
|
|
"6013","3508","nanmin on arrays with non numeric data.
|
|
I m unsure whether this behaviour of min with non numeric dtypes) is actually supported, docstrings seem to suggest not) but there are unexpected results, on 1.7.1 .
|
|
Strangely ?
|
|
), if the dtype is not object it min et al, raise .
|
|
From discussion on [this pandas SO question] http stackoverflow.com a 17497460 1240268).","bug"
|
|
"6013","3511","Incompatible numpy.float64.
|
|
round ).
|
|
.
|
|
i.e.
|
|
.)
|
|
is expected to return an ., but for numpy floats, it returns a numpy float.","bug"
|
|
"6013","3512","Invalid testcase from gh 3410. gh 3410 introduced a test failure, am I the only one seeing it?
|
|
I think the buffer test case is simply wrong, though I am not sure why it works on Travis, maybe it is not run.
|
|
For all I see buffer should be equivalent to memoryview.
|
|
m d w just tagging, if you remember or got a minute to remove it.
|
|
This is what I mean .","bug"
|
|
"6013","3515","like functions not recognizing broadcast objects shape.
|
|
I expected `np.ones like bdc)` to be the same as `np.ones bdc.shape)` but it isn t the former doesn t work.
|
|
.","bug"
|
|
"6013","3519","Convert WarningManager uses to warnings.catch warnings.. Now that all supported Python versions have the `with` statement WarningManager should be dispensed with and deprecated.","bug"
|
|
"6013","3523","Wrong dtypes returned from var method.. .
|
|
There are other problems of this sort.","bug"
|
|
"6013","3525","Installing Numpy on Mac 10.8 fails.
|
|
Trying to install numpy using pip install numpy .
|
|
fails with the error below.
|
|
I tried getting more info on scipy on irc.freenode.net.
|
|
and was suggested to post a bug here.
|
|
Python 2.7.5 was installed using brew.
|
|
Also tried to install via pip 2.7 install numpy and i get the same error.
|
|
Original intention was to install gdal via brew install gdal which has dependency on numpy and hence trying to install this.
|
|
compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 include python2.7 c Traceback most recent call last) File <string> , line 16, in <module> File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy setup.py , line 214, in <module> setup package ) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy setup.py , line 207, in setup package configuration=configuration ) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils core.py , line 186, in setup return old setup new attr) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command egg info.py , line 8, in run self.run command build src ) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command build src.py , line 152, in run self.build sources ) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command build src.py , line 163, in build sources self.build library sources libname info) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command build src.py , line 298, in build library sources sources = self.generate sources sources, lib name, build info)) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command build src.py , line 385, in generate sources source = func extension, build dir) File numpy core setup.py , line 646, in get mathlib info st = config cmd.try link int main void) { return 0;} ) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils command config.py , line 251, in try link libraries, library dirs, lang) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command config.py , line 149, in link libraries, library dirs, lang)) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command config.py , line 89, in wrap method ret = mth self,) args)) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils command config.py , line 143, in link src, obj) = self.
|
|
compile body, headers, include dirs, lang) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command config.py , line 99, in compile body, headers, include dirs, lang)) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils command config.py , line 89, in wrap method ret = mth self,) args)) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 distutils command config.py , line 138, in compile self.compiler.compile [src], include dirs=include dirs) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils ccompiler.py , line 203, in CCompiler compile self.
|
|
compile obj, src, ext, cc args, extra postargs, pp opts) File private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy numpy distutils unixccompiler.py , line 34, in UnixCCompiler compile self.spawn self.compiler so cc args [src, o , obj] TypeError coercing to Unicode need string or buffer, list found Command python setup.py egg info failed with error code 1 in private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy Exception information Traceback most recent call last) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 site packages pip 1.3.1 py2.7.egg pip basecommand.py , line 139, in main status = self.run options, args) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 site packages pip 1.3.1 py2.7.egg pip commands install.py , line 266, in run requirement set.prepare files finder, force root egg info=self.bundle, bundle=self.bundle) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 site packages pip 1.3.1 py2.7.egg pip req.py , line 1057, in prepare files req to install.run egg info ) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 site packages pip 1.3.1 py2.7.egg pip req.py , line 236, in run egg info command desc= python setup.py egg info ) File usr local Cellar python 2.7.5 Frameworks Python.framework Versions 2.7 lib python2.7 site packages pip 1.3.1 py2.7.egg pip util.py , line 662, in call subprocess command desc, proc.returncode, cwd)) InstallationError Command python setup.py egg info failed with error code 1 in private var folders 61 zlk3fs0x6 j5zdqm3t m0hww0000gp T pip build aagarwal numpy","bug"
|
|
"6013","3528","Inconsistent behavior of getitem on arrays with length 0 dimensions.
|
|
Fancy indexing a zero length dimension with a zero length array may work or fail, depending on whether there the array is 1d or has more dimensions This works np.zeros 0,))[[]] np.zeros 0, 1))[ , ] None of this works np.zeros 0, 1))[[]] np.zeros 0, 0))[[]] np.zeros 0, 1))[[], ] Those raise a very strange IndexError I never asked for index 0!)
|
|
IndexError Traceback most recent call last) <ipython input 101 ae785fa98259> in <module> ) > 1 np.zeros 0, 1))[[]] IndexError index 0 is out of bounds for axis 0 with size 0 I think it would be best if the second set of lines would have behavior identical to the first one, since it reduces the need for special case handling in user code.
|
|
Am I missing some other rule that would be broken by such a change?","bug"
|
|
"6013","3530","ENH Register NumPy scalar types with ABCs from Python s `numbers` module.
|
|
NumPy scalar type hierarchy is documented here http docs.scipy.org doc numpy reference arrays.scalars.html Python has a `numbers` module since Python 2.6, which implements numeric ABCs http docs.python.org 2 library numbers.html http docs.python.org 3 library numbers.html Inheriting from these ABCs will make it possible to do `isinstance )` checks on NumPy scalars, which seems to be a valuable feature.
|
|
According to our IRC discussion with seberg, it might be as easy as to add similar calls at the right place for the correct types .
|
|
I believe that the correct place might be `numpy core numerictypes.py`.
|
|
Unfortunately, not being a NumPy core expert, I m really clueless as to whether this is the right strategy or not, so I m not attaching a patch.
|
|
Thanks for your consideration, Yury.","bug"
|
|
"6013","3532","busday count return incorrect value.
|
|
See test in comments.. numpy 1.7.1. np.busday count 2013 02 02 , 2013 04 04 ) return 43 days... should be 44 days","bug"
|
|
"6013","3536","Non masked value treated as masked.
|
|
Steps to reproduce a = np.arange 6).reshape 3, 2) array [[0, 1], [2, 3], [4, 5]]) b = np.ma.masked where a > 5, a) print b.mask False b[b.mask] = 999 print b [[ 999 999] [ 2 3] [ 4 5]] I am passing some masked arrays to a rebin function with averages.
|
|
The returned array is then assigned a NoData value where a mask exists and written to a file.
|
|
For now I implemented a check to verify the count of unmasked values before I assign NoData if count == array size I make no assignment).
|
|
Is there a different method to handle this?
|
|
Windows 7 64bit Python 2.7 32bit Numpy 1.7.0 also verified on 1.7.1)","bug"
|
|
"6013","3540","subtypes and round ).
|
|
round ) does not consistently return an ndarray subtype test results are similar on windows and linux64 C \Program Files x86)\console2>c \Python27\python Python 2.7.3 default, Apr 10 2012, 23 31 26) [MSC v.1500 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.version.version 1.7.0 >>> class C np.ndarray) ... pass ... >>> a=np.arange 4).view C) >>> a C [0, 1, 2, 3]) >>> a.round decimals=0) C [0, 1, 2, 3]) >>> a.round decimals= 1) array [0, 0, 0, 0]) >>> a.round decimals=1) C [0, 1, 2, 3]) >>> a=np.arange 4,dtype=float).view C) >>> a.round decimals=1) array [ 0., 1., 2., 3.])
|
|
>>> a.round decimals=0) C [ 0., 1., 2., 3.])
|
|
>>> a.round decimals= 1) array [ 0., 0., 0., 0.])","bug"
|
|
"6013","3542","cumsum ) optional argument dtype not properly honored.
|
|
The optional argument dtype of the NumPy function cumsum ) seems to be not respected on some platforms.
|
|
The bug became apparent when writing a C extension.
|
|
Configuration NumPy 1.7.0, Python 2.7.3, OS Mandriva 10, 32bits, Steps to reproduce 1.
|
|
Write a simple C extension and compile it with the usual python setup.py build 2.
|
|
Minimal snipet from the C extension PyObject hello PyObject self, PyObject args) { PyObject obj; if !PyArg ParseTuple args, O , obj)) return NULL; if PyArray Check obj)) { printf Received an array OK.\n ); PyArray Descr desc = PyArray DESCR obj); if desc >type num != NPY INT32 ) printf Failed First element is NOT a NPY INT32.\n ); else printf OK First element is a NPY INT32.\n ); } else { printf Not an array!\n ); } int sts = 0; return Py BuildValue i , sts); } 3.
|
|
In Python import numpy as np import my module as mm a=np.array 0,dtype=np.int32) a.resize 20) a[ ]=4 b=a.cumsum dtype=np.int32) mm.hello a) mm.hello b) print b.dtype 4.
|
|
Results The last line abve prints dtype int32 ), as expected, but the check made in the C code fails only with b not with a !!
|
|
), indicating that internally we are not dealing with NPY INT32 elements.
|
|
Note that we deal with several platforms, and this was only seen on this one Mandriva 10, 32bits).","bug"
|
|
"6013","3543","masked arrays do not operate commutatively with numpy scalars.
|
|
When performing an operation with a masked array and a numpy scalar the order of the operands will change the results.
|
|
Example .
|
|
Output .
|
|
I assume this occurs because numpy.int64.
|
|
mult and the like do not take into account masked values.
|
|
This differs from ndarray objects where the order of the operands does not matter.","bug"
|
|
"6013","3544","BUG str and repr special methods blow up memory usage.
|
|
.
|
|
Calculated memory usage 43200 21600.
|
|
8) 43200 21600) 8398080000.0 bits > 0.9776651859283 GB","bug"
|
|
"6013","3545","Segmentation fault using numpy.linalg.eig h) function with OpenBLAS linked library.
|
|
On the latest version 1.8.0.dev 90ececa, when using OpenBLAS as BLAS LAPACK library, the functions numpy.linalg.eig h) cause a segmentation fault on matrices with a dimension larger than 1290.
|
|
This does not occur using the standard BLAS LAPACK libraries on version 1.7.1 or when using the same function on the scipy.linalg package linked to OpenBLAS, which seems to suggest a bug on Numpy s end.
|
|
Sample code import numpy import scipy mat = numpy.random.random 1291,1291)) 1290 works fine.
|
|
numpy.linalg.eig mat) This segfaults scipy.linalg.eig mat) This works","bug"
|
|
"6013","3548","BUG Assignment of complex rank zero arrays produces misleading error.
|
|
This all makes sense to me .
|
|
This also makes sense .
|
|
This does not make sense .
|
|
The problem is not that array 4j) is a sequence ; array 3) works fine.
|
|
The problem is that it s complex and you re trying to put it in a real array.
|
|
Took me a long time to figure out why `scipy.signal.impulse array [1, 1]), array [1, 1]), 15))` works but `scipy.signal.impulse array [1, 1]), array [1, 1]), 15 0j))` does not.","bug"
|
|
"6013","3551","BUG deprecated support for int ndarrays in condlists for select ) .
|
|
MILESTONE 1.9 BLOCKERS Deprecated at suggestion of njsmith and gbb.
|
|
Please see discussion here https github.com numpy numpy pull 3537","bug"
|
|
"6013","3553","Inconsistent flag setting behaviors.
|
|
If we create a matrix of size 1,1) with different orders, the transpose operation gives inconsistent flag setting behaviors 1) If we do A = np.zeros 1,1)), then A.flags has C CONTIGUOUS True and F CONTIGUOUS False, and A.T.flags still has C CONTIGUOUS True and F CONTIGUOUS False NO flag change takes place.
|
|
2) If we do A = np.zeros 1,1), order= F ), then A.flags has C CONTIGUOUS False and F CONTIGUOUS True, and A.T.flags has C CONTIGUOUS True and F CONTIGUOUS False there are flag change taking place.
|
|
It s a marginal case and actually both flags could be true, but the current behavior makes it pretty inconsistent when converting from one order to another.","bug"
|
|
"6013","3554","Redundency in description for loadtext documentation.
|
|
This issue concerns with documentation for [numpy.loadtext] https github.com numpy numpy blob v1.7.0 numpy lib npyio.py L649) For the parameter `unpack`, it says so that arguments may be unpacked using ``x, y, z = loadtxt ...)``.
|
|
The way I see it, all `unpack` does is, it transposes the array.
|
|
Why is the above line necessary?
|
|
Please do feel free to correct me if I am wrong.
|
|
I feel like I might be missing something obvious here.","bug"
|
|
"6013","3555","Polynomial roots computation calls unnecessarily general LAPACK routine.
|
|
The companion matrix is already upper Hessenberg, so it should be possible to use a specialized routine to find its eigenvalues, rather than the general non symmetric eigenvalue routine.
|
|
In particular, lib polynomial.py L227 should call d,z)hseqr rather than eigvals, which calls d,z)geev.","enhancement"
|
|
"6013","3557","numpy distutils support for MinGW is broken under Python 3.
|
|
Here is the traceback .
|
|
This is numpy 1.7.1.","bug"
|
|
"6013","3561","BUG Incorrect assignment to recarray items.
|
|
Assigning to field of recarray is not working correctly.
|
|
Example script is below.
|
|
.","bug"
|
|
"6013","3562","f2py Unable to expose parameters from xe2x80x9cusedxe2x80x9d modules.
|
|
I have tried to put the problem in the first paragraph.
|
|
The rest shows a basic example of the problem.
|
|
I am attempting to compile a module that contains a `USE` statement pointing to another, more general, module.
|
|
I would prefer to keep the used module separate so that it can be used in several places as a set of general settings.
|
|
When I complie the two modules using f2py, everything works correctly from the fortran side, but fromt he python side `USE` appears to be ignored.
|
|
If I allow f2py to generate a signature file, the file contains a `USE` statement as is appropriate, but if I complete the compilation and import from the resulting library, the parameters from the used module are not available in the module that contains the use statement.
|
|
Below are two modules illustrating the situation MODULE test INTEGER, PARAMETER a = 1 END MODULE test MODULE test2 USE test INTEGER, PARAMETER b = 2 END MODULE test2 In order to show the intermediate step I ran `f2py h test.pyf test.f90 test2.f90`.
|
|
The following signature file is generated; note that the test2 module contains `use test` !
|
|
f90 !
|
|
Note the context of this file is case sensitive.
|
|
python module test !
|
|
in interface !
|
|
in test module test !
|
|
in test test.f90 integer, parameter,optional a=1 end module test module test2 !
|
|
in test test2.f90 use test integer, parameter,optional b=2 end module test2 end interface end python module test !
|
|
This file was auto generated with f2py version 2).
|
|
!
|
|
See http cens.ioc.ee projects f2py2e If I now compile with `f2py fcompiler=gfortran c test.pyf test.f90 test2.f90` I obtain test.so same as running `f2py fcompiler=gfortran m test c test.f90 test2.f90` without creating the signature file first).
|
|
Importing from this library in python exposes test.test.a and test.test2.b, but does not expose test.test2.a as can be seen here In [1] import test In [2] print test.test.a 1 In [3] print test.test2.b 2 In [4] print test.test2.a AttributeError Traceback most recent call last) users solbrig svn checkouts inversion satmet branches solbrig rootpath data users GeoIPS src test <ipython input 4 bffcf464e408> in <module> ) > 1 print test.test2.a AttributeError a Just to illustrate that `b` is defined properly in test2 from the perspective of fortran, the following code uses test2 and prings both `a` and `b` SUBROUTINE run test ) USE test2 IMPLICIT NONE print , a = , a print , b = , b END SUBROUTINE run test After compiling with `f2py m run test c test.f90 test2.f90 run test.f90` and obtaining run test.so, run test can be imported in python and works as expected In [1] import run test In [2] run test.run test ) a = 1 b = 2 Please let me know if there is any other information that I can provide or if there is a solution for this issue.","bug"
|
|
"6013","3563","Confusing documentation for remainder and fmod.
|
|
Current numpy documentation for fmod says that This is the NumPy implementation of the Python modulo operator But this is confusing, since fmod and give different results fmod 3,2) > 1 3 2 > 1 Furthermore, the np.remainder documentation makes no mention of fmod, nor of sign conventions.
|
|
This is bad, since users coming from Matlab may use remainder to translate the Matlab rem , while the two things are not the same.","enhancement"
|
|
"6013","3571","include this patch for building with openblas?.
|
|
Maybe I m going about it the wrong way, but I have to apply this patch each time I want to be able to build numpy with openblas.
|
|
Is there a reason not to add it to numpy distutils?
|
|
https github.com akesandgren numpy commit 363339dd3a9826f3e3e7dc4248c258d3c4dfcd7c I can then build fine with this in `site.cfg` .","enhancement"
|
|
"6013","3573","Installation bug on Python 3.3.
|
|
I m trying to install NumPy 1.7.1 for Python 3.3 using pip install numpy However, I get the following error after a while error numpy.egg info dependency links.txt Operation not supported Is this a bug or am I doing something wrong?
|
|
If it matters, I m using virtualenv as I do not have root permission on this computer.
|
|
The end of the error log Command home jluttine .virtualenvs bayespy 3.3 bin python3.3 c import setuptools; file = home jluttine .virtualenvs bayespy 3.3 build numpy setup.py ;exec compile open file ).read ).replace \r\n , \n ), file , exec )) install record tmp pip 309wd8 record install record.txt single version externally managed install headers home jluttine .virtualenvs bayespy 3.3 include site python3.3 failed with error code 1 in home jluttine .virtualenvs bayespy 3.3 build numpy Exception information Traceback most recent call last) File home jluttine .virtualenvs bayespy 3.3 lib python3.3 site packages pip 1.2.1 py3.3.egg pip basecommand.py , line 107, in main status = self.run options, args) File home jluttine .virtualenvs bayespy 3.3 lib python3.3 site packages pip 1.2.1 py3.3.egg pip commands install.py , line 261, in run requirement set.install install options, global options) File home jluttine .virtualenvs bayespy 3.3 lib python3.3 site packages pip 1.2.1 py3.3.egg pip req.py , line 1166, in install requirement.install install options, global options) File home jluttine .virtualenvs bayespy 3.3 lib python3.3 site packages pip 1.2.1 py3.3.egg pip req.py , line 589, in install cwd=self.source dir, filter stdout=self.
|
|
filter install, show stdout=False) File home jluttine .virtualenvs bayespy 3.3 lib python3.3 site packages pip 1.2.1 py3.3.egg pip util.py , line 612, in call subprocess command desc, proc.returncode, cwd)) pip.exceptions.InstallationError Command home jluttine .virtualenvs bayespy 3.3 bin python3.3 c import setuptools; file = home jluttine .virtualenvs bayespy 3.3 build numpy setup.py ;exec compile open file ).read ).replace \r\n , \n ), file , exec )) install record tmp pip 309wd8 record install record.txt single version externally managed install headers home jluttine .virtualenvs bayespy 3.3 include site python3.3 failed with error code 1 in home jluttine .virtualenvs bayespy 3.3 build numpy","bug"
|
|
"6013","3580","numpy.bool fails isinstance x, numbers.Integral).
|
|
Numpy s bool type seems to fail an isinstance check with the Python numbers.Integral ABC .
|
|
I find this odd since Python s built in bool passes .
|
|
Is this a bug?
|
|
My system information is PYTHON Python 2.7.5 default, May 19 2013, 02 56 42) [GCC 4.2.1 Compatible Apple LLVM 4.2 clang 425.0.28)] on darwin NUMPY INSTALLED WITH PIP Downloading unpacking numpy from https pypi.python.org packages source n numpy numpy 1.7.1.tar.gz md5=0ab72b3b83528a7ae79c6df9042d61c6 Downloading numpy 1.7.1.tar.gz 2.8MB) 2.8MB downloaded Running setup.py egg info for package numpy Running from numpy source directory.
|
|
non existing path in numpy distutils site.cfg F2PY Version 2 blas opt info FOUND extra link args = [ Wl, framework , Wl,Accelerate ] define macros = [ NO ATLAS INFO , 3)] extra compile args = [ msse3 , I System Library Frameworks vecLib.framework Headers ] lapack opt info FOUND extra link args = [ Wl, framework , Wl,Accelerate ] define macros = [ NO ATLAS INFO , 3)] extra compile args = [ msse3 ]","bug"
|
|
"6013","3581","record array is made up of numpy.void, not numpy.record.
|
|
I raised this previously on the PyTables issue list, but it they have asked me to transfer to numpy.
|
|
https github.com PyTables PyTables issues 271) >>> import numpy as np >>> a = 1, 2, 3) >>> ra = np.rec.fromrecords [a]) >>> na = np.array [a], dtype= i8,i8,i8 ).view type=np.recarray) >>> na rec.array [ 1, 2, 3)], dtype=[ f0 , <i8 ), f1 , <i8 ), f2 , <i8 )]) >>> ra rec.array [ 1, 2, 3)], dtype=[ f0 , <i8 ), f1 , <i8 ), f2 , <i8 )]) >>> type ra) <class numpy.core.records.recarray > >>> type na) <class numpy.core.records.recarray > >>> ra[0] 1, 2, 3) >>> na[0] 1, 2, 3) >>> type ra[0]) <class numpy.core.records.record > >>> type na[0]) <class numpy.void > Any reason why .view can t return record arrays instead of sort of fake record arrays, or voidarrays?
|
|
Is there a cast missing in the code somewhere?","bug"
|
|
"6013","3582","BUG Wrong steps for GUFUNC arguments with repeated dimensions, affects umath linalg.
|
|
When creating a GUFUNC with a signature that has repeated dimensions for an argument, e.g.
|
|
umath linalg.solve has m,m), m,n) > m,n) , the steps argument passed on to the gufunc loop has incorrect strides for the argument with repeated dimensions.
|
|
In the solve example, one gets 9 total strides 3 main loop strides one for each argument), followed by two strides for the core dimensions of each of the arguments.
|
|
So steps[3] and steps[4] should be row and column strides of the first argument.
|
|
But for repeating dimensions as above, all the strides are equal to the last value, i.e.
|
|
steps[4] is correct, but steps[3] is wrong and equal to steps[4].
|
|
This has gone undetected in the umath linalg tests, where there are signatures with repeating values, either because the test cases fail to show the issue it s the case for solve), or because the tests are comparing results from several functions all having the same issue it s the case of det and the several eigenvalue functions, which rely on each other for correctness, and are all affected by the bug.)
|
|
Not sure of what goes on with svd, which has an instance also affected by the same issue.","bug"
|
|
"6013","3583","loadtxt ..., unpack=True) returns non contiguous arrays.
|
|
I was reading a simple two column float data set from a file with `numpy.loadtxt` numpy 1.6.2) and stumbled over non contiguous neither C nor F) arrays after unpacking the columns.
|
|
This seems strange behavior to me, I would have expected that `unpack=True` takes care of this otherwise what is the advantage over an appended `.T`?).
|
|
.
|
|
.","bug"
|
|
"6013","3585","Support anonymous mmap in numpy.memmap ).
|
|
Copied from scipy issues Original ticket http projects.scipy.org scipy ticket 1677 on 2012 06 14 by trac user iandavis, assigned to unknown.
|
|
As far as I can tell, numpy.memmap ) does not currently support anonymous mode MAP ANONYMOUS for C s mmap), where there is no backing file.
|
|
However, I ve found MAP ANON useful for creating sparse arrays larger than available memory, where most array entries are zero.
|
|
It s fast and efficient, and doesn t use any disk space or cause any disk IO) unless you outgrow RAM.
|
|
For certain applications, it has major advantages over scipy.sparse for instance, it can handle any shape and number of dimensions efficiently, while scipy.sparse only does 2 D matrices.
|
|
Anyway, I ve adapted numpy.memmap ) into the sparse zeros ) class below.
|
|
I don t understand what all the support code was for originally, so it s possible there are bugs places to improve.
|
|
I d propose this be added to either scipy.sparse, or directly to numpy itself.
|
|
I didn t see an obvious way to add this to numpy.memmap ) because it expects there to be a file name, but that would be a third possibility.
|
|
import numpy as np class sparse zeros np.ndarray) Copied from numpy.core.memmap v1.5.1) Provides a zeros ) like array backed by an anonymous mmap ).
|
|
Only pages with non zero values require memory storage.
|
|
If enough pages are written to, however, you ll still get swapping.
|
|
array priority = 100.0 def new subtype, shape, dtype=np.uint8, order= C ) Import here to minimize import numpy overhead import mmap descr = np.dtype dtype) if not isinstance shape, tuple) shape = shape,) bytes = descr.itemsize for k in shape bytes = k acc = mmap.ACCESS COPY mm = mmap.mmap 1, bytes, access=acc) self = np.ndarray.
|
|
new subtype, shape, dtype=descr, buffer=mm, order=order) self.
|
|
mmap = mm return self def array finalize self, obj) if hasattr obj, mmap ) self.
|
|
mmap = obj.
|
|
mmap else self.
|
|
mmap = None def flush self) pass def sync self) This method is deprecated, use `flush`.
|
|
warnings.warn Use ``flush``.
|
|
, DeprecationWarning) self.flush ) def close self) Close the memmap file.
|
|
Only do this when deleting the object.
|
|
if self.base is self.
|
|
mmap The python mmap probably causes flush on close, but we put this here for safety self.
|
|
mmap.flush ) self.
|
|
mmap.close ) self.
|
|
mmap = None def close self) Close the memmap file.
|
|
Does nothing.
|
|
warnings.warn ``close`` is deprecated on memmap arrays.
|
|
Use del , DeprecationWarning) def del self) We first check if we are the owner of the mmap, rather than a view, so deleting a view does not call close on the parent mmap if self.
|
|
mmap is self.base try First run tell ) to see whether file is open self.
|
|
mmap.tell ) except ValueError pass else self.
|
|
close )","enhancement"
|
|
"6013","3586","Pass function through scipy.interpolate.interp1d parameter fill value.
|
|
It would be nice if the fill value parameter of scipy.interpolate.interp1d accepted functions.
|
|
This would allow one to handle data outside of the range with other methods such as extrapolation.
|
|
Thanks!","enhancement"
|
|
"6013","3588","Printing masked object array generates AttributeError NotImplementedType object has no attribute view .
|
|
Dear all, I think this is a bug in NumPy, albeit an obscure one.
|
|
I have a numpy object array which contains two numpy arrays, of different lengths .
|
|
The array can be handled fine, and the member arrays also, but as soon as any printing operation `str )` or `repr )`) are called on the main array, it generates this AttributeError NotImplementedType object has no attribute view I have the numpy array as a pickle, included at the end of this bug report.
|
|
To exercise the bug just do this >>> import cPickle >>> aa=cPickle.load open array repr failing.pkl , rb )) >>> repr aa) This should show the problem clearly >>> type aa[0]) <class numpy.ma.core.MaskedArray > >>> aa[0].dtype dtype float64 ) >>> aa[0].shape 2,) >>> print aa[0] [64.21187757583947 63.1578947368421] >>> repr aa[0]) masked array data = [64.21187757583947 63.1578947368421],\n mask = [False False],\n fill value = nan)\n >>> type aa[1]) <class numpy.ma.core.MaskedArray > >>> aa[1].dtype dtype float64 ) >>> aa[1].shape 3,) >>> print aa[1] [73.3547875205356 73.81810530054845 74.28142308056131] >>> repr aa[1]) masked array data = [73.3547875205356 73.81810530054845 74.28142308056131],\n mask = [False False False],\n fill value = nan)\n >>> repr aa) Traceback most recent call last) File array repr failing.py , line 18, in <module> repr aa) File usr lib python2.7 site packages numpy core numeric.py , line 1409, in array repr , , array ) File usr lib python2.7 site packages numpy core arrayprint.py , line 445, in array2string separator, prefix, formatter=formatter) File usr lib python2.7 site packages numpy core arrayprint.py , line 248, in array2string int IntegerFormat data), File usr lib python2.7 site packages numpy core arrayprint.py , line 638, in init max str len = max len str maximum.reduce data))), File usr lib python2.7 site packages numpy ma core.py , line 3579, in eq check = ndarray.
|
|
eq self.filled 0), odata).view type self)) My versions are python 2.7.5 numpy 1.7.1 numpy.ma 1.0 cPickle 1.71 Here is the pickle with the data cnumpy.core.multiarray reconstruct p1 cnumpy ndarray p2 I0 tS b tRp3 I1 I2 tcnumpy dtype p4 S O4 I0 I1 tRp5 I3 S | NNNI 1 I 1 I63 tbI00 lp6 cnumpy.ma.core mareconstruct p7 cnumpy.ma.core MaskedArray p8 g2 I0 tp9 S b tRp10 I1 I2 tg4 S f8 I0 I1 tRp11 I3 S < NNNI 1 I 1 I0 tbI00 S \x1f\xbf\xf6f\x8f\rP Cy\r\xe55\x94O S \x00\x00 cnumpy.core.multiarray scalar p12 g4 S f8 I0 I1 tRp13 I3 S < NNNI 1 I 1 I0 tbS \x00\x00\x00\x00\x00\x00\xf8\x7f tRp14 tbag7 g8 g2 g9 S b tRp15 I1 I3 tg11 I00 S \xadn\xb7\xd6\xb4VR \x8d\xa2U\xd6[tR m\xd6\xf3\xd5\x02\x92R S \x00\x00\x00 g12 g13 S \x00\x00\x00\x00\x00\x00\xf8\x7f tRp16 tbatb.","bug"
|
|
"6013","3589","PyArray MapIterBind leaks memory on failure.
|
|
when run the test indexing.py file leaks 9 MB of memory.
|
|
This seems to be due to PyArray MapIterBind not dereferencing the full iterator on failure, nor its two callers doing so.
|
|
dereferencing it in the fail label fixes the leak and the testsuite still passes but I can t judge the side effects.","bug"
|
|
"6013","3595","arccosh works for much larger numbers than arccos.
|
|
arccosh x) = i arccos x), so `np.arccos )` and `np.arccosh )` should produce the same output but with real and imaginary flipped .
|
|
but at 10 7, arccos is inaccurate .
|
|
and then gives up at 10 8 and outputs infs .
|
|
but arccosh keeps going to almost the full float range .
|
|
and [is still accurate] http www.wolframalpha.com input ?i=arccos[10 307]) so whatever algorithm is used for arccosh could presumably be used for arccos also, just with the imaginariness flipped.","bug"
|
|
"6013","3598","Build fails with Python 3.4a1.
|
|
End of build log gcc numpy core src multiarray arrayobject.c numpy core src multiarray arrayobject.c In function xe2x80x98array richcomparexe2x80x99 numpy core src multiarray arrayobject.c 1307 13 error ISO C90 forbids mixed declarations and code [ Werror=declaration after statement] numpy core src multiarray arrayobject.c 1364 13 error ISO C90 forbids mixed declarations and code [ Werror=declaration after statement] cc1 some warnings being treated as errors numpy core src multiarray arrayobject.c In function xc3xa2array richcomparexc3xa2 numpy core src multiarray arrayobject.c 1307 13 error ISO C90 forbids mixed declarations and code [ Werror=declaration after statement] numpy core src multiarray arrayobject.c 1364 13 error ISO C90 forbids mixed declarations and code [ Werror=declaration after statement] cc1 some warnings being treated as errors error Command gcc pthread Wno unused result Werror=declaration after statement DNDEBUG g fwrapv O3 Wall Wstrict prototypes fPIC DHAVE NPY CONFIG H=1 Inumpy core include Ibuild src.linux i686 3.4 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I home rgommers Code pythons include python3.4m c numpy core src multiarray arrayobject.c o build temp.linux i686 3.4 numpy core src multiarray arrayobject.o failed with exit status 1 With Python 3.3 these are the default distutils compile flags C compiler i686 linux gnu gcc pthread DNDEBUG g fwrapv O2 Wall Wstrict prototypes g fstack protector param=ssp buffer size=4 Wformat Werror=format security D FORTIFY SOURCE=2 fPIC With Python 3.4 this becomes C compiler gcc pthread Wno unused result Werror=declaration after statement DNDEBUG g fwrapv O3 Wall Wstrict prototypes fPIC http docs.python.org 3.4 whatsnew 3.4.html says No build and C API changes , so I d think the compile args shouldn t change.
|
|
But I can t find so quickly who decided this and why.
|
|
Fix in numpy or file bug against Python?
|
|
Or both?","bug"
|
|
"6013","3602","BUG?
|
|
should np.ones like call array wrap on function return?.
|
|
Is there a reason ``np.diff,np.ones like,np.empty like`` do not call `` array wrap `` on exit; am I doing something wrong, or is this a bug?
|
|
using numpy 1.7.1) Create a NDLike class which is a container for a ndarray .
|
|
A ufunc calls `` array `` on function enter, and `` array wrap `` on exit .
|
|
I have found that both ``np.diff, np.ones like, np.empty like`` DO NOT appear to call `` array wrap `` on function exit.
|
|
This only difference that I can see is that these are basically pieces of python code wrapping other code, rather than direct c level calls, e.g.
|
|
``np.ones like`` calls ``empty`` a c level call.
|
|
.
|
|
Is there a reason these do not call `` array wrap `` on exit; am I doing something wrong, or is this a bug?
|
|
Some context I have refactored pandas ``Series`` object from a ``ndarray`` sub class to a sub class that is common in the pandas hierarchy https github.com pydata pandas pull 3482), for ease of code use modification, etc.
|
|
This issue only comes up if the ``ndarrayness`` is somewhat abused in that a user is expecting that a ``Series`` object that goes into an ndarray comes out with another ``Series`` object.
|
|
which is the case for nearly every method ufunc by using either wrapped methods or the `` array array wrap `` machinery) Thanks for taking a look Jeff","enhancement"
|
|
"6013","3603","Improving numpy.gradient to be second order accurate over the full domain.
|
|
Currently `gradient` uses a second order accurate central finite difference for interior elements, and a first order accurate forward backwards) finite difference for the first last) element.
|
|
This causes an issue at the ends of the domain where the gradient appears to jump.
|
|
By including an additional mesh point of the first and last elements i.e.
|
|
a second order one sided finite difference) it is possible to make numerical differential second order accurate over the whole domain.
|
|
For example.
|
|
!
|
|
[second order accurate] https f.cloud.github.com assets 112847 945162 7c153044 02f9 11e3 98ae 3d7e6433c07a.png) The second order accurate finite difference were calculated as follows, .
|
|
For more information on one sided finite difference approximations, the paper, [ One sided finite difference approximations suitable for use with Richardson extrapolation ] ftp ftp.demec.ufpr.br CFD bibliografia MER Rahul Bhattacharyya 2006.pdf) gives a nice review.","bug"
|
|
"6013","3605","NumPy matrices `.nonzero )` returns matrices instead of arrays.
|
|
.
|
|
NumPy matrices `.nonzero )` return a tuple of matrices instead of arrays.
|
|
The [documentation] http docs.scipy.org doc numpy reference generated numpy.nonzero.html) states that it should be a tuple of arrays.
|
|
Code example .","bug"
|
|
"6013","3611","Nansum of all NaN or empty slices should return 0..
|
|
This is a change scheduled for 1.9.","bug"
|
|
"6013","3613","frombuffer returns incorrect result with non contiguous input array.
|
|
For non contiguous input arrays, numpy.frombuffer ignores the stride the array.
|
|
The output has the correct length, but it assumes a stride equal to the item size.
|
|
The following commands were tested under Numpy 1.7.1 and Python 3.3. a = numpy.array [0, 1, 2, 3], dtype= i8 ) numpy.frombuffer a[0 4 2], dtype= i4 ) The output is [0, 0, 1, 0], but was expected to be [0, 1, 2, 3].
|
|
I also tested under NumPy 1.6.1 and Python 2.7.
|
|
For that configuration an exception was raised TypeError expected a single segment buffer object .","bug"
|
|
"6013","3614","object with dtype property crashes numpy.
|
|
The following code crashes numpy latest master) .","bug"
|
|
"6013","3615","Problem with imports in polynomial.polynomial docstring examples.
|
|
In most of the functions in polynomial.polynomial you can read this line >>> from numpy import polynomial as P and then >>> P.polyline 1, 1) >>> sum = P.polyadd c1,c2); sum >>> P.polysub c1,c2) >>> P.polyder c) d dx) c) = 2 6x 12x 2 [...] https github.com numpy numpy blob v1.7.1 numpy polynomial polynomial.py but all these lines fail because all those members are part of the polynomial module in the polynomial package.
|
|
It should be therefore like >>> P.polynomial.polyline 1, 1) >>> sum = P.polynomial.polyadd c1,c2); sum >>> P.polynomial.polysub c1,c2) >>> P.polynomial.polyder c) d dx) c) = 2 6x 12x 2 [...] Nevertheless, this looks like a bit too long for me and I wonder if it would be possible to import some convenience functions using power series polynomials polynomial.polyfit and things like that) directly into the polynomial package namespace.
|
|
I wrote polynomial too many times already )","bug"
|
|
"6013","3620","Improve ``np.take`` to work correctly when indices=boolean array.
|
|
Hi, When you pass a boolean array as ``indices`` to ``np.take`` you probably want to do boolean indexing or masking).
|
|
``np.take`` however converts the boolean array into an array of zeros and ones and uses that as indexing array.
|
|
Which is probably not what the user expected.
|
|
This is very confusing if not wrong.
|
|
Confusing, because most users probably expect ``np.take`` to behave like `` getitem `` to accept index or boolean arrays.
|
|
And if that is not the case one should at least put the analogous method ``np.compress`` which works like ``np.take`` just with boolean arrays in the `See Also` Section of ``np.take``.","bug"
|
|
"6013","3621","Add numpy.phase in addition to numpy.angle.
|
|
I wanted to suggest adding a function which directly gives the phase of a complex valued number instead of the angle .
|
|
This is a simple enhancement, which I think would make numpy more consistent and offer the benefit of simply being faster in large loops, rather than going the `ang = numpy.angle z); phase = numpy.cos ang) 1j numpy.sin ang)` route or `numpy.cos 1j ang)`, for what it s worth).","enhancement"
|
|
"6013","3627","View of rows as strings regression in 1.7 dtype cast).
|
|
The following code, based on 1) and used in 2), works in numpy 1.6 and fails in numpy 1.7 and 1.8.0.dev 3bc6b96.
|
|
It uses a casts a 2D integer array to a string view of the rows.
|
|
It use it to determine unique vectors in an array.
|
|
<pre> import numpy as np print np.
|
|
version a=np.arange 12).reshape 4,3) a[2]= 3,4,5) print a This line is ok on np 1.6.1, error on np 1.7.1 is TypeError data type S24S24S24S24 not understood on numpy 1.7.1 unq, unq idx = np.unique a.view S d a.itemsize a.shape[1])) a.shape[0]), return index=True) print unq, unq idx print a[unq idx, ] < pre> 1) http www.mail archive.com numpy discussion scipy.org msg04176.html 2) https bitbucket.org deeplycloudy lmatools src aa52817e41e796e185f70cdb86ea1659859ab7d4 density tools.py Previously mentioned in issue 3159 and pull request 3170.","bug"
|
|
"6013","3634","Use the builtin BufferError for the buffer API.
|
|
Python provides and documents that the builtin BufferError should be used in the buffer interface, but numpy currently uses ValueErrors exclusively.
|
|
This may be limited to the get buffer functionality, or might also be relevant for some other funcitons I did not check yet).","bug"
|
|
"6013","3637","Remove numarray and oldnumeric modules..","bug"
|
|
"6013","3640","Is this an error in numpy doc on numpy.polyfit?.
|
|
On page http docs.scipy.org doc numpy reference generated numpy.polyfit.html It first stated Fit a polynomial p x) = p[0] x deg ... p[deg] of degree deg to points x, y).
|
|
Returns a vector of coefficients p that minimises the squared error.
|
|
And then later, it said in the equations x[0] n p[n] ... x[0] p[1] p[0] = y[0] x[1] n p[n] ... x[1] p[1] p[0] = y[1] ... x[k] n p[n] ... x[k] p[1] p[0] = y[k] Should the latter be consistent with the former, at least to reduce confusion?
|
|
Instead of going from p[n] to p[0], it should be p[0] to p[n].
|
|
Thanks!","bug"
|
|
"6013","3641","return mutiple field selection as ro view.
|
|
This transition is scheduled for 1.9","bug"
|
|
"6013","3643","Errors when comparing ndarrays with Ellipsis.
|
|
These fail but shouldn t. .
|
|
But these work.
|
|
Which seems odd since the other case is broken.
|
|
.","bug"
|
|
"6013","3646","BUG In Ellipsis matrix indexing..
|
|
I think this is a bug.
|
|
Otherwise it could that I don t understand how ellipses ares supposed to behave.
|
|
If we have .
|
|
Indexing with some `Ellipsis` s give weird results.
|
|
The first example makes sense .
|
|
But I don t understand why the shape is different here .","bug"
|
|
"6013","3648","numpy.rec.fromrecords fails to ensure dtype= O for sequences.
|
|
When building a recarray where some elements are sequences or numpy arrays, `numpy.rec.fromrecords` does not directly ensure that these obtain `dtype= O `.
|
|
This makes data disappear in the following .","bug"
|
|
"6013","3649","numpy.complex128 can easily overflow taking ratios of large quantities.
|
|
.
|
|
.
|
|
1e 200j This is OK .
|
|
0j Overflow Weird enough, using arrays of complex128 there is no overflow .
|
|
9.9999999999999998e 201j","bug"
|
|
"6013","3650","numpy.pad still breaks for pad widths of zero.
|
|
`numpy.pad` still seems to break for pad widths of zero.
|
|
I noticed this commit c6ce294f6feeebcdf67b238564c770387be933a5 but the problem seems to persist <pre><code> >>> print np.
|
|
version 1.9.0.dev 202e78d >>> a = np.arange 12).reshape 3,4) >>> np.pad a, 1,2), 2,1)), mode= constant ) case 1 ok >>> np.pad a, 0,0), 2,1)), mode= constant ) case 2 breaks < code>< pre> Traceback <pre><code> ValueError Traceback most recent call last) <ipython input 11 58abfba54dea> in <module> ) > 1 np.pad a, pad, mode= constant , constant values=99) home alistair .venvs pylab base local lib python2.7 site packages numpy lib arraypad.pyc in pad array, pad width, mode, kwargs) 1280 1281 narray = np.array array) > 1282 pad width = validate lengths narray, pad width) 1283 1284 allowedkwargs = { home alistair .venvs pylab base local lib python2.7 site packages numpy lib arraypad.pyc in validate lengths narray, number elements) 1083 if chk[0] < 0) or chk[1] < 0) 1084 fmt = s cannot contain negative values.
|
|
> 1085 raise ValueError fmt number elements,)) 1086 return normshp 1087 ValueError 0, 0), 2, 1)) cannot contain negative values.
|
|
< code>< pre> Case 2 works fine in 1.7.1","bug"
|
|
"6013","3656","FAIL Test numpy dot with different order C, F. Using the latest numpy maintenance 1.7.x code with msvc and Intel s MKL on Windows, the following test fails on all Python versions .
|
|
.
|
|
.","bug"
|
|
"6013","3659","numpy.i Macro array size conflicts with boost range.
|
|
The `array size` macro defined in `numpy.i` [here] https github.com numpy numpy blob 938f586d6f61654dec8e9e3f14084e98d86f60e9 doc swig numpy.i L43) and [here] https github.com numpy numpy blob 938f586d6f61654dec8e9e3f14084e98d86f60e9 doc swig numpy.i L57)) conflicts with the `boost range` headers from the [boost] http www.boost.org ) library, since they use `array size` as an identifier.
|
|
There is also a number of similarly named macros in the same file, which may presumably cause similar name clashes in the future.
|
|
I don t know if it would be possible to reimplement these as functions; otherwise I hope that they can be renamed to something less generic in order to avoid name clashes.","bug"
|
|
"6013","3660","numpy doc Is an URL field possible for public facing APIs?.
|
|
I m working on a project that is using https github.com numpy numpy blob master doc HOWTO DOCUMENT.rst.txt to document the controllers and such.
|
|
It would be excellent if there as a URL section too for hosted, public facing APIs so people know what URLs they can hit.
|
|
I thought about forking it, but I wonder if it would be useful to other projects as well?
|
|
Not sure.
|
|
Feel free to close if this has already been addressed; I searched through the issues but didn t see anything related.","bug"
|
|
"6013","3662","BUG arrays initialize to dtype object when multiplied by long..
|
|
When initializing a numpy array, I get strange behavior when I multiply it by a long int on the same line as it is created.
|
|
This is in python version 3.2.3 and numpy version 1.6.1 Compare .
|
|
to .
|
|
The second example is presumably) the desired behavior, as silent conversion of the dtype of the array seems quite dangerous.","bug"
|
|
"6013","3663","libnpysort.a and libnpymath.a linking problem PPC, MacOS X 10.5.8, Python 3.3).
|
|
La Dxc3xa9fense, le 30 08 2013 With respect to commit b350089f440fdfd28d7be1cfa23cb6b1b66fc1e4 on PPC, MacOS X 10.5.8, the libraries libnpysort.a and libnpymath.a will not link to create the numpy.core.multiarray extension, with error message ld warning in build temp.macosx 10.3 ppc 3.3 libnpymath.a, file is not of required architecture ld warning in build temp.macosx 10.3 ppc 3.3 libnpysort.a, file is not of required architecture It should be noted that these files are for the PPC architecture which is that of the rest of my Python stack), so the error message is odd I have not dug deeper yet).
|
|
The compile options were reported as follows compile options Inumpy core include Ibuild src.macosx 10.3 ppc 3.3 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I Library Frameworks Python.framework Versions 3.3 include python3.3m c compile options Inumpy core include Ibuild src.macosx 10.3 ppc 3.3 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I Library Frameworks Python.framework Versions 3.3 include python3.3m c Sure enough, trying to build scipy or bottleck) results in a failure, as a symbol is missing aheapsort bool in this case).
|
|
In may of this year all could be built correctly, so this appears to be a regression.
|
|
Merci Hubert Holin","bug"
|
|
"6013","3666","why support for axis= 1,2,3), i.e.
|
|
a tuple, is not documented?.
|
|
In NumPy 1.7.1 we have seen some answers in SO http stackoverflow.com a 18357674 832621 http stackoverflow.com a 18545281 832621 where they are suggesting to pass a tuple to the `axis` argument.
|
|
It does work but we cannot find where it is documented.
|
|
Do you have any plans to include this into the official documentation?
|
|
Thank you!","bug"
|
|
"6013","3667","numexpr.evaluate result type != numpy type on Windows.
|
|
Checking the type of an ndarray returned from a numexpr.evaluate expression) against a numpy type fails on Windows.
|
|
A ticket was raised against numexpr, but the maintainers suggested to raise it against numpy https code.google.com p numexpr issues detail?id=112).
|
|
To reproduce .
|
|
Expected output .
|
|
Actual output .
|
|
Tested on Win 7 Enterprise with the following environment 64bit python = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Numexpr version 2.1 NumPy version 1.6.2 Python version 2.7.2 default, Jun 12 2011, 14 24 46) [MSC v.1500 64 bit AMD64)] AMD Intel CPU?
|
|
True VML available?
|
|
False Detected cores 4 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = 32bit python = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Numexpr version 2.1 NumPy version 1.6.2 and 1.7.0) Python version 2.7.2 default, Jun 12 2011, 15 08 59) [MSC v.1500 32 bit Intel)] AMD Intel CPU?
|
|
True VML available?
|
|
False Detected cores 4 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Tested on Win XP with the following environment = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Numexpr version 2.2 NumPy version 1.7.1 Python version 2.7.3 default, Apr 10 2012, 23 31 26) [MSC v.1500 32 bit In tel)] AMD Intel CPU?
|
|
True VML available?
|
|
True VML MKL version Intel R) Math Kernel Library Version 11.1.0 Product Build 201 30711 for 32 bit applications Number of threads used by default 1 out of 1 detected cores) = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =","bug"
|
|
"6013","3669","Backport 3668 to 1.8.x..","bug"
|
|
"6013","3671","Travisbot is missing some tests.
|
|
On my platform, fedora 19, numpy.test ) runs several hundred more tests than Travisbot.
|
|
It would be good to track this down.","bug"
|
|
"6013","3677","np.meshgrid copy=False odd behavior.
|
|
np.meshgrid appears to have some odd behavior that I can t explain when copy=False .
|
|
X, Y should be views of a, b which we can see by modifying a, b).
|
|
.
|
|
yields .
|
|
However, Y.base is b returns False.
|
|
Where is Y coming from if it is not a view of b?
|
|
Either somewhere, something was copied which shouldn t happen because copy=False), or Y.base is not correct.","bug"
|
|
"6013","368","math functions fail confusingly on long integers and object arrays generally).
|
|
If you pass a long integer greater than sys.maxint to np.sqrt, then it blows up with a confusing error message .
|
|
This was noticed because it breaks SciPy s kendalltau function http mail.scipy.org pipermail scipy user 2012 July 032652.html It was hard to diagnose because the error message is so misleading.
|
|
It looks like what s going on is that ndarray conversion on very large long objects gives up and simply returns an object array.
|
|
So... that seems reasonable, not sure what else we could do .
|
|
And then when called on an object array, np.sqrt does its weird ad hoc fallback thing, and tries calling a .sqrt ) method on each object.
|
|
Which of course doesn t exist, since it s just something we made up.
|
|
But this is where the confusing error message comes from.)
|
|
In fact, our handling of object arrays is pretty broken all around we can t even take the square root of float objects .
|
|
The math module versions of sqrt and friends accept long integers .
|
|
Mostly this just works by calling PyFloat AsDouble, which goes via the float method if defined as it is for longs).
|
|
However, the math module does have special code for longs in some cases log in 2.7, maybe more in future versions, who knows) .
|
|
So in conclusion np.sqrt and friends, when operating on object arrays, should fall back to the stdlib math functions.
|
|
This would be in addition to the current .sqrt method fallback.
|
|
I guess the current fallback should probably be tried, since anyone defining our ad hoc methods is presumably doing so specifically because they want numpy to respect that.)","bug"
|
|
"6013","3680","Simd test failure on 32 bit windows.. .","bug"
|
|
"6013","3681","Test failure on 32 bit windows.
|
|
.","bug"
|
|
"6013","3682","Any method no longer return booleans..
|
|
The any and all functions may return reals instead of booleans.
|
|
The result was bool in 1.6.x .","bug"
|
|
"6013","3683","libpaths in numpy distutils system info.py may fail.. kkdd reports the following as a [commit comment] https github.com numpy numpy commit 96ea318b10e31c405798a3f96d93b418d188969d commitcomment 4001084) .","bug"
|
|
"6013","3687","Xerbla test fails for 1.8.0b1 linked with MKL, also OS X .
|
|
Reported by Neal Becker, compiled on Fedora 19 x86 64. .","bug"
|
|
"6013","369","Bug in numpy.where.
|
|
On a Linux machine > uname srvop Linux 2.6.18 308.8.2.el5 1 SMP Tue May 29 11 54 17 EDT 2012 x86 64 GNU Linux this example shows a problem with the where function Python 2.7.1 r271 86832, Dec 21 2010, 11 19 43) [GCC 4.1.2 20080704 Red Hat 4.1.2 48)] on linux2 Type help , copyright , credits or license for more information.
|
|
import numpy as np print np.
|
|
version 1.5.1 net = np.zeros 3, dtype= >f4 ) net[1] = 0.00458849 net[2] = 0.605202 max net = net.max ) test = np.where net <= 0., max net, net) print test [ 2.23910537e 35 4.58848989e 03 6.05202019e 01] When I specified the dtype for net as >f8 , test[0] was 3.46244974e 68.
|
|
It worked as expected i.e.
|
|
test[0] should be 0.605202) when I specified float max net) as the second argument to np.where.","bug"
|
|
"6013","370","Unpickling can corrupt an internal cache of bytes object py3).
|
|
The Python 3 interpreter has an internal cache of single byte bytes objects, from which objects are retrieved in a few operations, notably computing length 1 slices of bytes objects.
|
|
Pickling and unpickling certain arrays can give direct write access into that cache.
|
|
A simple way to reproduce the problem is .
|
|
This applies not only to master, but also to released versions, e.g.
|
|
to Ubuntu 12.04 s python3 numpy, which has numpy 1.6.1 and Python 3.2.3.
|
|
A similar problem has apparently been encountered before, as the test suite has `test pickle string overwrite )` in `numpy core tests test regression.py`, which triggers the cache corruption but doesn t detect it.","bug"
|
|
"6013","3702","All nanTest Error on OS X 10.6.8.
|
|
Reported by David Menxc3xa9ndez Hurtado ``` ====================================================================== FAIL test allnans test nanfunctions.TestNanFunctions Sum) Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib tests test nanfunctions.py , line 249, in test allnans assert len w) == 1, no warning raised ) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy testing utils.py , line 44, in assert raise AssertionError msg) AssertionError no warning raised","bug"
|
|
"6013","3703","test mode raw fails on OS X 10.8.
|
|
Seems to be non uniqueness of factorization.
|
|
.","bug"
|
|
"6013","3704","Printed message On entry to DGETRF, parameter number 4 had an illegal value .
|
|
On OS X 10.8. and 10.6.","bug"
|
|
"6013","3707","ENH new function proposal numpy.close.
|
|
I would like include a new function `numpy.close`.
|
|
The behavior would be identical to `numpy.allclose` but would apply the equality test element by element and return an array of results.
|
|
For example, def close a, b, atol=1e 8, rtol=1e 5) return np.abs a b) <= atol rtol np.abs b)) This would be useful for filtering array of small values which appear due to numerical rounding.
|
|
For example one use would be, Comparison with a float data = np.array [1e 30, 1e 20, 1e 10, 0.1, 1, 10]) data[np.where np.close data,0))] = 0. np.array [0., 0., 0., 0.1, 1, 10]) Comparison with another array same shape) data = np.array [1e 30, 1e 20, 1e 10, 0.1, 1, 10]) data[np.where np.close data,np.zeros data.shape)))] = 0. np.array [0., 0., 0., 0.1, 1, 10]) This function would allow the comparison of floating point numbers within a tolerance and would be of much more practical use than the existing comparison function when doing numerical work.
|
|
Moreover, `numpy.equal`, and `numpy.array equal` apply the boolean equals operator `==`) to each element without considering a tolerance.
|
|
For numerical applications tolerance is very important.
|
|
Do you agree this is a good idea?
|
|
If so I will prepare a pull request.","enhancement"
|
|
"6013","3708","savez compressed does not have documentation on the website.
|
|
It seems that unlike the previous request https github.com numpy numpy issues 2292 `savez compressed` now works and has documentation numpy.savez compressed file, args, kwds) xe2x94x82 xe2x94x82 Save several arrays into a single file in compressed ``.npz``xe2x94x82 xe2x94x82 format.
|
|
xe2x94x82 xe2x94x82 xe2x94x82 xe2x94x82 If keyword arguments are given, then filenames are taken fromxe2x94x82 xe2x94x82 the keywords.
|
|
xe2x94x82 xe2x94x82 If arguments are passed in with no keywords, then stored filexe2x94x82 xe2x94x82 names are xe2x94x82 xe2x94x82 arr 0, arr 1, etc.
|
|
xe2x94x82 xe2x94x82 xe2x94x82 xe2x94x82 Parameters xe2x94x82 xe2x94x82 xe2x94x82 xe2x94x82 file str However, the website http docs.scipy.org doc numpy reference generated numpy.savez.html does not have a proper link.","bug"
|
|
"6013","3710","Add PyArray SelectkindConverter to conversion utils.h.
|
|
Lack of include may cause compile error.","bug"
|
|
"6013","3713","numpy looks for MSVCR debug dll even when Visual Studio is not installed.
|
|
The debug versions of MSVCR are only available with an install of Visual Studio.
|
|
They cannot officially be downloaded separately.
|
|
I am running 64 bit Windows 7, with Mingw w64 as my C compiler.
|
|
I have the release version of the appropriate MSVCR installed.
|
|
However, I still get this message from numpy Cannot build msvcr library msvcr90d.dll not found This is not a fatal error but it causes a significant delay every time my program starts up.
|
|
I commented out the offending line in Python27 Lib site packages numpy distutils mingw32ccompiler.py so that only the release version of the MSVCR gets build, and now I don t have this issue.","bug"
|
|
"6013","3714","New behavior of np.ndindex in numpy 1.8.0dev.
|
|
The bottleneck package gives four unit test failures https github.com kwgoodman bottleneck issues 71) when run with numpy 1.8.0dev.
|
|
Two of the failures are due to this change in numpy 1.8.0dev .
|
|
whereas in numpy 1.7.1 .
|
|
Was this change intentional?","bug"
|
|
"6013","3718","KnownFailureTest should be in numpy.testing namespace.
|
|
Currently, the only way to mark a test knownfailure is to use the `knownfailureif` decorator.
|
|
However, quite often in generator tests, you d like to mark a test as a known failure depending on some condition, in the function body.
|
|
The decorators are not useful for this purpose, although it s possible to abuse them, `knownfailureif True, some message ) lambda None) )` It would be useful to have the `KnownFailureTest` exception in `numpy.testing` main namespace, so that it could be raised whenever necessary.
|
|
And maybe rename it to the more sensible `KnownFailure` the `Test` suffix seems to be imitated from `unittest.SkipTest` but that s a verb phrase where the suffix makes sense).
|
|
Similarly, it would be useful to have the `SkipTest` exception there, so that it wouldn t need to be imported from the `nose` module.","enhancement"
|
|
"6013","3719","Heap corruption when using np.cov and importing matplotlib.pyplot.
|
|
I have a code snippet that causes python to abort with a double free or corruption error.
|
|
All the code really does is run np.cov in a loop and average the results.
|
|
Interestingly, I can only reproduce the error when several things are done together matplotlib.pyplot must be imported the machine must be running the proprietary nvidia driver I think) I can t seem to reproduce the problem using gdb or valgrind either.
|
|
I know these kinds of bugs are difficult to troubleshoot but we have many machines with the same configuration and the only way a crash seems to be manifested is when using np.cov.
|
|
The graphics otherwise work and matplotlib otherwise appears to work.
|
|
Any insight into this would be much appreciated!
|
|
Here is the code .
|
|
Here is the error I get .
|
|
It only crashes about 1 50 runs so I run it in a loop from bash .
|
|
I have tried numpy 1.7.1 and the latest build checked out on August 22, 2013.
|
|
I am using matplotlib 1.2.0.","bug"
|
|
"6013","3722","Support single precision in all lapack lite routines.. A number of single precision functions look to have come in with the gufunc work.
|
|
It would be good to support them in all of the current routines.","enhancement"
|
|
"6013","3732","may share memory fails when comparing arrays with datetime dtypes.
|
|
e.g.
|
|
.
|
|
Produces .
|
|
and here s my version .
|
|
Sidenote also fails just comparing .","bug"
|
|
"6013","3733","np.percentile broken for vector `q` argument.
|
|
np.percentile claims to function with the `q` argument the percentiles to calculate) being either a scalar or an array like.
|
|
However, the initial lines of code do .
|
|
which of course breaks with a truth value of array undefined exception if `q` is in fact an ndarray.
|
|
`q` isn t converted to an array before usage, so if it s passed as a list value, these checks don t cause an exception, but only because of the sketchiness of Python s built in equality testing.
|
|
I ve gotten a report of a crash in the ` compute qth percentile` function that seems to be due to the lack of coercion.","bug"
|
|
"6013","3737","test scalarmath.TestBaseMath.test blocked ) triggers assertion failure.
|
|
test scalarmath.TestBaseMath.test blocked ) triggers assertion failure in NumPy 1.8.0 beta1 and 1.8.0 beta 2.
|
|
This test does not exist in NumPy 1.7.1 and all tests of NumPy 1.7.1 pass.
|
|
Results of test suite of NumPy 1.8.0 beta 2 .","bug"
|
|
"6013","3740","Incorrect dtype for mask of MaskedArray.
|
|
The dtype of the mask should, I think, be boolean, but isn t for structured arrays .","bug"
|
|
"6013","3742","savez ) bug on OS X 10.6.
|
|
With Python 2.6 from python.org 32 bit Intel) ====================================================================== ERROR test 1D test io.TestSavezLoad) Traceback most recent call last) File Users rgommers Code numpy numpy lib tests test io.py , line 122, in test 1D self.roundtrip a) File Users rgommers Code numpy numpy lib tests test io.py , line 144, in roundtrip assert equal arr, self.arr reloaded[ arr d n]) File Users rgommers Code numpy numpy lib npyio.py , line 250, in getitem return format.read array bytes) File Users rgommers Code numpy numpy lib format.py , line 474, in read array data = fp.read int read count dtype.itemsize)) File Library Frameworks Python.framework Versions 2.6 lib python2.6 zipfile.py , line 594, in read bytes = self.fileobj.read bytesToRead) TypeError integer argument expected, got long 6 test errors in total, all the same.
|
|
Both for current master and 1.8.x","bug"
|
|
"6013","3743","Hash changed signature in Python 3.3.. .
|
|
I [Christolph] could reproduce the weirdness outside the testsuite on 64 bit Python 3.2, 3.3 and 3.4.
|
|
It looks like the interpreter gets corrupted.
|
|
It could be due to numpy 1.8 using a wrong signature for the hash function [1] on win amd64 py3.x.
|
|
Since Python 3.2 the return value must be of type `Py hash t` defined as `Py ssize t`) [2], not `C long` as for Python <= 3.1.
|
|
[1] https github.com numpy numpy blob master numpy core src multiarray scalartypes.c.src L3038 [2] http docs.python.org 3.2 c api typeobj.html PyTypeObject.tp hash","bug"
|
|
"6013","3750","Need comprehensive tests of hashing.. Bugs due to the change of hash variable type dating to Python 3.2 slipped through without being detected.
|
|
We need to do better.","bug"
|
|
"6013","3751","WISH option for f2py modules not to allow setting attributes that weren t defined from Fortran.
|
|
I ve just spent ages tracking down a bug that was at least partially) caused by forgetting to list a module variable in the .pyf file I set the attribute in Python, but when Fortran tried to access it there was a segfault the variable was an allocatable array).
|
|
It would be helpful if there was an option to make the generated module override ` setattr ` and refuse to set attributes that aren t connected to the Fortran module.","bug"
|
|
"6013","3755","Raise warning when condition number is too high?.
|
|
In certain cases, when trying to solve singular linear systems, no feedback is given about the singular nature of the solution.
|
|
For example .
|
|
But, on other cases .
|
|
This issue was raised here http web.archive.org liveweb http www.walkingrandomly.com ?p=5092 A warning based on the condition number of the matrix has been suggested as a possible solution.","bug"
|
|
"6013","3756","Data type conversion falling over to python types causes opaque bugs .
|
|
Hi numpy folks, I have a function implementing a supergaussian function like so .
|
|
I m using the numpy math functions because this is used for curve fitting.
|
|
The way that numpy handles data types for very large integers causes opaque problems.
|
|
Basically, the denominator in my exponential gets too large for a numpy.int64, e.g.
|
|
.
|
|
is numpy.int64 but .
|
|
becomes int, presumably because it is too large for numpy.int64.
|
|
In the first case, the value in the .
|
|
call is a numpy type numpy.float64), but in the second case it becomes the python float instead.
|
|
This leads to the code throwing an error .
|
|
If I evaluate the expression in the exponential, you can see that .
|
|
returns type numpy.float64, but .
|
|
becomes a normal float.
|
|
This seems to be related to the other open issue 3409 but in this case it s actually causing a problem unless I work around by forcing the type to be numpy.float64.
|
|
I would suggest that this issue be worked around by having numbers too large for numpy.int64 fall back to numpy.float64 instead of a built in type, or do the latter with a warning error.
|
|
Thanks!","bug"
|
|
"6013","3758","Exception assignment destination is read only when reading from a read only array.
|
|
Some array methods fail when passed a immutable view of an array, although they should not try to write to it.
|
|
Tested with 1.6.1 and 1.7.1.
|
|
A short demonstration .","bug"
|
|
"6013","3759","== None seems to do is None for the array instead of returning element wise comparison.
|
|
May be related to 1608 I understand that None is quite different beast but since comparison to None is allowed, I do not see why a == None should just return False instead of returning a bool dtyped array with element wise comparisons as it does for other types.
|
|
See below that it simply returns False while comparing to None array contains.
|
|
.","bug"
|
|
"6013","3760","MinGW build broken in einsum.c.src.
|
|
Standard MinGW 3.4.5 build chain as used for releases gcc mno cygwin O2 Wall Wstrict prototypes DHAVE NPY CONFIG H=1 DNPY MINGW USE CUSTOM MSVCR D MSVCRT VERSION =0x0900 Inumpy\core\include Ibuild\src.win32 3.2\numpy\core\include numpy Inumpy\core\src\private Inumpy\core\src Inumpy\core Inumpy\core\src\npymath Inumpy\core\src\multiarray Inumpy\core\src\umath Inumpy\core\src\npysort Inumpy\core\include IC \Python32\include IC \Python32\PC c build\src.win32 3.2\numpy\core\src\multiarray\einsum.c o build\temp.win32 3.2\Release\build\src.win32 3.2\numpy\core\src\multiarray\einsum.o In file included from C Python32 include Python.h 112, from numpy\core\src\multiarray\einsum.c.src 12 C Python32 include pythonrun.h 172 warning function declaration isn t a prototype numpy\core\src\multiarray\einsum.c.src In function `double sum of products contig outstride0 one numpy\core\src\multiarray\einsum.c.src 1590 error ` m128d undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 1590 error Each undeclared identifier is reported only once numpy\core\src\multiarray\einsum.c.src 1590 error for each function it appears in.)
|
|
numpy\core\src\multiarray\einsum.c.src 1590 error syntax error before a numpy\core\src\multiarray\einsum.c.src 1120 warning implicit declaration of function ` mm prefetch numpy\core\src\multiarray\einsum.c.src 1120 error ` MM HINT T0 undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 1124 error `accum sse undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 1124 warning implicit declaration of function ` mm add pd numpy\core\src\multiarray\einsum.c.src 1124 warning implicit declaration of function ` mm load pd numpy\core\src\multiarray\einsum.c.src 1130 error `a undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 1130 warning implicit declaration of function ` mm shuffle pd numpy\core\src\multiarray\einsum.c.src 1130 warning implicit declaration of function ` MM SHUFFLE2 numpy\core\src\multiarray\einsum.c.src 1132 warning implicit declaration of function ` mm store sd numpy\core\src\multiarray\einsum.c.src 1165 warning implicit declaration of function ` mm loadu pd numpy\core\src\multiarray\einsum.c.src In function `double sum of products stride0 contig outcontig two numpy\core\src\multiarray\einsum.c.src 371 error ` m128d undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 371 error syntax error before a numpy\core\src\multiarray\einsum.c.src 424 error `value0 sse undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 424 warning implicit declaration of function ` mm set1 pd numpy\core\src\multiarray\einsum.c.src 432 error `a undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 432 warning implicit declaration of function ` mm mul pd numpy\core\src\multiarray\einsum.c.src 433 error `b undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 434 warning implicit declaration of function ` mm store pd numpy\core\src\multiarray\einsum.c.src 468 warning implicit declaration of function ` mm storeu pd numpy\core\src\multiarray\einsum.c.src In function `double sum of products contig contig outstride0 two numpy\core\src\multiarray\einsum.c.src 584 error ` m128d undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 584 error syntax error before a numpy\core\src\multiarray\einsum.c.src 651 error ` MM HINT T0 undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 655 error `a undeclared first use in this function) numpy\core\src\multiarray\einsum.c.src 656 error `accum sse undeclared first use in this function) error Command gcc mno cygwin O2 Wall Wstrict prototypes DHAVE NPY CONFIG H=1 DNPY MINGW USE CUSTOM MSVCR D MSVCRT VERSION =0x0900 Inumpy\core\include Ibuild\src.win32 3.2\numpy\core\include numpy Inumpy\core\src\private Inumpy\core\src Inumpy\core Inumpy\core\src\npymath Inumpy\core\src\multiarray Inumpy\core\src\umath Inumpy\core\src\npysort Inumpy\core\include IC \Python32\include IC \Python32\PC c build\src.win32 3.2\numpy\core\src\multiarray\einsum.c o build\temp.win32 3.2\Release\build\src.win32 3.2\numpy\core\src\multiarray\einsum.o failed with exit status 1","bug"
|
|
"6013","3762","Operation on masked array changes fill value.
|
|
I first raised this issue on stackoverflow see link on the bottom) Seems like the new masked array should inherit the fill value from the two masked arrays being summed?
|
|
Can someone explain to me this behavior of a numpy masked array?
|
|
It seems to change the fill value after applying the sum operation, which is confusing if you intend to use the filled result.
|
|
data=ones 5,5)) m=zeros 5,5),dtype=bool) Mask out row 3 m[3, ]=True arr=ma.masked array data,mask=m,fill value=nan) print arr print Fill value , arr.fill value print arr.filled ) farr=arr.sum axis=1) print farr print Fill value , farr.fill value print farr.filled ) I was expecting this print nansum arr.filled ),axis=1) Prints output [[1.0 1.0 1.0 1.0 1.0] [1.0 1.0 1.0 1.0 1.0] [1.0 1.0 1.0 1.0 1.0] [ ] [1.0 1.0 1.0 1.0 1.0]] Fill value nan [[ 1.
|
|
1.
|
|
1.
|
|
1.
|
|
1.]
|
|
[ 1.
|
|
1.
|
|
1.
|
|
1.
|
|
1.]
|
|
[ 1.
|
|
1.
|
|
1.
|
|
1.
|
|
1.]
|
|
[ nan nan nan nan nan] [ 1.
|
|
1.
|
|
1.
|
|
1.
|
|
1.]]
|
|
[5.0 5.0 5.0 5.0] Fill value 1e 20 [ 5.00000000e 00 5.00000000e 00 5.00000000e 00 1.00000000e 20 5.00000000e 00] [ 5.
|
|
5.
|
|
5. nan 5.]
|
|
http stackoverflow.com questions 18879272 why does sum operation on numpy masked array change fill value to 1e20","bug"
|
|
"6013","3763","linalg.norm with ord=numpy.inf does not work for empty array.
|
|
It may sound silly but it may be of interest to be able to compute the norm of [] so that one has not to catch this case before would make my code more readable).
|
|
At the moment numpy.linalg.norm numpy.array []),ord=) works with ord = 0,1,2,... giving the result 0.
|
|
Although here is a incosistency returning np.int in the case of ord = 0 and np.float in the others.
|
|
works not not with ord = numpy.inf since this computes abs x).max ) without checking if x is nonempty.
|
|
It would be very nice, if in the [] case also for ord = inf one would return 0, which would be consistent with matlab behaviour.","bug"
|
|
"6013","3764","Weird behaviour of string array == [].
|
|
This just bit me.
|
|
seberg, maybe you fixed it already?
|
|
.
|
|
Specifically I was doing `np.all string array == array like)` and somehow my tests were failing...)","bug"
|
|
"6013","3768","Complex128 alignment leads to scipy test failures..
|
|
The tests that fail check inplace fft and fail due to copies being made on win 32 when allocated data memory is not 16 byte aligned.
|
|
The change is numpy numpy c9bf9b0.
|
|
Discussion at https github.com scipy scipy issues 2890 issuecomment 24764007.","bug"
|
|
"6013","3771","Inconsistent testing with numpy.testing.
|
|
In using `numpy.testing` for our test suite at PyMC, we occasionally run into scenarios where a test will fail when run alongside the other tests in the test suite, but pass when running the test individually, or in the console.
|
|
For example, nosetests s Code pymc pymc tests results in this failure that I cannot ever replicate anywhere else ====================================================================== FAIL test likelihoods pymc.tests.test distributions.test inverse wishart) Traceback most recent call last) File Users fonnescj Code pymc pymc tests test distributions.py , line 1223, in test likelihoods assert array almost equal calculated, right answer, decimal=1) File Library Python 2.7 site packages numpy 1.9.0.dev fde3dee py2.7 macosx 10.8 x86 64.egg numpy testing utils.py , line 811, in assert array almost equal header= Arrays are not almost equal to d decimals decimal)) File Library Python 2.7 site packages numpy 1.9.0.dev fde3dee py2.7 macosx 10.8 x86 64.egg numpy testing utils.py , line 644, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 1 decimals mismatch 100.0 ) x array 1.7976931348623157e 308) y array 108.9433934464193) Running `test distributions` on its own passes.","bug"
|
|
"6013","3774","Support atlas 3.10.
|
|
With atlas 3.10 the library names have changed, either libsatlas for serial or libtatlas for threaded.","bug"
|
|
"6013","3778","Test failures on Numpy master 2.7 tp compare didn t return 1 or 2 for exception).
|
|
I get two test failures with Numpy at 135443768 on Python2.7 with Ubuntu 13.04.
|
|
Python3.3 is Ok. Interestingly, these do not show up on Travis .
|
|
Requires running tests with `numpy.test raise warnings= develop )`.
|
|
The `tp compare` error is probably a sign of botched exception handling.","bug"
|
|
"6013","378","np.insert fails with float32, float64 input.
|
|
Namely .
|
|
returns .
|
|
because `insert` has a `isinstance` that does not check for float32, float64, but only for `int`, `long`, `integer` .
|
|
Changing it to .
|
|
seems only an half assed solution though.
|
|
EDIT forgot to mention that I m using `numpy 1.8.0.dev f2f0ac0 20120725 py2.7 macosx 10.8 x86 64.egg`.
|
|
EDIT2 Probably a better solution for the `isinstance` part would be the following .","bug"
|
|
"6013","3783","numpy won t link against lib{s,t}atlas.so.
|
|
I ve tried to build numpy with ATLAS 3.10.
|
|
This version of [ATLAS] http math atlas.sourceforge.net atlas install node22.html) produces instead of `libf77blas.so`, `liblapack.so`, and other libraries just `lib{s,t}atlas.so`.
|
|
Unfortunately numpy is not able to accept this change which results into numpy not being linked against ATLAS libs).
|
|
I would provide patch for this but [code] https github.com numpy numpy blob master numpy distutils system info.py L991) which performs these tests is a little bit hard to read for me.
|
|
Simple workaround for this is to create original files by creating symlinks ln s libtatlas.so libptcblas.so ln s libtatlas.so libptf77blas.so and so on","bug"
|
|
"6013","3787","BUG Something is wrotten in the object copying.
|
|
Try this .
|
|
Assigning a scalar into an array in PyArray AssignArray calls PyArray AssignRawScalar, which I am not sure if what it calls.
|
|
And apparently it is missing the check whether input and output are identical.
|
|
I am not sure if this needs to be fixed simply in PyArray AssignRawScalar or even lower probably not but...)","bug"
|
|
"6013","3793","scalar int hashing broken on 64 bit python3.
|
|
.
|
|
on python2 it returns True, discovered on s390x where hash np.int64 4) returns zero because int arrtype hash casts the object to PyIntScalarObject which cuts of four bytes as its really PyLongScalarObject) which are zero on big endian http bugs.debian.org cgi bin bugreport.cgi?bug=724047 This is probably also the cause of the difference on amd64.
|
|
most likely also the cause scipy scipy 2930","bug"
|
|
"6013","3796","Python 3 array from buffer of unicode failure.
|
|
Reported by cgohlke here https github.com scipy scipy issues 2890 issuecomment 25139024 Here s standalone code for the Python 3 specific ValueError ndarray is not C contiguous .
|
|
It passes with numpy 1.7.1 but fails with 1.8.0.dev 5a0d09c.
|
|
The expected result is .","bug"
|
|
"6013","3797","[1.7.1] fcompiler gnu.py ignores the LDFLAGS from python2.7 config Makefile.
|
|
I ran into a scipy build failure that was caused by the fact that numpy s fcompiler gnu.py ignores the linker flags defined through LDFLAGS in Python s config Makefile .
|
|
As a result, required linker flags were not passed to gfortran .
|
|
A few modifications to gnu.py resolved this After the c arg flags function def ld flags self) Return detected flags from LDFLAGS from distutils import sysconfig try return sysconfig.get config vars )[ LDFLAGS ].split ) except KeyError return [] In GnuFCompiler.get flags linker.so I extend opt right after defining it opt = self.linker so[1 ] opt.extend GnuFCompiler.
|
|
ld flags self) ) and then of course one doesn t need to = shouldn t) add the arch flags specifically on Darwin) in Gnu95FCompiler.get flags linker so def get flags linker so self) flags = GnuFCompiler.get flags linker so self) return flags With these modifications, scipy 0.12.0 builds fine under 32 64 x86 universal python 2.7.5 on Mac OS X 10.6.8, and the standard test suite completes OK.","bug"
|
|
"6013","380","Bug in reshape for 0 stride arrays.
|
|
The following test fails as reported on the ML http thread.gmane.org gmane.comp.python.numeric.general 51227 focus=51240 .
|
|
In the same thread Sebastian Berg has proposed the following patch .","bug"
|
|
"6013","3801","bool arrays with bytes >1 propagate.
|
|
.
|
|
Do we consider this a bug?
|
|
Or do we push the requirement onto the person constructing the array to ensure that the bytes are constrained to `\x00` and `\x01`?
|
|
A client of mine has run into this when reading an HDF5 file with pandas PyTables.
|
|
I assume that PyTables reads in the bytes from the file then overlays the bool array on top of that memory directly.
|
|
HDF5 does not appear to have a native bool type, so PyTables just uses a `uint8` type and marks it as boolean with its own metadata.
|
|
I can construct a PyTables HDF5 file with bad boolean data, and it demonstrates the above behavior.","bug"
|
|
"6013","3803","mlab import fails..
|
|
The imports in `mlab.py` of the type .
|
|
fail when doing `from numpy.oldnumeric import mlab`.","bug"
|
|
"6013","3804","object array creation new conversion to int.
|
|
3696 introduced a suble change in behavior for record arrays .
|
|
before the PR it keept the numpy scalars .
|
|
now it converts them to python integers.
|
|
.
|
|
The reason seems to be that old code just copied the points of the object, while the new code goes through PyArray CopyInto which calls dtype >get element on each entry.
|
|
The get elemnt of int arrays converts it to python integers in python2.
|
|
This causes a test failure in scipy where a int type array now compares false to the np.int32 type array test stats.py) I m not sure whats the best way to fix that or if it even should be fixed.","bug"
|
|
"6013","381","prepare bootstrap.py doesn t seem to work.
|
|
I was trying to use `tools win32build ` and I read the README there, but this is what I got so far .
|
|
Looks like `svn` is missing, so I installed it and .
|
|
So it s probably assuming that NumPy is installed using svn.
|
|
Can it be, that these scripts are outdated, or that I am using it incorrectly?","bug"
|
|
"6013","3812","numpy ufunc shouldn t disable the default Python binop mechanism.
|
|
If a class has a ` numpy ufunc ` method defined, this disables the usual Python binary operation mechanism, because Numpy s ` add ` etc.
|
|
are implemented via Ufuncs.
|
|
Example .
|
|
The fix is probably to alter the methods in Numpy s `multiarray number.c` so that either i) the TypeError is caught and converted to NotImplemented, or ii) a special internal kwarg is passed to the Numpy ufuncs that instructs them to return NotImplemented rather than raising errors.
|
|
The problem with i) is that also other TypeErrors could end up being caught, which may be confusing.
|
|
ii) doesn t have this issue, and would allow fixing the issue that ufuncs return NotImplemented in some cases they shouldn t NotImplemented should ever appear only when dealing with Python s binop mechanism).","bug"
|
|
"6013","3817","reshape ) should have a way to keep a dimension unchanged..
|
|
I often find myself writing x.reshape x.shape[0], 1) and occasionally x.reshape prod x.shape[ 2]), x.shape[2], 1) It seems like a common sort of thing that should have some syntactic sugar added.
|
|
I m thinking something like the newaxis=None definition magic values to let you do this x.reshape keepaxis, 1) and x.reshape mergeaxis, mergeaxis, keepaxis, 1) respectively.
|
|
I could see using keepaxis = slice None) and mergeaxis = None or mergeaxis = Ellipsis the advantage of which is that you could give reshape a ` getitem ` method and with some more logic do x.reshape[ ,...] and x.reshape[...,..., ] Even without indexing notation, a little logic in reshape would make it easy to keep an existing axis while flattening those around it.
|
|
Thoughts?","bug"
|
|
"6013","382","Segault in Python 3.2 in Wine.
|
|
I didn t have time to investigate yet, so I am just posting it here so that I remember .","bug"
|
|
"6013","383","NumPy docs LaTeX Error Too deeply nested.
|
|
This is what I get when I try to build the docs .
|
|
I will keep investigating.
|
|
This suggests a deeper problem http www.tex.ac.uk cgi bin texfaq2html?label=toodeep","bug"
|
|
"6013","3831","Unsafe use of PEP 3118 buffers.
|
|
The current Numpy code base contains some calls to `PyObject AsWriteBuffer PyObject AsReadBuffer`.
|
|
These are unsafe for PEP 3118 buffers, since these Python routines release the buffer immediately after obtaining the pointer.
|
|
The correct fix would probably be to use a refcounted buffer object, which would also find use in other parts of Numpy currently, memoryview is used, but it contains other stuff).
|
|
There s an implementation in CPython source code http hg.python.org cpython file f6792f734fcc Objects memoryobject.c l7 that could be used.","bug"
|
|
"6013","3833","numpy.linalg.solve and scipy.linalg.solve and singular matrices.
|
|
From the docs, `solve` should raise a `LinAlgError` exception if `a` is singular.
|
|
With the example, a = array [1, 2, 3], [4, 5, 6], [7, 8, 9]]) b = array [1, 2, 3]) `solve a, b)` returns `array [ 0.33333333, 0.66666667, 0.
|
|
])` which is a solution but not the unique solution since `a` is singular in this case.
|
|
For example, `array [ 0.66666667, 1.33333333, 1.
|
|
])` is also a solution.","bug"
|
|
"6013","3836","datetime64 breaks equality and hash invariant..
|
|
It looks as though the `datetime64` dtype breaks the Python rule that `x == y` should imply `hash x) == hash y)`.
|
|
This broke a Pandas application that was grouping on dates, and then doing a dictionary lookup to find the lines of a `DataFrame` associated to a particular date.
|
|
.","bug"
|
|
"6013","3837","MachAr.
|
|
do init ) hangs under mod wsgi.
|
|
numpy version 1.6.2) I ve been investigating a complete process hang during `from sklearn.cluster import DBSCAN`, and I ve traced it to a call to `numpy.finfo float)` in the `scipy.linalg.decomp schur` module.
|
|
The actual hang occurs in `core machar.py` [line 227] https github.com numpy numpy blob master numpy core machar.py L227) temp = z t This line is within a `for` loop and occurs on the ` = 9` iteration.
|
|
At that point I m getting `repr z) = array [ 5.56268465e 309])`, and `repr t) = array [ 1.
|
|
])` I don t know how to get a more accurate representation of their actual state).
|
|
The next line is never reached the process hangs indefinitely at the `temp = z t` line.
|
|
The weirdest part is that the `from sklearn.cluster import DBSCAN` import works fine when I try it in a clean Python shell, but hangs when running in our web application.
|
|
We re using numpy very sparsely, and we re certainly not messing with its internals, so I have no idea how to narrow down the conflict that apparently occurs between numpy and the rest of our codebase.","bug"
|
|
"6013","3840","Error when equating a masked and non masked structured array.
|
|
I see the following error with the latest version of Numpy 1.8.0rc1) .
|
|
I think the correct behavior internally should be equivalent to .","bug"
|
|
"6013","3842","unclear Advanced indexing section in 1.8 release notes.
|
|
this section in the 1.8 release notes is not very clear Advanced indexing using `np.newaxis` It is now possible to use `np.newaxis` `None` together with index arrays instead of only in simple indices.
|
|
This means that ``array[np.newaxis, [0, 1]]`` will now work as expected.
|
|
Just from reading this I have no idea what is expected here.
|
|
An example or description will save people reading the release notes some headaches.
|
|
E.g.
|
|
... as expected and prepend an empty dimension to arrays shape and select the first two rows.
|
|
seberg, can you have a look?","bug"
|
|
"6013","3843","lack of highlights in 1.8 release notes.
|
|
I ve been looking through the release notes a bit and noticed a severe lack of highlights 1.8 has plenty interesting things, so what to move or mention in highlights?
|
|
I would suggest full and full like this has been a much user requested functions, it should get top billing.
|
|
partition well I wrote it ;) seriously, its also a function requested several times on the mailing list and it allows significant performance improvements in some third party software.
|
|
general summary of performance improvements we have many performance improvements, a quick summary of the improved areas belongs in the highlights.
|
|
we could also add a count of bugs fixed.
|
|
I can t judge the significance of rest of the changes, if you think your change should be a highlight, please speak up.","enhancement"
|
|
"6013","3844","test nonzero byteswap uses denormals, fails depending on icc floating point compiler flags..
|
|
The test output can be seen here https gist.github.com scicalculator 6772961 Numpy is compiled with python3 using icc and ifort.
|
|
It is linked to the intel MKL library.
|
|
The first error test nonzero byteswap) was actually seen with 1.7.1, but I attempted to install 1.8 to see if that corrected the problem it did not).
|
|
I wanted to bring this to your attention, but also wanted to ask if this is something that I should worry about.
|
|
compiler version .
|
|
I also tested numpy 1.7.1 and python2.7.3 resulting in the byteswap error with a newer build of icc have not tested 1.8).
|
|
.
|
|
python version .","bug"
|
|
"6013","3845","Bernstein polynomials.
|
|
How difficult would it be to implement Bernstein polynomials in NumPy in an analogous way as the current power series, Chebyshev...?
|
|
The problem I see is that two coefficients are needed, I don t know if that is a major hassle when using the polynomial template.","enhancement"
|
|
"6013","3846","np.median behavior broken for array subclasses in 1.8.0rc1.
|
|
Previously, when writing Numpy array subclasses, it was possible to override the behavior of ``np.mean`` and ``np.median`` by defining a ``mean`` method on the sub class.
|
|
For example, .
|
|
produced the following output .
|
|
However, in 1.8.0rc1, it looks like none of the methods shown above get called .
|
|
This means that it is no longer possible to override the behavior of ``median``, which makes things like quantity handing with units break.
|
|
For instance .
|
|
or .
|
|
It would be great if this could be fixed by 1.8.0 final, as this will otherwise break a lot of unit handing code.","bug"
|
|
"6013","3847","solve ) fails on 0 sized axis.
|
|
With current 1.8, 5a0d09c .
|
|
gives .
|
|
The result should be an array with shape ` 3, 0)`.
|
|
This is a regression from 1.7, hence marking as a blocker.
|
|
Thanks to Jens Jxc3xb8rgen Mortensen for the original report http mail.scipy.org pipermail numpy discussion 2013 October 067898.html","bug"
|
|
"6013","3849","np.nansum pandas.Series [1.0, np.nan])) breakage in current 1.8.
|
|
[copied from mailing list so we don t lose track of it] With pandas 0.12.0, this code .
|
|
works correctly using current master 74abfa27), but is broken with current maintenance 1.8.x f5fe91b9).
|
|
Here correctly means returns a scalar and broken means returns a `Series` .
|
|
I bet the difference is In master, nansum ultimately calls arr.sum ...), which will be intercepted by Series.sum.
|
|
In 1.8.x, nansum ultimately calls np.add.reduce ...), which can t be intercepted and will return the wrong thing.
|
|
AFAICT the np.add.reduce a, ...) call could just be replaced with a.sum ...), but I might be missing something... surely there must have been some reason it was written that way in the first place?","bug"
|
|
"6013","385","BUG format numpy.void ABC ), ) throws a RuntimeError.
|
|
Using version 1.6.2.
|
|
Stumbled upon this issue while using format on indexed record arrays.
|
|
>>> import numpy >>> numpy.
|
|
version 1.6.2 >>> format numpy.void ABC ), ) Traceback most recent call last) File <console> , line 1, in <module> RuntimeError maximum recursion depth exceeded while calling a Python object","bug"
|
|
"6013","3850","Decide on nanargmax nanargmin behaviour for 1.8.
|
|
Just a placeholder to remind us that we should make a decision either way on what nanargmax nanargmin will do in 1.8.
|
|
See this thread http mail.scipy.org pipermail numpy discussion 2013 October 067912.html","bug"
|
|
"6013","3857","Indexing behaviour change in numpy1.8.0.
|
|
This runs in numpy 1.7.1 np.array [1, 2, 3])[[1], ] but raises in numpy 1.8.
|
|
Was the change intentional?","bug"
|
|
"6013","3858","ERROR test big arrays test io.TestSavezLoad) on OS X Python 3.3.
|
|
Reported by Piet van Oostrum on the mailing list against 1.8.0rc1 on OS X with Python 3.3 ====================================================================== ERROR test big arrays test io.TestSavezLoad) Traceback most recent call last) File Library Frameworks Python.framework Versions 3.3 lib python3.3 site packages numpy testing decorators.py , line 146, in skipper func return f args, kwargs) File Library Frameworks Python.framework Versions 3.3 lib python3.3 site packages numpy lib tests test io.py , line 149, in test big arrays np.savez tmp, a=a) File Library Frameworks Python.framework Versions 3.3 lib python3.3 site packages numpy lib npyio.py , line 530, in savez savez file, args, kwds, False) File Library Frameworks Python.framework Versions 3.3 lib python3.3 site packages numpy lib npyio.py , line 589, in savez format.write array fid, np.asanyarray val)) File Library Frameworks Python.framework Versions 3.3 lib python3.3 site packages numpy lib format.py , line 417, in write array fp.write array.tostring C )) OSError [Errno 22] Invalid argument","bug"
|
|
"6013","3860","BUG ddof ignored in nanvar and nanstd with int input arrays 1.8.0rc1).
|
|
I assume the problem is that the line below does not contain ddof if mask is None return np.var arr, axis, dtype=dtype, out=out, keepdims=keepdims) Demo In [1] np.nanvar [1,2,3,4], ddof=1) Out[1] 1.25 In [2] np.nanvar [1,2,3,4], ddof=2) Out[2] 1.25 In [3] np.nanvar [1,2,3,4], ddof=3) Out[3] 1.25 In [4] np.nanvar [1,2,3,4.0], ddof=1) Out[4] 1.6666666666666667 Same problem with np.nanstd In [1] np.nanstd [1,2,3,4], ddof=1) Out[1] 1.1180339887498949 In [2] np.nanstd [1,2,3,4], ddof=2) Out[2] 1.1180339887498949 In [3] np.nanstd [1,2,3,4.0], ddof=2) Out[3] 1.5811388300841898","bug"
|
|
"6013","3863","Broken Link in Documentation.
|
|
I apologize if this is the wrong place to report this, but I don t know where else to do this.
|
|
Clicking on the link labeled var on this page http docs.scipy.org doc numpy reference routines.statistics.html takes you to the page http docs.scipy.org doc numpy reference generated numpy.var.html numpy.var However, that page says Internal Server Error .
|
|
Clearly the link is broken and needs to be fixed.","bug"
|
|
"6013","387","``sum`` method returns array with invalid C F contiguous flags.
|
|
.
|
|
This still works in 1.5.1.
|
|
Broken on the latest Git HEAD.
|
|
It affects all reductions of ndarray objects.","bug"
|
|
"6013","3871","npy math.h has invalid isinf for Solaris with SUNWspro12.2.
|
|
See https github.com scipy scipy issues 2894 For this Solaris C C compiler, `npy isinf` expands to a missing symbol.
|
|
According to the report, can be avoided by adding .","bug"
|
|
"6013","3876","np.linalg.solve failure.
|
|
.","bug"
|
|
"6013","3877","Bento build broken under Python 3.2.
|
|
2to3 d bscript to take care of the rename of .
|
|
to .. .
|
|
numpy 7457f156 Bento 2be31f57","bug"
|
|
"6013","3878","information loss with bytes type in array elements when accessing through array interface.
|
|
When storing bytes in np array, they cannot be consistently recovered, e.g., when using in the bytes argument to UUID.
|
|
I assume all of this is a victim of the 2to3 transition and some things still need to be sorted out.
|
|
The following is on Fedora 19, numpy 1.7.1 last stable), python 3.3.2 In [25] x = np.array b \0 16) In [26] x Out[26] array b , dtype= |S16 ) In [27] bytes x) Out[27] b \x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00 In [38] x = np.array [b \0 16]) In [39] x[0] Out[39] b Obviously, it would return the same, no matter how many \x00 bytes I stored in it In [40] x = np.array [b \0 16] 5) In [41] x Out[41] array [b , b , b , b , b ], dtype= |S16 ) In [42] x.tolist ) Out[42] [b , b , b , b , b ] Some various examples you can try yourself, numpy will drop all trailing zero bytes.
|
|
In [43] x = np.array [b \0 5 b 1 b \0 10]) In [44] x[0] Out[44] b \x00\x00\x00\x00\x001 Even more confusing hence is In [49] x = np.array [b \0 5 b 1 b \0 10]) In [50] bytes x[0]) Out[50] b \x00\x00\x00\x00\x001 In [51] bytes x) Out[51] b \x00\x00\x00\x00\x001\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00 It s obvious why it does that, but is should not.
|
|
Possibly I am overlooking a functionality but how can I recover all the bytes including trailing \x00 in each array location using the array interface?
|
|
I think the behavior numpy now has for bytes would be OK for strings but for bytes it should be returning the full data.
|
|
Similar to what you could do with void , but it should transparently return just all the bytes.
|
|
In [52] x = np.array [b \0 5 b 1 b \0 10], dtype=np.void) In [53] x Out[53] array [[ 0 0 0 0 0 49 0 0 0 0 0 0 0 0 0 0]], dtype= |V16 ) In [54] bytes x) Out[54] b \x00\x00\x00\x00\x001\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00 In [55] bytes x[0]) Out[55] b \x00\x00\x00\x00\x001\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00 Alexander","bug"
|
|
"6013","3879","tp compare warning on sort in Python 2.6 2.7 with OSX with object array of length >= 20 with objects that raise TypeError.
|
|
We cpcloud and I) were investigating this strange problem with pandas where we were getting `RuntimeWarning tp compare didn t return 1 or 2 for exception` only on OSX.
|
|
We narrowed down the issue all the way to this specific set of circumstances it may occur other ways, but this at least triggers it).
|
|
Given that it still fails with an Exception, not sure if it actually matters Definitely occurs on recent Mac OSX not sure if version of OSX matters).
|
|
Does not occur on Arch Linux 64 bit.
|
|
Python 2.6 or 2.7 doesn t occur with Python 3) Must be object array with length >= 20 and at least through 1,000,000) Must raise some kind of error with comparison doesn t matter if builtin class like `datetime.datetime` or a Cythonized cdef d class) or a pure Python class Here s an example .
|
|
generates an Exception and a warning about `tp compare` .
|
|
Same thing occurs if you create an array of datetimes integers.
|
|
We re not sure what actually causes the warning, but we re thinking the reason length matters has to do with the choice of sort algorithm .","bug"
|
|
"6013","3887","assert array) equal does not check the scalar shape?.
|
|
I guess this is very likely to be intentional or there are so many tests that it cannot should not be changed, so please feel free to just close the issue.
|
|
But I have to wonder why .
|
|
passes.
|
|
It gets a bit funnier with .","bug"
|
|
"6013","3889","Bento does not install numpyconfig.h.
|
|
This breaks the scipy build .
|
|
Python 2.7.3 Bento 2be31f5 numpy 7457f15","bug"
|
|
"6013","389","GenFromText Does not Seem to Work.
|
|
I have a file exactly); Data from Naive Bayes Classifier example by Eric Meisner November 22, 2003 http www.inf.u szeged.hu ormandi teaching mi2 02 naiveBayes example.pdf Color Type Origin Stolen Red Sports Domestic Yes Red Sports Domestic No Red Sports Domestic Yes Yellow Sports Domestic No Yellow Sports Imported Yes Yellow SUV Imported No Yellow SUV Imported Yes Yellow SUV Domestic No Red SUV Imported No Red Sports Imported Yes if i give it a = genfromtxt filename,dtype=None,names=True,comments= ,skip header=2) print a print a.dtype it gives .
|
|
if I only do a = genfromtxt filename,dtype=None,names=True,comments= ) print a print a.dtype I get .
|
|
Per my understanding it should be automatically skipped the beginning as those are marked as comments.
|
|
Thanks.","bug"
|
|
"6013","3892","in1d fails if second argument is a set.
|
|
numpy.in1d does not behave like ``in`` if the second argument is a set.
|
|
Is this a bug?
|
|
If not, perhaps a prominent warning would be helpful.
|
|
>>> s = { a , b } >>> s set [ a , b ]) >>> x = np.array [ a , b , c ]) >>> np.in1d x, s) array [False, False, False], dtype=bool) >>> l = list s) >>> np.in1d x, l) array [ True, True, False], dtype=bool) Python 2.7.4, NumPy 1.9.0.dev b91c711","bug"
|
|
"6013","3897","np.info doesn t work on Python 3.x.
|
|
>>> np.info np.sin) Traceback most recent call last) File <stdin> , line 1, in <module> File home rgommers .local lib python3.3 site packages numpy lib utils.py , line 556, in info elif isinstance object, types.InstanceType) check for call method AttributeError module object has no attribute InstanceType Originally reported at https github.com scipy scipy issues 2981","bug"
|
|
"6013","3901","Linalg routines give wrong results on OSX Accelerate for 1x1 problems only?).
|
|
Investigating the failures [reported] http permalink.gmane.org gmane.comp.python.scientific.devel 18313) on OSX led to the following It s a Numpy issue.
|
|
Bisection search for Numpy commits vs Scipy 0.13.0rc1 tests show that 9c00887ba6 is the first failing commit.
|
|
The immediately preceding 9bfa19b11f doesn t fail.
|
|
That commit replaces numpy.linalg lapack lite with the one that was added in the gufunc linalg pull request.
|
|
Something apparently goes very wrong on OSX.
|
|
Compiled with Accelerate.
|
|
I have a minimal failing test case .
|
|
Interestingly, 2x2 and bigger matrices seem to work OK.
|
|
Apparently, Numpy s test suite doesn t test 1x1 eigenproblems hah).
|
|
That s possibly not the only failure, maybe something else is also broken.
|
|
It s also probably not a bug in Accelerate, as scipy.linalg.eig does work correctly.
|
|
The reason seems to be somehow linking with Apple s Accelerate.
|
|
If I remove the Accelerate detection from `numpy distutils system info.py` and build with `ATLAS=None BLAS=None LAPACK=None` so that the f2c d files are actually used, the bug goes away.
|
|
What is less clear is why 9c00887ba6 makes things fail.
|
|
EDIT numpy.linalg functions were replaced via the gufunc linalg functions before this commit, unlike what I claimed earlier) If you want SSH access to a machine on which to debug this, ask rgommers)","bug"
|
|
"6013","3909","runtests.py missing some tests..
|
|
There should be more than 5200 tests run, but .
|
|
gives `` Ran 4637 tests in 274.319s ```","bug"
|
|
"6013","391","np.count nonzero ) should accept axis parameter.
|
|
Currently, [np.count nonzero )] http docs.scipy.org doc numpy reference generated numpy.count nonzero.html) accepts only the input array.
|
|
I propose adding `axis` parameter similar to that of functions np.sum ) and np.max ) etc.
|
|
.
|
|
I think the alternative ` A != 0).sum axis=rows or cols)` isn t as clear.","enhancement"
|
|
"6013","3912","Provide an easy way to check BLAS LAPACK linkage from python.
|
|
According to the answers to this [stackoverflow] http stackoverflow.com questions 9000164 how to check blas lapack linkage in numpy scipy) question the only way to check if Numpy is using the system BLAS LAPACK is to inspect the output of `ldd` on some library belonging to numpy hidden deep in the filesystem.
|
|
In fact I am not 100 sure that I interprete my results correctly.
|
|
It would be nice with something like >>>print numpy.
|
|
lapack version ) 3.4.2 >>>print numpy.
|
|
blas version ) OpenBLAS 0.2.8 or if we are using the fallback >>>print numpy.
|
|
lapack version ) fallback","bug"
|
|
"6013","3915","dot behavior on zero strides depends on dtype.
|
|
While trying to test how `dot` would behave on a very large array without actually constructing one, I found this discrepancy .
|
|
This is the expected result, but on floats it fails .
|
|
This is unfortunate, since it makes it hard to test whether `dot` behaves correctly on arrays of size > 2 31 1.","bug"
|
|
"6013","392","Regression in numpy.insert ) when axis != None.
|
|
I believe that the fix to Ticket 808 has caused a regression in numpy.insert ) This is the [commit] https github.com numpy numpy commit 2c04244da264cb1665d6162ae119d2f05ad65150) Previously np.insert [[1,1,1]], 0, [2,2,2], axis=0) gave array [[2, 2, 2], [1, 1, 1]]) now it gives array [[2, 2, 2], [2, 2, 2], [2, 2, 2], [1, 1, 1]]) I don t think is the correct result.
|
|
Also issue [ 378] https github.com numpy numpy issues 378) is a result of same change","bug"
|
|
"6013","3920","Bug in docstring for numpy.random random sample description is inaccurate..
|
|
The top of the docstring for the `numpy.random` module has this description, under the `Utility functions` heading .
|
|
The description for `random sample` seems inaccurate it suggests that it s possible to specify the range, where in reality the range is fixed to be `[0.0, 1.0)`.
|
|
In fact, since `random` and `random sample` are the same function, the descriptions should probably be identical too.
|
|
Thanks to zakora on the python IRC channel for pointing this out.)","bug"
|
|
"6013","3922","v0.3.0 10303 g61c8568 fails to build on s390 box.
|
|
.","bug"
|
|
"6013","3926","Pickling numpy.string ) fails with ValueError itemsize cannot be zero.
|
|
.","bug"
|
|
"6013","3927","masked array.mean axis= 1) returns np.float64 .
|
|
type masked array [0, 1] ).mean axis= 1)) type masked array [[0, 1]]).mean axis= 1)) The return type depends on the number of dimensions of the masked array which can hide bugs with ``np.float64`` not working like 0 dimensional ``masked array``, like for example accessing the attribute ``data`` Just realised that this is documented behaviour for ``.sum`` but not for ``.mean``, this is thus probably only a documentation issue.
|
|
Personally I would feel better if methods always returns the same type I m running `` numpy 1.8.0.dev e589c6e ``","bug"
|
|
"6013","3928","numpy.random.gamma parameterization not consistent with scipy.stats.distributions.gamma.
|
|
`numpy.random.gamma` uses a slightly different parameterization of the gamma distribution for its random number generator than scipy uses in `scipy.stats.distributions.gamma`.
|
|
Specifically, the scale parameter is the inverse of that in scipy.
|
|
While neither is better than the other, you may want them to be consistent with one another.","enhancement"
|
|
"6013","3930","Variable declared after statement.. .","bug"
|
|
"6013","3934","pickling recarray records destroys information.
|
|
.
|
|
Output .","bug"
|
|
"6013","3939","Unicode byteorder seems mostly broken.
|
|
Is it supposed to be possible to use non native unicode byteorder?
|
|
The unicode comparison functions cannot handle non native byteorder, however this also applies to the dtype transfer functions.
|
|
The copyswap functions do anticipate it, but only for 4 byte wide unicode and I think there can be 2 byte wide as a compile option?
|
|
), maybe that is why printing works... .","bug"
|
|
"6013","394","RuntimeWarning invalid value encountered in absolute .
|
|
Sometimes I am getting these errors when running tests .
|
|
These are caused by calling `allclose x, y)` with for example `x = array [1, nan, 2])` the array contains `nan`).
|
|
Then inside `allclose`, one calls `abs` on the array of `nan`s, which produces the warning.
|
|
It can be reproduced by simply doing .","bug"
|
|
"6013","3940","Runtests fails for testing modules.. .
|
|
This works if 6dfe864 is reverted.","bug"
|
|
"6013","3957","regenerate cython source for release?.
|
|
mtrand is currently generated with cython 0.19.
|
|
There have been two bugfix releases since https github.com cython cython blob master CHANGES.rst 0192 2013 10 13 It might be a good idea to regenerate our files.
|
|
Though we are already post the last rc and from a quick glance nothing looks really relevant, maybe its safer to keep what we have now?","bug"
|
|
"6013","396","TODO list for the 1.7.0 release.
|
|
This issue is to track what needs to be done before the 1.7.0 release.
|
|
I will just be updating the text here.
|
|
Issues to work on ============= Issues to fix http projects.scipy.org numpy ticket 2108 378, 392 see Nathaniel s comment below) 394 424 426 438 294 291 464 Also we need to fix all Debian build issues 406, 407, 408, 409, 410, 411, 412, 413, 414, 415 Access to SPARC 64 needed for http projects.scipy.org numpy ticket 2076 Fixed Issues that need to be merged =========================== Work is done here, it just needs to get reviewed merged, or more discussion needed.
|
|
Issues that need clarification http projects.scipy.org numpy ticket 2150 http projects.scipy.org numpy ticket 2101 Issues PRs that need merging 2696 this is a PR against maintenance 1.7.x) Fixed ==== These are fixed, but not yet back ported to the 1.7.x branch.
|
|
459 2707 Backported ========= All these PRs are fixed in master and back ported to the 1.7.x branch left here for reference).
|
|
http projects.scipy.org numpy ticket 2185 PR 395) http projects.scipy.org numpy ticket 2066 PR 397) http projects.scipy.org numpy ticket 2189 PR 397) http projects.scipy.org numpy ticket 2187 PR 401) http projects.scipy.org numpy ticket 1588 PR 405) 416 PR 417) 376 404 390 eebd7b2 432 429 431 430 399 420 451 440 this adds a deprecation, so I guess it also needs a short mention added to the release notes) 449 The above 4 issues are backported by 472.","bug"
|
|
"6013","3960","Add LDL decomposition.
|
|
Add support for the [LDL decomposition] http en.wikipedia.org wiki Cholesky LDL decomposition 2), which is a variant of the Cholesky decomposition that doesn t take any square roots faster).
|
|
LAPACK has a function for this called DPTTRF, so I m guessing that supporting this is just a matter of adding a wrapper for this function.
|
|
Also, if this is something that would fit better in Scipy, let me know.","enhancement"
|
|
"6013","3961","numpy.set string function is unsafe when using multiple embedded sub interpreters.
|
|
Our application uses multiple embedded python interpreters.
|
|
When starting or stopping new interpreters, everything with numpy seems to work fine except for the printing of arrays.
|
|
After starting a second interpreter that uses numpy and then printing an array in the first interpreter, we get an error .
|
|
I ve traced it down somewhat.
|
|
Another embedded interpreter does not share sys.modules but any static variables used in CPython extensions are shared across interpreters.
|
|
It seems to be that when the second interpreter imports numpy it imports numeric.py which has these two lines set string function array str, 0) set string function array repr, 1) That changes the static variables PyArray StrFunction and PyArrayReprFunction in arrayobject.c that are shared across interpreters, and the first interpreter somehow still references the original PyArray StrFunction and therefore fails on a print.
|
|
Note that if the first interpreter directly uses numpy.core.arrayprint.array2string or numpy.core.numeric.array str those still work fine, it only seems to be the str array) that is broken.
|
|
Tested with python 2.7.1 and numpy 1.7.1 and numpy 1.5.0 on CentOS 5.","bug"
|
|
"6013","3962","Python segfaults when printing array in IPython.
|
|
The code .
|
|
causes a segfault when run in IPython.
|
|
This appears to be caused by numpy core src umath ufunc object.c line 272 .
|
|
where \ extract\ pyvals begins .
|
|
specifically, the \ errobj line causes the segfault","bug"
|
|
"6013","3966","np.median doesn t preserve Fortran order.
|
|
Output array of median along an axis in multidimentional Numpy arrays defined with order= FORTRAN doesn t have FORTRAN order.
|
|
np.sum ), np.mean ) etc preserves this order, but np.median doesn t. To reproduce the issue try the following.
|
|
a=np.ones 10,11,12),order= FORTRAN ) np.isfortran a) Will return True np.isfortran np.mean a,axis=0)) Will return True np.isfortran np.sum a,axis=0)) Will return True But np.isfortran np.median a,axis=0)) Will return False !!!
|
|
Why is only np.median ) not preserving the fortran order?
|
|
Is this also the reason, unlike mean and sum, median along the first axis is slower for Fortran ordered arrays than C ordered arrays?","bug"
|
|
"6013","3967","Implement tp traverse.
|
|
Trying to squish the memory leaks reported by valgrind during in the test suit, I noticed that we do not have tp traverse implemented.
|
|
There are two things here First, tp traverse for the `base` attribute, and second for object arrays.
|
|
I guess we should implement both, or is there some reason against object array traversal?","bug"
|
|
"6013","3968","Python segfault when using umath.so on OS X 10.9.
|
|
I have a freshly built numpy on OS X 10.9, and am running into pretty regular crashes related to umath.so.
|
|
[Here is a Gist with the crash report] https gist.github.com fonnesbeck 31d372805647da7af6e3).
|
|
It seems to crash in a variety of ways, but I can reliably get it to crash by importing a module like Theano, which has numpy as a dependency.
|
|
I have re built all of my packages since upgrading to OSX 10.9.","bug"
|
|
"6013","3975","savetxt ignores newline on Python 2.7.
|
|
Under some circumstances savetxt ignores the newline character on Python 2.7 running on Windows If newline is set to \n , actually \n\r is written, because of the linesep conversion of file.write.
|
|
Changing the file mode to wb would solve that issue.","bug"
|
|
"6013","3977","1.8.0rc2 changes np.linalg.eigh ) behavior.
|
|
Hi, I found an unexpected difference from numpy 1.7.1 and 1.8.0rc2 with Python 3.3.2 on Ubuntu 12.04 amd64).
|
|
Here is the test program .
|
|
with numpy 1.7.1 1.7.1 [ 0.33333333 0.33333333 0.33333333 0.99999999] with numpy 1.8.0rc2 1.8.0rc2 [ 0.
|
|
0.
|
|
0.
|
|
0.]
|
|
Looking into the docs for np.linalg.eigh ), I see that it is for Hermitian or symmetric matrices.
|
|
One could argue, reasonably, that my test code above is broken.
|
|
However, apparently some code I am using 1 and which seems to be rather popular 2 , 3 ) depends on the old behavior of eigh ) it does not construct a full symmetric matrix but only the lower left part.
|
|
So, the old, broken, behavior is depended upon in the wild.
|
|
If numpy 1.8.0 wants to clean up this bad behavior, I think numpy should raise an exception rather than return wrong results.
|
|
Andrew 1 the quaternion from matrix ) function of Christoph Gohlke s transformations.py module http www.lfd.uci.edu gohlke code transformations.py.html 2 in use by the geometry stack of the Robot Operating System ROS) https github.com ros geometry 3 in use by PyMOL molecular visualization system https github.com Pymol Scripts Pymol script repo","bug"
|
|
"6013","398","Ticket 1588.
|
|
This issue is for the ticket http projects.scipy.org numpy ticket 1588 Because the trac server is unreliable frequently fails and I need to wait couple minutes for the database to be unlocked again see my email to the numpy list) and I need to keep track of my progress.
|
|
Summary of the problem ================== The following patch fixes it, but probably creates a leak sometimes .
|
|
This code was first introduced by 9405a8703f6df655c17cdefc02f9977f92d3b2d8 and the moved around later.
|
|
There is no problem with this commit.
|
|
The actual segfault was introduced between the commits 9a9f08e089ff49fccca1feac9620c2837f8c09bd good) and 64e30a7261e5a575a12beed1c3971f80779760f1 bad).
|
|
Overall diff of these is https github.com numpy numpy compare 9a9f08e089ff49fccca1feac9620c2837f8c09bd...64e30a7261e5a575a12beed1c3971f80779760f1 The problem might be in the changes in the `ctors.c` file.
|
|
At the moment, the correct fix without leaks) is unclear.
|
|
See the comments below for detailed analysis.","bug"
|
|
"6013","3983","asarray inconsistent behavior when working with complex128 datatype.
|
|
It appears that asarray and view behave differently.
|
|
I am passing a buffer from C of ` <dd ` data type to represent a complex128 record struct module format string).
|
|
Accessing the buffer with `np.asarray buf, complex128 )` coerces records to real, dropping the imaginary part.
|
|
Using `np.asarray buf).view complex128 )` has the expected behavior.
|
|
Steps to reproduce in python .","bug"
|
|
"6013","3985","numpy.distutils ignores paths in any site.cfg sections.
|
|
I just ran into this issue, which seemed vaguely familiar.
|
|
Seems I ran into a few years ago too.
|
|
http numpy discussion.10968.n7.nabble.com why doesn t numpy distutils find ATLAS td763.html I think my reading at the time was right.
|
|
Consider a typical for me) site.cfg .
|
|
These directories are then never looked at by the usual numpy.distutils get info functions AFAICT.
|
|
Actually, even if I specify site.cfg as .
|
|
These directories are not found.
|
|
.
|
|
The above looks like a bug to me.
|
|
Because the build info is correctly recorded, though I suppose unrelated to the above, which I was assuming is runtime info.
|
|
.
|
|
This suggests to me that it s more robust in package building to use this get info written at build time than the numpy.distutils which is common among packages that use this for support.
|
|
Should I try to patch this?","bug"
|
|
"6013","3988","f2py fails on callback.
|
|
I m trying to call f2py on the Fortran code from http geo.phys.spbu.ru tsyganenko Geopack 2008.html but get the following errors .
|
|
Please rephrase the issue title if it s too meaningless.","bug"
|
|
"6013","399","Backport Python 3.3 patches from master to the 1.7.0 branch.
|
|
Currently the master works in Python 3.3, but not the 1.7.0 branch.
|
|
The patches need to be backported.
|
|
As well as all the patches that fix release critical issues, see 396 for a full list.","bug"
|
|
"6013","3993","recarray attributes of type strings will create a ndarray of strings instead of a chararray.
|
|
Hi, I recently read on the Numpy doc about ``numpy.chararray`` > The chararray class exists for backwards compatibility with Numarray, it is not recommended for new development.
|
|
Starting from numpy 1.4, if one needs arrays of strings, it is recommended to use arrays of dtype object , string or unicode , and use the free functions in the numpy.char module for fast vectorized string operations.
|
|
Moreover, I found that ``numpy.recarray`` turns all Numpy strings dtypes into a ``chararray`` instead of a ``ndarray`` of strings.
|
|
Is there a reason?
|
|
Example .
|
|
whereas .
|
|
returns a ``ndarray`` of type ``|S7``.
|
|
Can you think it s a relevant idea to have the same numpy type for ``a[ name ]`` and ``reca.name``?
|
|
Thus, it would be possible to get rid of ``chararray`` in this case.
|
|
Thanks, Damien G.","bug"
|
|
"6013","3994","abs ) is slow for complex, add abs2 ).
|
|
See the following code run on Ubuntu 64bits) .
|
|
`abs )` is slow compared to the manual formula.
|
|
Is there any drawback in using `np.sqrt b.real 2 b.imag 2)` like possible overflow ?)
|
|
Also, it could be useful to add an `abs2 )` function which would return the value of `abs ) 2` but should be really faster for complex .
|
|
It is commonly used to compute the energy of a signal or to sort complex numbers by norm value, for example.","enhancement"
|
|
"6013","3995","setup fails to read .numpy site.cfg on win7 x64.
|
|
https github.com numpy numpy blob master numpy distutils system info.py L274 uses the HOME environment variable to find .numpy site.cfg, from which it takes the info about the blas libraries.
|
|
the HOME variable isn t defined by default on windows.
|
|
Perhaps an alternative would be to use os.path.expanduser ), which is on whidows expands to the USERPROFILE env var","bug"
|
|
"6013","3998","numpy.float ) returns non numpy python builtin float.
|
|
`numpy.float )` returns a non numpy python builtin float type same goes for `numpy.int )`) .
|
|
This is different to behavior of `dtype` options in array initialization .
|
|
I find this unexpected and surprising.
|
|
This may have little practical impact as it only affects initialization of single numbers, but it means that e.g.
|
|
numpy floating point operation error handling is circumvented in this case which is how I stumbled upon this) .","bug"
|
|
"6013","4006","When Numpy 1.8 Reference Guide and User Guide release.
|
|
Haven t seen Numpy 1.8 Reference Guide and User Guide.
|
|
So can I ask when to release.
|
|
Thanks!","bug"
|
|
"6013","4007","numpy.dot crash with numpy.float32 input.
|
|
A user of gensim fbkarsdorp reported crash segfault) with NumPy piskvorky gensim 131 The crash seems to have nothing to do with gensim, so I m transferring the issue here.
|
|
It happens in `dot` of matrix vector in single precision, on his OS X Maverick.","bug"
|
|
"6013","4008","Missing documentation on numpy.sum L,0).
|
|
Hi.
|
|
When adding two matrices, say A and B, with the operator, we get a matrix C with the same size as A and B .
|
|
When we don t have a couple of variables like A and B, but a list of matrices like .
|
|
And we want to add them up, using C=numpy.sum L), we get a scalar as a result, instead of a matrix C as before.
|
|
numpy.sum L) is equivalent to numpy.sum L,axis=None) There is indeed a way to get back the operator, using numpy.sum L,axis=0), but such feature is not documented in the reference http docs.scipy.org doc numpy reference generated numpy.sum.html).
|
|
Is this, indeed, an undocumented feature or something that s working by chance?
|
|
In the first case, it would be useful to add an example to the reference and mention the special case in the description of the parameter axis .
|
|
Thanks, Facundo","bug"
|
|
"6013","4009","numpy segfaults when providing structured array shapes as list.
|
|
For reproduction .
|
|
However, this works nicely for tuples .
|
|
Unfortunately I do not get much information from gdb, somehow gdb does not find the debugging symbols, event though I tried providing them .","bug"
|
|
"6013","4011","numpy fails test on Mac Mavericks and default python; causes segmentaion fault 11 for pandas.read csv.
|
|
I was getting a segmentaion fault 11 when I do read csv from pandas and the Issue was caused because of a numpy bug.
|
|
Please refer to this pandas issue link https github.com pydata pandas issues 5396 I also get the same error when I try to read a xls file in pandas.
|
|
>>> import numpy as np >>> np.test full ) Running unit tests for numpy NumPy version 1.9.0.dev 54d3559 NumPy is installed in Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy Python version 2.7.5 default, Nov 2 2013, 14 54 49) [GCC 4.2.1 Compatible Apple LLVM 5.0 clang 500.2.79)] nose version 1.3.0 .............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................S............................F.......................................................................................................................................................................................................................................................................................................................................................................................................K.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K................................................................................................................K...SK.S.......S.......................................................................................................................................................................................................................................................................................................................................................................................................................................................EEEEEEEEEEEEEE.EEEEE........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................S.........................................................................................................S............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K.................................................... ====================================================================== ERROR test assumed shape.TestAssumedShapeSumExample.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test callback.TestF77Callback.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test callback.TestF77Callback.test docstring Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test kind.TestKind.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test mixed.TestMixed.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test mixed.TestMixed.test docstring Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return character.TestF77ReturnCharacter.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return character.TestF90ReturnCharacter.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return complex.TestF77ReturnComplex.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return complex.TestF90ReturnComplex.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return integer.TestF77ReturnInteger.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return integer.TestF90ReturnInteger.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return logical.TestF77ReturnLogical.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return logical.TestF90ReturnLogical.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return real.TestF77ReturnReal.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test return real.TestF90ReturnReal.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 348, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test size.TestSizeSumExample.test all Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test size.TestSizeSumExample.test flatten Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== ERROR test size.TestSizeSumExample.test transpose Traceback most recent call last) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose case.py , line 381, in setUp try run self.inst, setup , setUp )) File opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages nose util.py , line 469, in try run return func ) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 353, in setUp module name=self.module name) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy f2py tests util.py , line 80, in wrapper raise ret ImportError dlopen var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so, 2) no suitable image found.
|
|
Did find var folders gy t bltp8x695bss5tzpvjw1m40000gn T tmpG7oeAJ test ext module 5403.so mach o, but wrong architecture ====================================================================== FAIL test empty test multiarray.TestStats) Traceback most recent call last) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy core tests test multiarray.py , line 2840, in test empty assert len w) > 0) File Library Python 2.7 site packages numpy 1.9.0.dev 54d3559 py2.7 macosx 10.9 intel.egg numpy testing utils.py , line 49, in assert raise AssertionError smsg) AssertionError Ran 5040 tests in 87.422s FAILED KNOWNFAIL=5, SKIP=6, errors=19, failures=1) <nose.result.TextTestResult run=5040 errors=19 failures=1>","bug"
|
|
"6013","4012","Inconsistent behaviour of np.ma.median.
|
|
The correct mask of the masked array returned by np.ma.median in the following case should be `[False False]`.
|
|
Can someone confirm this?
|
|
This blocks https github.com scikit learn scikit learn issues 2560 .","bug"
|
|
"6013","4013","f2py passes 0 to omitted optional arguments instead of omitting them.
|
|
Compiling this Fortran module .
|
|
with `f2py m foobar c foobar.f90` and trying to use it gives the following results .
|
|
I m using numpy 1.8.0","bug"
|
|
"6013","4014","Using SVD instead of Cholesky in multivariate gaussian.
|
|
Hi guys I was perusing the `multivariate gaussian` code in `mtrand.pyx`, and I noticed that [you have a TODO item forgotten there since 2005] https github.com numpy numpy blame master numpy random mtrand mtrand.pyx L4175) .
|
|
Cholesky is faster than SVD, and is also what Matlab uses.
|
|
I thought you might want to be reminded of this item.
|
|
Thanks for the great work, otherwise it s a blast talking with you guys and finding the amazing libs such as `randomkit` and `cephes` that you have!","enhancement"
|
|
"6013","4015","Array from memory view constructor is not PEP 3118 compliant.
|
|
According to [PEP 3118] http www.python.org dev peps pep 3118 the py buffer struct), If ndims is 0 indicating a scalar), then [shape member] must be NULL.
|
|
Similarly, [buffer protocol documentation] http docs.python.org 3 c api buffer.html complex arrays) says If ndim == 0, the memory location pointed to by buf is interpreted as a scalar of size itemsize.
|
|
In that case, both shape and strides are NULL.
|
|
However, [numpy array constructor] https github.com numpy numpy blob 066bdec38358fa22933247bd11d7df64ee4c632c numpy core src multiarray ctors.c L1321) sets ndim to 1 when shape == NULL.
|
|
As a result, scalar objects that present a PEP 3118 compliant buffer interface produce length 1 1d arrays instead of scalars.
|
|
.","bug"
|
|
"6013","4016","Documentation examples include >>> .
|
|
Currently, all documentation examples contain selectable `>>>` and other non code characters, preventing one from e.g.
|
|
simply copying the example into the interpreter or a python script, ready to edit to the user s desire.
|
|
It would be much more user friendly to either have these characters be not selectable, or just not present.
|
|
I don t really see the need for them.
|
|
Everyone knows it s Python and can be thrown at the interpreter.
|
|
I understand this is a massive request, although the right regex or more simple search and replace might cut it.","enhancement"
|
|
"6013","4018","Segmentation fault dealing with very large arrays.
|
|
I stumbled upon a segmentation fault in numpy v1.7.1, when I tried to calculate a linear function of two arrays like so .
|
|
I obviously do not have enough memory for this but numpy crashed with a hard segfault, though I am not sure where exactly in the code it happened.","bug"
|
|
"6013","402","ERROR test power zero test umath.TestPower) with python2.
|
|
debug flavour.
|
|
Hello, the test suite executed under debug flavor of python2.6 or 2.7 returns this error ====================================================================== ERROR test power zero test umath.TestPower) Traceback most recent call last) File tmp buildd python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 139, in test power zero assert complex equal np.power zero, 0 1j), cnan) RuntimeWarning invalid value encountered in power The built was run on a Debian unstable distribution.
|
|
For a fill log, you can refer to https buildd.debian.org status fetch.php?pkg=python numpy arch=i386 ver=1 3A1.7.0 b1 1 stamp=1346419177 Regards, Sandro","bug"
|
|
"6013","4023","Using numpy.f2py inside python fails because of numpy.distutils.exec command.
|
|
The following does not work on winXP with Python 2.7.3 and numpy1.7 source= subroutine test ) print , Hello world end subroutine import numpy.f2py as f2py f2py.compile source,modulename= untitled ) with the error Could not locate executable C Python27python.exe Executable C Python27python.exe does not exist There is a stackoverflow question about this http stackoverflow.com questions 19183904 f2py could not locate executable cpython27pythonw exe executable cpython27pyt Somehow exec command looses the slashes in the python executable path.","bug"
|
|
"6013","4026","np.ma.compressed ) returns masked array if no elements are masked.
|
|
Hi, According to the documentation in http docs.scipy.org doc numpy reference generated numpy.ma.MaskedArray.compressed.html The output of compressed ) is noted to be be not a masked array.
|
|
And hence the equivalent method ma.compressed ) is also expected to not return masked array.
|
|
But it is found to return masked array in the following case.
|
|
MA=np.ma.array [1,2,3]) np.ma.isMA np.ma.compressed MA)) This will output True !!
|
|
Note that, if you do mask atleast one element, then it will return an ordinary numpy array as expected.","bug"
|
|
"6013","403","numpy.i bug.
|
|
Hi, There seems to be a simple error in numpy.i, numpy doc swig numpy.i) in lines 1506 and 1529 require fortran array) should be !require fortran array)","bug"
|
|
"6013","4035","exotically large data types seem flaky.
|
|
For example `float128` .
|
|
Copied from scipy issue https github.com scipy scipy issues 3051 issuecomment 28155999","bug"
|
|
"6013","4036","Multiplying a sequence type by a np.float should raise an error.
|
|
Example .
|
|
The normal Python behavior is to raise a TypeError instead of silently truncating .
|
|
This happens on Python 2 and 3.","bug"
|
|
"6013","4041","Lots of f2py documentation is out of date or inaccurate.
|
|
News from 2002, links to websites that don t exist any more, instructions for checking out code from CVS etc [Hopefully I can work out how to attach pull requests to this issue]","bug"
|
|
"6013","4043","ma.asarray ) should pass through objects of the right type.
|
|
For performance and consistency with `np.asarray`, `ma.asarray` should pass through objects that are already of the right type .
|
|
Same applies to `ma.asanyarray`.","enhancement"
|
|
"6013","4048","numpy.dot creates a copy for arrays created with stride tricks, leading to MemoryError.
|
|
Run this script with `ulimit v 1000000` limit memory use to 1GB) .
|
|
`np.dot` tries to allocate a 16384 16384 array and fails with MemoryError For cases like this, even when there is enough memory to allocate the big array, it would be much faster re cache etc) to do the calculation using the existing array and an unoptimised algorithm than copying to a big array then using an optimised BLAS.","bug"
|
|
"6013","4050","Instillation Problem Mac OS X 10.7.
|
|
I am fairly new to programing and need numpy for another program and I cannot install it.
|
|
I get did this in my command line easy install upgrade numpy and get this huge error and I have no idea what it means could anyone help?
|
|
I think the main error is the last one cannot locate working compiler and I don t know what that means.
|
|
Traceback most recent call last) File anaconda bin easy install , line 6, in <module> sys.exit main )) File anaconda lib python2.7 site packages setuptools command easy install.py , line 1986, in main with ei usage lambda File anaconda lib python2.7 site packages setuptools command easy install.py , line 1967, in with ei usage return f ) File anaconda lib python2.7 site packages setuptools command easy install.py , line 1990, in <lambda> distclass=DistributionWithoutHelpCommands, kw File anaconda lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File anaconda lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File anaconda lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File anaconda lib python2.7 site packages setuptools command easy install.py , line 360, in run self.easy install spec, not self.no deps) File anaconda lib python2.7 site packages setuptools command easy install.py , line 604, in easy install return self.install item spec, dist.location, tmpdir, deps) File anaconda lib python2.7 site packages setuptools command easy install.py , line 634, in install item dists = self.install eggs spec, download, tmpdir) File anaconda lib python2.7 site packages setuptools command easy install.py , line 829, in install eggs return self.build and install setup script, setup base) File anaconda lib python2.7 site packages setuptools command easy install.py , line 1109, in build and install self.run setup setup script, setup base, args) File anaconda lib python2.7 site packages setuptools command easy install.py , line 1095, in run setup run setup setup script, args) File anaconda lib python2.7 site packages setuptools sandbox.py , line 33, in run setup lambda execfile File anaconda lib python2.7 site packages setuptools sandbox.py , line 81, in run return func ) File anaconda lib python2.7 site packages setuptools sandbox.py , line 35, in <lambda> { file setup script, name main } File setup.py , line 192, in <module> File setup.py , line 185, in setup package File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils core.py , line 169, in setup File anaconda lib python2.7 distutils core.py , line 152, in setup dist.run commands ) File anaconda lib python2.7 distutils dist.py , line 953, in run commands self.run command cmd) File anaconda lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File anaconda lib python2.7 site packages setuptools command bdist egg.py , line 172, in run self.run command egg info ) File anaconda lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File anaconda lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command egg info.py , line 10, in run File anaconda lib python2.7 distutils cmd.py , line 326, in run command self.distribution.run command command) File anaconda lib python2.7 distutils dist.py , line 972, in run command cmd obj.run ) File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command build src.py , line 153, in run File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command build src.py , line 164, in build sources File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command build src.py , line 299, in build library sources File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command build src.py , line 386, in generate sources File private var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy core setup.py , line 672, in get mathlib info File anaconda lib python2.7 distutils command config.py , line 248, in try link self.
|
|
check compiler ) File var folders hn 9b9tk1q90mz54dq1zqg3 yn40000gn T easy install 2oRjNu numpy 1.8.0 numpy distutils command config.py , line 46, in check compiler File anaconda lib python2.7 distutils command config.py , line 103, in check compiler customize compiler self.compiler) File anaconda lib python2.7 distutils sysconfig.py , line 170, in customize compiler osx support.customize compiler config vars) File anaconda lib python2.7 osx support.py , line 412, in customize compiler find appropriate compiler config vars) File anaconda lib python2.7 osx support.py , line 191, in find appropriate compiler Cannot locate working compiler ) SystemError Cannot locate working compiler","bug"
|
|
"6013","4054","WISH [option to] make np.floor ), np.ceil ), np.trunc ) etc return arrays with dtype=int.
|
|
I m trying to plot two quantities against each other but they are sampled at different intervals.
|
|
I tried `plt.plot array1, array2[np.floor array1time array2interval)])`.
|
|
However, `np.floor` returns a float array so I need to cast the result to an integer type first.
|
|
This seems a bit silly.
|
|
Is there any reason why someone would depend on the result of floor ) being a floating point number?","bug"
|
|
"6013","4059","numpy html 1.8.0.zip missing.
|
|
http docs.scipy.org doc contains Numpy 1.8.0 Reference Guide, [HTML zip], [PDF] [HTML zip] link refers to http docs.scipy.org doc numpy 1.8.0 numpy html 1.8.0.zip which is not available.","bug"
|
|
"6013","406","1.7.0b1 Unittest errors building on Debian armhf.
|
|
Hello, when building 1.7.0b1 on armhf the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=armhf ver=1 3A1.7.0 b1 1 stamp=1346425398 Regards, Sandro","bug"
|
|
"6013","4062","numpy.dot out of memory when multiplying big matrix.
|
|
I have a 2000 by 1,000,000 matrix A and want to calculate the 2000 by 2000 matrix .> B = numpy.dot A,A.T) but numpy just eats up all my memory, slows down my whole computer and crashes after a couple of hours.
|
|
I then rewrote the matrix multiplication to .> B = numpy.zeros 2000,2000) .> A.shape = 2000,10000,100) .> for M in numpy.rollaxis A,2) .>.. B = numpy.dot M,M.T) and it just runs fine in a couple of minutes.
|
|
I don t see the reason numpy needs so much memory for a matrix multiplication.
|
|
But I think at least it should not crash on that problem.","bug"
|
|
"6013","4063","mean ) and median )) should work with object arrays.
|
|
With NumPy 1.8, `mean )` started to break when calculating the global) mean of an array that contains objects arrays with an object `dtype`).
|
|
This also breaks `median )` on such arrays.
|
|
Here is an example .
|
|
Another example is case of numbers with uncertainties from the uncertainties package https github.com lebigot uncertainties issues 22).
|
|
I think that it would be better if NumPy did not assume that scalar results have a `dtype`, since arrays can contain objects that have a meaningful mean.
|
|
I believe that such objects should not be forced to have a `dtype`, which is obviously NumPy specific they even can t, for Python scalars like floats).
|
|
Furthermore, a `dtype` is in principle not necessary for the calculation of the mean of such objects, so it would look strange if they had to have one.
|
|
The problem is that `numpy.mean )` assumes that the intermediate result obtained has a `dtype` with a `type` attribute).
|
|
Therefore, I suggest that NumPy s `mean )` also handle arrays of objects that are not of the standard NumPy types their `dtype` is object, and they contain objects that have a meaningful mean, like ints, floats, numbers with uncertainties, etc.
|
|
).","bug"
|
|
"6013","4065","np.log 1) gives the wrong answer when using the out argument and masked arrays.
|
|
These should all be zero, right?
|
|
.","bug"
|
|
"6013","4068","squeeze and record arrays.
|
|
a.squeeze ) for a 1,) shaped record array is changing the shape to ).
|
|
Is this expected behaviour?
|
|
The code is distilled from test io s test 3d shaped dtype test.
|
|
Note a.squeeze ).shape is unexpectedly empty uname a Linux mycomputer 3.5.0 43 generic 66 Ubuntu SMP Wed Oct 23 12 01 49 UTC 2013 x86 64 x86 64 x86 64 GNU Linux python Python 2.7.6rc1 v2.7.6rc1 4913d0e9be30, Nov 4 2013, 09 59 57) [GCC 4.7.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.version.version 1.8.0.dev 4a9d23b >>> from numpy import dtype, array >>> dt = dtype [ name , S4 ), x , float), y , float), ... block , int, 2, 2, 3))]) >>> a = array [ aaaa , 1.0, 8.0, [[[1, 2, 3], [4, 5, 6]], ... [[7, 8, 9], [10, 11, 12]]])], ... dtype=dt) >>> a.shape 1,) >>> a.squeeze ).shape )","bug"
|
|
"6013","4069","Array is too big dimension check too restrictive for some stride tricks.
|
|
The dimension check on array creation seems too restrictive for certain stride tricks.
|
|
For example, when adding an extra dimension to an array to get a sliding window index import numpy as np a = np.arange 1e6).reshape 1000,1000) b = np.lib.stride tricks.as strided a, shape= 1000,1000,1000), strides= 8000, 8, 8)) This example throws an array is too big exception.
|
|
The constructor in core src multiarray ctors.c) used by stride tricks multiplies out all the dimensions for the new array.
|
|
When stride tricks are used, the storage size of the array is not the product of the dimensions.
|
|
A better check may be to multiply out the dimensions WITH strides if the highest index in all dimensions results in a storage index that s too large, then the array is too big .","bug"
|
|
"6013","407","1.7.0b1 Unittest errors building on Debian ia64.
|
|
Hello, when building 1.7.0b1 on ia64 the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=ia64 ver=1 3A1.7.0 b1 1 stamp=1346422426 Regards, Sandro","bug"
|
|
"6013","4072","Register class a subclass of NumPy array.
|
|
Some packages use `isinstance x, numpy.ndarray)` to check if a given object can be used as an `ndarray`.
|
|
This fails of course) for object from classes derived from `object` even if they implement all numpy methods and attributes.
|
|
It would be good to be able to register a class as a ndarray subclass using `abc`","bug"
|
|
"6013","4074","ImportError Need nose >=0.10.0 for tests.
|
|
Hi, I just recently tried installing the numpy.py and after using python c import numpy; numpy.test ) I received this error.
|
|
root beaglebone python c import numpy;numpy.test ) Running unit tests for numpy Traceback most recent call last) File <string> , line 1, in <module> File usr lib python2.7 site packages numpy testing nosetester.py , line 326, in test self.
|
|
show system info ) File usr lib python2.7 site packages numpy testing nosetester.py , line 187, in show system info nose = import nose ) File usr lib python2.7 site packages numpy testing nosetester.py , line 69, in import nose raise ImportError msg) ImportError Need nose >= 0.10.0 for tests see http somethingaboutorange.com mrl projects nose Can you help me in fixing this please?","bug"
|
|
"6013","408","1.7.0b1 Unittest errors building on Debian kfreebsd amd64.
|
|
Hello, when building 1.7.0b1 on Debian unstable kfreebsd amd64 the tests suite returns some error .
|
|
Ful build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=kfreebsd amd64 ver=1 3A1.7.0 b1 1 stamp=1346419487 Regards, Sandro","bug"
|
|
"6013","4082","libgdal 1.10.1 gdal 1.10.0 numpy 1.8.0 do not work.
|
|
Ran into an issue using the above combination of gdal and numpy.
|
|
See http codepad.org JYGfSCYG Pasted here for convenience .
|
|
Downgrading to numpy 1.7.0 gets me back to a working combination of gdal numpy","bug"
|
|
"6013","4083","Close the old numpy Trac bug tracker and all issues.
|
|
Redirect to github?.
|
|
When finding bug reports or searching for errors you often come across the old Trac instance of numpy and very very old bug reports many of which have not been closed or commented on for years.
|
|
Be good if it got shut down, issues closed and a note or redirect to github instead.","bug"
|
|
"6013","4084","Align=true in dtype constructor does not fully ensure alignment.
|
|
Suppose we create the following dtype >>> dt1 = numpy.dtype dict names=[ i1 , i2 ], formats=[numpy.int32, numpy.int32], ... offsets=[0,4], itemsize=12, aligned=True)) >>> dt1.alignment 4 >>> dt1.isalignedstruct True The problem here is that the itemsize ``12`` is not actually achievable in C code for these particular fields without an explicit padding.
|
|
But the ``isalignedstruct`` field reports that the struct is aligned, which is a bit misleading.
|
|
There is no way to know which one is true 1) ``itemsize`` is consistent with the fields that is, in C code you would not need any explicit alignment modifiers) 2) ``itemsize`` requires an explicit alignment attribute in C code e.g, ``itemsize=16`` would be achievable with an explicit `` align 16)`` or equivalent applied to the whole struct).
|
|
3) ``itemsize`` is not achievable in C code and requires padding.
|
|
So, is the current numpy behavior an intended one, or is it a bug?
|
|
I would find the following behavior more logical >>> dt1 = numpy.dtype dict names=[ i1 , i2 ], formats=[numpy.int32, numpy.int32], ... offsets=[0,4], itemsize=8, aligned=True)) >>> dt1.base alignment 4 >>> dt1.alignment 4 >>> dt1.isalignedstruct True >>> dt1 = numpy.dtype dict names=[ i1 , i2 ], formats=[numpy.int32, numpy.int32], ... offsets=[0,4], itemsize=12, aligned=True)) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError itemsize 12 for this dtype is not achievable without an explicit padding >>> dt1 = numpy.dtype dict names=[ i1 , i2 ], formats=[numpy.int32, numpy.int32], ... offsets=[0,4], itemsize=16, aligned=True)) >>> dt1.base alignment 4 >>> dt1.alignment 16 >>> dt1.isalignedstruct True It would be also convenient to have a keyword for the total alignment of the struct which will affect how it is placed in an encompassing struct) >>> dt1 = numpy.dtype dict names=[ i1 , i2 ], formats=[numpy.int32, numpy.int32], ... offsets=[0,4], itemsize=8, alignment=8, aligned=True)) >>> dt1.base alignment 4 >>> dt1.alignment 8 >>> dt1.isalignedstruct True","bug"
|
|
"6013","4085","random.random integers fails if upper bound is 2 31 or higher.
|
|
This code gives an error I tried numpy 1.7.1 and 1.8.0; I m using 64 bit versions in Windows 8.1 64 bit, with python 2.7.5, Anaconda 1.8.0 64 bit) x = numpy.random.random integers 2 31,size=1).astype numpy.uint64) Python int too large to convert to C long Note that the docs don t specify an upper bound on the range.
|
|
This behavior is confusing.
|
|
If 2 31 is replaced by 2 30 it works fine.","bug"
|
|
"6013","4086","MaskedArray.
|
|
getitem does not call array finalize .
|
|
``MaskedArray.
|
|
getitem `` does not call `` array finalize `` before returning the slice unlike ``ndarray.
|
|
getitem ``).
|
|
This causes issues for sub classes of ``MaskedArray``.
|
|
As a workaround, sub classes can overload `` update from`` but this is a hack.
|
|
This should be reasonably straightforward to fix, so if no one else does it soon I will try and open a pull request.","bug"
|
|
"6013","409","1.7.0b1 Unittest errors building on Debian kfreebsd i386.
|
|
Hello, when building 1.7.0b1 on Debian unstable kfreebsd i386 the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=kfreebsd i386 ver=1 3A1.7.0 b1 1 stamp=1346418853 Regards, Sandro","bug"
|
|
"6013","4091","strides wrong for dimension = 1 objects with relaxed strides.
|
|
if compiled with NPY RELAXED STRIDES CHECKING=1 >>> x =np.array 1,)) >>> print x.strides 9223372036854775807,) it goes into thise code in numpy core src multiarray ctors.c 3739 .","bug"
|
|
"6013","4092","Fix mis formatting in html output for `histogram2d` docstring.
|
|
In the online html docs of `histogram2d` [here] http docs.scipy.org doc numpy reference generated numpy.histogram2d.html) contain some mis formatting in the examples.
|
|
Looking at the code [here] https github.com numpy numpy blob master numpy lib twodim base.py L653) it seems the reason is that the examples contain `\n` in the string .
|
|
Anyone know how to fix this?","bug"
|
|
"6013","4093","Loading arrays from a compressed .npz file fails under Python 2.6.6.
|
|
Attempting to load an array of more than 64K from a compressed .npz file leads to the exception ValueError buffer is smaller than requested size occurring at lib python2.6 site packages numpy lib format.py , line 476, in read array The code below reproduces the error for me .
|
|
I believe what s going on here is that numpy lib format.py is trying to read the file in chunks of BUFFER SIZE = 2 18 bytes, however ZipExtFile in 2.6.6 has compreadsize set to 2 16, so less data is read than expected.
|
|
System Info Python version 2.6.6 OS RHEL 2.6.32 279.el6.x86 64 numpy version 1.8.0","bug"
|
|
"6013","4096","np.dot returning incorrect result with Accelerate on OS X.
|
|
There s a chance I m missing something here, but I can t see what so I figure I should report this The problem I m seeing is that I m getting what I think are incorrect results from np.dot for single precision arrays on OS X 10.6.8) with numpy 1.8.0) compiled with the default Accelerate framework.
|
|
If I recompile numpy with OpenBLAS the problem goes away, i.e.
|
|
the results appear to be the same within floating point precision).
|
|
Below is an example, it shows that the result for single and double precision are very different.
|
|
.","bug"
|
|
"6013","4097","Missing gufuncs in 1.8.0.
|
|
I just installed 1.8.0 with ``pip 3.3 install numpy user``, and I get lots of test failures like ====================================================================== ERROR test vector single test gufuncs linalg.TestSolve) Traceback most recent call last) File home rgommers .local lib python3.3 site packages numpy linalg tests test gufuncs linalg.py , line 206, in test vector single array b, dtype=single)) File home rgommers .local lib python3.3 site packages numpy linalg tests test gufuncs linalg.py , line 375, in do assert almost equal b, gula.matrix multiply a,x)) File home rgommers .local lib python3.3 site packages numpy linalg gufuncs linalg.py , line 304, in matrix multiply return impl.matrix multiply a,b, kwargs) AttributeError module object has no attribute matrix multiply 32 bit Linux system, gcc 4.8.1.
|
|
Haven t investigated in detail yet, maybe someone can easily tell what s going on here?","bug"
|
|
"6013","410","1.7.0b1 Unittest errors building on Debian mips.
|
|
Hello, when building 1.7.0b1 on Debian unstable mips the tests suite returns some error ``` running tests for 2.7 plain ...........................................................................................................................................................................................................................................................................................................S.....................E..E....................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K.....E..........E.............E.....E.EEE....E....E..................................................................................................................................SSS...................................................................................................................................................................................................................................................................................E...........................K................................................................EEE............E.............EEEEK......................K............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E....................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test searchsorted test multiarray.TestMethods) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test multiarray.py , line 791, in test searchsorted a.real = [0, 0, 1, 1, 0, 1, np.nan, np.nan, np.nan] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test sort test multiarray.TestMethods) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test multiarray.py , line 484, in test sort a.real = [np.nan, np.nan, np.nan, 1, 0, 1, 1, 0, 0] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test numeric.TestIsclose.test equal nan Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1265, in test equal nan assert array equal isclose nan, nan, equal nan=True), [True]) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in multiply) Traceback most recent call last) File usr lib python2.7 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1234, in test ip isclose yield assert array equal, isclose x, y), result) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [0], [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose 0, [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip none isclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1241, in tst none isclose assert not any isclose x, y)), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test masked arrays Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1274, in test masked arrays assert type x) == type isclose inf, x))) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 3648, in mul return multiply self, other) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 950, in call result = m da RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test nan any test umath.TestArctan2SpecialValues) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 411, in test nan any assert arctan2 isnan np.nan, np.inf) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 328, in assert arctan2 isnan assert np.isnan ncu.arctan2 x, y)), arctan s, s) is s, not nan x, y, ncu.arctan2 x, y))) RuntimeWarning invalid value encountered in arctan2 ====================================================================== ERROR test umath.test nextafter Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1123, in test nextafter return test nextafter np.float64) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterf Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1126, in test nextafterf return test nextafter np.float32) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterl Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1131, in test nextafterl return test nextafter np.longdouble) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in absolute) Traceback most recent call last) File usr lib python2.7 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 482, in test cabs inf nan ref = g x[i], y[i]) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 477, in g return np.abs np.complex a, b)) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, 0, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, 0, nan, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, inf, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, inf, nan, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR Ticket 396 Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib tests test regression.py , line 51, in test poly1d nan roots self.assertRaises np.linalg.LinAlgError,getattr,p, r ) File usr lib python2.7 unittest case.py , line 476, in assertRaises callableObj args, kwargs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib polynomial.py , line 1200, in getattr return roots self.coeffs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib polynomial.py , line 224, in roots A[0, ] = p[1 ] p[0] RuntimeWarning invalid value encountered in divide ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4785 tests in 2006.096s FAILED KNOWNFAIL=5, SKIP=5, errors=24) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy Python version 2.7.3 default, Aug 26 2012, 19 24 34) [GCC 4.6.3] nose version 1.1.2 running tests for 2.7 debug ...........................................................................................................................................................................................................................................................................................................S.............................................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K........................................E.E..............................................................................................................................................SSS...............................................................................................................................................................................................................................................................................................................K.................................................E..................................................K......................K.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 1986, in allclose return all less equal abs x y), atol rtol abs y))) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 1986, in allclose return all less equal abs x y), atol rtol abs y))) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test power zero test umath.TestPower) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 139, in test power zero assert complex equal np.power zero, 0 1j), cnan) RuntimeWarning invalid value encountered in power ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4792 tests in 2328.468s FAILED KNOWNFAIL=5, SKIP=5, errors=6) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy Python version 2.7.3 default, Aug 26 2012, 14 33 04) [GCC 4.6.3] nose version 1.1.2 [1518851 refs] running tests for 2.6 plain ...........................................................................................................................................................................................................................................................................................................S.....................E..E....................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K.....E..........E.............E.....E.EEE....E....E..................................................................................................................................SSS...................................................................................................................................................................................................................................................................................E...........................K................................................................EEE............E.............EEEEK......................K............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E....................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test searchsorted test multiarray.TestMethods) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test multiarray.py , line 791, in test searchsorted a.real = [0, 0, 1, 1, 0, 1, np.nan, np.nan, np.nan] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test sort test multiarray.TestMethods) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test multiarray.py , line 484, in test sort a.real = [np.nan, np.nan, np.nan, 1, 0, 1, 1, 0, 0] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test numeric.TestIsclose.test equal nan Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1265, in test equal nan assert array equal isclose nan, nan, equal nan=True), [True]) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in multiply) Traceback most recent call last) File usr lib python2.6 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1234, in test ip isclose yield assert array equal, isclose x, y), result) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [0], [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose 0, [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip none isclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1241, in tst none isclose assert not any isclose x, y)), msg x, y)) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test masked arrays Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1274, in test masked arrays assert type x) == type isclose inf, x))) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 3648, in mul return multiply self, other) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 950, in call result = m da RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test nan any test umath.TestArctan2SpecialValues) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 411, in test nan any assert arctan2 isnan np.nan, np.inf) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 328, in assert arctan2 isnan assert np.isnan ncu.arctan2 x, y)), arctan s, s) is s, not nan x, y, ncu.arctan2 x, y))) RuntimeWarning invalid value encountered in arctan2 ====================================================================== ERROR test umath.test nextafter Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1123, in test nextafter return test nextafter np.float64) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterf Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1126, in test nextafterf return test nextafter np.float32) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterl Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1131, in test nextafterl return test nextafter np.longdouble) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in absolute) Traceback most recent call last) File usr lib python2.6 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 482, in test cabs inf nan ref = g x[i], y[i]) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 477, in g return np.abs np.complex a, b)) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, 0, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, 0, nan, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, inf, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, inf, nan, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR Ticket 396 Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib tests test regression.py , line 51, in test poly1d nan roots self.assertRaises np.linalg.LinAlgError,getattr,p, r ) File usr lib python2.6 unittest.py , line 336, in failUnlessRaises callableObj args, kwargs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib polynomial.py , line 1200, in getattr return roots self.coeffs) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib polynomial.py , line 224, in roots A[0, ] = p[1 ] p[0] RuntimeWarning invalid value encountered in divide ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4785 tests in 1973.746s FAILED KNOWNFAIL=5, SKIP=5, errors=24) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy Python version 2.6.8 unknown, Jun 26 2012, 03 02 04) [GCC 4.6.3] nose version 1.1.2 running tests for 2.6 debug ...........................................................................................................................................................................................................................................................................................................S.............................................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K........................................E.E..............................................................................................................................................SSS...............................................................................................................................................................................................................................................................................................................K.................................................E..................................................K......................K.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd2 python numpy 1.7.0 b1 1 mips vVC1yE python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equ","bug"
|
|
"6013","4100","array True) == False.
|
|
NumPy behavior is inconsistent with that of python builtins .
|
|
but .","bug"
|
|
"6013","4101","Compiling numpy 1.7.1 1.8.0 for python 3.3 on Windows with VS2010 SDK fails with broken toolchain due to missing manifest.
|
|
I ran into this, and the solution found on [numpy discussion] http mail.scipy.org pipermail numpy discussion 2012 November 064485.html) worked for me, but does turn up on a search of the issues.
|
|
Summary Python <3.2 are compiled with VS2008 SDK v7.0, sdk for win7 NET 3.5.1 ) Python 3.3 ?)
|
|
is compiled with VS2010 SDK v7.1, sdk for win7 NET 4 ) It s often recommended that c extensions such as numpy) be compiled using the same compiler used to compile the python binaries, to ensure CRT compatability, so using VS2010 SDK 7.1 to compile numpy for python 3.3 is implied.
|
|
VS2010 msvc10) behaves differently with regards to manifest files then VS2008 msvc9) but the cpython distutils, which the numpy build process relies on, have not been altered to allow for this as of python 3.3.3).
|
|
a [cpython issue] http bugs.python.org issue16296) was filed in dec 2012, but has stalled.
|
|
The issue manifests as a broken toolchain exception when attempting to build numpy both 1.7.1 and 1.8.0), which results from the following compiler error message.
|
|
.
|
|
The fix suggested by Christoph Gohlke on the ml is to edit the file [ Lib distutils msvc9compiler.py 648 ] https github.com akheron cpython blob 157e2d3078abe02576dd8710a8a4104dd620ccbe Lib distutils msvc9compiler.py L648), Located under your python 3.3 install directory, and modify the line from .
|
|
to .
|
|
I ve tried, and it has worked for me.
|
|
Thanks goes to cgohlke, the intrepid crusader for windows pydata binaries.
|
|
There, that should give pagerank something to chew on.","bug"
|
|
"6013","4104","Boolean.
|
|
In python True should be 1 and not False as in your own bool implementation!","bug"
|
|
"6013","4107","allclose does not work for some dtypes.
|
|
.
|
|
It is also inefficient for exact type where it is equivalent to equal x, y).all ).","bug"
|
|
"6013","411","1.7.0b1 Unittest errors building on Debian powerpc.
|
|
Hello, when building 1.7.0b1 on Debian unstable powerpc the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=powerpc ver=1 3A1.7.0 b1 1 stamp=1346419716 Regards, Sandro","bug"
|
|
"6013","4116","pip requirements files support.
|
|
Disclaimer I am just learning python...
|
|
So hopefully the following makes sense ;) I have a requirements file in order to use with pip [Docs] http www.pip installer.org en latest cookbook.html requirements files) requirement.txt .
|
|
When I run .
|
|
the install will always fail See shortened output below).
|
|
What always works though is installing numpy first and then the requirements file .
|
|
I believe this is similar to this issue https github.com pypa pip issues 720 failing output .","bug"
|
|
"6013","4118","BUG?
|
|
np.fromfile on Py3k.
|
|
I ve translated some code we have from Python2 to Python3, and I think I ve encountered a strange bug with `np.fromfile`.
|
|
If I run the following code snippet .
|
|
I get this output in Python2 .
|
|
But this output in Python3 .
|
|
This is not good, because subsequent reads from the file will not occur at correct positions.
|
|
In other words, I have to add the ` ` to the file open parameter in order to get multiple calls to `fromfile` to read with correct positions.
|
|
I ve uploaded the test file here in case someone wants to try to replicate http faculty.washington.edu larsoner test.w I m running Ubuntu Linux 13.10 64 bit.
|
|
I had this problem on the system version, as well as on a version I just compiled myself with MKL support).","bug"
|
|
"6013","4119","Building numpy causes VirtualBox to start a VM.
|
|
This is by far the most weird bug I ever saw.
|
|
Trying to build numpy `pip install numpy`) causes VirtualBox to start my `arch` VM.
|
|
It happens at the stage where the build process says `building library npymath sources`.
|
|
Environment OSX Maverix clang 500.2.79 Homebrew Python 2.7.6 pip 1.4.1 VirtualBox 4.3.4","bug"
|
|
"6013","412","1.7.0b1 Unittest errors building on Debian sparc.
|
|
Hello, when building 1.7.0b1 on Debian unstable sparc the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=sparc ver=1 3A1.7.0 b1 1 stamp=1346421921 Regards, Sandro","bug"
|
|
"6013","4121","numpy.linalg.matrix rank missing from docs.
|
|
numpy.linalg.matrix rank does not seem to be listed at either http docs.scipy.org doc numpy reference routines.linalg.html or http docs.scipy.org doc scipy reference linalg.html .
|
|
However it does appear to exist and work.","bug"
|
|
"6013","4123","BUG lexsort segfault.
|
|
This is very minor, as it only happens when you do something wrong, but I ll report it anyway.
|
|
It was originally seen in a `pandas` context when fuzz testing, as `np.lexsort pd.DataFrame range 5))` segfaults.
|
|
It wasn t expected to give sensible results, but I wasn t expecting this hard a crash, either.
|
|
It turns out to be because of insufficient error checking on `numpy` s side .
|
|
gives .
|
|
in 1.9.0.dev 8a2728c.","bug"
|
|
"6013","4125","savetxt raises a misleading error if array s shape is not rectangular.
|
|
Hi, .
|
|
The error is misleading because all elements are made of floats.
|
|
I agree that the array type is object but it would be easier to debug if we can distinguish between an array with mixed types and an array uncorrectly shaped.
|
|
It took me half an hour to figure out my mistake.
|
|
I would be happy to hear your comments.
|
|
Thanks.","bug"
|
|
"6013","4126","Integer overflow using numpy.dot.
|
|
Uncertain whether this is the desired behaviour or a bug.
|
|
In my opinion the programmer should not have to consider whether the length of the arrays are too long.
|
|
Code .
|
|
Returns 32768 Integer overflow also happens when using other bit sizes of int.
|
|
It especially becomes a problem when using `csr matrix` and probably many more modules) that automatically sets the `dtype` to `np.int8` when dealing with a matrix of zeros and ones.","bug"
|
|
"6013","4129","Pickling MaskedArray subclass doesn t preserve extra attributes.
|
|
I m having trouble getting a MaskedArray subclass to round trip through pickle and preserve the extra subclass attributes.
|
|
I m not sure if this is a bug or a lack of understanding on my part.
|
|
Following is the strawman code I m using, which is basically a stripped down version of `numpy ma tests test subclassing.py`.
|
|
The only change is the addition of a `dtype` keyword arg in `MSubArray.
|
|
new ` otherwise pickling fails completely).
|
|
.
|
|
This produces .
|
|
This is related to the suggestion of pierregm in 4086 that we rework the way that astropy Table subclasses MaskedArray https github.com astropy astropy pull 1872).
|
|
I m using numpy 1.7.1 here.","bug"
|
|
"6013","413","1.7.0b1 Unittest errors building on Debian armel.
|
|
Hello, when building 1.7.0b1 on Debian unstable armel the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=armel ver=1 3A1.7.0 b1 1 stamp=1346468161 Regards, Sandro","bug"
|
|
"6013","4131","1.8 deprecations Follow up ticket.
|
|
Depending on the change, these are probably interesting for removal in 1.10. the non integer one probably later.
|
|
[ ] insert and delete changes 0 d input array and boolean indexing array) [ ] non integer scalar deprecations.
|
|
see test deprecations; PyArray PyIntAsIntp) [ ] deprecation of `arr.
|
|
index )` for non 0 d arrays.
|
|
see test deprecations) [ ] full and economic modes of qr factorization [ ] nditer `oa ndim == 0` for no `op axes`.
|
|
This needs changing in `numpy core src multiarray nditer constr.c`.
|
|
see also gh 3104) [ ] `NPY RELAXED STRIDES CHECKING` was introduced as a compile time option in 1.8 grep code for it).
|
|
Please extend the list or comment) with other deprecations introduced in 1.8.","bug"
|
|
"6013","4132","1.9 deprecations Follow up ticket.
|
|
Issue to gather deprecations in 1.9 and hints on the actions needed for them [ ] Indexing deprecations implemented by gh 3798 see PR and deprecation tests) Multiple `...` usage `a[..., ...]`).
|
|
Non integer array like usage as index `a[[1., 2]]`).
|
|
[ ] Sequence repetition with non integers `float64 1) [2]`).
|
|
see gh 4038) [ ] FutureWarning in array split for empty arrays being replaced.
|
|
see gh 4102) [ ] Future and DeprecationWarning for 1 d fancy indexing special cases see gh 4804) [ ] Comparison deprecations of objects and None.
|
|
see gh 3514 and also consider gh 4910) [ ] DEP deprecate pkgload 5245","bug"
|
|
"6013","4135","Check that gh 4109 is fixed and tests in place..
|
|
Especially the part about `arr[ 0 , 0 ]` like corner cases.
|
|
At the time of writing this, the indexing rework fixes it, but that one corner cases would slip past though I will probably restructure it a little so that it does not).","bug"
|
|
"6013","4136","sum ) bug with float96 under python3.4b1.
|
|
Standard python3.4 debug build on 32 bit linux, gcc 4.8.1.
|
|
To reproduce >>> import numpy as np >>> d = np.arange 1, 3, dtype=np.float96) >>> np.sum d) 3.0 >>> np.sum d[ 1]) 2.0 >>> d array [ 1.0, 2.0], dtype=float96) >>> d[ 1] array [ 2.0, 1.0], dtype=float96) >>> d.sum ) 3.0 >>> d[ 1].sum ) 2.0 Shows up in these tests ====================================================================== FAIL test sum test ufunc.TestUfunc) Traceback most recent call last) File home rgommers Code numpy numpy core tests test ufunc.py , line 332, in test sum assert almost equal np.sum d[ 1]), tgt) File home rgommers Code numpy numpy testing utils.py , line 477, in assert almost equal raise AssertionError build err msg )) AssertionError Arrays are not almost equal to 7 decimals ACTUAL 2.0 DESIRED 3.0 ====================================================================== FAIL test sum complex test ufunc.TestUfunc) Traceback most recent call last) File home rgommers Code numpy numpy testing utils.py , line 454, in assert almost equal assert almost equal actualr, desiredr, decimal=decimal) File home rgommers Code numpy numpy testing utils.py , line 461, in assert almost equal return assert array almost equal actual, desired, decimal, err msg) AssertionError Arrays are not almost equal to 7 decimals mismatch 100.0 ) x array 2.0, dtype=float96) y array 3.0, dtype=float96) During handling of the above exception, another exception occurred Traceback most recent call last) File home rgommers Code numpy numpy core tests test ufunc.py , line 353, in test sum complex assert almost equal np.sum d[ 1]), tgt) File home rgommers Code numpy numpy testing utils.py , line 457, in assert almost equal raise AssertionError build err msg )) AssertionError Arrays are not almost equal to 7 decimals ACTUAL 2 2j) DESIRED 3 3j) I haven t looked at the second error above when building the error message.","bug"
|
|
"6013","4138","segfault with non native endian memoryview in python 3.4b1.
|
|
I don t get a traceback for this one.
|
|
Works fine with python 3.3, raises ValueError like it s supposed to.
|
|
>>> assert sys.byteorder == little >>> x = np.array [1, 2, 3], dtype= >q ) >>> x = memoryview x) Printed to stdout python3.4 Python ceval.c 4257 call function Assertion ` x != void )0) !PyErr Occurred )) || x == void )0) PyErr Occurred )) failed.
|
|
Aborted core dumped) 32 bit linux, python3.4b1 debug build, gcc 4.8.1.
|
|
The above is a simplified version of a test that crashes when running the test suite ``TestNewBufferProtocol.test roundtrip`` from ``test multiarray.py``.
|
|
Maybe someone familiar with memoryview can easily tell if this is a numpy or a python issue?","bug"
|
|
"6013","4139","compiling code with f2py broken by python3.4b1.
|
|
I m looking into this one.
|
|
Looks like a regression in ``importlib``.
|
|
Prevents building scipy.
|
|
Running the f2py tests gives 20 errors all like this one ====================================================================== ERROR test size.TestSizeSumExample.test transpose Traceback most recent call last) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose case.py , line 382, in setUp try run self.inst, setup , setUp )) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose util.py , line 469, in try run return func ) File home rgommers Code numpy numpy f2py tests util.py , line 353, in setUp module name=self.module name) File home rgommers Code numpy numpy f2py tests util.py , line 80, in wrapper raise ret File home rgommers Code pythons lib python3.4 unittest case.py , line 57, in testPartExecutor yield File home rgommers Code pythons lib python3.4 unittest case.py , line 567, in run self.setUp ) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose case.py , line 382, in setUp try run self.inst, setup , setUp )) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose util.py , line 469, in try run return func ) File home rgommers Code numpy numpy f2py tests util.py , line 353, in setUp module name=self.module name) File home rgommers Code numpy numpy f2py tests util.py , line 80, in wrapper raise ret File home rgommers Code pythons lib python3.4 unittest case.py , line 57, in testPartExecutor yield File home rgommers Code pythons lib python3.4 unittest case.py , line 567, in run self.setUp ) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose case.py , line 382, in setUp try run self.inst, setup , setUp )) File home rgommers .local lib python3.4 site packages nose 1.3.0 py3.4.egg nose util.py , line 469, in try run return func ) File home rgommers Code numpy numpy f2py tests util.py , line 353, in setUp module name=self.module name) File home rgommers Code numpy numpy f2py tests util.py , line 74, in wrapper memo[key] = func a, kw) File home rgommers Code numpy numpy f2py tests util.py , line 135, in build module cmd[4 ], asstr out))) RuntimeError Running f2py failed [ m , test ext module 5403 , tmp tmp4jurdobm foo.f90 ] Reading .f2py f2cmap ... Mapping real kind=rk) to double Succesfully applied user defined changes from .f2py f2cmap running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building extension test ext module 5403 sources f2py options [] f2py > tmp tmpes 1ciqt src.linux i686 3.4 test ext module 5403module.c creating tmp tmpes 1ciqt creating tmp tmpes 1ciqt src.linux i686 3.4 Traceback most recent call last) File <string> , line 1, in <module> File home rgommers Code numpy numpy f2py f2py2e.py , line 594, in main run compile ) File home rgommers Code numpy numpy f2py f2py2e.py , line 580, in run compile setup ext modules = [ext]) File home rgommers Code numpy numpy distutils core.py , line 169, in setup return old setup new attr) File home rgommers Code pythons lib python3.4 distutils core.py , line 149, in setup dist.run commands ) File home rgommers Code pythons lib python3.4 distutils dist.py , line 955, in run commands self.run command cmd) File home rgommers Code pythons lib python3.4 distutils dist.py , line 974, in run command cmd obj.run ) File home rgommers Code numpy numpy distutils command build.py , line 39, in run old build.run self) File home rgommers Code pythons lib python3.4 distutils command build.py , line 126, in run self.run command cmd name) File home rgommers Code pythons lib python3.4 distutils cmd.py , line 313, in run command self.distribution.run command command) File home rgommers Code pythons lib python3.4 distutils dist.py , line 974, in run command cmd obj.run ) File home rgommers Code numpy numpy distutils command build src.py , line 153, in run self.build sources ) File home rgommers Code numpy numpy distutils command build src.py , line 170, in build sources self.build extension sources ext) File home rgommers Code numpy numpy distutils command build src.py , line 335, in build extension sources sources = self.f2py sources sources, ext) File home rgommers Code numpy numpy distutils command build src.py , line 594, in f2py sources [ m , ext name] f sources) File home rgommers Code numpy numpy f2py f2py2e.py , line 353, in run main imp.reload crackfortran) File home rgommers Code pythons lib python3.4 imp.py , line 315, in reload return importlib.reload module) File home rgommers Code pythons lib python3.4 importlib init .py , line 161, in reload methods.exec module) File <frozen importlib.
|
|
bootstrap> , line 1116, in exec AttributeError NoneType object has no attribute name","bug"
|
|
"6013","414","1.7.0b1 Unittest errors building on Debian mipsel.
|
|
Hello, when building 1.7.0b1 on Debian unstable mipsel the tests suite returns some error ``` running tests for 2.7 plain ...........................................................................................................................................................................................................................................................................................................S.....................E..E....................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K.....E..........E.............E.....E.EEE....E....E..................................................................................................................................SSS...................................................................................................................................................................................................................................................................................E...........................K................................................................EEE............E.............EEEEK......................K............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E....................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test searchsorted test multiarray.TestMethods) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test multiarray.py , line 791, in test searchsorted a.real = [0, 0, 1, 1, 0, 1, np.nan, np.nan, np.nan] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test sort test multiarray.TestMethods) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test multiarray.py , line 484, in test sort a.real = [np.nan, np.nan, np.nan, 1, 0, 1, 1, 0, 0] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test numeric.TestIsclose.test equal nan Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1265, in test equal nan assert array equal isclose nan, nan, equal nan=True), [True]) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in multiply) Traceback most recent call last) File usr lib python2.7 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1234, in test ip isclose yield assert array equal, isclose x, y), result) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [0], [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose 0, [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip none isclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1241, in tst none isclose assert not any isclose x, y)), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test masked arrays Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1274, in test masked arrays assert type x) == type isclose inf, x))) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 3648, in mul return multiply self, other) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 950, in call result = m da RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test nan any test umath.TestArctan2SpecialValues) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 411, in test nan any assert arctan2 isnan np.nan, np.inf) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 328, in assert arctan2 isnan assert np.isnan ncu.arctan2 x, y)), arctan s, s) is s, not nan x, y, ncu.arctan2 x, y))) RuntimeWarning invalid value encountered in arctan2 ====================================================================== ERROR test umath.test nextafter Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1123, in test nextafter return test nextafter np.float64) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterf Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1126, in test nextafterf return test nextafter np.float32) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterl Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1131, in test nextafterl return test nextafter np.longdouble) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in absolute) Traceback most recent call last) File usr lib python2.7 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 482, in test cabs inf nan ref = g x[i], y[i]) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 477, in g return np.abs np.complex a, b)) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, 0, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, 0, nan, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, inf, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, inf, nan, nan, False) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR Ticket 396 Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib tests test regression.py , line 51, in test poly1d nan roots self.assertRaises np.linalg.LinAlgError,getattr,p, r ) File usr lib python2.7 unittest case.py , line 476, in assertRaises callableObj args, kwargs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib polynomial.py , line 1200, in getattr return roots self.coeffs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy lib polynomial.py , line 224, in roots A[0, ] = p[1 ] p[0] RuntimeWarning invalid value encountered in divide ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4785 tests in 217.077s FAILED KNOWNFAIL=5, SKIP=5, errors=24) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy Python version 2.7.3 default, Aug 26 2012, 16 59 14) [GCC 4.6.3] nose version 1.1.2 running tests for 2.7 debug ...........................................................................................................................................................................................................................................................................................................S.............................................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K........................................E.E..............................................................................................................................................SSS...............................................................................................................................................................................................................................................................................................................K.................................................E..................................................K......................K.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 1986, in allclose return all less equal abs x y), atol rtol abs y))) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.7 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core numeric.py , line 1986, in allclose return all less equal abs x y), atol rtol abs y))) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test power zero test umath.TestPower) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy core tests test umath.py , line 139, in test power zero assert complex equal np.power zero, 0 1j), cnan) RuntimeWarning invalid value encountered in power ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4792 tests in 511.190s FAILED KNOWNFAIL=5, SKIP=5, errors=6) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.7 dist packages numpy Python version 2.7.3 default, Aug 26 2012, 14 11 05) [GCC 4.6.3] nose version 1.1.2 [1518408 refs] running tests for 2.6 plain ...........................................................................................................................................................................................................................................................................................................S.....................E..E....................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K.....E..........E.............E.....E.EEE....E....E..................................................................................................................................SSS...................................................................................................................................................................................................................................................................................E...........................K................................................................EEE............E.............EEEEK......................K............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E....................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test searchsorted test multiarray.TestMethods) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test multiarray.py , line 791, in test searchsorted a.real = [0, 0, 1, 1, 0, 1, np.nan, np.nan, np.nan] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test sort test multiarray.TestMethods) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test multiarray.py , line 484, in test sort a.real = [np.nan, np.nan, np.nan, 1, 0, 1, 1, 0, 0] RuntimeWarning invalid value encountered in add ====================================================================== ERROR test numeric.TestIsclose.test equal nan Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1265, in test equal nan assert array equal isclose nan, nan, equal nan=True), [True]) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in multiply) Traceback most recent call last) File usr lib python2.6 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1234, in test ip isclose yield assert array equal, isclose x, y), result) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose nan, [nan, nan, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [0], [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose 0, [1e 08, inf, inf, nan]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1245, in tst isclose allclose assert array equal isclose x, y).all ), allclose x, y), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test ip none isclose [nan, 0], [nan, inf]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1241, in tst none isclose assert not any isclose x, y)), msg x, y)) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2069, in isclose x = x ones like cond) RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test numeric.TestIsclose.test masked arrays Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test numeric.py , line 1274, in test masked arrays assert type x) == type isclose inf, x))) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core numeric.py , line 2070, in isclose y = y ones like cond) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 3648, in mul return multiply self, other) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 950, in call result = m da RuntimeWarning invalid value encountered in multiply ====================================================================== ERROR test nan any test umath.TestArctan2SpecialValues) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 411, in test nan any assert arctan2 isnan np.nan, np.inf) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 328, in assert arctan2 isnan assert np.isnan ncu.arctan2 x, y)), arctan s, s) is s, not nan x, y, ncu.arctan2 x, y))) RuntimeWarning invalid value encountered in arctan2 ====================================================================== ERROR test umath.test nextafter Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1123, in test nextafter return test nextafter np.float64) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterf Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1126, in test nextafterf return test nextafter np.float32) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR test umath.test nextafterl Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing decorators.py , line 215, in knownfailer return f args, kwargs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1131, in test nextafterl return test nextafter np.longdouble) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath.py , line 1118, in test nextafter assert np.isnan np.nextafter np.nan, one))) RuntimeWarning invalid value encountered in nextafter ====================================================================== ERROR Failure RuntimeWarning invalid value encountered in absolute) Traceback most recent call last) File usr lib python2.6 dist packages nose loader.py , line 264, in generate for test in g ) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 482, in test cabs inf nan ref = g x[i], y[i]) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 477, in g return np.abs np.complex a, b)) RuntimeWarning invalid value encountered in absolute ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, 0, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, 0, nan, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, nan, inf, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR test umath complex.TestCarg.test special values <ufunc arg >, inf, nan, nan, False) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy core tests test umath complex.py , line 556, in check real value assert almost equal f z1), x) RuntimeWarning invalid value encountered in arg ====================================================================== ERROR Ticket 396 Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib tests test regression.py , line 51, in test poly1d nan roots self.assertRaises np.linalg.LinAlgError,getattr,p, r ) File usr lib python2.6 unittest.py , line 336, in failUnlessRaises callableObj args, kwargs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib polynomial.py , line 1200, in getattr return roots self.coeffs) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy lib polynomial.py , line 224, in roots A[0, ] = p[1 ] p[0] RuntimeWarning invalid value encountered in divide ====================================================================== ERROR Test a special case for var Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma tests test core.py , line 2735, in test varstd specialcases = method out=nout) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 4778, in std dvar = sqrt dvar) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 849, in call m |= self.domain d) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy ma core.py , line 801, in call return umath.less x, self.critical value) RuntimeWarning invalid value encountered in less ====================================================================== ERROR test nan array test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing tests test utils.py , line 264, in test nan array self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars ====================================================================== ERROR test nan items test utils.TestApproxEqual) Traceback most recent call last) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing tests test utils.py , line 276, in test nan items self.
|
|
assert func anan, anan) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy testing utils.py , line 540, in assert approx equal sc desired = desired scale RuntimeWarning invalid value encountered in double scalars Ran 4785 tests in 197.979s FAILED KNOWNFAIL=5, SKIP=5, errors=24) Running unit tests for numpy NumPy version 1.7.0b1 NumPy is installed in build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib python2.6 dist packages numpy Python version 2.6.8 unknown, Jun 26 2012, 07 20 45) [GCC 4.6.3] nose version 1.1.2 running tests for 2.6 debug ...........................................................................................................................................................................................................................................................................................................S.............................................................................................................................S....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................K........................................E.E..............................................................................................................................................SSS...............................................................................................................................................................................................................................................................................................................K.................................................E..................................................K......................K.................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................E..........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................EE..............K.................................................... ====================================================================== ERROR test numeric.TestIsclose.test ip isclose allclose [1e 08, 1, 1000020.0000000099], [0, nan, 1000000.0]) Traceback most recent call last) File usr lib python2.6 dist packages nose case.py , line 197, in runTest self.test self.arg) File build buildd python numpy 1.7.0 b1 1 mipsel VxuHJc python numpy 1.7.0 b1 debian tmp usr lib py","bug"
|
|
"6013","4143","Integer Division Discrepancy between truediv and itruediv.
|
|
When dividing arrays of integers, the operator truediv) performs floating point division, while the = operator itruediv) does not.
|
|
.
|
|
This seems like a bug to me.
|
|
Is this intended?","bug"
|
|
"6013","4145","BUG Incorrect shape of broadcast result with the exponentiation operator ` `..
|
|
I haven t explored the full range of inputs that produce the error, but when `x` has shape ` n,)` and `y` has shape ` 1,1)`, the result of `x y` does not have the correct broadcast) shape.
|
|
It should be ` 1, n)`, but in all the numpy versions I ve tried 1.6.1, 1.7.1, 1.8.0, and 1.9.0.dev e7fe68a), the result has shape ` n,)`.
|
|
Also, in numpy 1.8.0 and 1.9.0.dev e7fe68a, the operation generates an unexpected `DeprecationWarning`.
|
|
The following shows an example.
|
|
Python was run with the command `python W always`.
|
|
.","bug"
|
|
"6013","4147","Enhance rank correlation for numpy.corrcoef and numpy.cov.
|
|
Can rank correlation and covariance be added to numpy.corrcoef and numpy.cov as an enhancement?
|
|
Rank correlation is used on non parametric data to measure significance.
|
|
We can add a parameter such as .
|
|
with parameters for pearsons default), spearman s roh, and kendall s tau.
|
|
Rank correlation and Covariance is available in R http stat.ethz.ch R manual R patched library stats html cor.html pearson","enhancement"
|
|
"6013","4148","numpy.cast does not work with struct dtypes.
|
|
Tested on ``numpy 1.8``.
|
|
If I have an array with a struct dtype, and attempt to cast it to the same dtype import numpy dtype = numpy.dtype dict names=[ val1 , pad ], formats=[numpy.int8, numpy.int8])) dtype = numpy.float32 arr = numpy.ones 5, 6), dtype) print numpy.cast[dtype] arr)) I get a cryptic error python 3.3.2) > python tt.py Traceback most recent call last) File tt.py , line 9, in <module> print numpy.cast[dtype] arr)) File Users bogdan .pyenv versions 3.3.2 lib python3.3 site packages numpy core numerictypes.py , line 866, in <lambda> cast[key] = lambda x, k=key array x, copy=False).astype k) TypeError expected an object with a buffer interface python 2.7.5) > python tt.py Traceback most recent call last) File tt.py , line 9, in <module> print numpy.cast[dtype] arr)) File Users bogdan .pyenv versions 2.7.5 lib python2.7 site packages numpy core numerictypes.py , line 866, in <lambda> cast[key] = lambda x, k=key array x, copy=False).astype k) TypeError expected a readable buffer object","bug"
|
|
"6013","415","1.7.0b1 Unittest errors building on Debian hurd i386.
|
|
Hello, when building 1.7.0b1 on Debian unstable hurd i386 the tests suite returns some error .
|
|
Full build log is available at https buildd.debian.org status fetch.php?pkg=python numpy arch=hurd i386 ver=1 3A1.7.0 b1 1 stamp=1346549189 Regards, Sandro","bug"
|
|
"6013","4150","triu almost broadcasts as I would expect.
|
|
It is documented to work only for `ndim=2` arrays, but in 1.8.0 it also works as you would expect for N, N, N) arrays it is treated as N 2d arrays each of which are triu d).
|
|
But it doesn t work for M, N, N) arrays where M != N, because the broadcasting is wrong so it raises an exception.
|
|
It would work like I d expect if the `shape[0], shape[1]` [here] https github.com numpy numpy blob master numpy lib twodim base.py L461) were changed to `shape[ 2], shape[ 1]`, although it is technically not wrong in its current form because it is not documented to work with `ndim>2` arrays.
|
|
I m not sure how this interacts with the idea of generalized ufuncs.
|
|
If `triu` were a generalized ufunc linear algebra function then I think it would be expected to behave in the way I mentioned.","bug"
|
|
"6013","4151","r l shift on uint64 scalar and python int failing.
|
|
64 bit ubuntu np.array 123, dtype= uint64 ) >> np.array 1, dtype= uint64 ) 61 np.array 123, dtype= uint64 ) >> 0 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError ufunc right shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe np.version.version 1.8.0.dev 4a9d23b platform.architecture ) 64bit , ELF )","bug"
|
|
"6013","4154","test error on linux Debian GNU Linux 7.3 wheezy).
|
|
python version 2.7.6 numpy version 1.8.0 ERROR test zeros test multiarray.TestCreation) Traceback most recent call last) File usr local lib python2.7 site packages numpy core tests test multiarray.py , line 425, in test zeros d = np.zeros 30 1024 2,), dtype=dt) MemoryError FAILED KNOWNFAIL=5, SKIP=6, errors=1) <nose.result.TextTestResult run=5009 errors=1 failures=0>","bug"
|
|
"6013","4155","`frompyfunc` requires `dtype=object` when used with `accumulate`.
|
|
np.frompyfunc lambda x,y x y,2,1).accumulate np.arange 10)) raises an error ValueError could not find a matching type for <lambda> vectorized).accumulate, requested type has type code l `accumulate` only works if the `dtype` is set explicitly to `object` either in the input, the `out` array, or as a `kwarg`).
|
|
np.frompyfunc lambda x,y x y,2,1).accumulate np.arange 10,dtype=object)) This behavior may be consistent with the note in its `doc` that `The returned ufunc always returns PyObject arrays.`.
|
|
Also when `vectorize` uses a `ufunc` generated by this function, it converts the args to object type, and then converts the output to a specified or deduced) `otype`.
|
|
However, is this error message required?
|
|
A straight call of the `ufunc` returns an object array without any warning or error message np.frompyfunc lambda x 2 x,1,1) np.arange 4)) array [0, 2, 4, 6], dtype=object) Or is this just documentation issue?
|
|
Does its ` doc ` need an added note?","bug"
|
|
"6013","4156","BUG linalg.norm discards imaginary elements of complex arrays.
|
|
Found this while implementing 4144 `np.linalg.norm` calls `asfarray` in one of its cases, which discards the imaginary part of complex arrays.
|
|
I ve no example to reproduce this, I just spotted it while perusing the code.","bug"
|
|
"6013","416","regenerate mtrand.c with Cython 0.17. cgohlke reported Another task for Python 3.3 regenerate numpy random mtrand mtrand.c with Cython 0.17.","bug"
|
|
"6013","4160","BUG numpy ufunc passes on only a single output argument.
|
|
I m trying to use ` numpy ufunc ` for the `Quantity` subclass in astropy which adds and propagates units associated with the array), and while mostly it is a joy compared to the previous ` array prepare ` etc tangle, for functions with two outputs there is a problem, in that only one output is passed on among the keyword arguments.
|
|
This can be seen from the following example .","bug"
|
|
"6013","4162","Add an ASCII String Type for Python 3 Compatibility.
|
|
The transition towards split string bytes types in Python 3 has the unfortunate side effect of breaking the following snippet .
|
|
which returns .
|
|
on Python 2 and .
|
|
on Python 3.
|
|
To aid the porting of Python 2 applications it would be nice if there was an ASCII string data type |A , perhaps?)
|
|
which will always give a string rather than a bytes array, with the restriction that only strings which can be encoded as ASCII can be assigned.","bug"
|
|
"6013","4164","Extend numpy ufunc to most numpy.core.multiarray functions?.
|
|
The ` numpy ufunc ` mechanism likely is easily extended to all relevant numpy.core.multiarray functions.
|
|
This would also enable several pure Python Numpy functions e.g.
|
|
hstack, vstack, ...) to work transparently with 3rd party classes as long as array asarray are not called, things should work out transparently.
|
|
It might also make sense to add a helper function to multiarraymodule so that the override mechanism can be invoked from Python side.
|
|
This could be a sensible alternative e.g.
|
|
in the case of linear algebra functions.","enhancement"
|
|
"6013","4181","The docstring of np.random.pareto Version 1.8.0) contains multiple critical errors.
|
|
The docstring of np.random.pareto Version 1.8.0) contains multiple errors.
|
|
In a nutshell np.random.pareto does not draw samples from the Pareto distribution shown under Notes in the docstring.
|
|
Instead it draws samples from a Lomax distribution also stated in the docstring).
|
|
However, the description in the docstring of how to convert those Lomax random variates to a Pareto distribution is wrong.
|
|
Further, the examples are faulty.
|
|
More detailed The function np.random.pareto draws random variates from a Lomax distribution with shape a and location m =1.
|
|
To convert those samples to a classical Pareto distribution with shape a and location m you have to add 1 and multiply by m, instead of adding m as stated in the docstring.
|
|
More specifically, the docstring says The Lomax or Pareto II distribution is a shifted Pareto distribution.
|
|
The classical Pareto distribution can be obtained from the Lomax distribution by adding the location parameter m, see below.
|
|
Instead, it should read [..] by adding 1 and multiplying my m, see below.
|
|
The example at the bottom therefore should read >>> a, m = 3., 1. shape and mode >>> s = np.random.pareto a, 1000) 1) m Maybe an example with m=10 makes it clearer >>> a, m = 3., 10. shape and mode >>> s = np.random.pareto a, 1000) 1) m Additionally, calling m the location parameter could be misleading.
|
|
For simple Pareto Type I) distributions it is usually referred to as the x min or the scale parameter of the distribution.
|
|
When discussing generalized Pareto distributions m is also called the scale parameter a constant factor) while mu is the location an additive term) [1].
|
|
I assume the misleading naming could have caused some confusion and lead to the errors described above.
|
|
Last but not least I think it might also cause confusion to call a function random.pareto while its meaning is shifted by 1 pareto . )
|
|
[1] http en.wikipedia.org wiki Generalized Pareto distribution","bug"
|
|
"6013","4185","Bug in resize of structured array with initial size = 0).
|
|
I ve tried to resize a record array that was first empty on purpose, I need it) V = np.zeros 0, dtype=[ a , np.float32, 1)]) print V.dtype print np.resize V,2) and I got the following error while it s working for regular array).
|
|
Traceback most recent call last) File test resize.py , line 10, in <module> print np.resize V,2) File usr locaL Cellar python 2.7.6 Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core fromnumeric.py , line 1053, in resize if not Na return mu.zeros new shape, a.dtype.char) TypeError Empty data type Tested on numpy 1.8.0, python 2.7.6, osx 10.9.1 and confirmed by another user.
|
|
Full script V = np.zeros 0, dtype=np.float32) print V.dtype print np.resize V,2) V = np.zeros 0, dtype=[ a , np.float32, 1)]) print V.dtype print np.resize V,2)","bug"
|
|
"6013","4188","numpy.random.choice with probs is 10x slower than equivalent numpy.random.multinomial.
|
|
Using numpy.random.choice to pick an integer between 0 and ``n`` seems the logical way to go.
|
|
It came as a surprise that doing so was so much slower than the less obvious ``np.random.multinomial 1,probs).argmax )``.
|
|
.
|
|
Numpy 1.8","bug"
|
|
"6013","4189","test file position after tofile fails on OSX.
|
|
The following failure is seen on OSX on master 4b4796a), on Python 2.7.
|
|
Also fails at v1.8.0.
|
|
It s unrelated to the recent fixes in the file positioning on Python 3 in fact, the test passes on Python 3.3 on the same system probably because the Python 3 code does some extra steps to seek the file handle).
|
|
So either the test is wrong, and the file handle positioning in this case is not actually guaranteed by C standard, the OSX implementation is faulty, or, there is some subtle bug in some code.
|
|
.","bug"
|
|
"6013","4192","openblas configuration assumes lapack is embedded.
|
|
lapack opt info checks for openblas and stops, this is wrong openblas is not guaranteed to have lapack embedded.
|
|
In linux distributions this is always the case so you end up with these type of errors on import numpy linalg lapack lite.so undefined symbol dpotrf","bug"
|
|
"6013","4195","Unable to compile SciPy Python 3, error in npy 3kcompat.h.
|
|
Hello, I m trying to compile SciPy using the most recent NumPy commit, 1.9.0.dev 0b85f91.
|
|
The file scipy sparse sparsetools csr wrap.cxx includes numpy core include numpy npy py3kcompat.h and this file is throwing an error with both the GCC compiler and the Intel compiler on Ubuntu 12.04.3.
|
|
The Intel compiler says that the return value type does not match the function type on lines 196 and 213 of npy 3kcompat.h, where ` 1` is returned.
|
|
Other places in the same function return `NULL`.
|
|
This was changed in commit 41690fcf9bff30cb3d49d8c147bca1f595bfa70b","bug"
|
|
"6013","4198","Resizing memmap arrays.
|
|
I m referring to this [discussion][1] on stackoverflow.
|
|
I m trying to resize memmap arrays just as it works with ordinary arrays.
|
|
.
|
|
I m not sure about implications and side effects, but for me it looks like, if `mmap` supports resizing, `np.memmap` also should.
|
|
I understand, that it might be hard to guarantee private access for a memmap.
|
|
But if access can be excluded by the programmer, it should at least be possible to override any check by hand.
|
|
And if resizing really shouldn t be possible, at least the [documentation][2] could be clear about this point.
|
|
[1] http stackoverflow.com questions 20932361 resizing numpy memmap arrays [2] http docs.scipy.org doc numpy reference generated numpy.memmap.resize.html","bug"
|
|
"6013","4201","Wrong dtype for unicode field in np.rec.fromarrays ).
|
|
In Python 2 or 3 with Numpy 1.7.1 there seems to be problem with `np.rec.fromarrays` creating a dtype format that is a factor of 4 too large .
|
|
It looks like the problem is [here] https github.com numpy numpy blob 7b2f20b406d27364c812f7a81a9c901afbd3600c numpy core records.py L535), where here `itemsize` is 12 for a 3 character unicode string with UCS 4 encoding .","bug"
|
|
"6013","4203","Adding LDFLAGS to NumPy build stage resulted in undefined reference.
|
|
I m building a new version of NumPy and previously I had always just set LD LIBRARY PATH for runtime execution.
|
|
This time, I wanted to test with some specific optimizations of different BLAS LAPACK implementations.
|
|
So I have 2 independent Python versions, and am using MKL on one and OpenBLAS on the other.
|
|
Python 2.7.6, NumPy 1.8.0.
|
|
Compiling with GCC 4.8.2.
|
|
When I issue the following command LDFLAGS= Wl, rpath, apps OpenBLAS 0.2.8 lib python setup.py build I get an error during a linking of the dotblas.so shared library Undefined reference to main ...
|
|
I noticed that it was attempting to create a shared library, but missing the shared flag so, adding that flag to LDFLAGS variable was a workaround, but it looks redundant for other linking steps.
|
|
fixed the issue when linking the shared library together.
|
|
There is no issue if LDFLAGS is not defined, but I can do repeat the build again if the information would be helpful.
|
|
I think that.
|
|
I m not sure if it is just the dotblas.so part or if other parts were afftect.
|
|
Hopefully this helps a developer, if need more information, let me know.
|
|
Not as familiar with distutils as the compilers themselves.","bug"
|
|
"6013","4209","numpy.irr returns nan when there is a valid answer in the range 1,0).
|
|
Illustrative test case numpy.irr [ 100,0,0,74]) should return 10 as the IRR.
|
|
The implementation of numpy.irr limits the solution range to [0, inf), but IRR is valid over the range 1, inf), so a complete implementation of IRR would return an answer other than nan when there is a valid answer in the range 1,0).
|
|
The current oimplementation treats those cases as nan, rather than providing the correct answer.","bug"
|
|
"6013","421","numpydoc Attributes on classes not displayed when using both properties and normal attributes.
|
|
When documenting at the same time properties and normal attributes, nothing is rendered at all.
|
|
e.g.
|
|
in the example below .
|
|
The docstring is transformed in the following reST docscrape sphinx.py 57, str member list) .
|
|
but this is not correct as a newline is missing after prop1","bug"
|
|
"6013","4216","BUG recfunctions.join by fails if the keys are not in the same order.
|
|
When joining two recarrays and the keys are not in the same order in the two recarrays the join by function fails, because it assumes the key fields to be in the same order.
|
|
Example from numpy.lib.recfunctions import rec join from numpy import recarray a = recarray 1,), [ key1 , int), key2 , int), x , int)]) b = recarray 1,), [ key2 , int), key1 , int), y , int)]) a[0] = 0,1,50) b[0] = 1,0,100) ab = rec join key1 , key2 ), a, b, jointype= inner , r1postfix= , r2postfix= ) This outputs TypeError invalid type promotion because the wrong keys are concatenated.
|
|
The problem is lines 910 911 in recfunctions.py r1k = drop fields r1, [n for n in r1names if n not in key]) r2k = drop fields r2, [n for n in r2names if n not in key]) After this the fields should also be sorted for both arrays.
|
|
Version Numpy 1.8, python 2.7.1","bug"
|
|
"6013","4217","digitize is not the inverse of histogram right limit treated differently).
|
|
I think `digitize` should change its behavior with respect to binning a value that matches the rightmost bin s right edge.
|
|
Such a value should go in the rightmost bin and not be classified as beyond the bounds .
|
|
This would make the output of `histogram` compatible to `digitize` currently it s not).
|
|
Example short explanation) .
|
|
Longish explanation I am assuming that `digitize` generally is considered to do the inverse operation of `histogram`, i.e.
|
|
whereas `histogram` creates bins the edges of the bins) and assigns values to bins, `digitize` quote from `digitize` docs) Returns the indices of the bins to which each value in input array belongs.
|
|
given the data values and the bin edges).
|
|
This reverse indexing behavior has also been referred to in this issue https github.com numpy numpy issues 990.
|
|
It seems like other numerical frameworks provide this functionality, too.
|
|
In the `histogram` specs http docs.scipy.org doc numpy reference generated numpy.histogram.html), the meaning of the bin edges is clarified.
|
|
By default, bins contain the left edge and do not contain the right edge, except for the rightmost bin, whose right edge belongs to the bin The last bin, however, is [3, 4], which includes 4. )
|
|
In contrast, the `digitize` specs state that Each index i returned is such that bins[i 1] <= x < bins[i] .
|
|
There is no comment on the rightmost boundary, so this general statement also applies to it.
|
|
In conclusion, whereas `histogram` always creates the rightmost bin edge in a way that it corresponds to the maximum data value, `digitize` does not count this data value as part of the histogram represented by the bin edges array.
|
|
Restoring compatibility between both functions requires manually incrementing the rightmost bin edge by an epsilon value.
|
|
I think it comes down to the question of what people expect `digitize` to do exactly.
|
|
I would also be fine with always manually correcting one boundary, but am pretty sure that that has not been the original idea behind `digitize`.","bug"
|
|
"6013","4218","Masked record) array makes silent copy numpy 1.8.0).
|
|
Just got to update my sources which were running fine on old numpy 1.6) and start facing the problem on masked record) array.
|
|
It seems that when the array is first addressed by its index, the record it returns is a silent copy of actual row so changing it does not change the array).
|
|
On the other hand, when addressing first the record column and then the row, the change is correctly propagated to the original array.
|
|
See the sample below .","bug"
|
|
"6013","4222","Allow stacking along arbitrary dimensions.
|
|
Currently, there are three functions `hstack`, `vstack` and `dstack` which contatenate along the first three dimensions, even if the arrays are of a lower dimension.
|
|
It should also be possible to do the concatenation on any arbitrary nth dimension, expanding the dimensions of the arguments as required.
|
|
The function `concatenate` does not expand the dimensions automatically.
|
|
This could be done either by changing `concatenate` to allow this or creating a new function `nstack`.
|
|
See [here] http stackoverflow.com questions 19636487 concatenation of numpy arrays of unknown dimension along arbitrary axis) for an example.","enhancement"
|
|
"6013","4224","searchsorted should get a `axis` kwarg.
|
|
Is there a technical reason why `searchsorted` only works with 1d arrays?
|
|
In a current problem, I have an array `v.shape = 61, 160, 320)` and an array `a.shape = 1, 160, 320)`.
|
|
It would be great if I could do `np.searchsorted a, v, axis=0)` to yield a ` 1, 160, 320)` array of the searched indices.","bug"
|
|
"6013","4225","np.log1p inf) returns NaN while it should return inf.
|
|
I am using numpy version 1.8.0 on windows 7 and here is the result for np.log1p In [2] np.log1p np.inf) Out[2] nan In [3] np.
|
|
version Out[3] 1.8.0","bug"
|
|
"6013","4228","BUG ma Inconsistencies of Count with 1D arrays.
|
|
When a masked array of one dimension is given as input with nomask, the count function gives the output as the length of array, no matter what the axis is.
|
|
I consider this to be an error.
|
|
However when the mask= False rather than nomask, the function seems to work ok.
|
|
The function is always consistent for the arrays greater than 1D, irrespective of mask and axis.
|
|
Example >>> a=np.ma.masked array [1, 2, 3]) >>> a.count axis=0) 3 >>> a.count axis=1) 3 >>> a.count axis=2) 3 .... and so on When mask is not nomask >>> a=np.ma.masked array [1, 2, 3],False) >>> a.count 0) array 3) >>> a.count 1) Traceback most recent call last) File <pyshell 137> , line 1, in <module> a.count 1) File usr local lib python2.7 dist packages numpy ma core.py , line 3992, in count n1 = np.size m, axis) File usr local lib python2.7 dist packages numpy core fromnumeric.py , line 2536, in size return a.shape[axis] IndexError tuple index out of range","bug"
|
|
"6013","4231","numpy doc minor item).
|
|
On http www.scipy.org scipylib building macosx.html there is reference to Snow Lion.
|
|
Actually it should be Mountain Lion; Snow is used WRT Snow Leopard.
|
|
Like I said, a nit.","bug"
|
|
"6013","4236","Array operations may give incorrect answer when NumPy is compiled with Intel compilers and linked against MKL.
|
|
Hi there, I m not sure if this is entirely appropriate because I haven t found a good test case.
|
|
I followed the following tutorial to build and install Numpy Scipy http software.intel.com en us articles numpyscipy with intel mkl The only thing I did differently from the tutorial was I specified fcompiler=intelem when building both NumPy and SciPy as opposed to SciPy only.
|
|
I thought it was appropriate but maybe that was my mistake?
|
|
What occurred next was that my array operations e.g.
|
|
dot products) were randomly giving me wrong answers.
|
|
The answer would occasionally be right, but sometimes in 50 of cases) it was off by a factor of 2 or 3; in other cases it would be off by several orders of magnitude.
|
|
The error occurs deep inside my program ForceBalance) and I can t seem to replicate it with a simple script.
|
|
However, I just confirmed that if I reinstall NumPy and SciPy with GCC compilers, the problem goes away.
|
|
Please let me know if you ve seen something like this before.
|
|
Thanks!","bug"
|
|
"6013","424","Mysterious test pareto failure on Travis.
|
|
See here for a discussion http mail.scipy.org pipermail numpy discussion 2012 September 063801.html The failure is caused by Travis switching to 32bit Ubuntu 12.04 recently.
|
|
Some change in Ubuntu breaks it.
|
|
Even this old commit 3882d65c42acf6d5fff8cc9b3f410bb3e49c8af8 still has the failure, so it s probably gcc related.
|
|
The test is probably not robust enough.","bug"
|
|
"6013","4240","ENH allow a duck typed non NDArray sub class to use the iterators.
|
|
needs pandas 0.13 to fail fails with numpy 1.6 thru 1.9 dev) .
|
|
Prior to pandas 0.13, ``Series`` WAS a sub class of ``NDArray``, so this worked as expected.
|
|
``Series`` respects and implements virtually all of the interface to an ``NDArray``.
|
|
This exception is coming from https github.com numpy numpy blob master numpy core src multiarray iterators.c L867 I suspect that the ``PyArray Check obj)`` is not entered because Series is not a direct sub class.
|
|
If this check were relaxed a bit or provide a way to emulate this type of check; which without hooking into the c interface I don t think it possible ATM).
|
|
This was probably done for perf reasons I would guess.
|
|
thanks.
|
|
Jeff related pandas issue https github.com pydata pandas issues 6168","bug"
|
|
"6013","4241","nansum.
|
|
For numpy 1.9 and greater nansum will return zero if the array is empty or full of NaNs.
|
|
However, 0 for the summation of NaNs is just plain wrong.
|
|
This function no longer works correctly.","bug"
|
|
"6013","4242","can t see recfunctions after import numpy .
|
|
can only see recfunctions after import numpy.lib.recfunctions http stackoverflow.com questions 21413947 adding a field to a structured numpy array 3","bug"
|
|
"6013","4245","Building numpy on 3.3 with MSVC 2010 does not work.
|
|
Reported on the distutils sig by Paul Moore.
|
|
.","bug"
|
|
"6013","4249","multiprocessing fails with masked arrays.
|
|
This is the code !
|
|
usr bin env python import numpy as np import multiprocessing a = np.ma.masked equal [0, 0], 0) pool = multiprocessing.Pool 2) pool.apply async np.median, a,)).get ) This is the problem Process PoolWorker 1 Traceback most recent call last) File usr lib python2.6 multiprocessing process.py , line 232, in bootstrap self.run ) File usr lib python2.6 multiprocessing process.py , line 88, in run self.
|
|
target self.
|
|
args, self.
|
|
kwargs) File usr lib python2.6 multiprocessing pool.py , line 71, in worker put job, i, result)) File usr lib python2.6 multiprocessing queues.py , line 366, in put return send obj) File opt cep lofar external lib python site packages numpy ma core.py , line 5473, in reduce self.
|
|
getstate )) File opt cep lofar external lib python site packages numpy ma core.py , line 5446, in getstate self.
|
|
fill value, AttributeError MaskedConstant object has no attribute fill value","bug"
|
|
"6013","4250","std ) corner case behaves differently on different windows pythons.
|
|
xref https github.com pydata pandas issues 6136 a failing test led us to a case where using `.std ddof=1)` on a singleton array returns some infinity on py2.6.6 64 np1.8.0 but returns `nan` on other windows pythons and and in linux at least as far as we know).
|
|
This may just be a facts of life numerical corner case.
|
|
Reporting just in case.
|
|
I have data to reliably reproduce the issue.
|
|
The numpy binary is from cgohlke s collection.","bug"
|
|
"6013","4251","numpy.allclose fails with smallest negative integer as 2nd argument.
|
|
The line https github.com numpy numpy blob 843792b6d349133114930911424772e6bbbc0c9c numpy core numeric.py L2157) .
|
|
abs y) can be negative if y contains the smallest negative integer.
|
|
The following demonstrates the problem .","bug"
|
|
"6013","4254","`PyArray CheckFromAny` with `NPY ARRAY NOTSWAPPED` changes the byte order of single byte types.
|
|
The following example demonstrates what s going on .
|
|
The change is caused by searchsorted calling `PyArray CheckFromAny` with the `NPY ARRAY NOTSWAPPED ` flag set [here] https github.com numpy numpy blob master numpy core src multiarray item selection.c L2109).
|
|
There the dtype gets its byteorder set to `NPY NATIVE`, even if it originally was set to `NPY IGNORE` as is the case with single byte types, see [here] https github.com numpy numpy blob master numpy core src multiarray ctors.c L1842).
|
|
After the byte order of `np.uint8` has been changed `np.core.test )` returns 21 failures, due to mistmatches between the expected and actual encodings of dtypes, e.g.
|
|
.
|
|
I have solved it within 4244 by adding an explicit check for `NPY IGNORE` in `PyArray CheckFromAny`, i.e.
|
|
.
|
|
While it seems to handle this particular case, I am unsure if other cases need to be handled, and whether this should be fixed in a separate PR.","bug"
|
|
"6013","4256","f2py sometimes generates python3 incompatible C wrappings.
|
|
f2py uses deprecated CPython functions notably `PyString FromStringAndSize`) when wrapping callbacks with string parameters.
|
|
An example is the FORTRAN 77 subroutine .
|
|
compiled with f2py c m bug test.f on Arch Linux, this is the python3 version of f2py with the Python wrapping .
|
|
Calling the Python2 version of `f2py` and running the script under Python 2.7 yields the correct answer.
|
|
However, under Python3 we get .
|
|
I tested this under Arch Linux fully updated) with Python 3.3.3 and Python 2.7.6.
|
|
In both cases I used Numpy 1.8.
|
|
I also tested this under Ubuntu Quantal) with Python 3.2.3 and Python 2.7.3.
|
|
In both cases I used Numpy 1.6.2.","bug"
|
|
"6013","4258","DeprecationWarning in Python 3 and Numpy 1.8.
|
|
I invoked `python` with ` Wall`, and following expression raised the `DeprecationWarning`, >>> np.divide .4, 1, dtype=float) main 1 DeprecationWarning Implicitly casting between incompatible kinds.
|
|
In a future numpy release, this will raise an error.
|
|
Use casting= unsafe if this is intentional.
|
|
Actually, this expression is from `scikit learn`, see [this].
|
|
Any idea what caused the warning, and how could we handle it?
|
|
[this] https github.com scikit learn scikit learn blob master sklearn utils fixes.py L211","bug"
|
|
"6013","4259","Support for Cray cross compilers?.
|
|
The error has to do with matching the version string, and should have a simple fix, see below.
|
|
The Cray cross compilers prefer to be invoked as `ftn`, `cc`, or `CC` for the corresponding Fortran, C, and C compilers.
|
|
The GNU compilers are included on all Cray systems, but they are accessed through the generic front ends, not their standard name.
|
|
When the Intel, Portland Group, or PathScale compilers are made available, they are also shared through their front ends.
|
|
I know that NumPy does not have great support for cross compiling, but I m wondering what some first steps toward supporting these compilers might look like.
|
|
A simple choice would be to add the corresponding executables to the list of `possible executables` in `distutils fcompiler .py`.
|
|
Another option would be to inherit the compilers and make them available as `cray gnu`, etc.
|
|
Yet another option would be to create some sort of plugin that lives outside of the NumPy source that injects the Cray compilers in to the NumPy distutils fcompiler module.
|
|
I m happy to submit a PR that implements the first two options or put together the third option), but I was curious what everybody else thinks before I proceed.","bug"
|
|
"6013","426","RuntimeWarning invalid value encountered in power.
|
|
See here http travis ci.org !
|
|
numpy numpy jobs 2337235 .
|
|
This is related to 394.","bug"
|
|
"6013","4260","improve accuracy of logaddexp.reduce.
|
|
scipy has the logsumexp function which avoids unnecessary calls to `log` and thus probably saves a couple ulp of rounding errors.
|
|
The underflow avoidance should already be present in numpys current version.
|
|
I don t see a reason why we couldn t do this directly in logaddexp.reduce.
|
|
The only possible issue is that the reduce loop is blocked to a few thousand elements by nditer so the maximum in the kernel is no global maximum.
|
|
But I don t think that makes a big difference in numeric terms as only the largest terms even contribute to the sum.
|
|
But I would appreciate input of an expert in this regard.
|
|
A prototype implementation is available here https github.com juliantaylor numpy tree logaddexp","enhancement"
|
|
"6013","4263","np.random.uniform documentation incorrect.
|
|
This is in reference to <http docs.scipy.org doc numpy reference generated numpy.random.uniform.html> Consider this code import numpy as np norm min = np.zeros 1) norm max = np.zeros 1) 10 print Input type {0}.
|
|
.format type norm min)) norm init = np.random.uniform low=norm min, high=norm max) print Output type {0}.
|
|
.format type norm init)) print norm init.tolist ) The documentation says the return type from this method is an `ndarray`, however, if the input arrays are single element, the result is a float.
|
|
Either the method is incorrect in returning a float I believe if you give it a single element array, that is what it should be returned), or else the documentation should note this.
|
|
Yes, this can be fixed with norm init = np.random.uniform low=norm min, high=norm max, size=norm min.size) but that does not address the contradiction above.","bug"
|
|
"6013","4264","Segmentation Fault with PyArray Converter in PyArg ParseTuple.
|
|
I m getting a SIGSEGV when I try to create a NumPy array from my arguments in my C extension.
|
|
Here s part of the code I m trying to use.
|
|
I get a seg fault on the PyArg ParseTuple line.
|
|
.
|
|
I followed how PyArray Converter is used in the documentation with the O argument specification.
|
|
I even tried passing just O to PyArg ParseTuple ) and using PyArray Converter later, but I then got a seg fault on the PyArray Converter line.
|
|
I might have the code set up wrong, because this seems like a fundamental construct, but I can t figure out what s wrong.
|
|
Thanks","bug"
|
|
"6013","4266","BUG numpy.histogramdd wrong result for entries very close to last boundary standard numpy.histogram succeeds there).
|
|
Numpy Version 1.8) See the following example for the precision where it happens 1) <code> In [68] numpy.histogramdd [0.9999995], bins=[[0, 0.5, 1]]) Out[68] array [ 1., 0.
|
|
]), [array [ 0. , 0.5, 1. ])])
|
|
< code> As you can see the entry lands in the wrong bin.
|
|
Lowering the value by 10 7 suddenly makes the jump to the last bin which is correct) 2) <code> In [69] numpy.histogramdd [0.9999994], bins=[[0, 0.5, 1]]) Out[69] array [ 0., 1.
|
|
]), [array [ 0. , 0.5, 1. ])])
|
|
< code> Interestingly, increasing the 2nd to last edge value to a high precision also helps 3) <code> In [88] numpy.histogramdd [0.9999999997], bins=[[0, 0.99999, 1]]) Out[88] array [ 0., 1.
|
|
]), [array [ 0. , 0.99999, 1. ])]
|
|
< code> Standard numpy.histogram behaves as expected to arbitrary high values close to 1 4) <code> In [71] numpy.histogram 0.99999999999, bins=[0, 0.5, 1]) Out[71] array [0, 1]), array [ 0. , 0.5, 1. ]))
|
|
< code> The whole issue is only there for the rightmost boundary, not for intermediate ones.
|
|
But as you can see in snippet 3), it has to do with the 2nd to last edge value aswell.
|
|
I hope it gets fixed soon.
|
|
The precision is not that high cosine of small angles can easily be in this regime actually for my work it is) and a wrong bin is populated... if unnoticed wrong results are the outcome.","bug"
|
|
"6013","4268","linalg.svd uses more memory than necessary.
|
|
Hello!
|
|
I just discovered that linalg.svd is using more memory than it needs to both Numpy and Scipy s implementations).
|
|
I haven t had a chance to look through the code but I suspect it has to do with temporarily storing a matrix in memory rather than immediately using the result in computation.
|
|
One knows this is the case because feeding a full dense matrix to scipy.sparse.lingalg.svds does not use even a fraction of the memory used by linalg.avd!
|
|
Anyways, I ran into the issue while working on my Macbook Pro with 16GB and I wanted to compute the SVD for some large matrices. )
|
|
Cheers, Trevor","bug"
|
|
"6013","4269","False warning for masked arrays RuntimeWarning divide by zero encountered in divide.
|
|
If I divide by a masked array with masked zeros I get a one time runtime warning.
|
|
This should not happen, because there is no such division since the zeros are masked print 1 np.ma.masked equal [1, 0], 0) usr bin ipython 1 RuntimeWarning divide by zero encountered in divide !
|
|
usr bin python [1 ]","bug"
|
|
"6013","427","Bug in sorting structured numpy array with more than 2 31 elements.
|
|
When using structured arrays of size more than 2 31, sorting doesn t work.
|
|
The sort function returns immediately without sorting.
|
|
Following is a simple test case.
|
|
.
|
|
The above has been tested using Python 2.7 and Numpy version 1.6.2 as well as the Numpy Git version a72ce7e.
|
|
The test was done on a 64bit linux system with 48G of memory.","bug"
|
|
"6013","4270","np.random.shuffle not work with non default dtypes.
|
|
I display the situation here .
|
|
I get the expected result .
|
|
After shuffle .
|
|
The result is .
|
|
Is this an issue, or `np.random.shuffle` is expected not work with other dtypes?
|
|
I use `numpy 1.7.1` and `python 2.7.5`.","bug"
|
|
"6013","4277","numpy.matrix.trace ) returns another matrix instead of a scalar.
|
|
Maybe there s some good reason for this that I m missing, but wouldn t it be better if `X.trace )` for a matrix `X` returned the trace as a scalar instead of a 1x1 matrix?","bug"
|
|
"6013","4278","matrix wart.
|
|
If the output of this product is a matrix, should it not be a column vector matrix rather than a row vector matrix?
|
|
.
|
|
This seems like a bug, although I do not have extensive knowledge of the behavior of `numpy.matrix` and this has probably already been discussed extensively.","bug"
|
|
"6013","428","numpy.concatenate fails with array size beyon 2 31 .
|
|
UPDATE Sorry for making this bug report., it seems the current git version already fixes this.
|
|
When the size of to the two arrays being concatenated gets bigger than 2 31, numpy.concatenate raises ValueError negative dimensions are not allowed .
|
|
The following is a simple test case.
|
|
.
|
|
Due to this bug numpy.unique also fails for array sizes greater than 2 31.
|
|
This bug has been tested on Python2.7, Numpy version 1.6.2.
|
|
Thee test system uses linux 64bit.","bug"
|
|
"6013","4280","maybe a problem with matrix rank.
|
|
.
|
|
The first and second values look reasonable but the third one looks wrong.
|
|
I think this is caused by feeding lapack a matrix with an `inf` and then returning 0 because lapack doesn t understand `inf`.
|
|
Or maybe this is not considered a numpy bug and the solution is just that I should know better than to try computing `np.linalg.matrix rank` of a matrix with an `inf`.","bug"
|
|
"6013","4283","poly1d indexing and length.
|
|
`poly1d` is instanciated with and internally uses a coefficient list ordered from rank k to rank 0, but reverses this order when the user accesses the coefficients with index notation.
|
|
Its length is defined to be its order, rather than the expected `order 1`.
|
|
While convenient, this is still inconsistent and therefore something which IMHO does not belong into a Python module reminded me of PHP and its array functions argument order madness, actually ).
|
|
It also breaks with some Python functionality which assumes that the default iterator of an indexed object always is an in order iterator and that the largest index is the object s length minus one.
|
|
For example, .
|
|
and of course, anyone who did not read the docs would assume that .
|
|
Changing this would break the API, so this bug is not solvable right now.
|
|
But if you in the long term want to revise any of the APIs, please consider changing this one as well!","bug"
|
|
"6013","4285","Using take ) to do In Place row column permutation results in incorrect behavior.
|
|
.
|
|
Starting with [this SO question] http scicomp.stackexchange.com questions 3229 quickly permute a large square matrix in python numpy), I arrived at take ) to a in place array permutation, however when the call is made with the clip argument to avoid creation of a backup in case of exception) the function can be seen to duplicate rows columns.
|
|
Exact test code demonstrating this issue was provided by [Stefano M] http scicomp.stackexchange.com users 1744 stefano m) from the SO post comment stream, and appears below.
|
|
.
|
|
The output is .
|
|
Also tested in numpy 1.6.2. and 1.7.1, where it seems broken as well.","bug"
|
|
"6013","4291","LinAlgError Array must not contain infs or NaNs.
|
|
I m having some weird behavior.
|
|
In my local environment the same code works as expected, and in my prod environment throws this error .
|
|
This is because a matrix in prod has nans, but the same matrix in local is correct.
|
|
So this makes me think that for some reason numpy is working differently in my prod environment.
|
|
I m running python 2.7.6 using GCC 4.2.1 on osx in my local environment.
|
|
And i m running python 2.7.3 using GCC 4.6.3 on ubuntu 12.04 in my prod environment.
|
|
Does anyone know which may be the difference in the numpy behavior?
|
|
Any help is appreciated!","bug"
|
|
"6013","4293","np.ma.array np.matrix 0)).compressed ) returns a np.matrix.
|
|
According to the [docs] http docs.scipy.org doc numpy reference generated numpy.ma.MaskedArray.compressed.html numpy.ma.MaskedArray.compressed) it should return an ndarray, although technically a `np.matrix` is an `ndarray` in the sense of inheritance so maybe this is OK. see https github.com scipy scipy issues 3324","bug"
|
|
"6013","4295","regression corrcoef doesn t work now in 1.9.0.dev with lists args.
|
|
.","bug"
|
|
"6013","4301","median of np.matrix is broken.
|
|
.","bug"
|
|
"6013","4308","problem with np.partition of np.matrix.
|
|
copypasting a comment from empeeu https github.com numpy numpy issues 4301 issuecomment 35223028 ========================= This actually seems to be a problem with partition.
|
|
.
|
|
median just so happens to use partition as part of its calculation.","bug"
|
|
"6013","4310","fancy indexing does not release GIL.
|
|
this harms parallization of code example for simple boolean indexing .
|
|
.
|
|
this time could be cut in half on a 2 core machine.","bug"
|
|
"6013","4311","Function to chain multiplications `dot` for many arrays.
|
|
I wrote myself a little helper function tat I constantly use.
|
|
The function chains calls of `dot` which makes long matrix multiplication much more readable.
|
|
I call the function `mdot` for multiple dot .
|
|
.
|
|
This is the implementation .
|
|
This is a really simple function but it helps me tremendously.
|
|
Are you interested in integrating something like this?","enhancement"
|
|
"6013","4312","PyArray CumSum Memory leak when using the out parameter.
|
|
Numpy version 1.8.0 Scipy version 0.13.2 When using the out parameter of the PyArray CumSum, it seems to cause a memory leak.
|
|
I might be missing somes Py XINCREF Py XDECREF, if so, I would be glad to know what I m doing wrong.
|
|
Here is a little gist showing the memory leak in question this particular example should not exceed 400Mb of memory leaks) https gist.github.com MarcCote 9053599","bug"
|
|
"6013","4314","Bus error on SPARC for test recfunctions.TestRecFunctions test find duplicates.
|
|
I am getting a bus error on Debian SPARC during testing of the recarray functions.
|
|
On standard Debian squeeze system `uname a`) Linux vagus 2.6.32 5 sparc64 smp 1 SMP Tue Sep 24 00 00 54 UTC 2013 sparc64 GNU Linux Running .
|
|
runtests.py t numpy lib tests test recfunctions.py TestRecFunctions verbose gives Building, see build.log...
|
|
Build OK Running unit tests for numpy NumPy version 1.9.0.dev 297f54b NumPy is installed in home matthew dev trees numpy build testenv lib python2.6 site packages numpy Python version 2.6.6 r266 84292, Dec 26 2010, 23 29 26) [GCC 4.4.5 20100913 prerelease)] nose version 1.3.0 test drop fields test recfunctions.TestRecFunctions) ... ok test find duplicates test recfunctions.TestRecFunctions) ...
|
|
Bus error","bug"
|
|
"6013","4317","Pickling unpickling a dtype resets isbuiltin flag.
|
|
This just happened to me with numpy 1.8.0 from Debian) .","bug"
|
|
"6013","4318","Update indexing docs.
|
|
For example this one is mostly oudated now, even if the basic advanced indexing logic somewhat stayed the same doc source reference internals.code explanations.rst mostly a note for myself)","bug"
|
|
"6013","4320","small boolean assignment incosistencies.
|
|
Usual assignment and now also advanced indexing assignments allow trailing ones in the values array such as .
|
|
However boolean indexing does not have any logic in place to ignore these.
|
|
Also for object assignments view based assignments and now also advanced indexing assignments through a temporary step) have some extra logic to allow sequences .","bug"
|
|
"6013","4323","Protect the wsave array in fftpack lite against concurrent access.
|
|
fftpacklite module.c releases the GIL.
|
|
The .
|
|
array must therefore be protected against concurrent access.
|
|
Only a minor change to fftpack.py is required.
|
|
Basically line 56 must be .
|
|
instead of .
|
|
And finally we need to put it back before returning from .
|
|
.
|
|
Unless we make this change, two Python threads calling numpy.fft.
|
|
functions concurrently can accidentally use the same work array.
|
|
The comment lines should be included so it s clear why a pop instead of an indexing is required.","bug"
|
|
"6013","4326","numpy needs libmath to build.. What I was trying to do I tried to upgrade numpy pip install upgrade numpy ... chaouche build numpy numpy numarray capi.c 198 undefined reference to `log10 collect2 ld returned 1 exit status error Command gcc pthread shared Wl, as needed Wl, no undefined Wl, z,relro Wl, O1 Wl, build id Wl, enable new dtags build temp.linux i686 2.7 numpy numarray capi.o L usr lib Lbuild temp.linux i686 2.7 lpython2.7 o build lib.linux i686 2.7 numpy numarray capi.so failed with exit status 1 Rolling back uninstall of numpy Replacing usr lib python2.7 site packages numpy 1.6.1 py2.7.egg info Command usr bin python c import setuptools; file = home chaouche build numpy setup.py ;exec compile open file ).read ).replace \r\n , \n ), file , exec )) install single version externally managed record tmp pip lppaDW record install record.txt failed with error code 1 Exception information Traceback most recent call last) ... InstallationError Command usr bin python c import setuptools; file = home chaouche build numpy setup.py ;exec compile open file ).read ).replace \r\n , \n ), file , exec )) install single version externally managed record tmp pip lppaDW record install record.txt failed with error code 1 What the problem was I eventually figured out that they didn t link to the math library where all those symbols where defined.
|
|
I found that out by manually trying the gcc command with the lm flag and it compiled fine.
|
|
What my question is What files do I need to edit in order to add this lm flag and install numpy correctly ??
|
|
I can t see any ext modules in setup arguments.","bug"
|
|
"6013","4327","Support of sparse arrays in np.histogram.
|
|
n, bins, patches = ax2.hist SRL[ ,0], 50, normed=1, facecolor= green , alpha=0.75) >>> SRL[ ,0] <4669x1 sparse matrix of type <type numpy.float64 > with 4669 stored elements in Compressed Sparse Row format> raise NotImplementedError adding a nonzero scalar to a NotImplementedError adding a nonzero scalar to a sparse matrix is not supported","bug"
|
|
"6013","4328","REGR indexing issue with numpy scalars on 32 bit linux.
|
|
On numpy master ONLY on 32 linux versions!
|
|
works fine on numpy 1.8.0 found originally here https groups.google.com forum !topic pydata 90QNz8EY74Q pandas issue https github.com pydata pandas issues 6410 .
|
|
.","bug"
|
|
"6013","4332","BUG Masked scalar comparison returns float.
|
|
.
|
|
Moreover, the result has float dtype even if the masked scalar is an int .
|
|
The problem is not present in `oldnumeric` .
|
|
.","bug"
|
|
"6013","4337","Display of 0 D datetime64 arrays is odd.. .","bug"
|
|
"6013","434","ENH CreateSortedStridePerm suboptimal.
|
|
The CreateSortedStridePerm does not work correctly when it comes to non C ordered arrays that include a 1 dimensional axis somewhere in the middle and are not contiguous as this is typically optimized anyways).
|
|
Its a corner case, but since I noticed thought I would post it.
|
|
For these cases it will fail to sort the array and this results unnecessarily slow operations for some things.
|
|
IE `e = np.empty 400,400,400)[ 2, 2, 2]; e[ ,0, ] = 5`) is much faster then `e[ ,0 1, ] = 5`.
|
|
Also this means that `e.copy k )` is not Fortran order, which maybe it never was, but seems not what one would expect.","bug"
|
|
"6013","4340","Running np.lib.test ) causes a failure in np.core.test ).
|
|
With numpy built from master I am seeing the following weird behavior .
|
|
Apparently `lib` is tested after `core`, because running the full test suite once raises no errors, although it will if I launch it again.
|
|
My build is a little shabby MinGW under Windows), and the statement All test pass when I run `np.core.test )` is not fully true, as I always get 2 failures from `test umath.TestComplexFunctions` about precision loss with `sinh`, but this should be unrelated to what I am seeing now.
|
|
Can someone confirm it?","bug"
|
|
"6013","4341","genfromtxt won t unpack if dtype=None.
|
|
Not sure if this is a bug or a feature but if I use f data = np.genfromtxt file.dat , dtype=None, unpack=True) where `file.dat` is a standard data file with several columns ie http pastebin.com ihEW2dxS) the result is not transposed as one would expect having set `unpack=True`.
|
|
For example, the line `f data = np.loadtxt file.dat , unpack=True)` works as expected.","bug"
|
|
"6013","4345","DOC Documentation unavailable at docs.scipy.org and does not compile.
|
|
The numpy documentation hosted at https docs.scipy.org doc numpy user install.html is unavailable.
|
|
This is the host Google returns for all the numpy related searches I ve made.
|
|
I m also unable to compile the numpy documentation for myself due an error from Sphinx; I ve reported that bug to [sphinx dev] https groups.google.com forum !topic sphinx dev NFdR mztURc).
|
|
This error does not occur for the documentation for numpy v1.7.1, only master.","bug"
|
|
"6013","4346","DOC latest doc does not compile.
|
|
I m unable to compile the latest documentation, though v1.7.1 works on my machine.
|
|
I m using OS X 10.9.1, Homebrew provided Python 2.7.6.
|
|
The error I get is a stacktrace from Sphinx, which I ve filed at https groups.google.com forum !topic sphinx dev NFdR mztURc.","bug"
|
|
"6013","4352","numpy any and all applied to object arrays should return booleans..
|
|
Currently in 1.9 devel .","bug"
|
|
"6013","4354","Buggy byte bounds ) in numpy lib utils.py.
|
|
See discussion at 3732.
|
|
It fails for datetime arrays.
|
|
.","bug"
|
|
"6013","4355","ENH Equivalents of newer array constructors are not implemented in ma.
|
|
The following constructors are missing from ma 1. full 2. full like 3. ones like 4. zeros like","enhancement"
|
|
"6013","4356","BUG Many ndarray methods are not specialized for ma.MaskedArray .
|
|
Since ma.MaskedArray inherits from ndarray, all ndarray methods are inherited, but with exception of ufuncs, such inheritance is ill advised and may lead to silent data corruption .
|
|
The list of such methods includes 1. choose 2. dumps 3. dump 4. item 5. partition 6. argpartition 7. fill 8. searchsorted 9. dot I would classify this issue as a bug, but a fix can be as simple as making these methods raise NotImplementedError.","bug"
|
|
"6013","4357","SystemError when printing name of dtype with user defined type object.
|
|
.","bug"
|
|
"6013","4359","BUG Infinite recursion from str.format on a flex array element .
|
|
.","bug"
|
|
"6013","4360","Unexpected resetting of flags when using newaxis.
|
|
When using np.newaxis, the array flags of the new array are set to say that the array is not C or Fortran contiguous, even when it should be.
|
|
For example import numpy as np from numpy.random import rand Make an array that is C contiguous and F contiguous A = rand 10) A2 = A[...,np.newaxis] print A2.flags The flags are shown to be C CONTIGUOUS False F CONTIGUOUS False OWNDATA False WRITEABLE True ALIGNED True UPDATEIFCOPY False Even though the actual memory layout of the array is both C and Fortran contiguous.
|
|
I can verify that this behavior appears on Linux and Windows installations of NumPy 1.8.","bug"
|
|
"6013","4363","DOC Document ma.filled behavior with non scalar fill value .
|
|
The ma.filled function and equivalent MaskedArray method are documented as follows .
|
|
However, the current implementation using np.copyto .., where=mask)) allows any shape that can be broadcast to the shape of self for fill value.
|
|
This is a useful property and should be documented.","bug"
|
|
"6013","4368","distutils mingw32 mess in msvcr versions.
|
|
Hi, I made a simple extension and used numpy.distutils to package it.
|
|
Windows binary was built on wine with mingw32.
|
|
Dependency viewer showed that resulting .pyd depends on kernel32.dll msvcrt.dll libgcc s dw2 1.dll msvcr90.dll when I tried to distribute it libgcc s dw2 1.dll and msvcrt.dll were not found on host system win7).
|
|
libgcc s dw2 1.dll is linked in due to using g as linker even though my code is pure C, setting compiler.linker so[0] = gcc solved this problem.
|
|
msvcrt.dll is used as default runtime library by mingw32, currently numpy.distutils tries to replace it by setting D MSVCRT VERSION =0x0900 and lmsvcr90.
|
|
In my case this creates additional correct) dependency on msvcr90.dll but does not remove msvcrt.dll.
|
|
To achieve correct linking only) to msvcr90.dll, compiler specs need to be modified, since linking to msvcrt.dll happens implicitly in the internals.
|
|
look at the first comment here http www.mingw.org wiki HOWTO Use the GCC specs file comment 106) Please consider the pull request 4378.
|
|
It generates and uses the needed specs modifications.
|
|
.
|
|
.
|
|
.","bug"
|
|
"6013","437","Document Contiguity of result of astype changed intentionalContiguity of result of astype changed.
|
|
See the thread ` [Numpy discussion] Contiguity of result of astype changed intentional?` on the numpy list.
|
|
What has to be done is to get it reviewed by Mark and then submit a PR with the docs change.","bug"
|
|
"6013","4371","Add density keyword to histogramdd and histogram2d..
|
|
This is needed for compatibility with `histogram`, which deprecated `normed` in favor of `density`.","bug"
|
|
"6013","4373","Remove doc cython and doc pyrex from numpy.. Cython does a good job with numpy these days, pyrex is obsolete.","bug"
|
|
"6013","4374","Move doc swig to tools swig..
|
|
There are a few pull requests on doc swig that should go in before this move.","bug"
|
|
"6013","4379","BUG Possibly) Problems assigning complex numbers to array elements.
|
|
I was trying to assign a complex number value to an element in an array, but got some strange errors.
|
|
I posted a question on stackoverflow http stackoverflow.com questions 22016847 assigning complex values to numpy arrays ) and it seems I m not alone in finding the situation a bit confusing.
|
|
Comments with numbers denote points of interest .
|
|
It seems to me like `x` and `y` are both one element arrays, identical except that `y.dtype` is `complex` while is `x.dtype` is `float`.
|
|
Line 2 works just fine, but 3 gives `ValueError setting an array element with a sequence.`.
|
|
I do not understand why `y` would be a sequence if `x` is not.
|
|
From the stackoverflow answer, I m aware that I can work around this strange behaviour by slicing `C[0,0 1] = y`, and this works for both 2 and 3, but I don t understand why it is required in one case but not the other.
|
|
Also, changing 1 to `C = zeros 2,2), dtype=complex)`, which seems like a reasonable thing to do when I want to store complex values in `C`, changes the error given to `TypeError can t convert complex to float`.
|
|
From stackoverflow, it appears this is a known bug?
|
|
I don t know, but I thinks it s worth mentioning that as well.","bug"
|
|
"6013","438","document the change to .base semantics introduced by Add maskna= and ownmaskna= parameters to np.asarray and friends 6c0ad59).
|
|
Per comment https github.com numpy numpy commit 6c0ad59d384606ccf2a0afa20fb9d8a15ddd7255 commitcomment 1819948 this change should be documented and added into the release notes.","bug"
|
|
"6013","4380","weird behavior of np.double np.asarray [3])).
|
|
Is this behavior of `np.double` desired?
|
|
I m wondering about the middle case which I would have expected to be `array [3.
|
|
])` instead of `3.0`.
|
|
I assume there is nothing special about `double` or `3` and that this is probably not a real issue, so it can be closed soon.
|
|
.","bug"
|
|
"6013","4381","BUG not assigning full precision values to longdouble scalars.
|
|
If you try to set a longdouble scalar with a string, it appears to go through Python s normal string parsing and therefore gets converted to a C double rather than a long double In [19] numpy.longdouble 1.12345678901234567890 ) Out[19] 1.1234567890123456912 Note the loss of those last couple digits.
|
|
In [20] numpy.float 1.12345678901234567890 ) Out[20] 1.1234567890123457 In [21] numpy.longdouble 1.12345678901234567890 ) numpy.float 1.12345678901234567890 ) Out[21] 0.0 And so internally they are identical.
|
|
One solution to this is a special case for longdoubles and to use strtold ) to do the conversion.
|
|
Cheers, Scott","bug"
|
|
"6013","4382","parse flags fails with whitespaces in directories.
|
|
Scipy failed to build today when using a self compiled version of numpy, being in a folder that had some whitespaces in it.
|
|
I could track the issue back to numpy s parse flags method, which fails if there are whitespaces in the directory.
|
|
The following 5 lines of python code will allow to reproduce the issue from numpy.distutils.npy pkg config import parse flags d = parse flags L my folder without whitespaces in it ) print d d = parse flags L my folder with whitespaces in it ) print d","bug"
|
|
"6013","4384","int32 64 cast issue.
|
|
This is related to 4328, I think .
|
|
with configuration .","bug"
|
|
"6013","4393","load ) fails with AttributeError Expression object has no attribute getChildNodes .
|
|
I m having some trouble using np.load ).
|
|
It raises an exception each time I try to access one of the stored arrays.
|
|
AttributeError Expression object has no attribute getChildNodes I first noticed this while using the current release 1.8.0) installed via pip install numpy .
|
|
Hoping it was a bug, I checked out the git repository and installed via pip install e .
|
|
I don t think there were any errors, and NumPy generally seems to be working fine.
|
|
<pre> In [1] np.
|
|
version Out[1] 1.9.0.dev f1aab14 In [2] x = np.arange 5) In [3] np.savez compressed test.npz , x=x) In [4] z = np.load test.npz ) In [5] x2 = z[ x ] AttributeError Traceback most recent call last) <ipython input 5 03f06b1ca330> in <module> ) > 1 x2 = z[ x ] home me git numpy numpy lib npyio.pyc in getitem self, key) 246 if magic == format.MAGIC PREFIX 247 bytes = self.zip.open key) > 248 return format.read array bytes) 249 else 250 return self.zip.read key) home me git numpy numpy lib format.pyc in read array fp) 444 msg = only support version 1,0) of file format, not r 445 raise ValueError msg version,)) > 446 shape, fortran order, dtype = read array header 1 0 fp) 447 if len shape) == 0 448 count = 1 home me git numpy numpy lib format.pyc in read array header 1 0 fp) 332 descr dtype.descr 333 try > 334 d = safe eval header) 335 except SyntaxError as e 336 msg = Cannot parse header r\nException r home me git numpy numpy lib utils.pyc in safe eval source) 1137 raise 1138 try > 1139 return walker.visit ast) 1140 except SyntaxError as err 1141 raise home me git numpy numpy lib utils.pyc in visit self, node, kw) 985 cls = node.
|
|
class 986 meth = getattr self, visit cls.
|
|
name , self.default) > 987 return meth node, kw) 988 989 def default self, node, kw) home me git numpy numpy lib utils.pyc in visitExpression self, node, kw) 992 993 def visitExpression self, node, kw) > 994 for child in node.getChildNodes ) 995 return self.visit child, kw) 996 AttributeError Expression object has no attribute getChildNodes < pre> I have a decent amount of test failures 154) but most of them show the same attribute error.
|
|
Full log http pastebin.com BSHZ2tS2","bug"
|
|
"6013","4397","Deprecate ndim > 2 behavior in dot for 1.9.
|
|
Given that numpy is moving toward allowing its linear algebra routines to operate on k, m,n) arrays where the array represents a stack of k, m, n) matrices, it is important to provide a matrix multiply that can operate on such objects without unnecessary contortions.
|
|
The current behavior for np.dot with ndim > 2 doesn t provide this.
|
|
As the first step to providing it, the ndim > 2 behavior should be deprecated in the next release so that it can be changed sooner rather than later.
|
|
I would suggest that this be a 1.9 blocker.","enhancement"
|
|
"6013","4398","Increase NPY MAXARGS to more than 32.
|
|
It is quite frequent for some applications numexpr, but others too) to hit the NPY MAXARGS limit.
|
|
You can find a report about this problem here https github.com PyTables PyTables issues 286 Making this number larger say 256) would alleviate the issue a lot.
|
|
PR 226 tries to tackle the problem, but probably just increasing the value would be enough.
|
|
There has been a recent discussion in the numpy mailing list too http mail.scipy.org pipermail numpy discussion 2014 February 069266.html","enhancement"
|
|
"6013","442","Concatenate with axis=None now requires equal number of array elements.
|
|
Concatenate with axis=None used to work with arrays of any shape <pre> In [51] np.
|
|
version Out[51] 1.6.1 In [52] r4 = range 4) In [53] r3 = range 3) In [54] np.concatenate r4, r3), None) Out[54] array [0, 1, 2, 3, 0, 1, 2]) < pre> but now <pre> In [46] np.
|
|
version Out[46] 1.7.0rc1.dev ea23de8 In [47] np.concatenate r4, r3), None) ValueError Traceback most recent call last) Users mb312 tmp <ipython input 47 e354b8880702> in <module> ) > 1 np.concatenate r4, r3), None) ValueError all the input arrays must have same number of elements < pre> The change requiring the same number of elements appears to have been added explicitly by Mark in commit 9194b3af , in multiarraymodule.c at around line 477 Figure out the final concatenated shape starting from the first array s shape.
|
|
for iarrays = 1; iarrays < narrays; iarrays) { if PyArray SIZE arrays[iarrays]) != shape[1]) { PyErr SetString PyExc ValueError, all the input arrays must have same number of elements ); return NULL; } } Travis agreed this was a regression on the mailing list.","bug"
|
|
"6013","4422","ma.sort ..., endwith=True) broken with nans.
|
|
A masked array containing `nan` currently sorts such that `nan`s appear at the end of the array, violating `endwith=True` .
|
|
Apart from its own contract, this breaks consistency within the current implementation of `np.ma.median` which should probably not rely on `sort` anyway), returning .
|
|
not that I m sure what it means to take a median when there are `nan`s present) Perhaps `minimum fill value` needs an alternative for the `sort` case because of the convention that `nan` comes last.","bug"
|
|
"6013","4425","BUG ufunc in place results in an incorrect mask set.
|
|
Two bugs seem to be present when considering masked arrays) The mask of the resulting array does not update with the following `np.add a, b, out=a)` The behaviour of the ufunc and the object operator don t agree.
|
|
Possibly this is as intended??
|
|
however, this difference of behaviour is not documented as far as I can tell `a b` does not give the same values as `np.add a, b)` one uses the mask to ignore, the other doesn t ignore masked values).
|
|
Not limited to `numpy.add`, those tested and observed as being effected include `add`, `multiply`, `divide`, `multiply`.
|
|
.
|
|
.","bug"
|
|
"6013","4426","fail to instal numpy.
|
|
I have this message when installing numpy, any ideas?
|
|
!
|
|
[numpyfailed] https f.cloud.github.com assets 6840376 2310018 110544bc a2de 11e3 8552 3a3a1363c960.png)","bug"
|
|
"6013","4427","Typos in doc of seterrcall.
|
|
`seterrcall` claims that two arguments are passed to the callback function in call mode) The first is the type of error one of xe2x80x9cdividexe2x80x9d, xe2x80x9coverxe2x80x9d, xe2x80x9cunderxe2x80x9d, or xe2x80x9cinvalidxe2x80x9d), and the second is the status flag.
|
|
The flag is a byte, whose least significant bits indicate the status [0 0 0 0 invalid over under invalid].
|
|
In other words, flags = divide 2 over 4 under 8 invalid.
|
|
1.
|
|
In case of over underflow, the first argument is overflow or underflow instead of over or under ).
|
|
2.
|
|
The flags byte should be [0 0 0 0 invalid overflow underflow divide] ; moreover it is unclear from the docs whether there can be multiple flags set simultaneously and, if this is indeed possible, what is then the value of the first argument?
|
|
).","bug"
|
|
"6013","4436","Masked array filled method ignores fill value .
|
|
.","bug"
|
|
"6013","4438","numpy.linalg.lstsq gives empty residual.
|
|
python version 2.7.6 numpy version 1.8.0 Here is a short example .
|
|
Since `a` is 2x2 matrix of rank 2, I expect the residual to be 0 instead of empty list according to the doc http docs.scipy.org doc numpy reference generated numpy.linalg.lstsq.html).
|
|
But here is what I get .","bug"
|
|
"6013","4439","Function Request.
|
|
Where How do you request an additional function for numpy?","enhancement"
|
|
"6013","4440","Timedelta does not allow assignment from float.
|
|
When trying gh 4434, I ran into a force cast difference between timedelta force casting and timedelta item assignment.
|
|
There is also a discrepancy for `float64` vs. python floats .
|
|
and .
|
|
I could easily hack around this by adding an ellipsis, but this all seems a bit weird...","bug"
|
|
"6013","4441","Got an error message polyfit takes at least 3 arguments 3 given).
|
|
Here is a screen shot of the output.
|
|
http imgur.com Aepim6e Evidently, the error is that I didn t set the degree, but the error message is not informative.","bug"
|
|
"6013","4442","Windows 32 bit Intel MKL segfault in pinv on 1.8.0 from canopy.
|
|
This is a mystery to me.
|
|
Statsmodels had a report of a segfault [1] coming from `np.linalg.pinv` on 32 bit windows from canopy.
|
|
There are no NaNs or infs in the array.
|
|
Files to reproduce https docs.google.com file d 0Bzz ZaP wS HOTJra3ZJM1d6ckk edit Code to reproduce np.linalg.pinv X) This crashes the interpreter.
|
|
We don t have the necessary setup to reproduce, though enthought reportedly has.
|
|
[1] https github.com statsmodels statsmodels issues 1443","bug"
|
|
"6013","4443","Inconsistent behaviour of recarrays.
|
|
Currently 1.9.0.dev and before), in the recarray docs, the first way to convert an existing structured array into a recarray is to make a view.
|
|
However, this lead to different results >>> a = np.recarray 10, dtype=[ field , float)]) >>> type a[0]) numpy.core.records.record >>> b = np.empty 10, dtype=[ field , float)]).view np.recarray) >>> type b[0]) numpy.void In the second case, attribute access for scalars fails >>> a[0].field 6.9120073664173429e 310 >>> b[0].field AttributeError numpy.void object has no attribute field","bug"
|
|
"6013","4444","Recarray indexing returns numpy.void for nested dtypes.
|
|
In the following example NumPy 1.8), recarray indexing return numpy.void >>> dtype = [ name , S10 ), pos , [ x , float), y , float), z , float)])] >>> source = np.recarray 10, dtype=dtype) >>> type source[0]) numpy.core.records.record >>> type source[ 1].pos) numpy.core.records.record >>> type source[0].pos) numpy.void I would have expected numpy.core.records.record instead.As a result, the following would be nice if it worked expression fails >>> source[0].pos.x AttributeError numpy.void object has no attribute x I know that source.pos.x[0] works ok)","bug"
|
|
"6013","4445","Masked array median is incorrect.
|
|
Running numpy 1.8.0 on python 2.7.5, sometimes the median of a masked array is obviously incorrect.
|
|
Example python Python 2.7.5 default, Sep 24 2013, 10 24 28) [GCC 4.4.7 20120313 Red Hat 4.4.7 3)] on linux2 Type help , copyright , credits or license for more information >>> import numpy as np >>> np.version.version 1.8.0 >>> db = np.array [ [7, 1, ], [ 1, 7]], dtype=float) >>> mb = np.ma.masked values db, 1) >>> np.mean mb) 7.0 >>> np.median mb) masked >>> np.min mb) 7.0 >>> np.max mb) 7.0 In some cases the median is returned as a float that is outside the bounds of [min, max].","bug"
|
|
"6013","4452","BUG Masked array addition does not work when dtype=object.
|
|
.","bug"
|
|
"6013","4461","Masked arrays and apply over axes.
|
|
Cannot get apply over axes to work on masked arrays.
|
|
Have tried both np.apply over axes and np.ma.apply over axes.
|
|
>>> x = np.array [1,2,3]) >>> x = np.masked where x==3, x) >>> np.ma.apply over axes np.ma.mean, x, axes= 0)) array [ 2. ])
|
|
This should be 1.5.
|
|
Have tried all combinations of np.ma and regular np functions.
|
|
Can t get it to work.
|
|
Obviously the example given does not require apply over axes, but used it for simplicity.)
|
|
Using numpy version 1.8","bug"
|
|
"6013","4464","ENH Implement matmul function.
|
|
The operator PEP gh 4351) describes two operations that would be a valuable addition to numpy independently of the main purpose of the PEP.
|
|
Most of the functionality is already available in various places `np.dot`, `np.linalg.matrix power`, etc.)
|
|
but no single function conforms to the PEP.","enhancement"
|
|
"6013","4465","BUG ndarray.put disrespects flags.writeable = False.
|
|
It seems that `ndarray.put` disrespects writeable flag unlike other fancy indexing setters .
|
|
At least 1.8.0 is affected, but given that `PyArray PutTo` doesn t include `PyArray FailUnlessWriteable` in [master] https github.com numpy numpy blob master numpy core src multiarray item selection.c L248), I d say `master` is affected too.","bug"
|
|
"6013","4467","Advanced Indexing TODO.
|
|
Some issues for me to fix in advanced indexing may get more) Axis reversing is currently possible in the dummy iterator used only for allocation of the output array.
|
|
But if reversing is used, stride getting won t work.","enhancement"
|
|
"6013","4469","BUG Numpy does not recognize c as PEP 3118 buffer format string.
|
|
One way to demonstrate the bug without writing a C extension is .
|
|
http legacy.python.org dev peps pep 3118 additions to the struct string syntax http docs.python.org 2 library struct.html format characters","bug"
|
|
"6013","4470","BUG np.equal does not work where == does .
|
|
.","bug"
|
|
"6013","4472","Busdaycalendar should accept list of holiday functions.
|
|
[Busdaycalendar] http docs.scipy.org doc numpy reference generated numpy.busdaycalendar.html numpy.busdaycalendar ) currently accepts a list of dates holidays array like of datetime64[D]) to account for public holidays.
|
|
It would be more convenient to be able to specify a list of functions representing a certain holiday.
|
|
Otherwise one has to generate huge lists of holidays see `UsBday.gen cal`) and if a date outside this range is selected, the business calendar will not exclude the public holiday as expected.
|
|
.","bug"
|
|
"6013","4476","Masked array view fails if structured dtype has datetime component.
|
|
A view as `numpy.ma.MaskedArray` fails if the array has a structured dtype, including at least one part that is `datetime64`, as follows .","bug"
|
|
"6013","4478","numpy.ma.core.mvoid.
|
|
repr and str fail on structured arrays with vector members.
|
|
`numpy.ma.core.mvoid.
|
|
repr ` and `.
|
|
str `) are unable to handle structured arrays with vector members.
|
|
See below, tested against the latest git revision .
|
|
It fails only on display .
|
|
so the bug is in `numpy.ma.core.mvoid.
|
|
repr ` and its twin brother `numpy.ma.core.mvoid.
|
|
str `.
|
|
numpy.ma.core.MaskedArray.
|
|
repr has no problem .
|
|
This is against the latest git revision at the time of writing .","bug"
|
|
"6013","448","nditer multi index gives incorrect value for 0 dim arrays.
|
|
From discussion in 445) .
|
|
The multi index here should be ` )`, not ` 0,)`.","bug"
|
|
"6013","4481","Cannot save very large sparse matrix.
|
|
I have a very large matrix the full matrix size 35390104 x 28005, currently I am saving them as two small part files, if I tried to combine them and save them as one large sparse matrix I got the following error.
|
|
The total file size is around 15GB.
|
|
Traceback most recent call last) File home felix test.py , line 58, in <module> main sys.argv[1 ]) File home felix test.py , line 49, in main save npz matrix, tmp combined matrix.npz ) File home felix test.py , line 22, in save npz np.savez output, row=row, col=col, data=data, shape=shape) File usr local lib python2.7 dist packages numpy lib npyio.py , line 530, in savez savez file, args, kwds, False) File usr local lib python2.7 dist packages numpy lib npyio.py , line 592, in savez zip.write tmpfile, arcname=fname) File usr lib python2.7 zipfile.py , line 1102, in write zinfo.file size)) struct.error L format requires 0 <= number <= 4294967295 I am using 1.8.0 version of numpy, is there a work around of the zipfile limitation?","bug"
|
|
"6013","4482","multiply ) and inner ) non commutative for scalar matrix of objects.
|
|
Similar to 2469, ``mutiply`` and ``inner`` are not commutative for scalar matrix of objects.
|
|
``` Python 2.7.5 default, Feb 27 2014, 19 39 55) [GCC 4.8.1] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> arr = np.matrix [1, 2], dtype=object) >>> np.multiply 3, arr) NotImplemented >>> np.multiply arr, 3) matrix [[3, 6]], dtype=object) >>> np.inner 3, arr) NotImplemented >>> np.inner arr, 3) matrix [[3, 6]], dtype=object)","bug"
|
|
"6013","4484","class fromnxfunction does not accept scalars.
|
|
class fromnxfunction.
|
|
call does not accept scalars only ndarrays, lists and tuples), unlike the numpy counterparts.","bug"
|
|
"6013","4485","obscure einsum bug.
|
|
I have ran into an obscure einsum bug.
|
|
Things which I have been able to pinpoint, is that there is an interaction between the mixed datatypes of the arguments, and the small and identical axes sizes.
|
|
Making both types identical, or changing the p axis to 4, means the problem disappears.
|
|
The nature of the error is that in the output of es, es[0,0] is equal to es[1,0], which must be some sort of indexing error in einsum.
|
|
However, all other entries appear to be computed correctly.
|
|
<code> import numpy as np A = np.random.rand 3,3).astype np.float64) B = np.random.rand 3,3,64,64).astype np.float32) es = np.einsum cl,cpxy >lpxy , A,B) tp = np.tensordot A,B, axes= 0,0)) print np.allclose es[0,0], es[1,0]) print np.allclose es, tp) < code> Note I am using numpy 1.8.0, Anaconda python win7 64bit","bug"
|
|
"6013","4486","Core dump and SystemError when casting from and to high precision datetime64 objects, respectively.
|
|
Casting to ` <M8[fs] ` causes a `SystemError`.
|
|
Casting from ` <M8[fs] ` causes a core dump.
|
|
This happens with either `numpy 1.7.1` or with the latest git revision.
|
|
See below .","bug"
|
|
"6013","4489","Make random.multivariate normal check for negative eigenvalues.. See discussion at 3938.
|
|
The current SVD can be used in the check.","bug"
|
|
"6013","4492","Rewrite Pyrex documentation for Cython.
|
|
The current documentation comes from numpybook and is located in `doc source user c info.python as glue.rst`.","bug"
|
|
"6013","4495","void types print differently in python3 and python2.. See test in 388.
|
|
In Python2 .
|
|
In Python3 .
|
|
We need to decide which form is preferable.
|
|
I m leaning towards the second at this point...","bug"
|
|
"6013","4499","DOC fix examples in numpy.polynomial.polynomial module.
|
|
Almost all the doc examples in numpy.polynomial.polynomial begin with >>> import numpy.polynomial as P but it should actually be >>> import numpy.polynomial.polynomial as P","bug"
|
|
"6013","450","structured dtype of the form np.dtype [ name , str ]) are confusing.
|
|
It is relatively intuitive to want to create a structured array with str string for a dtype, which is converted to a S0 type.
|
|
I think this should be detected as an error instead ?","bug"
|
|
"6013","4516","numpy failed to build on ppc64le archi.
|
|
trying to build numpy on fedora for ppc64le archi.
|
|
numpy 1.8.0 5.fc21.src.rpm) the build is failing with following log extract === ... Traceback most recent call last) File setup.py , line 192, in <module> setup package ) ...
|
|
File numpy core setup common.py , line 305, in long double representation raise ValueError Unrecognized format s) saw) ValueError Unrecognized format [ 001 , 043 , 105 , 147 , 211 , 253 , 315 , 357 , 000 , 000 , 000 , 124 , 064 , 157 , 235 , 301 , 000 , 000 , 000 , 000 , 000 , 000 , 000 , 0 === related fedora bugzilla https bugzilla.redhat.com show bug.cgi?id=1078354","bug"
|
|
"6013","4519","numpy.f2py produces wrong results for arrays of strings.
|
|
I have tried several variants with arrays of fortran character strings and none of them did work fixed variable string length and array dimension).
|
|
I even tried to use iso c binding with 2 dimensional character kind=C CHAR) array, but this seemed to produces strided strings row column major reordering of the data).
|
|
I could not really find anything helpful related to this problem.
|
|
In http cens.ioc.ee projects f2py2e README.html there s a remark from 2004 that this should work.
|
|
So here a simple example .
|
|
Obviously calling this function from Fortran works fine.
|
|
I can compile it using > f2p m Fortran c chararraytest.f90 But with my small test program I obtain strange results .
|
|
I get the following output .
|
|
When I use fixed character length strings, e.g.
|
|
character len=20), the results look similar.
|
|
System Debian testing, 64 bit \> uname rvmo 3.13 1 amd64 1 SMP Debian 3.13.5 1 2014 03 04) x86 64 GNU Linux Python 2.7.6 NumPy 1.7.1 \> dpkg list python numpy [...] python numpy 1.7.1 5 amd64 I also tried an installation of numpy 1.8 on another 64 bit Linux machine.","bug"
|
|
"6013","4520","Strange behavior of np.ma.masked where .
|
|
When the mask and array shapes do not match and the mask is always false, no exception is thrown, probably a warning could be good.
|
|
Consider the example x = np.arange 10) y = np.arange 5) z = np.ma.masked where y > 6, x)","bug"
|
|
"6013","4521","numpy.histogram vs. numpy.histogram2d.
|
|
Minor, but `numpy.histogram` deprecates the `normed` parameter and introduces a new `density` parameter.
|
|
`numpy.histogram2d` should use the same interface.
|
|
Currently, I don t have the time to fix this...","bug"
|
|
"6013","4523","Where is the npy format documented?.
|
|
The docstring for [`np.save`] https github.com numpy numpy blob master numpy lib npyio.py L404) says .
|
|
Where s that?","bug"
|
|
"6013","4526","numpy.random.poisson documentation undocumented functionality.
|
|
The numpy.random.possion documentation says that the lam parameter has to be a float.
|
|
However lam can also be an array of floats.
|
|
This allows to quickly generate random poisson distributed numbers with lambdas.
|
|
There is a discussion about this at stackoverflow http stackoverflow.com questions 16134591 python numpy scipy draw poisson random values with different lambda I just stumbled accross this when I was trying to figure out an efficient way to do this, because I also was mislead by the documentation.","bug"
|
|
"6013","4527","Segfault.
|
|
from numpy.core.multiarray import result type result type None) Found it while playing around with a fuzzer.","bug"
|
|
"6013","4528","Segfault bus error.
|
|
Here s a legit one that works on master built using clang on OS X) from numpy.core.multiarray import einsum einsum None, None, None, None)","bug"
|
|
"6013","453","error in financial.npv ?.
|
|
Contrary to what the doc says, the sum is done from t=1 to M. I m not a finance expert, but it seems wrong to me, and I would tend to agree with the doc.
|
|
As for the financial.irr function, it returns the right result.
|
|
Correct code would be .","bug"
|
|
"6013","4530","numpy.random inconsistency of random numbers from other implementations.
|
|
Hi everyone, I would like to point out something that could either be called a feature or bug.
|
|
For a science analysis I want to use the deterministic nature of the Mersenne Twister algorithm in order to produce the same set of random numbers using different languages and libraries.
|
|
I decided to try the Mersenne Twister generator MT19937ar.
|
|
I test three libraries that I think) implement this algorithm, the GNU Scientific Library gsl rng mt19937, gsl rng uniform and ROOT s TRandom3 .Uniform ) in C , and numpy.random.RandomState, numpy.random.random sample ) with numpy.
|
|
Given the same input seed, gsl and ROOT give the same outputs a[i].
|
|
However, numpy gives output b[i]=a[2i], in other words it produces every other number produced by the C .
|
|
It is as if the MT has two state changes in python for every one in gsl and ROOT.
|
|
Here is my python code .
|
|
and my C code .","bug"
|
|
"6013","4537","ma.sum ) does not support keepdims argument.
|
|
.","bug"
|
|
"6013","4539","uninitialized use in clongdouble einsum.
|
|
.
|
|
causes this valgrind error .","bug"
|
|
"6013","4543","in python 3 numpy.savetxt can not save matrices with unicode data.
|
|
numpy.savetxt uses under the hood the numpy.compat.py3k.asbytes utility function.
|
|
In python 3 and above the function forces decoding of the data into latin1 .
|
|
.
|
|
It would be better to allow overriding the default encoding for the function or even default it to utf 8 ).","bug"
|
|
"6013","4546","generic scalar type should inherit from numbers.Number rather than object.
|
|
I was recently trying to cross two streams packages where one was yielding Numpy number types `np.int32`, etc.)
|
|
and another was doing some type checking against Python s [`numbers`] http docs.python.org 2 library numbers.html) ABCs and there was much complaining until I recasted some of the arguments.
|
|
It would be highly convienent if the basic numeric type `numpy.core.numerictypes.generic`, I think) inherited from `numbers.Number` rather than `object`.
|
|
Python s `numbers` was only added in 2.6 though, so if ongoing support of 2.4 2.5 is a priority c est la vie .","bug"
|
|
"6013","4552","error range dtype.
|
|
It is nearly universal in scientific data to have data with error ranges.
|
|
However, there is currently no easy way to deal with data with error ranges in python.
|
|
You generally need to either keep a second array or carry a lot of extra values around.
|
|
There is also no established way to do mathematics on data with error ranges.
|
|
I think it would be useful if python provided dtypes for dealing with data with error ranges.
|
|
It would probably be something like the complex dtype, which could be thought of as holding two floating point values per element.
|
|
There would probably be two dtypes, one which holds two floating point values per element for cases where the error above and below is the same), and a second which holds three floating point values per element for cases where the error above and below is different).","bug"
|
|
"6013","4554","Polynomial.fit ) fails for multi dimensional coefficient arrays.
|
|
`x, y = np.random.rand 100), np.random.rand 100, 25)` `fit = np.polynomial.polynomial.polyfit x, y, 2)` works fine, but `fit = np.polynomial.Polynomial.fit x, y, 2)` fails with `ValueError Coefficient array is not 1 d` from `polyutils.py`.","bug"
|
|
"6013","4559","r [array of objects, ] working when adding scalar but failing with object.
|
|
`r []` works when adding a scalar to an array of objects .
|
|
However, when adding another object, it fails .
|
|
This is unexpected.","bug"
|
|
"6013","456","1.7 deprecations Follow up ticket.
|
|
In 1.9 or thereabouts it might depend on how long it ends up being between releases etc.
|
|
), we should follow through and make the various changes that were warned about deprecated in 1.7.
|
|
These include PyArray Diagonal should begin returning a read only view on the original array and another ticket like this should be filed to remind us to switch this to a read write view at some later date) A similar change needs to happen for structured dtype column subsetting.
|
|
The ufunc casting mode should be switch to same kind by changing NPY DEFAULT ASSIGN CASTING), and the wacky `NPY INTERNAL UNSAFE CASTING BUT WARN UNLESS SAME KIND` casting rule should be removed.
|
|
451) Specifying a custom string formatter with a format array attribute is deprecated not sure what this means, needs looking up) The macros in old defines.h are deprecated and will be removed in the next minor release >= 1.8).
|
|
Disallow passing invalid axis= arguments to concatenate when working with single dimensional arrays 440) Invalid number of bytes in types, such as h100 , see gh 294.","bug"
|
|
"6013","4560","DOC Improve ma.asarray documentation.
|
|
> ma.asarray has an order keyword which is not used, nor available in masked array.
|
|
Also, ma.asanyarray doesn t have an order keyword, but is documented as having one.
|
|
See charris comment at gh 4045.","bug"
|
|
"6013","4563","Inconsistent behavior of ufuncs with 0d arrays.
|
|
0d arrays are a special case for ufuncs.
|
|
Indeed, rather than returning a 0d array, ufuncs return scalars in case of a 0d array inputs.
|
|
For example .
|
|
returns `<type numpy.float64 >`, while .
|
|
returns `<type numpy.ndarray >`.
|
|
Ideally ufuncs should return 0d arrays in case of 0d array inputs, and scalars in case of scalar inputs.
|
|
array>scalar) returns an array in every cases except in the case of a 0d array, which in my opinion breaks the Zen of Python.
|
|
One of the consequences of the behavior was that `x[x>0]` was giving an error in the case of 0d arrays, which was solved by PR 3798 by allowing scalar boolean to be used as an index for 0d array.","bug"
|
|
"6013","4564","BUG .flat on ndarray subclasses does not return subclass items.
|
|
In astropy, we have defined an array subclass `Quantity` that carries units.
|
|
One method that does not always keep subclass information is `.flat` .
|
|
Right now, we are working around this by overriding `.flat` and using our own iterator, similar to what is done in `MaskedArray` with `MaskedIterator`.
|
|
In principle, though, it would seem that the iterator should not just return python float, but perhaps pass through some hook that can be overridden possibly ` array finalize `).","bug"
|
|
"6013","4571","numpy.random.binomial 1,nan) = 9223372036854775807. should be a nan, I figure...","bug"
|
|
"6013","4575","numpy.fromstring unable to parse values > 2 32 on 32 bit arch.
|
|
Steps to reproduce run the following command on a 32 bit x86 python .
|
|
and the incorrect output is .
|
|
If the same command is executed on a 64 bit machine, the output appears correct .
|
|
Versions tested numpy 1.8.1 Python 2.7.3 default, Feb 27 2014, 20 00 17) [GCC 4.6.3] on linux2","bug"
|
|
"6013","4579","Array C API Documentation for PyArray Newshape missing parameter.
|
|
While writing C code to reshape an array, I call the PyArray Newshape function, passing arguments of the types listed in the Array API documentation http docs.scipy.org doc numpy 1.8.0 reference c api.array.html).
|
|
During compile, I get the following error error too few arguments to function After digging around in the source code, I found that instead of the function signature listed in the documentation, PyObject PyArray Newshape PyArrayObject self, PyArray Dims newshape) the source code specifies the following interface PyObject PyArray Newshape PyArrayObject self, PyArray Dims newshape,NPY ORDER order) with NPY ORDER explained at http docs.scipy.org doc numpy 1.6.0 reference c api.array.html NPY ORDER) Once I updated my code to include an appropriate enum value for NPY ORDER, the program compiled just fine.
|
|
I propose that the documentation be updated to include this third parameter.","bug"
|
|
"6013","4581","Allowing setting denormal handling.
|
|
It would be really nice if numpy had a function that could set the DAZ denormals are zero) and the FTZ flush to zero) flags.
|
|
Setting these flags causes a zero to be stored or loaded instead of a denormal number.
|
|
This can dramatically effect the performance, for instance, in a simulation of mine, using float32 instead of float64 causes the run time to be more than twice as long.
|
|
Setting these flags using a tiny custom python extension cuts the time for float32 to 75 of the time used for float64.
|
|
It would be nice to be able to make this trade off without requiring it at all times as setting these flags using compiler options does.","enhancement"
|
|
"6013","4582","numpy.ma.zeros ones empty don t work correctly with keyword arguments.
|
|
The zeros ones empty constructors for numpy.ma don t like to be called with keyword arguments `shape`, `dtype` .","bug"
|
|
"6013","4583","ValueError when initialising structured masked array having datetime64.
|
|
When initialising a structured masked array with `empty`, `zeros`, or `ones`, in the `numpy.ma` module, a `ValueError` is raised, as illustrated below .","bug"
|
|
"6013","4592","Nan are converted to int with slicing.
|
|
Related to 1578, nan s are still converted to maxint when it is assigned by slicing.
|
|
.","bug"
|
|
"6013","4593","BUG 1.9.0 dev unknown on OSX 10.9 error in test callback.TestF77Callback.test string callback.
|
|
I noticed that there is an error in the test `test callback.TestF77Callback.test string callback` when I have installed numpy via homebrew for python 2.x.
|
|
This error shows up for brewed python and for system python.
|
|
I maintain Homebrew python, and especially numpy scipy, so I wanted to assure that 1.9.0 of numpy will work, once released.
|
|
And now I am puzzled about that error message.
|
|
Is numpy actually trying to compile something with gfortran during the tests?
|
|
Note that in homebrew we use `gfortran 4.8.2` at the moment.
|
|
The gist of the output is at https gist.github.com samueljohn 10008974.","bug"
|
|
"6013","4595","ma.masked does not have fill value attribute.
|
|
.
|
|
and as a consequence .","bug"
|
|
"6013","4597","Cannot compile numpy on Mac OSX .
|
|
Hi, I am trying to compile numpy on Mac OSX Mavericks linked to Intel MKL but I get the following error python setup.py config compiler=intelem build clib compiler=intelem build ext compiler=intelem icc O3 g fPIC fp model strict fomit frame pointer openmp xhost shared build temp.macosx 10.6 intel 2.7 numpy core src dummymodule.o Lbuild temp.macosx 10.6 intel 2.7 o build lib.macosx 10.6 intel 2.7 numpy core dummy.so icc command line warning 10006 ignoring unknown option shared Undefined symbols for architecture x86 64 Py InitModule4 64 , referenced from init dummy in dummymodule.o main , referenced from start in crt1.10.6.o ld symbol s) not found for architecture x86 64 icc command line warning 10006 ignoring unknown option shared Undefined symbols for architecture x86 64 Py InitModule4 64 , referenced from init dummy in dummymodule.o main , referenced from start in crt1.10.6.o ld symbol s) not found for architecture x86 64 error Command icc O3 g fPIC fp model strict fomit frame pointer openmp xhost shared build temp.macosx 10.6 intel 2.7 numpy core src dummymodule.o Lbuild temp.macosx 10.6 intel 2.7 o build lib.macosx 10.6 intel 2.7 numpy core dummy.so failed with exit status 1","bug"
|
|
"6013","4598","dstack tup) not equivalent to concatenate tup, axis=2).
|
|
The documentation for dstack says .
|
|
However, in Numpy 1.6.1, i get .","bug"
|
|
"6013","4599","Missing identity for logaddexp.
|
|
Should `logaddexp` and `logaddexp2` have identities defined for reduce?
|
|
In [1] x = np.array [ 1, 1, 1], dtype=float) In [2] np.logaddexp2.reduce x) Out[2] 0.58496250072115619 In [3] np.logaddexp2.reduce x[ 2]) matches np.logaddexp2 x[ 2]) Out[3] 0.0 In [4] np.logaddexp2.reduce x[ 1]) Out[4] 1.0 In [5] np.logaddexp2.reduce x[ 0]) ValueError zero size array to reduction operation logaddexp2 which has no identity I was expecting ` inf` when reducing an empty array.","bug"
|
|
"6013","4600","Bug with NumPy `loadtxt )` and unicode strings.
|
|
Please, refer to this question posted in StackOverflow http stackoverflow.com q 22936790 832621 The OP uses windows and `ISO 8859` text file created by linux with very long lines, with `CRLF` line terminators.
|
|
When reading into NumPy, except the first line which contains labels with special characters, usually only the greek mu) Python 2.7.6, Numpy 1.8.0, this works perfectly data = np.loadtxt input file.txt , skiprows=1) Python 3.4.0, Numpy 1.8.0, gives an error np.loadtxt input file.txt , skiprows=1) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python3.4 site packages numpy lib npyio.py , line 796, in loadtxt next fh) File usr lib python3.4 codecs.py , line 313, in decode result, consumed) = self.
|
|
buffer decode data, self.errors, final) UnicodeDecodeError utf 8 codec can t decode byte 0xb5 in position 4158 invalid start byte It worked with `genfromtxt )`.","bug"
|
|
"6013","4601","trapz returns wrong values when x is out of order.
|
|
Internally, trapz uses np.diff ) to calculate dx.
|
|
This makes the assumption that the x values are increasing monotonically.
|
|
This leads to wrong values when x does not have this property .
|
|
This limitation of trapz should be fixed by sorting both y,x values before proceeding.
|
|
I can sugest two possible courses of action Either properly document trapz, indicating this limitation, and add a new safe trapz that fixes this.
|
|
Or do the exact opposite, have trapz sort the input values, and add a fast trapz that doesn t.","bug"
|
|
"6013","4604","None in a timedelta array is converted to a very small number.. {python} np.array [np.timedelta64 1, s ), np.timedelta64 2, s ), None], dtype=np.timedelta64) results in {python} array [ 1, 2, 9223372036854775808], dtype= timedelta64[s] ) rather than {python} array [ 1, 2, NaN], dtype= timedelta64[s] ) Other dtypes work like this {python} np.array [1, 2, None], dtype=np.float64) >>>array [ 1., 2., nan]) or throw an error {python} np.array [1, 2, None], dtype=np.int64) TypeError Traceback most recent call last) <ipython input 98 e16efa488157> in <module> ) > 1 np.array [1, 2, None], dtype=np.int64) TypeError long ) argument must be a string or a number, not NoneType","bug"
|
|
"6013","4605","Critical issue when subtracting two numpy arrays.
|
|
I ve discovered a rather strange but definitely critical bug when subtracting two numpy arrays with ndim>1 Try the following example <code>x = np.zeros arsize,2)) res = x[arsize 2] res[ ]= 1,2) x =res Expected result a 2D array of length arsize with entries 1, 2) everywhere except at arsize 2, where it should be 0.,0.).
|
|
However, this fails for arsize > 4097 in 2D), because the value of res is updated to 0.,0.)
|
|
BEFORE the whole array is traversed, and thus nothing gets subtracted from the second half!
|
|
Result for arsize>4097 half of the array correctly has the value 1, 2), the rest still has 0.,0.).
|
|
The same is true for 3 dimensions and arsize> 2732, so it happens when the total number of elements is > 8196.
|
|
4D 2049, 1D seems fine for arbitrarily large numbers of arsize.)
|
|
We have been able to reproduce this with numpy versions 1.7.1 and 1.8.1 on MacOSX 10.8.5 and 10.9, as well as version 1.8.1 on Ubuntu.","bug"
|
|
"6013","4607","BUG 1.8.1 openblas not found.
|
|
After back porting 4191 into 1.8.1, openblas is not found on systems by default.
|
|
Backward compatibility logic is wrong.
|
|
.","bug"
|
|
"6013","4608","boolean datatype in numpy array is not shown correctly.
|
|
.
|
|
`np.array [ 1.0, 2.0), 13.0, 77.0), 11.0, 30.0), 5.0, 10.0), 3.0, 4.1)], dtype=[ x , <f8 ), y , bool)])` Generates `array [ 1.0, True), 13.0, True), 11.0, True), 5.0, True), 3.0, True)], dtype=[ x , <f8 ), y , ?
|
|
)])` Where the `bool` type is show by `?` I am using `numpy 1.8.1`","bug"
|
|
"6013","4610","Polynomial.py Optimization.
|
|
polyval in Polynomial.py is about 5 times slower than an implementation based on reduce.
|
|
I would suggest replacing the following lines 672 and 673) for i in range len p)) y = x y p[i] return y with return reduce lambda y, c y x c, p )","bug"
|
|
"6013","4613","Py DECREF causing segfault in arrayobject.c.
|
|
The following code causes a segfault .
|
|
I m using `Version 1 1.8.1 1` from the debian sid repos.
|
|
This worked with older versions of numpy, and I think the culprit is https github.com numpy numpy blob maintenance 1.8.x numpy core src multiarray arrayobject.c L1401 L1413.
|
|
Here is my stack trace .","bug"
|
|
"6013","4616","Function rank ) return bad value.
|
|
Hi, for matrix [[1, 2], [ 1, 2]] rank ) returns 2, but real value is 1.
|
|
It is very big issue for Control Systems Engineers we couldn t detect controlability and observabiliti of MIMO system.","bug"
|
|
"6013","4620","numpy v1.8.1) gradient exception with shapes 1, N) or N,1).
|
|
Hi, When i try to compute 1D gradient of arrays with shapes 1, N) or N,1) an indexerror exception is raised.
|
|
Simple example i have used Python 2.7.6 default, Mar 22 2014, 15 40 47) [GCC 4.8.2] Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> test=numpy.random.rand 5) >>> numpy.gradient test) array [ 0.44777757, 0.16156607, 0.08967521, 0.03061334, 0.36522252]) >>> test=numpy.random.rand 5).reshape 5,1)) >>> numpy.gradient test) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.7 dist packages numpy lib function base.py , line 918, in gradient out[slice1] = f[slice2] f[slice3]) IndexError index 1 is out of bounds for axis 1 with size 1 Regards, G.","bug"
|
|
"6013","4623","link to f2py manual pdf is broken .
|
|
From http docs.scipy.org, the link to the html version of the guide is fine http docs.scipy.org doc numpy dev f2py But the link to the pdf version http docs.scipy.org doc numpy dev f2py 1.8.1.pdf gives a file not found error The requested URL doc numpy dev f2py 1.8.1.pdf was not found on this server.","bug"
|
|
"6013","4624","inconsistent behaviour.
|
|
x = np.array [1, 2]) b = 2 x =float b) produces x = [0, 1] integer division ?!)
|
|
x = x float b) produces x = [0.5, 1]","bug"
|
|
"6013","4625","reading datetimes from an npz saved in python 2.7 fails on python 3.4.
|
|
I have npz files saved in python 2.7, that contain among other things) numpy arrays of datetime objects.
|
|
Reading those back in python 2.7 is no problem.
|
|
When I attempt to do the same thing in python 3.4, I get the following error .
|
|
I vaguely understand the error comes from encoding handling.
|
|
I have read hints on the web the pickle.load call needs to be done with `encoding=latin1`, but I have no idea where to do this, and I worry about breaking things.","bug"
|
|
"6013","4627","``ma.apply along axis`` regression.
|
|
After merging gh 4463, a ``scipy.stats.mstats.trim`` is failing https github.com scipy scipy issues 3554 It looks to me like the expected result for this test is correct, but the mask has one row too many of the mask set to true import numpy as np from scipy import stats x = np.ma.arange 110).reshape 11, 10) x[1] = np.ma.masked trimx = stats.mstats.trim x, 0.1,0.2), relative=True, axis=0) np.testing.assert equal trimx.
|
|
mask.ravel ), [1] 20 [0] 70 [1] 20) trimx = stats.mstats.trim x.T, 0.1,0.2), relative=True, axis= 1) np.testing.assert equal trimx.T.
|
|
mask.ravel ), [1] 20 [0] 70 [1] 20)","bug"
|
|
"6013","4631","BUG polynomial UnboundLocalError local variable quo referenced before assignment.
|
|
An error was reported on stackoverflow http stackoverflow.com questions 23186008 numpy namespace issue).
|
|
My answer, repeated here, describes the problem.
|
|
In the implementation of ` rtruediv ` at https github.com numpy numpy blob master numpy polynomial polybase.py L421, you can see that if `len self.coef)` is not 1, the statement return self.
|
|
class quo, self.domain, self.window) is executed, but `quo` is not defined.
|
|
Here s a simple example that generates the error In [1] import numpy.polynomial.polynomial as npp In [2] p1 = npp.Polynomial [1,2]) In [3] p2 = npp.Polynomial [3,4]) In [4] npp.polydiv p1, p2) UnboundLocalError Traceback most recent call last) <ipython input 4 00c25f56fb20> in <module> ) > 1 npp.polydiv p1, p2) Users warren anaconda lib python2.7 site packages numpy polynomial polynomial.pyc in polydiv c1, c2) 406 len2 = len c2) 407 if len2 == 1 > 408 return c1 c2[ 1], c1[ 1] 0 409 elif len1 < len2 410 return c1[ 1] 0, c1 Users warren anaconda lib python2.7 site packages numpy polynomial polynomial.pyc in rtruediv self, other) UnboundLocalError local variable quo referenced before assignment","bug"
|
|
"6013","4634","Segfault when ctr c during pathological fft.
|
|
On Debian unstable, numpy 1.8.1 built from sources .
|
|
2 16 1 is a size where the fft goes back to O N 2) behaviour prime size).","bug"
|
|
"6013","4636","regression division by int32 0) and by int64 0) are not consistent .
|
|
Numpy 1.8 or 1.8.1, I don t know) changed the behaviour for python float np.int32 0) and np.float value) np.int32 0).
|
|
I am not sure which is the best behaviour 1.7 or 1.8) but I think it should be consistent between int32 and int64.
|
|
Using np.float64 instead of np.float ) does not raise.
|
|
.
|
|
.","bug"
|
|
"6013","4638","BUG overflow on integer ops on 32 bit.
|
|
numpy 1.8.1 32 bit linux .
|
|
64 bit linux .
|
|
So it seems that when the ``int32`` overflows it should then upcast to ``int64`` if needed; interesting that this works on 64 bit.
|
|
related issue https github.com pydata pandas issues 6915","bug"
|
|
"6013","464","1.7.0b2 Unittest errors building on Debian amd64.
|
|
Hello, when building 1.7.0b2 on Debian unstable amd64 the tests suite returns some error .
|
|
That happend even with beta1 but forgot to report for amd64 architecture; it s shown also on the other Debian architectures.
|
|
Regards, Sandro","bug"
|
|
"6013","4642","Segfault when using numpy.searchsorted with recarrays in multithreaded environment.
|
|
recarrays have a dtype.flags of 0 which allows numpy to release the GIL when using the NPY BEGIN THREADS DESCR macro.
|
|
We suspect that PyArray SearchSorted in item selection.c uses VOID compare for comparison of recarray elements which requires Python API functions.
|
|
Python code to reproduce under Windows 7 x64 using Python 2.7.5 x64 and Numpy 1.8.1 or 1.7.1) .
|
|
Thanks and best regards Martin Gfeller martin.gfeller swisscom.com) and Fabio Scala fabio.scala swisscom.com)","bug"
|
|
"6013","4643","Undo more problematic deprecations for 1.9 release.
|
|
It seems that the multiple ellipsis deprecation and the boolean unary and binary ` ` deprecations are causing more trouble in scipy than they are worth.
|
|
So they should be reverted the boolean) or ripped out the ellipsis) from the 1.9.x release.","bug"
|
|
"6013","4645","[abrt] numpy f2py capi maps.py 378 getpydocsign KeyError void .
|
|
Reported in Fedora here https bugzilla.redhat.com show bug.cgi?id=1091994 .","bug"
|
|
"6013","4646","numpy.corrcoef asks for bias ddof.
|
|
Not a real issue, but why does numpy.corrcoef ask for bias ddof?
|
|
I don t think it changes the value anyway.
|
|
The covariance and variance are proportional to 1 N ddof), but since they appear in the same power in the nominator and denominator of the formula for the correlation coefficent they cancel each other.
|
|
Beside of pathologic cases where N < ddof they don t play a role here, afais.
|
|
Why not just calculate the cov with ddof = 0 or may be better N 1?)
|
|
and don t ask for it in corrcoef?","bug"
|
|
"6013","4647","linalg.norm needs a `keepdims` keyword.. Self explanatory.","enhancement"
|
|
"6013","465","np.array does not always return the order asked for.
|
|
Is the behavior below expected?
|
|
It looks like a bug to me <pre> >>> a = np.ones 10,) >>> np.array a, order= C , copy=False, ndmin=2).flags C CONTIGUOUS False F CONTIGUOUS True OWNDATA False WRITEABLE True ALIGNED True UPDATEIFCOPY False >>> np.array a, order= C , copy=True, ndmin=2).flags C CONTIGUOUS False F CONTIGUOUS True OWNDATA False WRITEABLE True ALIGNED True UPDATEIFCOPY False < pre>","bug"
|
|
"6013","4650","add useful array searching functions .
|
|
i have a couple generic functions for searching arrays by value, and creating slices based on value bounds in the scikit rf package.
|
|
are these of any use to numpy?
|
|
if so i can make a PR.
|
|
https github.com scikit rf scikit rf blob 472c1365260d32c99fcc39c8d933d046245b77f3 skrf util.py L73 L141","enhancement"
|
|
"6013","4651","Using too many indices throws SystemError instead of IndexError when second index is 1. .
|
|
That should be an `IndexError` instead.","bug"
|
|
"6013","4652","Memory leak in combination with Theano.
|
|
Hi everyone, I found a memory leak writing code using the [theano] http deeplearning.net software theano ) module and I was able to trace it back to numpy.
|
|
[I already posted on the theano mailing list about it] https groups.google.com forum !topic theano dev 3gI S9ns9oE) Sample code to reproduce the leak Required modules Numpy [commit] https github.com numpy numpy commit 0c9f285f38ea4d143c5e79badd0d36cb808242a6) `0c9f285f38ea4d143c5e79badd0d36cb808242a6`) Scipy [commit] https github.com scipy scipy commit d9a8c214c4cea75f7d240957a646b562cf081f8b) `d9a8c214c4cea75f7d240957a646b562cf081f8b`) Theano [commit] https github.com Theano Theano commit 26d913091e56b3e80e7c562175b856a90dd782e4) `26d913091e56b3e80e7c562175b856a90dd782e4`) for profile decorator memory profiler and psutils I also use OpenBlas .
|
|
For more logs https gist.github.com ogh b6694c8eb7793b454dee With the same configuration except for numpy switched out with a version from last September everything works fine without any leak.
|
|
tried the November version since I was using it on a server where I did not have the memory leak) The version of numpy that works without a leak [commit] https github.com numpy numpy tree 135443768a24ab4fbfd4fa5c8fc40f27d2e25c96)`135443768a24ab4fbfd4fa5c8fc40f27d2e25c96` Can you give me any advice on how to narrow down the problem further?","bug"
|
|
"6013","4653","Memory leaks with np.random.shuffle.
|
|
Hi everyone, I ve found a nasty memory leak while using `np.random.shuffle` import numpy as np profile def shuffle leaktest ) x = np.random.randn 1000, 1000) for ii in xrange 1000) np.random.shuffle x) del x if name == main print numpy version np.
|
|
version shuffle leaktest ) ` python m memory profiler shuffle leaktest.py` numpy version 1.9.0.dev fae89b0 Filename shuffle leaktest.py Line Mem usage Increment Line Contents ================================================ 3 110.535 MiB 0.000 MiB profile 4 def shuffle leaktest ) 5 118.195 MiB 7.660 MiB x = np.random.randn 1000, 1000) 6 337.770 MiB 219.574 MiB for ii in xrange 1000) 7 337.770 MiB 0.000 MiB np.random.shuffle x) 8 337.770 MiB 0.000 MiB del x Using `git bisect` I ve tracked down the first instance of memory leakage to commit 607863d7387e80dab4b064856b598a8c86e9bee4.
|
|
No leak numpy version 1.9.0.dev c7a30d5 Filename shuffle leaktest.py Line Mem usage Increment Line Contents ================================================ 3 103.633 MiB 0.000 MiB profile 4 def shuffle leaktest ) 5 111.285 MiB 7.652 MiB x = np.random.randn 1000, 1000) 6 111.305 MiB 0.020 MiB for ii in xrange 1000) 7 111.305 MiB 0.000 MiB np.random.shuffle x) 8 103.672 MiB 7.633 MiB del x Leak numpy version 1.9.0.dev 607863d Filename shuffle leaktest.py Line Mem usage Increment Line Contents ================================================ 3 112.664 MiB 0.000 MiB profile 4 def shuffle leaktest ) 5 120.316 MiB 7.652 MiB x = np.random.randn 1000, 1000) 6 183.031 MiB 62.715 MiB for ii in xrange 1000) 7 183.031 MiB 0.000 MiB np.random.shuffle x) 8 175.398 MiB 7.633 MiB del x However, this leak seems to be much less pronounced than the one I m currently seeing in master.
|
|
It seems that things suddenly got a lot worse with commit 9f8771accdc11a83dc928a99bd0ba48fe7bcca89 Small leak numpy version 1.9.0.dev f57c77b Filename shuffle leaktest.py Line Mem usage Increment Line Contents ================================================ 3 87.219 MiB 0.000 MiB profile 4 def shuffle leaktest ) 5 94.871 MiB 7.652 MiB x = np.random.randn 1000, 1000) 6 157.789 MiB 62.918 MiB for ii in xrange 1000) 7 157.789 MiB 0.000 MiB np.random.shuffle x) 8 150.156 MiB 7.633 MiB del x Big leak numpy version 1.9.0.dev 9f8771a Filename shuffle leaktest.py Line Mem usage Increment Line Contents ================================================ 3 118.914 MiB 0.000 MiB profile 4 def shuffle leaktest ) 5 126.566 MiB 7.652 MiB x = np.random.randn 1000, 1000) 6 346.133 MiB 219.566 MiB for ii in xrange 1000) 7 346.133 MiB 0.000 MiB np.random.shuffle x) 8 346.133 MiB 0.000 MiB del x","bug"
|
|
"6013","4655","numpy.fft.fft rfft etc routines are not threadsafe.
|
|
It appears that numpy.fft.fftpack.
|
|
raw fft is doing internal caching in a non threadsafe manner, causing data corruption if FFTs are done on similarly sized arrays from multiple threads at the same time.
|
|
The following sample code can be used to reproduce the issue .","bug"
|
|
"6013","4657","Maximum recursion depth exceeded error with Intel compiler.
|
|
Trying to build numpy 1.8.1 using Intel compiler results in RuntimeError maximum recursion depth exceeded while calling a Python object Setup Fedora 20 Python 3.3.2 Intel Parallel Studio XE 2013 SP1 Update 2 Compiling takes place in an virutalenv environment site.cfg contains .
|
|
Output of the command python setup.py config compiler=intelem fcompiler=intelem .
|
|
Output of command python setup.py build compiler=intelem fcompiler=intelem .","bug"
|
|
"6013","4658","Bug in behavior of np.random.negative binomial.
|
|
Hi, I haven t delved into the code, but I think the random number generation in `np.random.negative binomial` is screwed.
|
|
A simple test shows that, e.g.
|
|
the numerical mean is not the same as what is analytically expected even for large empirical samples .
|
|
The last two numbers should be the same ).
|
|
See http en.wikipedia.org wiki Negative binomial distribution, cited in the doc string of `random.negative binomial`.","bug"
|
|
"6013","4659","argmin ) for boolean arrays could be much faster, no?.
|
|
With NumPy 1.8.1, `argmin )` on a boolean array seems to go through all the elements of the array an array of 1000000 true values takes the same time as the same array but where the first element is false.
|
|
In principle, the calculation could stop as soon as the first false value is encountered, no?
|
|
Code .
|
|
The potential speed up is big.","bug"
|
|
"6013","4660","Should np.ma.masked be hashable?.
|
|
It seems that in python 2.7 I can do `a = {np.ma.masked AAA }` while in python3 this code throws ` TypeError unhashable type MaskedConstant ` numpy 1.8. in both cases).","bug"
|
|
"6013","4661","array creation from ctypes.c double fails.
|
|
.
|
|
this works but gives you a object array which is probably not what one wants numpy.array [[3,3,3], ctypes.c double 3) ), ctypes.c double 3) )]) both examples worked in 1.6 but not in newer numpy, but also in 1.6 you got object arrays","bug"
|
|
"6013","4663","Switch to routing allocations through Python s allocators on py35.
|
|
Python 3.5 s memory allocator system now exposes calloc http bugs.python.org issue21233 ...which means that if we are building for 3.5, we can route all allocations through Python s memory allocator system, and get memory tracing profiling out of the box.
|
|
So we should do this.","enhancement"
|
|
"6013","4664","Travis tests are busted..
|
|
I suspect there has been a change in the configuration of some of the build machines, but maybe something else is broken.
|
|
The end of the traceback is .
|
|
Where `old install mod` is python s `distutils.command.install`.","bug"
|
|
"6013","467","1.7.0b2 Unittest errors building on Debian mips.
|
|
Hello, when building 1.7.0b2 on Debian unstable mips the tests suite returns some RuntimeWarning .
|
|
Full builld log at https buildd.debian.org status fetch.php?pkg=python numpy arch=mips ver=1 3A1.7.0 b2 1 stamp=1348964750","bug"
|
|
"6013","4672","numpy.random.RandomState cannot handle integers > 20 digits.
|
|
I think I ve stumbled upon a either a coding or documentation bug.
|
|
from numpy.random import RandomState testing state = RandomState 1234567891011121314) Works without problems.
|
|
However adding one more digit breaks with the following error testing state = RandomState 12345678910111213141) ValueError Traceback most recent call last) <ipython input 2 346870f94b6d> in <module> ) > 1 testing state = RandomState 12345678910111213141) Library Frameworks EPD64.framework Versions 7.3 lib python2.7 site packages numpy random mtrand.so in mtrand.RandomState.
|
|
init numpy random mtrand mtrand.c 5217) ) Library Frameworks EPD64.framework Versions 7.3 lib python2.7 site packages numpy random mtrand.so in mtrand.RandomState.seed numpy random mtrand mtrand.c 5468) ) ValueError object of too small depth for desired array From the docs http docs.scipy.org doc numpy reference generated numpy.random.RandomState.html) seed {None, int, array like}, optional Random seed initializing the pseudo random number generator.
|
|
Can be an integer, an array or other sequence) of integers of any length, or None the default).
|
|
numpy.
|
|
version 1.8.0 I think this is probably one where either the any length in the docs should be changed to say < 20 probably simplest) or RandomState to actually take any sequence of integers.","bug"
|
|
"6013","4673","Spurious failures in creating npy common.h.
|
|
When creating npy common.h I see many errors like configtest.c 5 16 error size of array xe2x80x98test arrayxe2x80x99 is negative One source of these is the test for complex compatability in numpy core setup.py complex def = struct { s x; s y;} type, type) res = config cmd.check type size complex def, expected=2 expected[type]) This works better for me res = config cmd.check type size complex def, expected=[2 t for t in expected[type]]) Does this need to be a pull request?","bug"
|
|
"6013","4674","Standardise docstrings in random module.
|
|
As a followup to PR 4641 more standardisation in docstrings in the random module would be nice descriptions for other) arguments of many functions do not include the default value an optional keyword is present in many functions docstrings but not all where it applies returns should also be standardized use the same template) as there are many possibly raises too","enhancement"
|
|
"6013","4675","Test error for scipy 0.15.0.dev 5d197ed.
|
|
The first bad commit is d8fd28389adb491e24b7cdc25cd1b20f539310c3, the isolated error is .","bug"
|
|
"6013","4679","BUG numpy tile does not copy.
|
|
For single repetitions, tile returns the same reference.
|
|
More intuitive behavior would be to return a copy regardless of number of repetitions.
|
|
.","bug"
|
|
"6013","468","1.7.0b2 Unittest errors building on Debian mipsel.
|
|
Hello, when building 1.7.0b2 on Debian unstable mipsel the tests suite returns some RuntimeWarning .
|
|
Full builld log at https buildd.debian.org status fetch.php?pkg=python numpy arch=mipsel ver=1 3A1.7.0 b2 1 stamp=1348958298","bug"
|
|
"6013","4681","Change RELAXED STRIDES to not intentionally mangle them up.
|
|
Currently when RELAXED STRIDES is active, new arrays which have a dimension with size 1 get a stride of `NPY MAX INTP` set to help debugging since this will almost certainly trigger such bugs for any input array with such a shape, and not juts when one happens to have an input array with funny strides).
|
|
In release versions we should probably not do this.
|
|
So for 1.10 I would suggest creating a RELAXED STRIDES=2 setting where we actually still scramble up the stride, but as default with RELAXED STRIDES=1 don t in the release.
|
|
This requires `ctors.h` around line 3718 ` For testing purpose only `) to be changed.","bug"
|
|
"6013","4683","poor performance of multidimensional masked median.
|
|
due to the use of apply over axis masked median operations gets incredibly slow when computing them along a small dimension of a large array .
|
|
a way out is using bottlenecks nanmedian which is on par with our non masked performance and better for the slow axis) but I think we should think of ways to improve this in numpy.
|
|
I can think of following improvements for median implement apply over axis in C, should be relatively easy but will only give us 50 boost at best.
|
|
it should still be worthwhile as there are likely more users of that function out there.
|
|
extend partition to broadcast the index, then one could compute the right index from the mask and do everything in one partition call.
|
|
Probably worthwhile but it will make the partition interface more complicated and would probably imply rewritting our pretty ugly sorting code.
|
|
probably the simplest would be add a check on the size of the axis and use sort instead of partition as currently done in ma.median but then instead of selecting via apply along axis, one selects the elements to average or not) via fancy indexing always select the middle two and if the non masked count is odd duplicate one of the elements so mean ) does nothing) add a MaskedPartition C Api function see if we can integrate bottleneck directly somehow any more options I m might be overlooking?","enhancement"
|
|
"6013","4685","DEPR why are bool comparison on object arrays deprecated now in 1.9?.
|
|
depr warning just started showing up we test against numpy master in pandas master) this seems like a reasonable comparison In fact I don t check ``is`` here on purpose as I only want an element wise comparison and pandas handles the nan issue separately) .
|
|
or should I be masking the ``np.nan`` first?
|
|
e.g.
|
|
we are ultimately doing this .","bug"
|
|
"6013","4686","Extrange behaviour with numpy boolean values.
|
|
.
|
|
Obviously, question is why can t be used is True statement into comparison ?","bug"
|
|
"6013","4689","Importing scipy.ndimage raises RuntimeError on Python 3.
|
|
I noticed problems with [builds] https travis ci.org ufo kit concert jobs 24726108) on Python 3.
|
|
Whenever, `scipy.ndimage` is imported, NumPy raises a `RuntimeError`.
|
|
The problem is related to [these] https github.com numpy numpy blob master numpy import tools.py L114 L117) lines, in which you modify the dictionary while iterating over it.
|
|
The Python [docs] https docs.python.org 2 library stdtypes.html dict.iterkeys) are pretty clear, that this causes undefined behaviour.
|
|
Why this only happens on Python 3 and with this particular package, I have no idea.","bug"
|
|
"6013","469","np.copy should have order= K by default for backward compatibility.
|
|
In current tree <pre> >>> a = np.ones 10, 10)) >>> np.copy a.T).flags C CONTIGUOUS True F CONTIGUOUS False OWNDATA True WRITEABLE True ALIGNED True UPDATEIFCOPY False < pre> In numpy 1.6 <pre> >>> a = np.ones 10, 10)) >>> np.copy a.T).flags C CONTIGUOUS False F CONTIGUOUS True OWNDATA True WRITEABLE True ALIGNED True UPDATEIFCOPY False < pre> I suggest to put order= K by default in numpy 1.7 maybe with a FutureWarning) to ensure backward compat.","bug"
|
|
"6013","4690","Array with 2000 named columns ValueError header does not fit inside 65536 bytes.
|
|
It seems that arrays that have a dtype that contains too many names `[ col0 , float), ]`) cannot be written to file with `numpy.save )` see the traceback in the PS).
|
|
Could this limitation be removed?
|
|
It is really a pity to have users change a lot of code just because they add columns to their array I moved from about 1000 to about 2000 named columns). )
|
|
PS Here is the traceback .","bug"
|
|
"6013","4693","raising numpy even number array by 2.
|
|
Hi, I don t know what is going on but I could see that my numpy is returning a strange result when I try this >>> array [2]) 2 array [0]) >>> array [4]) 2 array [0]) any idea about what is causing this?","bug"
|
|
"6013","4694","Numerical stability.
|
|
It is no numerical stability in the NumPy version1.6.2 but perhaps also in future fersions ).
|
|
<pre> >>> d=array [ 253., 253., 253., 253., 253., 253., 252., 253., 252., 252., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 252., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 252., 251., 254., 254., 254., 252., 252., 253., 253., 254., 253., 253., 253., 254., 253., 252., 253., 254., 253., 252., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 252., 253., 254., 252., 252., 253., 253., 253., 253., 253., 252., 253., 253., 253., 252., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 254., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 254., 253., 253., 253., 253., 252., 252., 253., 253., 253., 253., 254., 253., 253., 253., 253., 253., 253., 253., 253., 253., 253., 252., 253., 253., 253., 253., 252., 253., 253., 253., 253., 252., 253., 253., 252., 252., 252., 252., 252., 253., 252., 253., 252., 253., 253., 252., 253., 253., 252., 252., 252., 252., 252., 253., 253., 253., 252., 252., 251., 252., 252., 252., 252., 252., 252., 253., 252., 252., 252., 252., 252., 252., 253., 253., 253., 252., 252., 252., 253., 253., 252.
|
|
], dtype=float32) >>> d.sum ) len d) == d.mean ) True it is wrong when used huge number of numbers last digits can be lost >>> d 2).mean ) d.mean ) 2 0.072070897236699238 unreal bullshit happens dispersion can t be negative >>> x = 0.
|
|
>>> for i, v in enumerate d, 1) ... x = v x) i ... >>> x 252.90442890442881 >>> d.mean ) 252.9044289044289 x is the numerically stable d.mean ) >>> y = 0.
|
|
>>> for i, v in enumerate d 2, 1) ... y = v y) i ... >>> y 63960.853146853144 >>> d 2).mean ) 63960.578088578091 >>> y x 2 0.20298737785924459 >>> d.std ) 2 0.20298681725988854 at least d.std ) have some stability < pre>","bug"
|
|
"6013","4695","numpy quad and maple.
|
|
I am not quite sure if this is a bug or not.
|
|
I did the following integral using numpy.quad quad lambda h np.exp .5 2334.0090702455204 1936.9610182100055 5 log10 h) 12.5 2132.5498892927189) log10 h) log10 h)),0,inf) and I got the following answer 1.8368139214123403e 126, 3.3631976081491865e 126).
|
|
I tried the integral with maple and mathematica and I got the following 2.643019766 10 127) Could you please check if this is a bug or a problem some where else.
|
|
Cheers Ihab","bug"
|
|
"6013","4696","Change numpy.bitwise and.identity from 1 to None.
|
|
I would normally expect .
|
|
to be .
|
|
The current behaviour with .
|
|
looks very suprising to me .
|
|
I also doubt its usefulness Why use .
|
|
to only process the last bit?
|
|
.
|
|
instead of ., also making it inconsistent with ., see also http isntagramm.com questions 21050875 numpy bitwise and reduce behaving unexpectedly Is it correct that just setting identity to .
|
|
would fix these cases?","bug"
|
|
"6013","4698","numpy 1.9.0.dev 02008f7.win amd64 py2.7 test failures.
|
|
.
|
|
The first two tests also fail on win amd64 py3.4.
|
|
The MaskedArray test failures should be easy to fix by testing for `np.int64` or casting the result in the `count` function to int.","bug"
|
|
"6013","470","mmap objects used in base should expose a filename attribute.
|
|
In numpy <= 1.6 it is possible to inspect arrays to find out if they come from memmapped files and which file <pre> >>> import numpy as np >>> a = np.ones 10, 10)) >>> np.save a.npy , a) >>> b = np.load a.npy , mmap mode= r ) >>> c = np.asarray b) >>> c.base is b True >>> c.base.filename home varoquau a.npy < pre> In 1.7b2, the base of c is now a reference to the Python mmap object, and thus it is now impossible to get back to original filename.
|
|
This may be an important usecase in parallel computing e.g.
|
|
with multiprocessing, or IPython) to control if data is shared between processes or not.
|
|
<pre> >>> import numpy as np >>> a = np.ones 10, 10)) >>> np.save a.npy , a) >>> b = np.load a.npy , mmap mode= r ) >>> c = np.asarray b) >>> c.base <mmap.mmap object at 0x353cae8> < pre> I propose to use as a base in the numpy memmap objects something that keeps tracks of the filename and maybe the offset, although it can be retrieved with a bit of magic).
|
|
The simplest option would probably to simply use a subclass of mmap.mmap.
|
|
I, GaelVaroquaux, volunteer to implement such a patch, if the strategy above is deemed as acceptable.","enhancement"
|
|
"6013","4701","np.asscalar should pass through scalars.
|
|
Suggestion It would be very convenient if ``np.asscalar )`` would pass through scalars.
|
|
The use case is to ensure that an argument is a scalar, regardless what is passed in as long as it behaves as a scalar).
|
|
.
|
|
Presently these fail as the implementation simply looks for the ``.item )`` method, which scalars do not possess .
|
|
.python np.asscalar np.asarray x)) .python def asscalar a) try return a.item ) except AttributeError, e return np.asarray a).item ) ```","bug"
|
|
"6013","4702","inconsistent string formating.
|
|
At string formatting, the behaviour of 1x1 matrices is different, whether one uses `str.format` or ` ` formatting.
|
|
.
|
|
Maybe provide custom formatting for matrices and not just `str ...)`.
|
|
See also, solved issue https github.com numpy numpy issues 2271 for scalars.","bug"
|
|
"6013","4703","segfault in numpy when doing pcolormesh through basemap matplotlib).
|
|
I am running Arch Linux on a Macbook Pro with numpy 1.8.1 and python 3.4, I am running this script http scipy central.org item 75 1 ibcao international bathymetric chart of the arctic ocean class for python and scipy which plots a large map using Basemap.
|
|
I recompiled python and numpy with debugging symbols and I get the below attached backtrace in gdb.
|
|
The offending line in my script is https github.com gauteh ibcao py blob master ibcao.py L151 ``` 0 0x00007ffff74b57a0 in memmove ssse3 back ) from usr lib libc.so.6 1 0x00007ffff4bbdca8 in memmove len=4, src=0x7fffc3bf4888, dest=<optimized out>) at usr include bits string3.h 57 2 strided to contig size4 dst=<optimized out>, dst stride=<optimized out>, src=0x7fffc3bf4888 <error Cannot access memory at address 0x7fffc3bf4888>, src stride=40, N=1161, NPY UNUSED TAGGEDsrc itemsize=<optimized out>, NPY UNUSED TAGGEDdata=0x0) at numpy core src multiarray lowlevel strided loops.c.src 144 3 0x00007ffff4b76219 in PyArray CopyAsFlat dst=0x2a30220, src=0x489, order= 1010874232) at numpy core src multiarray ctors.c 2566 4 0x00007ffff4bf98c1 in PyArray Flatten a=0x154ed40, order=NPY CORDER) at numpy core src multiarray shape.c 1039 5 0x00007ffff4bf99c1 in PyArray Ravel arr=0x154ed40, order=NPY CORDER) at numpy core src multiarray shape.c 1011 6 0x00007ffff4bd7209 in array ravel self=0x154ed40, args=<optimized out>, kwds=<optimized out>) at numpy core src multiarray methods.c 2182 7 0x00007ffff7a5e8b5 in call function oparg=<optimized out>, pp stack=0x7fffffffdd10) at Python ceval.c 4227 8 PyEval EvalFrameEx f=0x1966c98, throwflag=<optimized out>) at Python ceval.c 2829 9 0x00007ffff7a5ff49 in PyEval EvalCodeEx co=0x2a30224, globals=0x1, locals=0x1966e10, args=0x7fffe40f3470, argcount=1161, kws=0x2a30220, kws entry=0x7ffff7f89060, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0) at Python ceval.c 3578 10 0x00007ffff79d1cd6 in function call func=0x7fffec86e9d8, arg=0x7fffe40f3458, kw=0x7fffe3eb1308) at Objects funcobject.c 632 11 0x00007ffff79a9d48 in PyObject Call func=func entry=0x7fffec86e9d8, arg=arg entry=0x7fffe40f3458, kw=kw entry=0x7fffe3eb1308) at Objects abstract.c 2067 12 0x00007ffff7a583e1 in ext do call nk=<optimized out>, na=1, flags=<optimized out>, pp stack=0x7fffffffdfc8, func=0x7fffec86e9d8) at Python ceval.c 4551 13 PyEval EvalFrameEx f=0x13bcbe8, throwflag=<optimized out>) at Python ceval.c 2869 14 0x00007ffff7a5ff49 in PyEval EvalCodeEx co=0x2a30224, co entry=0x7fffec830810, globals=0x0, locals=0x13bcd60, args=0x7ffff7f2d5b8, argcount=1161, argcount entry=3, kws=0x2a30220, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0) at Python ceval.c 3578 15 0x00007ffff7a5d4e5 in fast function nk=0, na=3, n=<optimized out>, pp stack=0x7fffffffe1c0, func=<optimized out>) at Python ceval.c 4334 16 call function oparg=<optimized out>, pp stack=0x7fffffffe1c0) at Python ceval.c 4252 17 PyEval EvalFrameEx f=0x7ffff7f2d438, throwflag=<optimized out>) at Python ceval.c 2829 18 0x00007ffff7a5ff49 in PyEval EvalCodeEx co=0x2a30224, co entry=0x7ffff7ef95d0, globals=0x0, locals=0x7ffff7f2d5b0, args=0x0, argcount=1161, argcount entry=0, kws=0x2a30220, kws entry=0x0, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0) at Python ceval.c 3578 19 0x00007ffff7a5ffeb in PyEval EvalCode co=co entry=0x7ffff7ef95d0, globals=globals entry=0x7ffff7f2a488, locals=locals entry=0x7ffff7f2a488) at Python ceval.c 773 20 0x00007ffff7a7c174 in run mod mod=mod entry=0x6837c0, filename=filename entry=0x7ffff7e1a2f0, globals=globals entry=0x7ffff7f2a488, locals=locals entry=0x7ffff7f2a488, flags=flags entry=0x7fffffffe440, arena=arena entry=0x631980) at Python pythonrun.c 2175 21 0x00007ffff7a7e3a5 in PyRun FileExFlags fp=0x690ea0, filename str=<optimized out>, start=<optimized out>, globals=0x7ffff7f2a488, locals=0x7ffff7f2a488, closeit=1, flags=0x7fffffffe440) at Python pythonrun.c 2128 22 0x00007ffff7a7f363 in PyRun SimpleFileExFlags fp=0x2a30224, filename=0x7ffff7f22440 ibcao.py , closeit=1, flags=0x28) at Python pythonrun.c 1601 23 0x00007ffff7a9543c in run file p cf=<optimized out>, filename=<optimized out>, fp=<optimized out>) at Modules main.c 319 24 Py Main argc= 136207696, argv=0x7ffff7f22420) at Modules main.c 751 25 0x0000000000400af6 in main argc=2, argv=<optimized out>) at .
|
|
Modules python.c 69 xc2xb4xc2xb4xc2xb4","bug"
|
|
"6013","4704","Difference in behavior between python and numpy array add subtract assignment operator.
|
|
Summary There is a difference in how the add subtract assignment operators work between normal Python ints and int64s in Numpy arrays that leads to potentially unexpected and inconsistent results.
|
|
Problem Consider the following code, in which a normal Python int is typecast to a float in a new variable .
|
|
As expected, the type of y has now been cast to float, due to the addition of a float to an int.
|
|
Now, using an assignment operator on an int gives a similar result it typecasts x as a float.
|
|
I assume it needs to create new memory to cope, but it s now a float to the user.
|
|
.
|
|
Now consider the following operation using a numpy array, in which assignment of a new variable y) results in typecasting the int64 to a float64, yielding the expected result.
|
|
.
|
|
A user may reasonably be expected to perform the add assignment operation using a numpy array with the exact same results as the parallel operation on a standard int and as shown immediately above in the numpy array, when not using an assignment operator.
|
|
The following example illustrates the issue .
|
|
Expected result A user may reasonably expect the result of this operation to be equivalent to assignment to a new variable, since this behavior is paralleled in normal Python ints.
|
|
One would expect the following as a continuation from above .
|
|
Actual result Instead of new memory being created to allocate the new dtype, numpy does this instead .
|
|
Workarounds Obviously, the workaround is to typecast the original numpy array in the type consistent with the types with which it is expected to work in this case, float64) .
|
|
Comments This is normally not a problem when you have overt control over the output type.
|
|
But this issue was encountered in the following context, in which the user would have to explicitly re type the resultant vector as floats instead of ints or avoid confusingly using assignment operators in this case but NOT a similar non Numpy Python case .
|
|
It turns out that this issue affects add assignment and subtract assignment operators but not multiply shown below) and divide assignment operators, for which it is happy to typecast the existing numpy array as a float .
|
|
System Python 2.7.6 from MacPorts on Mac OS X 10.9.2.","bug"
|
|
"6013","4706","Travis added python 3.4.
|
|
I m not sure what the optimal change to `.travis.yml` is what with the chroot and all, but python 3.4 is now available on Travis as a standard venv, so I guess we should add it.","bug"
|
|
"6013","4709","No way to mark NotImplemented for comparison operations on subclasses.
|
|
In astropy, we have a `Quantity` class which is a subclass of `ndarray` for physical quantities.
|
|
One of its features is that if you compare two of them and the units aren t equivalent, they compare to false.
|
|
In Numpy master but not Numpy 1.8.x), as of 9b8f6c72 seberg), it seems there might no longer be a way to implement this, since raising an exception in ` array prepare ` bubbles up through `ndarray.richcompare` it currently emits a `DeprecationWarning`, but I understand will eventually pass the original exception through).
|
|
Since ` array prepare ` is only able to return an instance `ndarray` or subclass), there doesn t seem to be a way to flag to the comparison operator that what we really want is `NotImplemented`.
|
|
Overriding the ` eq ` operators, etc., doesn t seem to be sufficient, since it can t handle the case where there is a regular) array on the left hand side.
|
|
Here s a minimal script to reproduce the issue the actual code is much more complex, but this boils it down to the essence of the problem) .
|
|
The output on Numpy 1.8.x is .
|
|
The output on Numpy master is .
|
|
If you uncomment the ` eq ` function above, the first test passes, but the second still fails.
|
|
Is there a workaround here, or something I m missing?
|
|
mhvk, astrofrog","bug"
|
|
"6013","471","Offset is not updated when slicing memmaps.
|
|
<pre> >>> import numpy as np >>> a = np.ones 10, 10)) >>> np.save a.npy , a) >>> b = np.load a.npy , mmap mode= r ) >>> c = b[3 ] >>> c.offset 80 >>> b.offset 80 < pre> The offset of c is wrong it does not take in account the fact that the slicing has changed the offset.","bug"
|
|
"6013","4714","ENH Random processes.
|
|
numpy provides samples from random distributions.
|
|
However, another common way to get numbers from a random distribution is using a random or stochastic) process.
|
|
To quote Wolfram, A random process models the progression of a system over time, where the evolution is random rather than deterministic .
|
|
A classic example is a random walk, where the individual steps are random values with a particular distribution.
|
|
https reference.wolfram.com mathematica guide RandomProcesses.html http en.wikipedia.org wiki Stochastic process A common example is the poisson process, which is used to approximate things as diverse as radioactive decay events, the arrival times of people to a store, photon detection events, web server request times, and the times of neural events.
|
|
http en.wikipedia.org wiki Poisson process I think it would be nice if numpy also provided some commonly used random process generators in addition to the random distribution generators.
|
|
Probably a good set of starting processes would be the Poisson process, Wiener process, and Ornsteinxe2x80x93Uhlenbeck process.
|
|
http en.wikipedia.org wiki Wiener process http en.wikipedia.org wiki Ornstein E2 80 93Uhlenbeck process","enhancement"
|
|
"6013","4715","Tolerance threshold limit in linalg.py .
|
|
Quick Question What is the tol set for linalg for matrix inversion in linalg.py ?
|
|
Usecase Given a non invertible non full rank) matrix, reduce the matrix to its independent features.
|
|
Towards the same, I am building on the code here http stackoverflow.com questions 13312498 how to find degenerate rows columns in a covariance matrix and , post that feeding the reduced matrix which is then converted back to a pandas dataframe to fit a Logit model.
|
|
Here s the stack trace I checked that X is a full rank matrix prior to running the code y is the outcome var x is the regressor var Traceback most recent call last) File home ekta Python scripts DYB logistic regression iab list.py , line 275, in <module> test scores = smf.Logit y,X,missing= drop ).fit ) File usr local lib python2.7 site packages statsmodels 0.5.0 py2.7 linux i686.egg statsmodels discrete discrete model.py , line 1186, in fit disp=disp, callback=callback, kwargs) File usr local lib python2.7 site packages statsmodels 0.5.0 py2.7 linux i686.egg statsmodels discrete discrete model.py , line 164, in fit disp=disp, callback=callback, kwargs) File usr local lib python2.7 site packages statsmodels 0.5.0 py2.7 linux i686.egg statsmodels base model.py , line 357, in fit hess=hess) File usr local lib python2.7 site packages statsmodels 0.5.0 py2.7 linux i686.egg statsmodels base model.py , line 405, in fit mle newton newparams = oldparams np.dot np.linalg.inv H), File usr local lib python2.7 site packages numpy linalg linalg.py , line 445, in inv return wrap solve a, identity a.shape[0], dtype=a.dtype))) File usr local lib python2.7 site packages numpy linalg linalg.py , line 328, in solve raise LinAlgError, Singular matrix LinAlgError Singular matrix","bug"
|
|
"6013","4717","np.sum ) returns wrong answer.
|
|
The following code returns two different sums .
|
|
NumPy s answer is incorrect, maybe due to some overflow in np.float32.
|
|
import numpy as np x = np.zeros 123456789, dtype=np.float32) x = x 123456789.987654321 print np.sum x) print sum x)","bug"
|
|
"6013","4718","histogram with array valued weights.
|
|
Sometimes, I have a sample of data which I want to divide into bins and then sum an associated vector over all samples within each bin.
|
|
This could be accomplished by creating a histogram with array valued weights.
|
|
Perhaps there are other ways to do it, but this seems to be the best conceptual match to what I want to do, and is a very natural extension of `np.histogram`.
|
|
In the particular example that I am dealing with at the moment, I want to re weight my samples according to a 4th order polynomial of an associated quantity.
|
|
That is, I have an array `hist variable` with shape ` n,)`, and another array `reweighting variable` also of shape ` n,)`.
|
|
My weights are some polynomial function of `reweighting variable` .
|
|
This works fine, as long as `a0`, `a1`, etc.
|
|
are known and fixed.
|
|
If I instead want to vary `a0`, `a1`, etc., then I can create five different histograms .
|
|
And then I can choose any `a0`, `a1`, etc.
|
|
that I like, and get my weighted histogram with the function `weighted counts`, without performing the expensive `np.histogram` step again.
|
|
This works, but it is not ideal.
|
|
The reason that it is not ideal is that I must re do the binning step of `np.histogram` many times.
|
|
This is expensive.
|
|
I ve tried to create a solution using `np.digitize` and fancy indexing, but that is also quite slow.
|
|
What I wish I could do is .
|
|
Can this be implemented efficiently?
|
|
It seems like summing array valued weights shouldn t be drastically slower than summing scalar valued weights.","enhancement"
|
|
"6013","4720","Vonmises problem for kappa slightly bigger than threshold 1e 8.
|
|
When sampling using vonmises I get stuck for kappa slightly bigger than 1e 8, no returned value and I need to kill the process.
|
|
Code snippet that shows the problem import numpy as np np.random.vonmises 0., 1.1e 8) From what I can see the threshold for I guess?)
|
|
truncating to a uniform dist over pi, pi is 1e 8. https github.com numpy numpy blob master numpy random mtrand distributions.c I think an efficient solution would be to instead implement an efficient uniform rejection sampler for small kappa.
|
|
That our first sample on pi,pi would be rejected is very unlikely.","bug"
|
|
"6013","4723","fails to compile on AIX.
|
|
I am building numpy on AIX 6.1 using the xlc compilers.
|
|
They are failing with errors like this creating build temp.aix 6.1 2.7 build src.aix 6.1 2.7 numpy core src npysort compile options Ibuild src.aix 6.1 2.7 numpy core src private Inumpy core include Ibuild src.aix 6.1 2.7 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I profit freeware python aix python2.7.6 include python2.7 Ibuild src.aix 6.1 2.7 numpy core src private Ibuild src.aix 6.1 2.7 numpy core src private Ibuild src.aix 6.1 2.7 numpy core src private Ibuild src.aix 6.1 2.7 numpy core src private c xlc r build src.aix 6.1 2.7 numpy core src npysort heapsort.c usr include stdio.h , line 528.12 1506 343 S) Redeclaration of fgetpos64 differs from previous declaration on line 323 of usr include stdio.h .
|
|
usr include stdio.h , line 528.12 1506 377 I) The type long long of parameter 2 differs from the previous type long restrict .
|
|
usr include stdio.h , line 531.12 1506 343 S) Redeclaration of fseeko64 differs from previous declaration on line 471 of usr include stdio.h .
|
|
The problem is that on AIX as well as other platforms, Python.h contains system level defines that are used in standard libraries.
|
|
This particular problem is in stdio.h.
|
|
On AIX, Python is built with DLARGE FILES and this define is in Python.h actually pyconfig.h included in Python.h.)
|
|
Basically, since Python.h has these system level defines, Python.h should be included before any other include so that the defines are seen first before any other includes are processed.
|
|
If you look at all of the Python source code, you will see that this is the case.
|
|
In fact, most of the numpy source also does this and it probably is a know fact that it should be done.
|
|
However, the following numpy source files do not include Python.h first or in some cases at all.
|
|
numpy core src multiarray methods.c numpy core src umath test rational.c.src numpy core src umath operand flag test.c.src The following don t have include Python.h at all and need to have for them to compile numpy core src npysort heapsort.c.src numpy core src npysort quicksort.c.src numpy core src npysort mergesort.c.src Once I moved the include Python.h to be the first include or added it if not there, numpy compiled.","bug"
|
|
"6013","4724","argsort does not work for multidimensional arrays.
|
|
consider the following code a = np.random.random [5,5]) ind = np.argsort a, axis=1) a sorted = a[ind] np.sort a, axis=1) now a sorted and a should be both sorted along the 1 axis.
|
|
However not even the shapes are identical anymore.
|
|
a is still of 5,5) while a sorted is 5,5,5).
|
|
If this is the intended behaviour, can someone tell me why?","bug"
|
|
"6013","4725","arrays could support getsizeof ).
|
|
sys.getsizeof ) uses the `` sizeof `` method, if defined, to return a Python object s full memory footprint including owned data areas that lay outside of the basic object structure).
|
|
Currently .
|
|
On the other hand, since the underlying buffer can be shared by several objects, it s not obvious how non misleading the answer would be.","bug"
|
|
"6013","4726","dead link in generalized ufunc documentation.
|
|
http docs.scipy.org doc numpy reference c api.generalized ufuncs.html has a link to http scipy.org scipy numpy wiki GeneralLoopingFunctions, but that URL returns a 404.","bug"
|
|
"6013","4728","Current mingw fails to compile cpucaps nsis plugin.
|
|
I m trying to build scipy installers for Windows.
|
|
I have installed mingw 3.4.5 using the steps I recorded here https github.com numpy numpy wiki building scipy superpack I downloaded and install scons 2.3.1 from http www.scons.org download.php I tried to build cpucaps with .
|
|
This gave me a long error trace starting with .
|
|
and ending with .
|
|
I m not experienced with MinGW, and couldn t see anything obvious with a quick google search; any clues as to what I should try next?","bug"
|
|
"6013","4729","Creating an array from a sequence of memoryviews fails under Python 3.4.
|
|
Consider the following code snippet .
|
|
It will succeed under 2.7 and 3.3, but fail under 3.4 .
|
|
I m not sure this is supposed to be officially supported, but it s used by numba in order to create record instances.","bug"
|
|
"6013","473","memory error with strings in structured arrays.
|
|
The following code writes to memory blocks beyond the limits of the array, resulting in memory errors in both released and dev version of numpy I ve tried 1.6.2 and 1.8.0.dev) .
|
|
The problem I believe) is that passing `str` to `dtype` results in a zero length string, and this is not checked on assignment.","bug"
|
|
"6013","4730","numpy.conjugate and numpy.ndarray.conjugate behave inconsistently.
|
|
If x is an ndarray of objects that define a method conjugate , then numpy.conjugate x) takes the conjugate of each element, but x.conjugate ) does nothing.
|
|
>>> numpy.version.git revision 697316a867a32e9f72778a29226c9febbf867ee8 >>> numpy.version.version 1.7.1 >>> numpy.version.release True >>> class Object object) ... def conjugate self, args, kwds) ... print Calling conjugate ... return self ... >>> o = Object ) >>> o.conjugate ) Calling conjugate < main .Object object at 0x18db1c90> >>> oo = numpy.array [o,o]) >>> numpy.conjugate oo) Calling conjugate Calling conjugate array [< main .Object object at 0x18db1c90>, < main .Object object at 0x18db1c90>], dtype=object) >>> oo.conjugate ) array [< main .Object object at 0x18db1c90>, < main .Object object at 0x18db1c90>], dtype=object)","bug"
|
|
"6013","4732","Release numpy wheel ?.
|
|
Are you planning to release numpy as wheel on pypi Linux Windows Mac) ?
|
|
So users won t need to compile numpy each time they build a new virtualenv.
|
|
Note sure we could use custom wheel repo localy or remotely) but it would be nice to have them on pypi.","bug"
|
|
"6013","4733","svd bug.
|
|
tracking down the `matrix rank` bug report in http mail.scipy.org pipermail numpy discussion 2014 May 070193.html I find .","bug"
|
|
"6013","4736","problem of numpy.fft.fft2 example .
|
|
http docs.scipy.org doc numpy reference generated numpy.fft.fft2.html I run the example,, however, what I got is different from what you shown in the example.
|
|
I got array [[ 50.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j ], [ 12.5 17.20477401j, 0.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j ], [ 12.5 4.0614962j , 0.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j ], [ 12.5 4.0614962j , 0.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j ], [ 12.5 17.20477401j, 0.0 0.j , 0.0 0.j , 0.0 0.j , 0.0 0.j ]]) I don t know what happened?
|
|
Thanks a lot.","bug"
|
|
"6013","4737","test fail on Mac OS 10.9.3 python 3.4.0 ImportError dlopen ...) =mage found..
|
|
I have installed numpy using `pip` but full tests fail.
|
|
Here is the full log https gist.github.com SashaOv 0403c4ef38af00ecaf0d","bug"
|
|
"6013","474","OSError dropped silently in ctypeslib loadlibrary for defective .so.
|
|
When calling numpy.ctypeslib.load library mylib , . )
|
|
load library looks for both mylib.so and mylib.pyd files.
|
|
If mylib.so exists but raises an OSError, and mylib.pyd does not exist the returned error is OSError mylib.pyd cannot open shared object file No such file or directory when it should be for example) OSError mylib.so undefined symbol QJacobianN See e.g.
|
|
http www.jmodelica.org 4688 comment 2154","bug"
|
|
"6013","4741","BUG In numpy.i, ARGOUTVIEWM typemaps require fragment NumPy Utilities..
|
|
In numpy.i, NumPy 1.8.1), the argout typemaps of all the ARGOUTVIEWM typemap suites make use of the free cap function, defined in the NumPy Utilities fragment, which is not listed in the fragment argument of the typemaps.
|
|
As a result, the free cap function is not included in the source generated code by SWIG, and the code do not compile unless another typemap including the missing fragment has been used) demo wrap.cpp 3097 72 error xe2x80x98free capxe2x80x99 was not declared in this scope This is easily fixed by adding NumPy Utilities to the fragment argument of the typemaps.","bug"
|
|
"6013","4742","numpy.i contains duplicated code..
|
|
In numpy.i, NumPy 1.8.1), the 4 typemap suites for ARGOUTVIEWM for 4D arrays are duplicated sed n 2737,2887p numpy.i > part0.i sed n 2889,3039p numpy.i > part1.i diff part0.i part1.i Seems that the first part can be removed and that it contains a bug at line 2827, overriden by the second part).","bug"
|
|
"6013","4743","Please add support for OpenRISC or1k architecture.
|
|
Forwarding the bug from Debian https bugs.debian.org cgi bin bugreport.cgi?bug=749195 Hi, The patch provided against version 1.8.1 1) allows to get this package compiled in OpenRISC or1k architecture that Christian and I are creating in Debian.
|
|
After installing python numpy and python nose, most of the unit tests seem to pass, and the ones that do not pass might be due to underlying problems in blas or others e.g.
|
|
threading issues in the platform) that we still have to work out.
|
|
Cheers.","bug"
|
|
"6013","4747","PyObject AsWriteBuffer does not work on Fortran order array in Python 3.
|
|
This snippet of code works for me in Python 2.7, but fails in Python 3.4 with the following error .
|
|
I m using numpy 1.8.1 in both cases.
|
|
.","bug"
|
|
"6013","4752","BUG cannot specify second output to ufunc as keyword argument.
|
|
Currently, it does not seem possible to pass on an array to catch the second output of a `ufunc` via a keyword parameter .
|
|
One can do `np.modf 1.333, out=a1)`, which works as expected.
|
|
If one gives instead `np.modf 1.333, out2=a2)`, the first part is assigned to `a2`.
|
|
Note that the documentation suggests to use `out1, out2`, but any characters after `out` simply seem to be skipped.)
|
|
It would seem good to have the option of giving these as keyword arguments.
|
|
One solution might be to have it be consistent with the usage in ` numpy ufunc `, i.e., one would give `np.modf values, out= a1, a2))`.","bug"
|
|
"6013","4755","meshgrid silently ignores unknown arguments..
|
|
The ``meshgrid`` function silently ignores arguments.
|
|
.","bug"
|
|
"6013","4763","RandomState ctor accesses dev urandom, a significant bottleneck when copying RandomStates.
|
|
The RandomState ctor function in numpy.random.
|
|
init makes an call to construct a new RandomState object without an explicit seed.
|
|
The unseeded call results in an access to dev urandom which is wildly expensive.
|
|
Given that the purpose of this function to add in pickling, so that the internal state will be updated to reflect the state of the pickled RandomState, this expensive access actually don t do anything.
|
|
Because the state of the newly created RandomState is irrelevant it ll be updated by pickling), its safe to use the fast cached seed value of 0.
|
|
An example of updating the function is below as well as a nosetest backed unit test that verifies behavior.
|
|
In order to incorporate into master branch the code in init .py in random simply needs to have seed=0 added to it.
|
|
.
|
|
Unit test to verify behavior works .","bug"
|
|
"6013","4766","NumPy ignores array priority when the right hand side operand is a subclass of ndarray.
|
|
The following code does not output sub class radd as expected since NumPy ignores ` array priority ` when the right hand side operand is a subclass of ndarray.
|
|
.
|
|
The problem is [line 282 in number.c] https github.com numpy numpy blob master numpy core src multiarray number.c L282) where a branch uses `PyArray Check )` to check the right hand side operand `m2`.
|
|
If `PyArray Check m2)` returns `True`, which is the case when `m2` is a subclass of ndarray, ` array priority ` is ignored.
|
|
Now, this is not a problem when the left hand side operand is an ndarray since in CPython a subclass always takes precedence over its superclass.
|
|
However, a NumPy scalar is not a sub super class of ndarray as the example code above illustrates.
|
|
There is a very simple fix change `PyArray Check m2)` to `PyArray CheckExact m2)`.
|
|
However, this fix has the side effect that the operation will fail with NotImplementedError if the subclass doesnxe2x80x99t implement the operator, which is a change to the current behavior.","bug"
|
|
"6013","4767","percentile for masked array.
|
|
Hello, I would like to know if it is possible to calculate a percentile when an input array is a masked array.
|
|
Natalia","enhancement"
|
|
"6013","4771","Memory leak in numpy version 1.8.0 .
|
|
Hi, I ve noticed what I believe to be a memory leak when using numpy.dot to take the dot product of elements of an array.
|
|
Below is some example code which highlights the issue.
|
|
import numpy as np e= np.zeros 400, 1024 2048)) for i in range len e)) g = np.dot e[i],e[i])","bug"
|
|
"6013","4774","byteswap regression in v1.8.0b1 149 2 leads to Bus error on sparcs.
|
|
with 1.8.x on sparcs pytables started to experience a bus error didn t happen with 1.7.1).
|
|
bisect fun seems to lead to commit 99cb95f7379a95f978fd04d183fde2d262b262a6 ENH optimize byteswapping via intrinsics Byteswapping is implemented in hardware on x86 cpus which is more than double as fast than the existing implementation.
|
|
The masked version used for 32bit plays better with compiler pattern matching so it can be optimized also by non gcc clang compilers.
|
|
.","bug"
|
|
"6013","4775","f2py doesn t update .mod file.
|
|
Hello, When `f2py c foo.f90 m bar` and a foo.mod exists.
|
|
F2py doesn t update the .mod file, and what worse, it used the .mod file implicitly.
|
|
If I changed something in the .f90 file, the .mod file will not be consistent with the .f90 file.
|
|
The .mod file is an intermediate file, as the same as .o, so if we don t use it explicitly on a command line, it should be regenerated, at least, it should not be used.
|
|
The module variable with `parameter` attribute will not be written in the .o file, but in the .mod file.
|
|
I have just encounter an odd bug because f2py implicitly used inconsistent .mod file.","bug"
|
|
"6013","4776","Multiprocessing hangs on Mac OS X without MKL.
|
|
This bug occurs on Mac OS X 10.9 but does not seem to occur on Linux and Windows) with both Python 2.7 and Python 3.3 at least for numpy 1.8.1 not compiled with MKL.
|
|
The following piece of code reproduces the bug .
|
|
The above code hangs in the `workers.map` call when executing `np.dot np.eye n), np.eye n))`.
|
|
If the sequential call to `compute` is ommitted, the code runs as it should.
|
|
Also, if the argument `n` is reduced sufficiently < 100 on the computer which the bug was detected on), the code runs as it should.
|
|
Again, the bug is not present when numpy is compiled with MKL.","bug"
|
|
"6013","4778","PR 4606 introduces regression into astropy table join.
|
|
The numpy commit e47a50efc284c677ba4d0337e11dc8514fca7e5b causes a failure of an astropy test that works fine with all previous numpy versions through 1.5.
|
|
This was determined with git bisect on numpy dev.
|
|
The failure manifests as an apparent leak of memory bytes between two tables during a table join operation.
|
|
This is demonstrated below.
|
|
Apologies that I haven t had time to make a simpler test to reproduce without using the `astropy.table.np utils.join` function, but perhaps the symptoms may be obvious enough?
|
|
If not I can try harder to figure out where this is happening within the `join` function.
|
|
This is discussed in https github.com astropy astropy issues 2281. .
|
|
The expected first value of the first row is ` M101 `.
|
|
It appears that the final three characters of the first value in the `t2` table are leaking into the `t3` table the `516` characters from `NGC3516`).","bug"
|
|
"6013","4779","arr.dtype.type has different hashes.
|
|
Not sure if this is supposed to be like this.
|
|
original issue here https github.com pydata pandas issues 7332 cross post to numexpr here https code.google.com p numexpr issues detail?id=126 essentially in pandas were looking up a ``dtype,type`` in a cython dictionary turns out that for ``int64`` and ``int32`` but NOT ``int64`` on 32 bit platforms), the hashes are DIFFERENT, but same for other dtypes including ``float64``).
|
|
Is this maybe an implementation detail on ``numexpr`` and or incorrect usage of ``dtype.type`` and or invalid guarantees on this object?
|
|
FYI, we switched to using ``dtype.name`` for the lookup and no issues.
|
|
.
|
|
For the floats the same though .","bug"
|
|
"6013","478","append bug masked array.
|
|
The source .
|
|
should not be modified by .
|
|
.
|
|
Output .
|
|
using v1.6.1)","bug"
|
|
"6013","4780","Wrong shape of array in sparse.mean axis=0).
|
|
For example.
|
|
X = np.random.rand 10, 10) print X.mean axis=0).shape 10,) csr = sparse.csr matrix X) print csr.mean axis=0).shape 1, 10) Is this a feature or a bug?","bug"
|
|
"6013","4785","numpy.unique inconsistent behavior on array of tuples.
|
|
I have noticed that numpy.unique have inconsistent behavior on array of tuples depending on whether return inverse flag is set .
|
|
The output is .
|
|
It seems setting `return inverse` will cause the array of tuples to be flattened, otherwise the array will not be flattened.
|
|
Could this be fixed so numpy.unique would have consistent behavior ideally, not flatten array of tuples)?
|
|
Thanks!
|
|
python version 2.7.6 numpy version 1.8.1","bug"
|
|
"6013","4787","Numpy.log2 returns incorrect result.
|
|
Would anyone know why numpy is returning 2.9999999999999996 instead of 3 for numpy.log2 8) ?
|
|
This is only happening in Windows) Python 3.4.1 |Anaconda 2.0.0 64 bit)| default, May 19 2014, 13 02 30) [MSC v.1600 64 bit AMD64)] IPython 2.1.0 An enhanced Interactive Python.
|
|
C \Anaconda3\python.exe 3.4.1 |Anaconda 2.0.0 64 bit)| default, May 19 2014, 13 02 30) [MSC v.1600 64 bit AMD64)] pylab Using matplotlib backend Qt4Agg Populating the interactive namespace from numpy and matplotlib np.log2 8) Out[3] 2.9999999999999996 np.
|
|
version Out[4] 1.8.1 np.log2 2) Out[5] 1.0 np.log2 4) Out[6] 2.0 np.log2 8) Out[7] 2.9999999999999996 np.log2 16) Out[8] 4.0","bug"
|
|
"6013","479","f2py fails creating an interface file.
|
|
Somewhere in the code a wrong variable is passed down to an exec statement.
|
|
This is not a pull request as I believe someone should look into it and potentially also replace the actual `exec )` call.
|
|
In the meanwhile, this ugly hack fixes it .
|
|
Additional information Numpy Version .
|
|
Library Used on [mtspec] http wwwprof.uniandes.edu.co gprieto software mwlib.html) library v3.1 if anyone wants to reproduce.
|
|
Call and Traceback .","bug"
|
|
"6013","4791","Let `fromiter` handle object arrays.
|
|
In 2006, Tim Hochberg questioned the reason [why fromiter does not handle object arrays] https github.com numpy numpy blob master numpy core src multiarray ctors.c L3580 3588) .
|
|
http mail.scipy.org pipermail numpy discussion 2006 November 024747.html teoliphant responded .
|
|
http mail.scipy.org pipermail numpy discussion 2006 November 024841.html I couldn t find any further discussion of this issue since 2006.
|
|
Is this analysis of what needs to be done to `PyArray FromIter` still sound?","bug"
|
|
"6013","4795","MSVC fails to build 32 bit numpy 1.9.0b2.
|
|
The Visual Studio 2008 and 2010 compilers for 32 bit fail to build numpy 1.9.0b2 .
|
|
The 64 bit builds succeed.
|
|
The compiler error is .","bug"
|
|
"6013","4797","Error creating rpm file with setuptools.
|
|
I am trying to create a custom RPM file but bdist is not able to create the spec file.
|
|
Throws that error rpmbuild SOURCES numpy 1.8.1 python2.7 setup.py bdist rpm spec only requires python27 Running from numpy source directory.
|
|
usr lib python2.7 distutils dist.py 267 UserWarning Unknown distribution option test suite warnings.warn msg)","bug"
|
|
"6013","4798","numpy.load ) can t load matplotlib sample data goog.npy in Python 3.
|
|
A bug report.
|
|
New to GitHub, I don t know it s proper to put this here this way.
|
|
See also in [this] http stackoverflow.com questions 24105148 load npy file in python 3 comment37242668 24117216) stackoverflow post.
|
|
`numpy.load )` can t load `matplotlib sample data goog.npy` in Python 3.4, but works in Python 2.7.
|
|
NumPy version `1.8,1` .
|
|
and exception raised .
|
|
Someone on stackoverflow suggested to modify function `read array` in `numpy lib format.py` from .
|
|
to .
|
|
but didn t work.
|
|
The `.npy` file can be found [here] https github.com matplotlib sample data blob master goog.npy).
|
|
The full traceback I got .","bug"
|
|
"6013","4799","Missing data point in histogram2d when using float32 arrays.
|
|
I am getting an incorrect count from the histogram2d function when using arrays with dtype=numpy.float32.
|
|
Specifically, the histogram2d fails to count one of the data points when I use bins=100.
|
|
If on the other hand I put dtype=numpy.float64, then I get an output of 3 from both function calls below.
|
|
.
|
|
Information eyurtsev puddles Desktop uname a Linux puddles 3.13.0 24 generic 47 Ubuntu SMP Fri May 2 23 30 00 UTC 2014 x86 64 x86 64 x86 64 GNU Linux eyurtsev puddles Desktop ipython c import numpy; print numpy.
|
|
version 1.8.1 Python 2.7.6 default, Mar 22 2014, 22 59 56) Please let me know if there is any more information I can provide that will be helpful.","bug"
|
|
"6013","4802","Unexpected behavior of compound assignment =.
|
|
Hello together, the = assignment seems to break for matrices > 90,90).
|
|
Reproduction import numpy as np for a in xrange 200) px=np.arange a) px=np.random.random a) dx = np.repeat px[np.newaxis, ], a, axis=0) np.repeat px[ , np.newaxis], a, axis=1) dx2 = np.repeat px[np.newaxis, ], a, axis=0) dx2 = dx2.T if len np.nonzero dx dx2)[0]) != 0 print a break I m using `numpy` version 1.8.1.
|
|
Cheers, Stephan Edit I tested it on different machines It gives out 91 on`numpy` version 1.6.2, 1.8.0 and 1.8.1.
|
|
3 on 1.3.0","bug"
|
|
"6013","4805","missing documentation for np.full full like.
|
|
`np.full` and `np.full like` don t seem to exist in the compiled documentation including the version on the web) except for the announcement of their introduction `release.html`).","bug"
|
|
"6013","4806","Record access on non existing fields no error issued.
|
|
When accessing multiple records with `arr[[ field0 , field1 , ]]`, using a non existing field is silent .
|
|
This is inconsistent with accessing a single field .
|
|
Furthermore, it creates error messages that can be harder to make sense of, like in .
|
|
ValueError need more than 1 value to unpack).
|
|
I can t find any documentation on the current behavior either it s not in http docs.scipy.org doc numpy reference arrays.indexing.html record access).
|
|
I would suggest that non existing fields raise an exception, for consistency least surprise and practicality reasons.","bug"
|
|
"6013","4807","numpy.corrcoef not to be used optional variable ddof.
|
|
while in the function numpy.cov it has a meaning to specify ddof, since numpy.corrcoef return the correlation coefficient that are the normalized covariance coefficient, ddof is irrelevant.
|
|
If I did not misunderstood the meaning of numpy.cov and numpy.corrcoef, I would suggest to remove the optional argument ddof in the function numpy.corrcoef.","bug"
|
|
"6013","481","segfault in IsWriteable when data is not owned.
|
|
The following, albeit slightly contrived, example causes a segfault in IsWriteable.
|
|
foo.pyx .
|
|
test.py .
|
|
`python test.py` > segmentation fault Here s IsWriteable with unnecessary bits removed).
|
|
.
|
|
a.setflags write=True) works because ap == a base == NULL and we hit the first condition base is NULL or we own the data) b.setflags write=True) doesn t work because ap == b base == a b.flags[ OWNDATA ] == False so we go into the loop base doesn t own the data either, so we try and find the base of a, which is NULL, and PyArray Check NULL) is a segfault.
|
|
I think the fix is .
|
|
But I m not sure which of returning True or False is correct.","bug"
|
|
"6013","4810","scipy.special.hankel1 0, 10 10) return `nan`?.
|
|
>>> import scipy.special >>> scipy.special.hankel1 0, 10 8) 3.2060295340412074e 05 7.3063911655217044e 05j) >>> scipy.special.hankel1 0, 10 9) 2.4687471886269185e 05 5.2104226538976152e 06j) >>> scipy.special.hankel1 0, 10 10) nan nan j) I m not sure but i think this should be done by either Mathematica or MATLAB but i got nan when use scipy.","bug"
|
|
"6013","4811","Enhancement request add ndmin option to genfromtxt.
|
|
A couple years ago, `loadtxt` was given the argument `ndmin`.
|
|
`genfromtxt` should also have this option.
|
|
See, for example, http stackoverflow.com questions 24247103 numpy genfromtxt collapse recarray when data file has only one row","enhancement"
|
|
"6013","4813","numpy.linalg hangs with multiprocessing and OpenBLAS Linux.
|
|
Consider the following code .
|
|
If numpy uses OpenBLAS, it hangs forever consuming 100 CPU with OPENBLAS NUM THREADS>1.
|
|
If export OPENBLAS NUM THREADS=1, it works OK.
|
|
I am testing this code on 32 core machine, if this matters.
|
|
It may be related to issue 654 numpy version 1.8.1 python version 3.4.1 openblas version 0.2.8 ubuntu 14.04","bug"
|
|
"6013","4814","Memory bloat using numpy.ma.median Py 2.7.4, Np 1.7.1).
|
|
I ve got a large 7000,180,3600) masked array that I ve been sending to numpy.ma.median.
|
|
Prior to the function call memory used on the system is 20Gb I have a bunch of other variables I m also dealing with), and upon the function call memory quickly increases to 200 and then 300Gb!
|
|
without even returning a result I kill the job before it completes) Is this expected behaviour?
|
|
I ve tried using the overwrite input=True option, however this doesn t seem to resolve anything.
|
|
I m using python 2.7.4 and numpy 1.7.1","bug"
|
|
"6013","4817","Array creation in `numpy.mat` style.
|
|
Is there any function similar to `numpy.mat` to create an `numpy.ndarray`?
|
|
I mean something equivolent to np.array np.mat .2 .7 .1; .3 .5 .2; .1 .1 .9 )) which is a little bit verbose.
|
|
And np.array [[.2, .7, .1], [.3, .5, .2], [.1, .1, .9]]) sounds daunting for beginners.
|
|
This might seem trivial as `mkarry = lambda s np.array`), but when I do some demonstrations to people who have never used Python, it matters.
|
|
Is there such a function in `numpy`?
|
|
If no, is it sensible to add one in `numpy`?","bug"
|
|
"6013","4818","Allow linspace to take vectors as parameters.
|
|
It would be very useful to be able to send start, end, and step size to linspace as vectors, and return a 2d array.
|
|
This is particularly useful for image warping.
|
|
Here s my bespoke implementation, it could probably be tweaked and used here def linspace2d start, stop, iris height) step = stop start) iris height width = len start) x = np.empty width, 1)) y = np.arange 0,iris height) xx,yy = np.meshgrid x, y) y = mx c return yy step start","bug"
|
|
"6013","4819","numpy.corrcoef returns 1 when applied on 2 matrices..
|
|
Please consider the following behaviour I understand that the input is wrong but the function returns 1 without any error messages) <pre> In [1] import numpy as np In [2] x = np.random.randn 20, 10, 10) In [3] y = np.random.randn 20, 10, 10) In [4] np.corrcoef x, y) Out[4] 1 < pre>","bug"
|
|
"6013","4820","dtype code crashes interpreter.
|
|
The current dtype creation and comparison codepaths lead to crashes if feed a non known dtype This is from https github.com pydata pandas pull 7217 python 2.7.7, win7, 64bit) .
|
|
and .
|
|
As pandas will introduce a category dtype inplemented in python https github.com pydata pandas pull 7217) this crashes could happen if one is working with the data.","bug"
|
|
"6013","4824","convert datetime64 from to vectorized dates.
|
|
In our applications, we often represent datetime as a np.array with dtype `[ YY , <i4 ), MM , <i4 ), DD , <i4 ), hh , <i4 ), mm , <i4 ), ss , <f8 )]` Can this be converted to datetime64, using vector methods?
|
|
Of course, converting to string and reparsing is not an option, it would be awfully slow.","bug"
|
|
"6013","4829","ENH add numerically robust polygcd function for numerical calculation of polynomial greatest common divisor.
|
|
GCD for polynomials is an important operation for computing sums of polynomial ratios A x) B x) C x) D x).
|
|
The naive approach to computing GCD B x),D x)) is numerically ill conditioned .
|
|
I figured there was an easy workaround that would be more robust, but apparently there isn t. There is some promising literature on the subject see below), but I m woefully underskilled for porting these algorithms to numpy, much less understanding how they work http math.univ lille1.fr bbecker ano pub 1997 ano369.pdf http www.mathcs.emory.edu boito newfifthspecial.pdf http www.mathcs.emory.edu boito thesis.pdf http www.mathcs.emory.edu boito fast gcd wls.m [matlab code] https who.rocq.inria.fr Jan.Elias pdf je masterthesis.pdf http ieeexplore.ieee.org xpls abs all.jsp?arnumber=875462 http www.mmrc.iss.ac.cn pub mm21.pdf zhi1.pdf http arxiv.org pdf 1207.0630v2.pdf http www.neiu.edu zzeng http www.neiu.edu zzeng uvgcd.pdf http citeseerx.ist.psu.edu viewdoc summary?doi=10.1.1.27.4881","bug"
|
|
"6013","4830","numpy.ndarray.ctypes.data returns invalid pointer.
|
|
I suspect passing a numpy.ndarray.ctypes.data pointer to a C function using the c types package actually only passes a 32 bit pointer.
|
|
This caused some indeterministic behaviour for me it obviously works depending on the allocated address itself), so the bug cannot be reproduced easily.
|
|
The API states it returns this as a A pointer to the memory area of the array as a Python integer .
|
|
I believe it should instead return ctypes.c size t in order to be compatible with 64 bit addressing?
|
|
Current work around When I used data as obj) with a ctypes.c void p the problem is resolved.","bug"
|
|
"6013","4833","Just beginning.
|
|
How do I simply install the PyLibNIDAQmx to a windows 7 computer","bug"
|
|
"6013","4834","Segmentation fault when using memory allocation tracker.
|
|
I use Python 2.7.6, numpy 1.8.1 and cython 0.20.2 under Red Hat 6.5.
|
|
I tried out the code in numpy tools allocation tracking, that is, I installed the extension alloc hook) and then copied track allocations.py into my own code and used it according to the example in the main block.
|
|
But the moment I allocate a numpy array I get a segmentation fault.
|
|
In case it is related, here is the extension compile output .
|
|
thouis charris","bug"
|
|
"6013","4835","add iterative methods and reordering algorithms.
|
|
I noticed that numpy doesn t have any iterative methods for solving linear systems Jacobi, Gauss Seidel, SOR, conjugate gradient...) and was interested in adding it.
|
|
on top of that, there are no reordering algorithms to reduce the bandwidth of a matrix Reverse )Cuthill McKee, the greedy algorithm I can t remember the name of...), although this may be more suitable for scipy.linalg as R)CM is best operated on sparse matrices.
|
|
I d be willing to add this too. )","enhancement"
|
|
"6013","4836","glibc detected double free or corruption error on numpy.percentile.
|
|
Hi, I ve got this error while trying to run `np.percentile` on a series of images in a loop .
|
|
So I tried running it in different ways .
|
|
But, I have no problem in running it outside loop .
|
|
It s just when I used `np.percentile` in loop then things got messy.
|
|
I used numpy 1.9.0.dev c24cc4e in this case, but on my other machine with numpy 1.8.0 has no problem at all.
|
|
Best regards, Habib","bug"
|
|
"6013","4838","Issues when compiling using MKL on Mac OS X. I ve managed to compile using Intel MKL on Mac OS X with Python 3.4 using the following shell script .
|
|
Issue 1 Patching `numpy distutils intelccompiler.py` should not be required.
|
|
And the documentation for setting the compiler flags and `site.cfg` is almost completely absent.
|
|
Issue 2 `export DYLD LIBRARY PATH` should not be required and according to this page https software.intel.com en us articles linking and redistributing intel dynamic libaries on mac os x it can be fixed by using `install name tool`.
|
|
Issue 3 There are a few test errors, see the [build test log] https gist.github.com AndreasMadsen 5f01871d4b1d0c60178a).","bug"
|
|
"6013","4846","masked arrays with subfields throws ValueError on accessing array elements.
|
|
Consider the following program import numpy as np import numpy.ma as ma if name == main works Y = ma.array np.array [ 1,2)], dtype=[ ,int), ,float)]), mask=[ True, False)]) print Y[0] breaks Y = ma.array np.array [ 1,[2,3])], dtype=[ ,int), ,float, 2,))]), mask=[ True, False, True))]) print Y[0] The 2nd construction of the array Y tries to create a numpy array from the mask, which doesn t quite work .
|
|
Is there a workaround for the time being?
|
|
Thanks!","bug"
|
|
"6013","4848","np.linalg.
|
|
umath linalg.svd m has wrong complex type signature.
|
|
numpy.linalg.svd ) works well in the case of fat 2d arrays, but numpy.linalg.cond ) reports an error about signature .
|
|
I don t know how to fix it, so I simply avoid the Fat case.
|
|
def my cond a, p=None) Computes the condition number of a matrix 2D Array).
|
|
Call numpy.linalg.cond and avoid its Bug in the case of fat 2D Arrays.
|
|
See also numpy.linalg.cond nrows, ncols = a.shape cn = np.linalg.cond a.T, p) if nrows < ncols else np.linalg.cond a, p) return cn EDIT seberg) The error is that `np.linalg.
|
|
umath linalg.svd m` should have a complex >float signature just like `svd n`.","bug"
|
|
"6013","485","unittest failures under OS X Trac 1).
|
|
Original ticket http projects.scipy.org numpy ticket 1 on 2006 01 06 by trac user a.h.jaffe ..., assigned to unknown.
|
|
Python 2.4.1 Ippolitto s framework), OS X 10.4.1, gcc 3.3, numpy 0.9.3.1840 Sorry for the weird wikified markup...) In [5] numpy.test 1,1) Found 2 tests for numpy.core.umath Found 21 tests for numpy.core.ma Found 6 tests for numpy.core.records Found 3 tests for numpy.distutils.misc util Found 3 tests for numpy.lib.getlimits Found 9 tests for numpy.lib.twodim base Found 44 tests for numpy.lib.shape base Found 4 tests for numpy.lib.index tricks Found 42 tests for numpy.lib.type check Found 3 tests for numpy.dft.helper Found 6 tests for numpy.core.defmatrix Found 33 tests for numpy.lib.function base Found 0 tests for main ............F..............................................................................................................................F...FF................................. ====================================================================== FAIL Test of masked element Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test ma.py , line 508, in check testMasked self.failUnlessRaises Exception, lambda x,y x y, masked, masked) AssertionError Exception not raised ====================================================================== FAIL check basic numpy.core.defmatrix.test defmatrix.test algebra) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test defmatrix.py , line 111, in check basic assert allclose mA i).A, B) AssertionError ====================================================================== FAIL check basic numpy.core.defmatrix.test defmatrix.test properties) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test defmatrix.py , line 34, in check basic assert allclose linalg.inv A), mA.I) AssertionError ====================================================================== FAIL check comparisons numpy.core.defmatrix.test defmatrix.test properties) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy core tests test defmatrix.py , line 50, in check comparisons assert all mB == matrix A 0.1)) AssertionError Ran 178 tests in 1.136s FAILED failures=4) Out[5] <unittest.TextTestRunner object at 0x16a7850>","bug"
|
|
"6013","4851","numpy.linalg.norm doesn t work well with int16 inputs.
|
|
ex In [18] norm 1000000 np.ones 2).astype int16 )) Out[18] nan In [19] norm 1000000 np.ones 2)) Out[19] 1414213.562373095","bug"
|
|
"6013","4853","Universal OS X build segfault 11 in arch i386 mode.
|
|
I am getting a segfault 11 when I do a numpy.frombuffer on a buffer that s not 16 byte aligned when I start Python on the i386 architecture in OS X.
|
|
The exact same script works on the 64 bit architecture arch i386 python Python 2.7.6 v2.7.6 3a1db0d2747e, Nov 10 2013, 00 42 54) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> b = 00 str np.arange 20).astype np.float64).data) >>> np.version.version 1.10.0.dev 251acc0 >>> np.frombuffer b[2 ]) Segmentation fault 11 and on the 64 bit architecture python Python 2.7.6 v2.7.6 3a1db0d2747e, Nov 10 2013, 00 42 54) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> b = 00 str np.arange 20).astype np.float64).data) >>> np.version.version 1.10.0.dev 251acc0 >>> np.frombuffer b[2 ]) array [ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13., 14., 15., 16., 17., 18., 19.])
|
|
OS X version is 10.9.
|
|
I found this on 1.8.0, tested on 1.9 and master branch with the same result.
|
|
A colleague reports that 1.7 does not have the problem and that the problem is reproduced on his machine with 1.8.
|
|
I am guessing that it s SSE related since the problem only happens if the buffer size is greater than 32 bytes and that s probably when the conditional code decides to use SSE instead of the regular instructions for copying.","bug"
|
|
"6013","4855","BUG subclass with numpy ufunc segfaults with recarray comparison.
|
|
If I define a `ndarray` subclass that implements ` numpy ufunc `, it segfaults whenever it contains a record array and a comparison is done see below).
|
|
Note that any direct calls to `ufunc` give the usual `TypeError`, and the segfault can be avoided by defining one s own ` eq `, ` ne `.
|
|
.","bug"
|
|
"6013","4857","Record arrays with fields that contain multiple objects will segfault on array.resize.
|
|
This short script will always segfault for me.
|
|
.","bug"
|
|
"6013","4858","v1.8.1 runtests.py throw 1 error cannot load test array from pyobj ext.so.
|
|
compile under ubuntu with openblas 0.2.9 .","bug"
|
|
"6013","4859","triu generates nan when called on matrices with infinite values.
|
|
In numpy 1.8.1, I ve just found some strange behaviour.
|
|
Summary when calling np.triu on a matrix, inf entries in the discarded region get converted to nan.
|
|
This broke my code unexpectedly.
|
|
.
|
|
Also posted on stackexchange https stackoverflow.com questions 24678932 numpy triu generates nan when called on matrices with infinite values","bug"
|
|
"6013","486","linalg failures cholesky decomposition) Trac 2).
|
|
Original ticket http projects.scipy.org numpy ticket 2 on 2006 01 06 by trac user a.h.jaffe ..., assigned to atmention pearu.
|
|
Problem with cholesky decomposition under 0.9.3.1840 but problem existed at least since 1837), 2.4.1, OS X framework build of 2.4.1 Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy linalg linalg.py in cholesky decomposition a) 127 if results[ info ] > 0 128 raise LinAlgError, Matrix is not positive definite Cholesky decomposition cannot be computed > 129 return transpose triu a,k=0)).copy ) 130 NameError global name triu is not defined triu ) is found in numpy lib twodim base; I suppose this means it s not being imported correctly here.","bug"
|
|
"6013","4860","Inconsistent handling of arguments to numpy.any ).
|
|
I m using numpy version 1.8.1 on Ubuntu 14.
|
|
I ve run into the following problem >>> a = 1, 2, 3) >>> np.asarray a).any ) >>> True >>> a = 1 , 2 , 3 ) >>> np.asarray a).any ) Traceback most recent call last) File <stdin> , line 1, in <module> File usr lib python2.7 dist packages numpy core methods.py , line 33, in any keepdims=keepdims) TypeError cannot perform reduce with flexible type This seems like a bug or at the very least, this inconsistency is not apparent from the documentation.","bug"
|
|
"6013","4862","The array wrap interface is inconsistent.
|
|
I think that certain functions should return the same type as the input, if they implement ` array wrap `, but aren t necessarily subclass of `ndarray`.
|
|
The pandas `Series` class is an instance of this.
|
|
We d like it if functions like `where`, `diff`, and a few others returned a `Series` object similar to how ufuncs work.
|
|
Here are the current functions that don t behave in the desired way `zeros like` `ones like` `empty like` `sort` `diff` `where` Here s a related pandas issue https github.com pydata pandas issues 6539 There may be others as well.
|
|
Here s an example .","bug"
|
|
"6013","4863","numpy.unique numpy1.8, python3) doesnt work with return index=True flag.
|
|
Script below doesnt give expected output in numpy 1.8.0, while it does in numpy 1.6.1 numpy 1.7.1 .","bug"
|
|
"6013","4864","Bug misfeature in NumPy s buffer interface implementation for array scalars.
|
|
.
|
|
Originally reported here https bitbucket.org mpi4py mpi4py issue 6 incorrect reduce result on python 3","bug"
|
|
"6013","4867","BUG polyval does not substitute variable names.
|
|
For example .
|
|
I would expect the result of polyval to produce the same output as .
|
|
since x is being substituted by y","bug"
|
|
"6013","4868","In place operations on 0 strided arrays.
|
|
I think 0 strided arrays could and should behave more consistently with their non 0 strided counter part.
|
|
The following is the actual output .
|
|
I understand why it is acting like this, but I think it would be more consistent if it acted like this .","bug"
|
|
"6013","487","Zero rank arrays are printed with full precision Trac 3).
|
|
Original ticket http projects.scipy.org numpy ticket 3 on 2006 01 08 by trac user ndarray ..., assigned to trac user sasha.
|
|
In svn revision 1855 >>> from numpy import >>> print array [4.7599999999999998]) [ 4.76] >>> print array 4.7599999999999998) 4.7599999999999998","bug"
|
|
"6013","4876","numpy.min accepts non integer axis argument.
|
|
import numpy numpy.min 1,0.5) returns the value 1.
|
|
It seems to me that `min )` should throw an exception if the axis argument is not an integer.","bug"
|
|
"6013","4879","Error unpickling unicode in Python 3.
|
|
Numpy 1.8.1 used with Python 3 gives an error when unpickling a numpy unicode object which was pickled with Python 2.
|
|
The bug is in the numpy.core.multiarray.scalar dtype,string) routine which is used to unpickle this type of numpy object.
|
|
In Python 3 passing the second argument of scalar ) as a string causes an error TypeError initializing object must be a string ).
|
|
The scalar ) call works in Python 3 only if the second argument is a byte array.
|
|
In Python 2 the scalar ) routine works with a string as the second argument and also with a byte array as second argument).
|
|
The error is in file numpy core src multiarray multiarraymodule.c in the array scalar ) routine at line 1874 where it uses PyString Check obj) on the second argument then gives the error initializing object must be a string .
|
|
In Python 3 this accepts only a byte array, while in Python 2 it accepts a string.
|
|
To check for a string in Python 3 I believe requires PyUnicode Check ) while checking for a byte array uses PyBytes Check ) and PyString Check ) from Python 2 has been eliminated.
|
|
I m not clear on how the numpy code compiles using Python 3 with that PyString Check ).
|
|
Here is a test case that demonstrates the bug.
|
|
It uses Python 2 to create the pickle string, and Python 3 to unpickle it.
|
|
It is necessary to create the pickle string with Python 2 because the handling of unicode by numpy changes in Python 3.
|
|
Pickle is documented as being backwards compatible between all Python versions.
|
|
This error causes PyTables, an HDF5 file interface to load unicode string data incorrectly returning a pickle string instead of unicode string due to the error reported here).
|
|
Test case.
|
|
python2.7.5 >>> import numpy >>> numpy.
|
|
version 1.6.2 >>> import pickle >>> p = pickle.dumps numpy.unicode0 abc )) >>> pickle.loads p) u abc >>> p cnumpy.core.multiarray\nscalar\np0\n cnumpy\ndtype\np1\n S U3 \np2\nI0\nI1\ntp3\nRp4\n I3\nS < \np5\nNNNI12\nI4\nI0\ntp6\nbS a\\x00\\x00\\x00b\\x00\\x00\\x00c\\x00\\x00\\x00 \np7\ntp8\nRp9\n.
|
|
python3.4 >>> import numpy, pickle >>> numpy.
|
|
version 1.8.1rc1 >>> p = b cnumpy.core.multiarray\nscalar\np0\n cnumpy\ndtype\np1\n S U3 \np2\nI0\nI1\ntp3\nRp4\n I3\nS < \np5\nNNNI12\nI4\nI0\ntp6\nbS a\\x00\\x00\\x00b\\x00\\x00\\x00c\\x00\\x00\\x00 \np7\ntp8\nRp9\n.
|
|
>>> pickle.loads p) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError initializing object must be a string The unpickle operation pickle.loads p)) results in the following numpy call which is causing the error >>> numpy.core.multiarray.scalar numpy.dtype U3 ), a\x00\x00\x00b\x00\x00\x00c\x00\x00\x00 ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError initializing object must be a string If we change the second argument of the scalar ) call to a byte array it works correctly >>> numpy.core.multiarray.scalar numpy.dtype U3 ), b a\x00\x00\x00b\x00\x00\x00c\x00\x00\x00 ) abc But the second argument is encoded in the pickle string as a string, not a byte array, so the numpy scalar ) routine must accept a string as the second argument.
|
|
Here is the associate PyTables bug report https github.com PyTables PyTables issues 368","bug"
|
|
"6013","488","numpy.distutils include dir patch when using site.cfg patch included) Trac 4).
|
|
Original ticket http projects.scipy.org numpy ticket 4 on 2006 01 22 by trac user strawman, assigned to pearu.
|
|
!python Index numpy distutils command build ext.py =================================================================== numpy distutils command build ext.py revision 1923) numpy distutils command build ext.py working copy) 32,6 32,18 def finalize options self) old build ext.finalize options self) if self.distribution.include dirs is not None Make sure we re not affected by an apparent) Python distutils bug which, when self.include dirs is present, ignores self.distribution.include dirs.
|
|
This case arises when using setup.cfg or .pydistutils.cfg or distutils.cfg) file with include dirs present, which then causes distutils to ignore the directories specified by numpy in self.distribution.include dirs.
|
|
for d in self.distribution.include dirs if d not in self.include dirs print inserting dir ,d self.include dirs.insert 0,d) self.set undefined options config fc , fcompiler , fcompiler )) return","bug"
|
|
"6013","4884","Change Raise and assertion in doc strings..
|
|
In the file `numpy testing utils.py`, there are several instances of Raise an assertion .
|
|
These end up in the docs here, I think, http docs.scipy.org doc numpy reference generated numpy.testing.assert array equal.html I propose these instances be changed to either Raise an AssertionError Raise an exception I think the first one is the best choice.
|
|
I ll be happy to fix and send a pull request.","bug"
|
|
"6013","4885","ipython ` pylab=qt` option breaks f2py routine.
|
|
Note This issue was first posted on the ipython issue tracker [here] https github.com ipython ipython issues 6158), but was refereed here because `f2py` is part of `numpy`.
|
|
Summary This is a strange one I am calling a fortran routine that is compiled with `f2py` and reads numbers from a file.
|
|
The number in the file is for example 449.99996900000002.
|
|
A python script `example.py` calls this routine and gets that number as return value.
|
|
Now when I run `.
|
|
example.py`, everything works I get `449.99996900`).
|
|
When I go to `ipython` and call ` run example.py` everything works.
|
|
When I go to `ipython pylab=tk` and call ` run example.py` everything works.
|
|
When I go to `ipython pylab=osx` and call ` run example.py` everything works.
|
|
When I go to `ipython pylab=qt` and call ` run example.py` I get `449.00000000`!!
|
|
Details I m running `ipython 2.1.0`, `numpy 1.8.1`, gcc version 4.8.2.
|
|
The fortran code `examplemodule.f90` is compiled with f2py m examplemodule c examplemodule.f90 And here are the files, see also [this gist] https gist.github.com a6073758 b36a50e91e3c206bcfa6) example.py !
|
|
usr bin env python import examplemodule res = examplemodule.testsub ) print { 13.8f} .format res)) examplemodule.f90 program test implicit none double precision result call testsub result) write , ) result = ,result end program test subroutine testsub result) implicit none integer nf,i double precision,intent out) result double precision, allocatable table ) open unit=1,file= example.dat ) nf = 1 read 1, ) nf allocate table nf)) do i=1,nf read 1, ) table i) write , ) table ,i, )= ,table i) enddo result = table 3) end subroutine test sub example.dat 9 100.000000 200.000000 449.999969 499.999969 850.000000 999.999939 1300.00000 2999.99976 7000.00000","bug"
|
|
"6013","489","Eigenvalue hangs Trac 5).
|
|
Original ticket http projects.scipy.org numpy ticket 5 on 2006 01 23 by trac user boisgera ..., assigned to unknown.
|
|
The numpy eigenvalues method may hang forever.
|
|
A platform that exhibits this behavior numpy 0.9.4 from the tar.gz) Linux Mandriva gcc 4.0.1).
|
|
no optimized BLAS version used.
|
|
Greg Landrum pointed out that it may be a gcc 4.0 related problem and proposed a workaround to add the option ffloat store to CFLAGS.
|
|
Works for me !","bug"
|
|
"6013","4890","DOC some missing functions.
|
|
I was surprised that `ascontiguousarray` and `asarray chkfinite` are not in http docs.scipy.org doc numpy reference routines.array manipulation.html changing kind of array.","bug"
|
|
"6013","4893","numpy.argmax and numpy.argmin don t have an out keyword argument to specify the output array..
|
|
This looks like a very easy fix, as numpy.argmax and numpy.argmin just call numpy.ndarray.argmax and numpy.ndarray.argmin, both of which do allow the output array to be specified.","enhancement"
|
|
"6013","4894","1.9.0b1 Strange FutureWarning for object array containing np.float32 .
|
|
In Numpy 1.9.0b1, I get a FutureWarning when I create and print an object array containing None and a np.float32 or np.float64 object.
|
|
With np.float, it works ok.
|
|
Platform is Windows 7 64 bit, Python is 2.7.5. .
|
|
Best regards, Martin Gfeller","bug"
|
|
"6013","4895","Masked array mean reports spurious floating point error underflow).
|
|
When numpy is set to raise exception on floating point errors, taking the mean of masked arrays sometimes raises spurious errors, even when no values are masked.
|
|
For example .","bug"
|
|
"6013","4896","Numpy fails to build with MSVC 2013 because of incomplete C99 support.
|
|
I m trying to build Numpy with MSVC 2013 but it fails when building npy math.c with .
|
|
MSVC 2013 has complex.h, but evidently it is incomplete.
|
|
How can I disable the use of C99 complex?","bug"
|
|
"6013","4898","DOC is masked.
|
|
I couldn t find `is masked` https github.com numpy numpy blob maintenance 1.9.x numpy ma core.py L5833) on http docs.scipy.org doc numpy reference routines.ma.html.
|
|
SciPy is using this wrong and having bugs.","bug"
|
|
"6013","490","Python API calls inside of LOOP BEGIN THREADS LOOP END THREADS Trac 6).
|
|
Original ticket http projects.scipy.org numpy ticket 6 on 2006 01 24 by dmcooke, assigned to dmcooke.
|
|
ufuncobject.c uses LOOP BEGIN THREADS LOOP END THREADS inside of PyUFunc GenericFunction to release the GIL; however, it calls the Python API inside of that like Py DECREF).
|
|
Should be fixed at some point.","bug"
|
|
"6013","4900","Issues importing numpy.
|
|
Admittedly I am very new to numpy and python.
|
|
I have installed MacPorts and added the XCode Commandline to the terminal and using the suggested command sudo port install py27 numpy py27 scipy py27 matplotlib py27 ipython notebook py27 pandas py27 sympy py27 nose I have tired to install numpy.
|
|
However, whenever I try to import numpy within the terminal I get the following error Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy init .py , line 137, in <module> import add newdocs File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib init .py , line 4, in <module> from type check import File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core init .py , line 5, in <module> import multiarray ImportError dlopen Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core multiarray.so no matching architecture in universal wrapper My version of python according to terminal when I call python) is >>> print sys.version) 2.7.6 v2.7.6 3a1db0d2747e, Nov 10 2013, 00 42 54) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] Any guidance on this one?
|
|
What have I done wrong?","bug"
|
|
"6013","4903","NaN handling in npy ObjectMax ) and npy ObjectMin ) in Python 2.
|
|
On Python 2, when one of the arguments is NaN, these ultimately dispatch to `default 3way compare )`, yielding the comparison of the addresses and, thus, an arbitrary result.
|
|
I propose to eliminate the ifdef check for `NPY PY3K` at [numpy core src umath funcs.inc.src 59] .. tree master numpy core src umath funcs.inc.src) and use the first branch.
|
|
`PyObject RichCompareBool )` is available in 2.6 and 2.7, so this should cause no portability issues.
|
|
As an example, consider that the result of .
|
|
will be `1` or ` 1` depending on the addresses of the two arguments.
|
|
The effect will be that the return value of `np.maximum )` or `np.minimum )` might change for temporary arrays like, for example, when used from Pandas) even in cases when they have identical contents.
|
|
That is, for `DafaFrame` `df` of floats using the `object` `dtype`, the result of .
|
|
or any operation that causes reallocation of NumPy data, might change run to run.","bug"
|
|
"6013","4909","test failures when building win32 binaries with mingw based on gcc 4.8. our current win32 binaries are built with mingw based on gcc 3.4.1 and succeed all their tests.
|
|
When one builds with mingw based on gcc 4.8 via the binaries from the official mingw site) or with mingw64 one gets a few test failures probably related to accuracy of the libm math functions .
|
|
the last one is especially worrying as the results should be exactly the same, but they are to about 40 different by about one ulp.
|
|
The issue does not occur when one only runs the random testsuite and not everything.
|
|
I don t know what is happening here, could it be rounding mode changes that are somehow sticky to a thread?
|
|
But to my knowledge numpy does not change the rounding mode anywere.
|
|
I have also updated the atlas version used using carls binaries from here https nipy.bic.berkeley.edu scipy installers atlas builds maybe it now plays around with rounding mode?","bug"
|
|
"6013","491","Segfaults when trying to build an array using the description in array protocol Trac 7).
|
|
Original ticket http projects.scipy.org numpy ticket 7 on 2006 01 30 by trac user faltet ..., assigned to unknown.
|
|
Apparently something wrong went in latest version of numpy SVN In [4] numpy.
|
|
version Out[4] 0.9.5.2021 In [5] numpy.empty 1,), dtype= i8 ) Out[5] array [ 5193749191705487200], dtype=int64) In [6] numpy.empty 1,), dtype=[ p , i8 )]) Segmentation fault This usually worked well with older versions In [7] numpy.
|
|
version Out[7] 0.9.5.1980 In [8] numpy.empty 1,), dtype=[ p , i8 )]) Out[8] array [ 5192306666809247592L,)], dtype= void,8)) Thanks","bug"
|
|
"6013","4911","meshgrid should return a tuple instead of a list.
|
|
The original motivation of this report is the fact that the interpolation functions in [`scipy.interpolate`][interpolate] treat output grid arguments specially when they are given as tuples.
|
|
The explicit conversion is done by function ` ndim coords from arrays` in [interpnd.pyx][interpnd].
|
|
However, it seems that `tuple` is a more suitable type for the iterable returned by `meshgrid` anyway, because of the immutability and the stronger connotations of the order of the elements.
|
|
Another option will be to return a generator.
|
|
I do not know if this will create some incompatibilities with existing functions in scipy.
|
|
[interpolate] http docs.scipy.org doc scipy reference interpolate.html [interpnd] https github.com scipy scipy blob master scipy interpolate interpnd.pyx","bug"
|
|
"6013","4912","Running `setup.py egg info` starts to compile everything.
|
|
This is very annoying especially when you want to use pip to just download numpy.
|
|
`egg info` doesn t need to compile anything","bug"
|
|
"6013","4914","Automatic creation of new axis where necessary.
|
|
A lot of the confusion between matrix operations and tensor operations seems to come down to the differences between one dimensional tensors, and matrices that single dimensional.
|
|
The squeeze method and atleast 2d function are great for helping sort this out, but would it be possible to create automatic broadcasting such that if a shape n,) and shape [ n,1) or 1,n) ] array need to interact, they can be automatically broadcast?
|
|
i.e.
|
|
automatically apply squeeze or atleast 2d or c )","enhancement"
|
|
"6013","4915","Datetime scalar None comparison.
|
|
Comparing a datetime array with None using np.equal, but at some point also `==`), will give that `NaT` is equal to `None`.
|
|
Doing the same with the datetime scalar does currently always evaluate to False hardcoded, but array == None used to be just False).","bug"
|
|
"6013","492","arange doesn t handle step size of 0.1 properly Trac 8).
|
|
Original ticket http projects.scipy.org numpy ticket 8 on 2006 02 08 by trac user mtreiber ..., assigned to trac user sasha.
|
|
arange 1,3,1) returns array [1,2]) which is proper but arange 1,1.3,.1) returns array [1.,1.1,1.2,1.3]) which should have ended at 1.2.","bug"
|
|
"6013","4922","docs are misleading for numpy matrix flatten.
|
|
I have the pleasure to work with sums along axes of scipy sparse matrices which conveniently give `np.matrix` as output, but sadly I find myself searching to convert this most elegant object to a plain ndarray for compatibility with some old code.
|
|
The functions `flatten )` or `ravel )` were suggested, but the matrix is too noble to be fazed by these attempted debasements .
|
|
Under the most charitable technical interpretation, these docs could be argued to not be incorrect, but I feel that they would be improved by noting that they return numpy matrix objects.
|
|
So while in the exalted `np.matrix` land it would not be inconsistent to have a copy of the input array, flattened to one dimension that has ndim=2, some clarification could be helpful for those of us unfortunate enough to have to stray into the realm of base ndarrays.
|
|
http docs.scipy.org doc numpy reference generated numpy.matrix.ravel.html http docs.scipy.org doc numpy reference generated numpy.matrix.flatten.html","bug"
|
|
"6013","4923","concatenate error could be more helpful.
|
|
.
|
|
Actually the problem is that the first argument needs to be a tuple, and it s interpreting `[2]` as the axis argument.
|
|
.
|
|
I d fix it myself, but it s a C module and I don t have any idea how to work with those.","bug"
|
|
"6013","493","Extensive output from setup.py for invalid command Trac 9).
|
|
Original ticket http projects.scipy.org numpy ticket 9 on 2006 02 11 by matthew brett, assigned to unknown.
|
|
When running python setup.py invalid command distutils runs extensive library etc checks and outputs to stdout) before coming up with the expected error invalid command invalid command","bug"
|
|
"6013","4933","Undefined variable in numpy compat inspect.py.
|
|
.","bug"
|
|
"6013","4934","Audit Fixmes in numpy lib, numpy lib tests.
|
|
Just so, a task.","bug"
|
|
"6013","4939","savetxt cannot handle unicode objects.
|
|
I apologise if I m using this incorrectly, but I could not find anything relevant in the issue list, and as far as I can work out, this really should work.
|
|
If I pass an array with unicode objects instead of string objects) into savetxt, it bombs out, trying to run the unicode object through the asbytes function.
|
|
It does not change things if the dtype of the array is U instead of S .
|
|
>>> import numpy as np; np.savetxt test.csv , np.array [ this wxc3xb6rks , u this dxc3xb6s not ]), fmt=[ s ]) UnicodeDecodeError Traceback most recent call last) <ipython input 5 6a3125624190> in <module> ) > 1 import numpy as np; np.savetxt test.csv , np.array [ this wxc3xb6rks , u this dxc3xb6s not ]), fmt=[ s ]) UnicodeDecodeError ascii codec can t decode byte 0xc3 in position 6 ordinal not in range 128)","bug"
|
|
"6013","494","All libraries silently ignored when invalid library included in site.cfg Trac 10).
|
|
Original ticket http projects.scipy.org numpy ticket 10 on 2006 02 11 by matthew brett, assigned to unknown.
|
|
Given this site.cfg file [atlas] atlas libs = lapack, f77blas, cblas, atlas, wrong lib numpy setup rejects the whole list of libraries silently because of this test in system info.py, round line 483 if len found libs) == len libs) It would seem better to generate and error or warning here.","bug"
|
|
"6013","4940","Inconsistent advanced indexing behavior.
|
|
The following comments are partially based on impressions triggered by the following part of the NumPy documentation >For the discussion below, when the selection object is not a tuple, it will be referred to as if it had been promoted to a 1 tuple, which will be called the selection tuple.
|
|
To avoid confusion, by wrap into a tuple I mean a > a,), whereas by converting to a tuple, I mean a >tuple a).
|
|
Let s start with some array .
|
|
If an indexing object is not a tuple, it is wrapped in a tuple .
|
|
If the indexing object is a list, we have advanced indexing .
|
|
Providing the trivial tuple wrap ourselves, we get the same result .
|
|
If the indexing object has several levels of nesting, so does the result .
|
|
Relying on the automatic wrapping into a tuple, we might want to remove the explicit tuple construction .
|
|
This reveals the inconsistency the result has changed as if the outer list were actually reinterpreted as a tuple.
|
|
This theory is consistent with the following behavior .
|
|
Explicitly wrapping into a tuple makes it work .
|
|
Alternatively, explicitly converting to an array also works .
|
|
It would be logical to convert the object to a tuple only if the result obtained by wrapping the object was not suitable for indexing.
|
|
Even better would be to actually not backtrack too much while trying to make sense of the indexing object, as it makes the behavior really hard to understand.
|
|
The following behavior seems to be the instance of the same problem .
|
|
In my opinion either the order in which various interpretations are tried should be strict or, even better but problematic due to backwards compatibility) raise an error if there is too much ambiguity.","bug"
|
|
"6013","4942","matrix squeeze axis error message.
|
|
This error message is misleading, but probably low priority to fix.
|
|
.","bug"
|
|
"6013","4944","Numpy and Scipy mailing lists down?.
|
|
Recent activity looks sparse.","bug"
|
|
"6013","4945","Should numpy matrix methods use super instead of hardcoding ndarray?.
|
|
https github.com numpy numpy pull 4943 discussion r15767060","bug"
|
|
"6013","495","On entry to DGESDD parameter number 12 had an illegal value Trac 11).
|
|
Original ticket http projects.scipy.org numpy ticket 11 on 2006 02 14 by trac user An.Ghysels , assigned to unknown.
|
|
When executing the example program below, this error is reported On entry to DGESDD parameter number 12 had an illegal value import Numeric a = Numeric.array [[170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 221.77542342278846, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0], [0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 221.77542342278846, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0], [0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 221.77542342278846, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 170.75402535544072, 0.0, 0.0, 225.82919075776189, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539, 0.0, 0.0, 42.862124266509539], [0.0, 430.62651003812533, 41.792976261168995, 0.0, 1259.6122873233351, 0.32177351244113606, 0.0, 13.36660991807039, 42.049877912133709, 0.0, 204.61808135206056, 533.23531889447372, 0.0, 354.34987991088957, 1213.5508784482399, 0.0, 360.02513402386285, 264.18482608285728, 0.0, 1164.4846853541821, 573.35004204216273, 0.0, 3.3687618972243789, 247.65562798081521, 0.0, 330.88431180069279, 637.34968410286876, 0.0, 49.382582103036889, 82.50998360363846, 0.0, 278.33150404531671, 111.60515879785029, 0.0, 277.65817036476915, 67.837740583176029, 0.0, 267.16419246051566, 48.394572451248251, 0.0, 50.791293683944247, 51.52642859127608, 0.0, 180.08561594004769, 140.9740251392497, 0.0, 7.2068492193429829, 224.07561576835661, 0.0, 126.50646324602697, 307.9083724186716, 0.0, 36.369495477658766, 277.25140003592162, 0.0, 130.56720073436699, 194.37020408319739, 0.0, 245.23706859884516, 114.71684651655534, 0.0, 287.59601822564008, 30.254737923758171, 0.0, 219.96125781850782, 213.21884927332093], [ 430.62651003812533, 0.0, 420.87549320170541, 1259.6122873233351, 0.0, 493.82674840511885, 13.36660991807039, 0.0, 98.587025268296301, 204.61808135206056, 0.0, 55.489703871464329, 354.34987991088957, 0.0, 391.07516261665461, 360.02513402386285, 0.0, 446.33836389824495, 1164.4846853541821, 0.0, 532.28807175890859, 3.3687618972243789, 0.0, 393.47945655437366, 330.88431180069279, 0.0, 1047.2383550696129, 49.382582103036889, 0.0, 177.41536555551616, 278.33150404531671, 0.0, 83.948745322340017, 277.65817036476915, 0.0, 0.47488937435946671, 267.16419246051566, 0.0, 197.07325852650698, 50.791293683944247, 0.0, 294.07696627271463, 180.08561594004769, 0.0, 179.28714085008767, 7.2068492193429829, 0.0, 213.42766128203638, 126.50646324602697, 0.0, 10.398643639607206, 36.369495477658766, 0.0, 112.30214379978842, 130.56720073436699, 0.0, 169.1947494234181, 245.23706859884516, 0.0, 214.12602324446641, 287.59601822564008, 0.0, 43.410947442619687, 219.96125781850782, 0.0, 48.024737062654474], [ 41.792976261168995, 420.87549320170541, 0.0, 0.32177351244113606, 493.82674840511885, 0.0, 42.049877912133709, 98.587025268296301, 0.0, 533.23531889447372, 55.489703871464329, 0.0, 1213.5508784482399, 391.07516261665461, 0.0, 264.18482608285728, 446.33836389824495, 0.0, 573.35004204216273, 532.28807175890859, 0.0, 247.65562798081521, 393.47945655437366, 0.0, 637.34968410286876, 1047.2383550696129, 0.0, 82.50998360363846, 177.41536555551616, 0.0, 111.60515879785029, 83.948745322340017, 0.0, 67.837740583176029, 0.47488937435946671, 0.0, 48.394572451248251, 197.07325852650698, 0.0, 51.52642859127608, 294.07696627271463, 0.0, 140.9740251392497, 179.28714085008767, 0.0, 224.07561576835661, 213.42766128203638, 0.0, 307.9083724186716, 10.398643639607206, 0.0, 277.25140003592162, 112.30214379978842, 0.0, 194.37020408319739, 169.1947494234181, 0.0, 114.71684651655534, 214.12602324446641, 0.0, 30.254737923758171, 43.410947442619687, 0.0, 213.21884927332093, 48.024737062654474, 0.0], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.32510025124324898, 0.090094508708536913, 0.0096604312147142782, 0.18990650081038948, 0.17751749084309748, 0.085185950204932845, 0.87154155835399882, 0.054359708997793632, 0.046565479474067598, 0.065756856832268326, 0.011981790361578728, 0.065125904181317848, 0.087082371378642762, 0.0083215616331960285, 0.03214658958171767, 0.036631252059709157, 0.0059101877891677172, 0.035268699171737042, 0.062138469748309222, 0.0079123046459139579, 0.019005252682406232, 0.097879591589747836, 0.0068113442757990949, 0.04157394578947958, 0.048877277857657664, 0.0095205253784917988, 0.050165612290965216, 0.012071637362169693, 0.024974309269839103, 0.060596515628860481, 0.013088927902777991, 0.0039172984671967575, 0.014509764611549151, 0.016670447901027022, 0.0085389583740601883, 0.026767862563499219], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.11650992405489624, 0.37482781806225113, 0.046875457563222221, 0.21276431343628147, 0.19905349942999806, 0.20000872884200915, 0.018472847401720501, 0.80574629245012919, 0.043012728293284702, 0.032825432616607871, 0.0041538590727980378, 0.016032592487849294, 0.067676723916331227, 0.062375886746517652, 0.031117185410451158, 0.080095199388261082, 0.0081672507159460152, 0.054722459451910434, 0.083904683105385902, 0.033041121789696454, 0.09533166403629216, 0.06388298815883342, 0.045078213745835365, 0.080539415943076675, 0.051123886635364713, 0.014233287778992864, 0.055440622543443091, 0.031057232210541449, 0.068010163761109063, 0.0024071254242400749, 0.003579854853134955, 0.0074585905247618357, 0.012405272202638384, 0.061458656894372125, 0.026120894525804555, 0.034131319813977254], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.079835885111927948, 0.060646634732822204, 0.31568784700215913, 0.050311587049335083, 0.23066000569338763, 0.078677418859569692, 0.071990131427402376, 0.0082261490864987061, 0.87199189294332657, 0.013113647895808694, 0.012900875882415555, 0.00779587610907742, 0.047660855557435211, 0.0093471548691131407, 0.020417243146460665, 0.00060551348114890188, 0.014606600671345471, 0.0085964613445962332, 0.033252553061769483, 0.019650721503385805, 0.080074471502201816, 0.010616595936770761, 0.014125156064385583, 0.051587658743459573, 0.034832607406817713, 0.019102029910907897, 0.033627480266987787, 0.067124817083111271, 0.024317897438388292, 0.12751542009269773, 0.078208367933773032, 0.024915809666991265, 0.084038697646346847, 0.060715870974357475, 0.022820975867133705, 0.079565070447924505], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.29004963712954257, 0.012261252938769786, 0.0062789884007416049, 0.049708573634465757, 0.01261874152661888, 0.031109112207436156, 0.058027421202097119, 0.012911082384004575, 0.0019820485263956525, 0.93356611344163998, 0.0092599170597686895, 0.0068614291077244679, 0.06756034164892892, 0.003379397574379796, 0.0037932916651307008, 0.074411650158645135, 0.0063215813019909415, 0.0037320180354829115, 0.055126742449367022, 0.00086250597435033149, 0.004431847132994533, 0.053163927720069393, 0.0018932637371669825, 0.0061785828038079284, 0.059888427909861687, 0.0057207828578721863, 0.0065700184590033244, 0.074904025674203387, 0.0090658632660735247, 0.0037677426325675545, 0.07226080455364553, 0.0016511964236465887, 0.00036410444129583985, 0.078464561360813451, 0.0024417830709526664, 0.0013513766797819532], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.068710119968431863, 0.22337798334246595, 0.096214854480634979, 0.071783275048526649, 0.01752794860250332, 0.32478597173397061, 0.012891487173122653, 0.052858016755790799, 0.03023850716304035, 0.026963753813476071, 0.82286497950019033, 0.014496129237455633, 0.068471710264450925, 0.11796398661859284, 0.01041230187317066, 0.1017459036776498, 0.14481373477328791, 0.0040768882168735181, 0.13365312374120175, 0.06509833695191572, 0.044353491485514686, 0.11728667362364974, 0.098518238278684792, 0.040404194331980677, 0.084001743622437716, 0.13492345013203907, 0.033673253040785549, 0.049294776898832826, 0.025062328566380243, 0.019320809848823072, 0.011539504053008701, 0.052618491525471323, 0.022052029517878661, 0.092890534533088306, 0.045943154597400365, 0.009543512537812826], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.085260988920047839, 0.10175372339118355, 0.24193312213887255, 0.13652310339378623, 0.28476345378294193, 0.057046548034366283, 0.090460465776911206, 0.030357303345990894, 0.016018553590191303, 0.012960190784293384, 0.022601245654490568, 0.84368049902425091, 0.059526029931170461, 0.02754176984781195, 0.11155958839703042, 0.0023899604819377232, 0.020731546482803204, 0.12051328335305733, 0.051299544308392578, 0.01507146080705165, 0.06876832483884418, 0.0072204549121113485, 0.021717119715250758, 0.10413055394533667, 0.052562889775231172, 0.015181924501716421, 0.12126031344078217, 0.093196847332956453, 0.009969861482812134, 0.02862057082034368, 0.10854699299166912, 0.0085790969196554157, 0.036016730420122503, 0.083954998828133454, 0.011258401634175342, 0.027091158866373803], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.34339757321583503, 0.016932970234292379, 0.059234505404575047, 0.10398561799795601, 0.12062350257594916, 0.13155643783377718, 0.092105334820038298, 0.031469829969327005, 0.041683109273144968, 0.079362023630575915, 0.055435868851650881, 0.027376300026135954, 0.88616730998476712, 0.03568021450030251, 0.010439856582020271, 0.10423905534703566, 0.040607146458298646, 0.014702772017018197, 0.010539085516860358, 0.0031075331406771453, 0.0103814868080719, 0.018116230812278547, 0.022605069547346514, 0.0038768581016597452, 0.0093269453073015259, 0.031670509761095811, 0.011373112125473693, 0.048690747674277508, 0.038156679821423789, 0.045256559314892683, 0.027466325024138788, 0.0077179662387096603, 0.020411491181879033, 0.070016541672189889, 0.0080713338882542618, 0.022358184703521397], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0031473469738122375, 0.27557655760708372, 0.071734242798491402, 0.062907652355945073, 0.10248618614344952, 0.21479253340144738, 0.025821211959498391, 0.024410001484469285, 0.0040251815401680041, 0.0072100021012626277, 0.13933429012223117, 0.025931521351436688, 0.028577164241861516, 0.88009245026586003, 0.031461985202770687, 0.055868728721076552, 0.11299343350128194, 0.040175156288153692, 0.087342591323832153, 0.044416853294807476, 0.015208925325952732, 0.083029999796197851, 0.087878217780151299, 0.0097266817443105898, 0.055950063405627033, 0.087795267636821217, 0.00045656209069132027, 0.032141559942645455, 0.028668979125194408, 0.018992674627719724, 0.0099407201816196639, 0.014378113876938131, 0.01535307071222768, 0.058645385914556793, 0.019492507943820687, 0.032510870041434117], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.087609718831054256, 0.039722872750273995, 0.27201195783081972, 0.0048460755550285796, 0.2252438735710576, 0.043962583878735008, 0.069150443026297359, 0.006620988065477432, 0.004613418640433492, 0.015148276161647265, 0.0077564718748684095, 0.13383379116859212, 0.046044204098810007, 0.0024800603271526805, 0.88651439524680975, 0.0028671953624208437, 0.012305928485150759, 0.12911982513520084, 0.02503850631664516, 0.025854524478462874, 0.032844420001441665, 0.014637191031564136, 0.010689718918031601, 0.060521920669186935, 0.02708003539277214, 0.024747217171226913, 0.081718602155595066, 0.058057395299139773, 0.038292342423708936, 0.014677077966244199, 0.067671587424233626, 0.040420505326931061, 0.049180728884190568, 0.052763428523974663, 0.034555340535033147, 0.0640300689163701], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.3161022085553164, 0.065547824880205316, 0.057648144087746191, 0.0086764213180828709, 0.030862793363176277, 0.2281957886016841, 0.035656618889967601, 0.086504843487136496, 0.0032152690665496969, 0.080194916935045699, 0.073696781445012016, 0.030332751885940843, 0.098992689344650886, 0.032374145577837889, 0.018434182627302419, 0.86442744787050341, 0.051290981795762317, 0.016564874830249456, 0.017671037559172459, 0.0039312957419170074, 0.029544345619097464, 0.018963502786894126, 0.019064784517537261, 0.034681364823909544, 0.017161590035475793, 0.04461576638537585, 0.034307299221021277, 0.11874709611570639, 0.066436720496898374, 0.0059382259223088015, 0.095795416305390582, 0.012372824878436968, 0.0093632126116712806, 0.14708787335309942, 0.017111743600107462, 4.2569750504701827e 05], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.06700593675775425, 0.23136138986781135, 0.063712443846574832, 0.093584030690680761, 0.0045921528791418621, 0.22588979034039955, 0.0021251860934468876, 0.031891568491527356, 0.027348597772714409, 0.022954713327448573, 0.1427744380649936, 0.0036173263715984097, 0.054229845778569132, 0.093826839236796217, 0.0024786727727553149, 0.075281345595047344, 0.87774108397746997, 0.011988467858416667, 0.093370329138301886, 0.073088828390810129, 0.048465382083621424, 0.079412076864968234, 0.087884183160703344, 0.042517211538443143, 0.058266596565529946, 0.12275968873830137, 0.032408576477247465, 0.03446617209994924, 0.015705317857377856, 0.010979435301371299, 0.0071124000508946865, 0.073025764063160428, 0.015032993869570454, 0.065707805743264458, 0.064614050209901075, 0.0037250362895699307], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.12672225734055267, 0.034757465404424308, 0.275554700894524, 0.05782173763284907, 0.30221672056348647, 0.030212381518629518, 0.091688583797049084, 0.02281630267555454, 0.014194704150195636, 0.022552164795819424, 0.0058039169147673505, 0.15595658990722516, 0.061305904744609875, 0.014363981368704313, 0.13550275273812892, 0.0060108865380550771, 0.014649910019252742, 0.83963068148812081, 0.026527561628297752, 0.041026983583731448, 0.0060209624818630047, 0.023698897004070085, 0.011391730554913399, 0.043023225530232023, 0.029740781113473815, 0.038997904270707491, 0.07494164338529298, 0.070774869337096938, 0.0652075620140639, 0.0084189139487017006, 0.082524936319324393, 0.069531265480388502, 0.047152746546243379, 0.064588808189114263, 0.058030266365496207, 0.077041468687475703], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.16545629651704294, 0.0015455181757288463, 0.14719428692430039, 0.02210773362736259, 0.2264851043891821, 0.37640311237169843, 0.024010227644478726, 0.10581143310438024, 0.085123689800728652, 0.027540847462088933, 0.14751170599006441, 0.039875173345157729, 0.050828501291771022, 0.087501795719336847, 0.011813139716352013, 0.049677504254822882, 0.10672292501542768, 0.021279901234230819, 0.74393481692553298, 0.0042850844995657236, 0.035403658399295247, 0.19927720535973109, 0.054722065004839046, 0.0093675962882219209, 0.19904096252640452, 0.085515073755787263, 0.0058400900954178603, 0.069957867273674618, 0.10929599409948909, 0.087762063433024839, 0.12383367066196911, 0.02158092122281445, 0.045532219101575004, 0.013216011399374325, 0.024630715344883966, 0.044827902815710842], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.071422476480266353, 0.24633830474744617, 0.020841403198716649, 0.16924712632955138, 0.040935334321633071, 0.026932696115954825, 0.028521210855039468, 0.014942818685587722, 0.02284478195748918, 0.014623790998746094, 0.054881452624487258, 0.027144977532503955, 0.020837752207886487, 0.028511449624990602, 0.039871242987752561, 0.0096881409323014457, 0.065443310432900095, 0.059825977597617708, 0.0094213071220171658, 0.90277605356782276, 0.06701976557985527, 0.018398810171723561, 0.060549937736027307, 0.054500096801200534, 0.0075954821011281785, 0.095877074349584854, 0.033278969895942, 0.0033594678278839321, 0.12606364892755781, 0.011467043405061257, 0.0046222043860283174, 0.13283732511942029, 0.00304967682752438, 0.0032801723256451433, 0.11826606564145367, 0.042375988801265287], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.11958844532098392, 0.072351676900421963, 0.25018448782224051, 0.28994940988181883, 0.17594733558696349, 0.11805998247561224, 0.047923153446423193, 0.082862334534991727, 0.10201970900301718, 0.024252350019352494, 0.019608967407939704, 0.019078359654267334, 0.033329325001589798, 0.063477533909340272, 0.0032118183960113491, 0.014305731915032916, 0.00076635731490655955, 0.037113459441088348, 0.019849906711959863, 0.055532779216800554, 0.82154321972818933, 0.034071186220017091, 0.0081049173704038826, 0.15758115631817809, 0.015611903352200689, 0.051637919492761503, 0.12162887856674508, 0.0056349165909263116, 0.10702678216315714, 0.043355196659827185, 0.0067346945021561355, 0.11687647967691461, 0.058562786961473948, 0.0066129810126587391, 0.092045162475333786, 0.0086020598960724771], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.21484124632695192, 0.052811422283027563, 0.10849446958204191, 0.10670390999197386, 0.069428297555161028, 0.34331892502117073, 0.07134312069836643, 0.11474352923406111, 0.038176601945858368, 0.036494969894762419, 0.12171537440137883, 0.0080825684857459036, 0.011568656363313669, 0.062950153871974224, 0.01009972999620622, 0.040920294691357488, 0.086405884739523944, 0.0046211605085952033, 0.2117586133663113, 0.0014163419368031637, 0.038882013064760128, 0.81090650027602229, 0.038404400744706295, 0.032196735436514953, 0.15953332343790347, 0.072103887346056009, 0.025539731158122952, 0.022770796672326406, 0.099832577181081794, 0.031813543911609746, 0.063991654669331918, 0.019104456470478742, 0.026653958946681561, 0.024042183727234333, 0.024243076443025906, 0.018758411985033766], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.034454017514052818, 0.25926001721680397, 0.032499119896111817, 0.04832498349162144, 0.033653938082518674, 0.11536257299174932, 0.0010020815398081465, 0.015173361438648163, 0.014043897029317635, 0.011774171208795323, 0.10455476701241567, 0.0017686993267944763, 0.02777575705786646, 0.079462229654083824, 0.001384168523701758, 0.038501372889466298, 0.094101333696971726, 0.0063029606983289787, 0.047689538614396947, 0.069066363505483286, 0.024965824201501792, 0.040529675488452793, 0.92348042604526148, 0.021889113247824301, 0.029755010640757251, 0.094446892146164876, 0.016660551560493845, 0.017604183060548281, 0.039825228145442343, 0.0055773112869422917, 0.0036212072572169318, 0.069139026246525265, 0.0076736905659914069, 0.033570580787090458, 0.064797268900203694, 0.0020285049011975363], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.081732396369739974, 0.10392798778593387, 0.23601181361405768, 0.32011678213724293, 0.063645190151599418, 0.11197591916742317, 0.012752326030367375, 0.088407315658327615, 0.070990690966664174, 0.018113831236170726, 0.026398815716402846, 0.054525380752704605, 0.010072256121923017, 0.069248192698145722, 0.017742135324014509, 0.014250083126495981, 0.0080836391200192776, 0.010834182045440113, 0.036820790165328722, 0.046665776601819746, 0.17271750474252001, 0.02935094564810227, 0.015308229726753587, 0.83427390213808394, 0.033299680629390808, 0.042978004893838566, 0.13813372832632415, 0.027833556998592603, 0.096715539239714751, 0.011559088151734858, 0.032257118526259405, 0.10643640873941726, 0.048769967805204678, 0.023633797284872061, 0.082223641382392063, 0.017318144667123926], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.18905669879349823, 0.027399359771030654, 0.10912617111516049, 0.0026070947712699888, 0.21536097969391854, 0.2469979778046513, 0.016693095488241654, 0.060561780022956528, 0.075063257630440361, 0.037534854269774869, 0.10304176596424768, 0.047798083832475621, 0.025287005152436315, 0.065623950696939437, 0.017909867825517737, 0.0097376891490180417, 0.075354195183933395, 0.02565784382167231, 0.2036372887213477, 0.0054030565765140334, 0.020025165141843278, 0.1521966716596837, 0.041508912877600342, 0.0053627602580016706, 0.83367857484891772, 0.058984036445618962, 0.018852623312413285, 0.090874441130511741, 0.071650990800480957, 0.081062021045853236, 0.13008186852551051, 0.014443574100955548, 0.037117591012904148, 0.051235445333073608, 0.015297952897572952, 0.040184950908530338], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.10739374389908826, 0.21143658222547282, 0.032923331707800423, 0.20416962020359544, 0.064023441828012909, 0.15333784663085989, 0.020594886580077049, 0.048617991839352739, 0.038916481804317518, 0.029116366951856022, 0.11731964081390987, 0.018370393932919792, 0.058097891974605048, 0.058985493827926831, 0.032996442577064924, 0.065652564979334221, 0.10972040767393405, 0.055892438702222366, 0.06471540199276217, 0.096548587188468024, 0.089652401329197703, 0.046944041127737264, 0.079376236853422891, 0.075301504100163119, 0.039046875821587924, 0.86616705537690442, 0.050956067485500203, 0.023980854875135228, 0.074745169235369266, 0.0004660917869522658, 0.0018856673712654357, 0.12426398250875741, 0.0092244642377077666, 0.048153962514321517, 0.10641238871700881, 0.035911037034667491], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.036731614518160739, 0.1075205885569102, 0.23558591367654313, 0.27999617270166566, 0.034584312223391298, 0.098003475598275727, 0.016595115182363802, 0.075201335120305815, 0.046070780475915112, 0.0098581257905631676, 0.026372325234497458, 0.083568414184625628, 0.0096574227799698532, 0.059993085095937607, 0.042982853383536579, 0.011397123871162618, 0.01207178630773002, 0.024105329925397028, 0.042506062608685335, 0.030697250945190959, 0.14702821197730612, 0.01996100066936686, 0.01778454219302052, 0.15200817179455772, 0.04022634036782876, 0.027894798581744342, 0.86376433643279382, 0.047877607680238896, 0.0697740071381132, 0.0022245182391070228, 0.055642558531097591, 0.077479372006024685, 0.044433993238061693, 0.042048276626895351, 0.058513921613937127, 0.0077982895823179871], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.22612463154041834, 0.13654154455084949, 0.023451969619194664, 0.12756255159061272, 0.28242475407533613, 0.10645752070451639, 0.034384913036154777, 0.075987114760196539, 0.075625178899177525, 0.060788772566623864, 0.009306342271622723, 0.10102342028258322, 0.023794520788497368, 0.017846946679619387, 0.049394237950116299, 0.10027216125777727, 0.0025552654151554427, 0.054690030990331442, 0.079964017194919079, 0.012240054315620211, 0.026661926768264198, 0.022430545173317391, 0.013595262697871556, 0.062402717705355215, 0.096630821396038652, 0.0092941122514754414, 0.076326248808063227, 0.81879729361459974, 0.03127384393221487, 0.097033420516041427, 0.1859721085255705, 0.004639466399621487, 0.024703852329344692, 0.18476717979820464, 0.011456240677596159, 0.043226640436438903], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.065732080729859976, 0.26912019939831899, 0.11103014227719796, 0.22768167001637751, 0.075539686416519244, 0.30183677906874301, 0.058109235620714829, 0.073958813608769228, 0.014498267137668939, 0.0032713719775990038, 0.040625819826027816, 0.057064381726775588, 0.019078444504028489, 0.038190902772407893, 0.07522244167053288, 0.064100061474925127, 0.0026553817759992098, 0.10374793814924758, 0.1214168961943472, 0.1190324329931377, 0.077575833742307018, 0.12357253323683032, 0.0310063771026362, 0.059658644815411707, 0.07912430331458313, 0.061607380406491816, 0.029317662384690951, 0.044587897338066604, 0.76128047958667755, 0.034531828103574733, 0.016883026818836463, 0.18873363130651599, 0.022545565899307295, 0.078895320921700318, 0.16952267200976334, 0.078744889322185527], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.025117159435091452, 0.10183874836108175, 0.33721203219833562, 0.20108306297376854, 0.15855636328726366, 0.052640662594620108, 0.05303139038929993, 0.05085727867706822, 0.11376729213764686, 0.0017883802071889565, 0.02380243327196991, 0.0099434359087682972, 0.034296912083607412, 0.042231221391335769, 0.051554385326581069, 0.0066041547483632205, 0.016079449080568325, 0.056048113409363463, 0.045786958537002882, 0.0070520062087290379, 0.023513449541744037, 0.0058732767481875772, 0.01928374272419282, 0.0034984051092211009, 0.045335153063165101, 0.0056662943623768227, 0.0024369245745177325, 0.069249823409530856, 0.028151643359901009, 0.85701246331816583, 0.080598373298115075, 0.032504253208164559, 0.088571665173095601, 0.06186916528459159, 0.022162313079781747, 0.11782609734427189], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.19666425704423654, 0.13428403229080854, 0.059636743088830761, 0.13443602964008122, 0.33333467786235399, 0.010705783760035116, 0.042587913499792192, 0.048178834067721547, 0.095350162742058278, 0.05162288613506251, 0.027591346254373875, 0.10899397313592504, 0.0041626513683698692, 0.039281937468562476, 0.051378856961643161, 0.071121570463210035, 0.024058919695462644, 0.058928572564561903, 0.12730978071879495, 0.013069811160549602, 0.017333556752692346, 0.056754363715111729, 0.02695322206132687, 0.058863446469862267, 0.12945998828896052, 0.012171563295678318, 0.076302936617313216, 0.18026484644292062, 0.003463455619715246, 0.11700085986846996, 0.80164764867993998, 0.00082208663636848468, 0.035559376262906339, 0.1696365503798864, 0.005103642584523318, 0.053548416779697475], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.11819082272538373, 0.22348981963538839, 0.057555058051220215, 0.29813354975964157, 0.12088759738772784, 0.11414513960946064, 0.055198478093871788, 0.0037904577244906716, 0.036164832143700837, 0.021641258768801595, 0.024293604183508542, 0.052976753994816876, 0.024875203965608418, 0.012637546811891822, 0.075655079218666002, 0.00230480416902068, 0.050199758093123881, 0.11122809797723453, 0.043904875613723762, 0.12467165322058819, 0.11489710473003868, 0.057227386722911534, 0.050353016610178561, 0.092573399429820319, 0.030684374456585405, 0.10833276009738832, 0.05474150414562208, 0.015980357483714399, 0.19578784017128248, 0.024997592328406419, 0.010820471807565877, 0.8071217057894956, 0.010003738773647431, 0.024483217126943731, 0.16730885570214751, 0.080105659714651084], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.010986636665461812, 0.019951762384532105, 0.29800737794062976, 0.032903240139511, 0.043801125957549016, 0.064289520770117375, 0.014167712730243549, 0.0079011934092044194, 0.077473938334021261, 0.0014768610577427493, 0.0045441593926499029, 0.050216936714495869, 0.0092657688844838427, 0.0067963070574515606, 0.058690136049437798, 0.0008695362540858361, 0.003620586012917721, 0.058215137739755569, 0.0095299945924690131, 0.00084829423542340579, 0.060335416621030133, 0.00016894849494501182, 0.0040249347510682926, 0.054865179792430163, 0.0096143621432712139, 0.00099132268919397207, 0.053037129729062729, 0.015987258439034771, 0.0016743033867850188, 0.081525642839685108, 0.018615007564290024, 0.002228799729950734, 0.93017035826095518, 0.01435300897923711, 0.00097433085815661123, 0.073513941730289328], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.254153931610549, 0.1481986128225746, 0.014566291616895981, 0.13297449287672591, 0.24571719585087165, 0.21839901894742353, 0.032072853983566907, 0.11152959209499592, 0.059128772038680825, 0.070557927448724286, 0.050308759429622131, 0.099098443448829482, 0.051637507024578443, 0.0044080891235505678, 0.050545836600440888, 0.13456661104156312, 0.031850463963338703, 0.053731524355596132, 0.029183943681014016, 0.012155600665512837, 0.038663598075307538, 0.017970861190053616, 0.00010732758195097639, 0.070474210021641642, 0.062857286194914139, 0.033408817974854239, 0.081487106547128096, 0.19006976075599966, 0.063810249751907452, 0.081712079817537212, 0.18058636927188851, 0.010928065794532779, 0.014487021286762368, 0.79059512897888473, 0.019178532794664645, 0.034840118575435827], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.092644345231347625, 0.23891408871197523, 0.051431123907500917, 0.238638941168166, 0.083739232713895981, 0.10852551549473033, 0.045458546743840728, 0.01479214112913135, 0.027899133046377766, 0.016263017346290695, 0.026398383229464216, 0.043732287724809665, 0.01686770674022282, 0.00071384593747159434, 0.061952047919768195, 0.0068278769978893728, 0.04901296378696568, 0.090534832488984182, 0.042168435363124861, 0.11337977575140709, 0.091155679256147273, 0.052191394315892835, 0.051392461639565651, 0.073217378238910835, 0.029012637257507665, 0.0967086593581543, 0.042819265287938182, 0.015379903801844154, 0.17552996111219699, 0.02124342773539847, 0.0093501570681039127, 0.1693831259976053, 0.009198888853749471, 0.024409203962223805, 0.85105848206509827, 0.065524362693893515], [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.054199419476311191, 0.060727965641383184, 0.31963556662738896, 0.19394648696424369, 0.050714901941691641, 0.034919380135750015, 0.011463243913632785, 0.053848892451885878, 0.057584632813064987, 0.011780565449391843, 0.015552015741267683, 0.074549833791843392, 0.0085894643582694729, 0.042032418065020401, 0.095702458539136531, 0.0087772737520931815, 0.0042240237742239483, 0.11383959506392499, 0.020921152202947279, 0.029635887852503671, 0.0024343258190647411, 0.018551785265531463, 0.0086827884531995216, 0.0032957268202219096, 0.018680096433392048, 0.027344624100179689, 0.021138988070898784, 0.013694069474375991, 0.060592117662382892, 0.094088597016022721, 0.015849472135165941, 0.066588978995730172, 0.073444381579192725, 0.011439944503132481, 0.051621397457875841, 0.8857648346383804]]) import LinearAlgebra LinearAlgebra.singular value decomposition a, True)","bug"
|
|
"6013","4950","numpy.percentile gives INCORRECT results on some cases.
|
|
numpy version ` 1.9.0.dev 7bab957 ` .
|
|
when run `np.percentile x, q = np.linspace 0, 100, 5))` expected `>>> array [ 0., 1., 2., 2., 6.
|
|
])` but got `>>> array [ 2., 1., 2., 2., 6.
|
|
])` However running `np.percentile x, q = np.linspace 1, 100, 5))` gives the correct result `>>> array [ 0., 1., 2., 2., 6.
|
|
])`","bug"
|
|
"6013","4952","numpy ufuncs don t have qualname .
|
|
Numpy ufuncs have a `` name ``, but don t have a `` qualname `` as defined in PEP 3155.
|
|
Similarly, they don t seem to define a `` module ``.
|
|
note that for numpy s builtin ufuncs, the `` qualname `` should simply be the same as the `` name ``, e.g.
|
|
``np.sin.
|
|
qualname == sin ``)","bug"
|
|
"6013","4954","IndexError in numpy.select.
|
|
Try the following .
|
|
This gives an IndexError numpy\lib\function base.py in select condlist, choicelist, default) 777 pfac = asarray 1) 778 for k in range n2 1) > 779 pfac = pfac asarray choicelist[k]) 780 if type S) in ScalarType 781 S = S ones asarray pfac).shape, type S)) IndexError index out of bounds There s no problem when using numpy.extract instead of numpy.select but the output should be the same.","bug"
|
|
"6013","4955","Structured array fields with str dtype have length 0, not 1.
|
|
See [this SO question] http stackoverflow.com questions 25219344 numpy set values in structured array based on other values in structured array) for the original source.
|
|
I have reproduced this with current master, the original report was on 1.8.1 .
|
|
What seems to be going on is that, when using `str` as the field type, it is created with type `S` which apparently means `S0`, not `S1`.
|
|
Assigning any length to the string field makes it work .
|
|
This does not happen with regular arrays .","bug"
|
|
"6013","4959","Do not warn when invalid values are masked in a MaskedArray.
|
|
When doing a mathematical operation on a masked array, and some elements are invalid for the particular mathematical operation, numpy should not issue a warning if all invalid elements are masked.
|
|
This cannot be handled by `numpy.seterr` I want numpy to warn or raise when performing an invalid operation on any non masked value, but to be silent when performing an invalid operation on a masked value.
|
|
.","bug"
|
|
"6013","496","test unicode causes test ) warning Trac 12).
|
|
Original ticket http projects.scipy.org numpy ticket 12 on 2006 02 15 by trac user a.h.jaffe , assigned to unknown.
|
|
If I run with higher verbosity there are lots of other warnings, but this one persists in normal mode)In [9] import numpy) NB test unicode.py does seem to exist in the correct place.
|
|
In [10] numpy.
|
|
version Out[10] 0.9.5.2106 In [11] numpy.test ) Found 11 tests for numpy.core.umath Found 8 tests for numpy.lib.arraysetops Found 26 tests for numpy.core.ma Found 6 tests for numpy.core.records Found 14 tests for numpy.core.numeric Found 4 tests for numpy.distutils.misc util Found 3 tests for numpy.lib.getlimits Found 30 tests for numpy.core.numerictypes Found 9 tests for numpy.lib.twodim base Found 1 tests for numpy.core.oldnumeric Found 44 tests for numpy.lib.shape base Found 4 tests for numpy.lib.index tricks Found 42 tests for numpy.lib.type check Found 3 tests for numpy.dft.helper Warning !!
|
|
FAILURE importing tests for <module numpy.core.multiarray from ...c.egg numpy core multiarray.so > Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy 0.0.0 py2.4 macosx 10.4 ppc.egg numpy core tests test multiarray.py 195 ImportError No module named test unicode in ?)
|
|
Found 7 tests for numpy.core.defmatrix Found 33 tests for numpy.lib.function base Found 0 tests for main .......................................................................................................................................................................................................................................................
|
|
Ran 247 tests in 1.309s OK Out[11] <unittest.TextTestRunner object at 0x74d650>","bug"
|
|
"6013","4962","confusing error message should give more information?.
|
|
Just to be clear I m not asking for help setting up numpy.
|
|
I ll just uninstall numpy, scipy, cython, and matplotlib, then reinstall all of them from their most recent github sources.
|
|
My question is about this weird error message I get, with a screwed up numpy configuration.
|
|
The error message does not include a full traceback, and I don t know which module it s talking about.
|
|
I suspect it could even be from matplotlib somehow.
|
|
Where exactly does it come from, and how could it be made clearer to help people fix their bad numpy installations?
|
|
.","bug"
|
|
"6013","4963","feature request set data method for errorbar.
|
|
It would be awesome to have a `set data` method for an errorbar object.
|
|
Using that method for a line object is very convenient when I make complicated plots with labels, text, and so forth.
|
|
Clearing the entire plot and redrawing everything just to plot some different data is a pain, and I am unable to figure out how to clear just the errobar lines.
|
|
Even if I could, however, a set data method would still be more convenient.","enhancement"
|
|
"6013","4964","Bento does not find cblas, does not build dotblas.
|
|
cournape It seems like `bld.env.HAS CBLAS` is never true.
|
|
The build output also shows a failure to find atlas.","bug"
|
|
"6013","4965","numpy.einsum new feature request repeated output subscripts as diagonal.
|
|
I think that the following new feature would make `numpy.einsum` even more powerful useful awesome than it already is.
|
|
Moreover, the change should not interfere with existing code, it would preserve the minimalistic spirit of `numpy.einsum`, and the new functionality would integrate in a seamless intuitive manner for the users.
|
|
In short, the new feature would allow for repeated subscripts to appear in the output part of the `subscripts` parameter i.e., on the right hand side of ` >`).
|
|
The corresponding dimensions in the resulting `ndarray` would only be filled along their diagonal, leaving the off diagonal entries to the default value for this `dtype` typically zero).
|
|
Note that the current behavior is to raise an exception when repeated output subscripts are being used.
|
|
This is simplest to describe with an example involving the dual behavior of `numpy.diag`.
|
|
.
|
|
By opposition to `numpy.diag`, the approach generalizes to higher dimensions `einsum iii >i , A)` extracts the diagonal of a 3 D array, and `einsum i >iii , v)` would build a diagonal 3 D array.
|
|
The proposed behavior really starts to shine in more intricate cases.
|
|
.
|
|
If this is not convincing enough, replace `eye 2)` by `eye P w ab.shape[1])` and replace `eye 3)` by `eye P y wxab.shape[0])`, then imagine more dimensions and repeated indices...
|
|
The new notation would allow for crisper codes and reduce the opportunities for dumb mistakes.
|
|
For those who wonder, the above computation amounts to P X=x,Y=y,Z=z|A=a,B=b) = \sum w P W=w|A=a,B=b) P X=x|A=a) P Y=y|W=w,X=x,A=a,B=b) P Z=z|Y=y) with P X=x|A=a)=\delta {xa} and P Z=z|Y=y)=\delta {zy} using LaTeX notation, and \delta {ij} is [Kronecker s delta] http en.wikipedia.org wiki Kronecker delta)).","enhancement"
|
|
"6013","4966","irfft2 fails when the last dimension has a dimension of size 1.
|
|
This applies when the number of output points is not specified for example >>>> numpy.fft.irfft2 np.random.random 10,1))) fails rfft outputs n 2 1 points, and irfft reverse this with n 1) 2.
|
|
This clearly fails when n is 1 See fftpack.py line 416","bug"
|
|
"6013","4967","index tricks fill diagonal, wrap=True needed in examples.
|
|
https github.com numpy numpy blob master numpy lib index tricks.py The doc for `fill diagonal` has a couple of examples that are meant to illustrate the `wrap=True` feature.
|
|
But that parameter is missing.
|
|
>>> fill diagonal a, 4) should read >>> fill diagonal a, 4, wrap=True) for lines 738 and 747 tall matrices wrap) and wide)","bug"
|
|
"6013","497","Revision 2164 build broken with Visual Studio and MinGW GCC Trac 13).
|
|
Original ticket http projects.scipy.org numpy ticket 13 on 2006 02 24 by alberts, assigned to unknown.
|
|
The build is broken as of revision 2164 and probably earlier), at least on Windows.
|
|
Build logs attached.","bug"
|
|
"6013","4972","cann t go to numpy.scipy.org site or can t view any documentation sites.
|
|
Can t go to or refer any documentation docs.scipy.org doc ) in numpy and can t view numpy.scipy.org","bug"
|
|
"6013","4973","Reference leak in convert pyobject to datetime metadata ).
|
|
The temporary object `ascii` in that function is never decref ed.
|
|
in datetime.c, lines 1892 )","bug"
|
|
"6013","4974","Reference leak in convert pyobject to timedelta ).
|
|
The temporary object `bytes` in that function is never decref ed.
|
|
in datetime.c, lines 2541 )","bug"
|
|
"6013","4978","Randomly changing results.
|
|
By calculating the eigen problem using numpy.linalg.eig ) I get deviating results changing more or less randomly for big matrices see output below in the code in attached file).
|
|
The probability for errors increases with the size of matrices and when more calculations are done around it for that reason I calculate multiple things in between).
|
|
In the real application, I have a bigger matrices and instead of an expected unity matrix I get something really different.
|
|
This error occured on my Laptop using python 2.7.3, 2.7.8 and 3.2.3 for linux ubuntu); I was NOT able to reproduce it on a cluster using the same version as well as on other computers... code sorry for the long file) .
|
|
.
|
|
gdb run gdb usr bin python GNU gdb Ubuntu Linaro 7.4 2012.04 0ubuntu2.1) 7.4 2012.04 Copyright C) 2012 Free Software Foundation, Inc. License GPLv3 GNU GPL version 3 or later <http gnu.org licenses gpl.html> This is free software you are free to change and redistribute it.
|
|
There is NO WARRANTY, to the extent permitted by law.
|
|
Type show copying and show warranty for details.
|
|
This GDB was configured as i686 linux gnu .
|
|
For bug reporting instructions, please see <http bugs.launchpad.net gdb linaro >... Reading symbols from usr bin python... no debugging symbols found)...done.
|
|
gdb) run bug3.py Starting program usr bin python bug3.py [Thread debugging using libthread db enabled] Using host libthread db library lib i386 linux gnu libthread db.so.1 .
|
|
0.0 2.8914368725 2.8914368725 [Inferior 1 process 7925) exited normally]","bug"
|
|
"6013","4979","NumPy s Py buffer format string doesn t correctly handle sub struct padding when the entire struct is a multiple of 4..
|
|
Consider np.dtype [ a , [ s , np.int16), t , np.int8)]), x , np.int8)]) which gives a buffer format string ` T{T{h s b t } a b x } ` which is missing the = none alignment spec although internally numpy packs this into 4 bytes).
|
|
Adding another field or otherwise making the entire struct odd in size) np.dtype [ a , [ s , np.int16), t , np.int8)]), x , np.int8), y , np.int8)]) correctly gives ` T{T{=h s b t } a b x b y } `","bug"
|
|
"6013","498","f2py numpy 0.9.5 segfault Trac 14).
|
|
Original ticket http projects.scipy.org numpy ticket 14 on 2006 02 24 by unknown, assigned to pearu.
|
|
I wrapped a fortran program with the f2py which comes with numpy 0.9.5 and I get a segfault on import.
|
|
To create the wrapper I did f2py c m pythia pythia6326.f skip pygaus pygau2 This successfully creates pythia.so.
|
|
As soon as I import pythia.so I get the segfault python Python 2.4.2 2, Nov 20 2005, 17 04 48) [GCC 4.0.3 20051111 prerelease) Debian 4.0.2 4)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import pythia Segmentation fault I get the same behaviour on 32 and 64 bit AMD processors I have not tried Intel).
|
|
The fortran program is here http www.thep.lu.se torbjorn pythia pythia6326.f","bug"
|
|
"6013","4980","Boolean array indexing fails silently .
|
|
Indexing with a boolean array of length shorter than the indexed array seems to fail silently .
|
|
Is this behaviour intended?
|
|
It seems preferable to raise an error if you try to index with a boolean array of a shape that cannot be broadcast to the same shape as the indexed array.
|
|
For what it s worth, I just spent 2 hours of debugging to find this was the culprit.
|
|
I m using numpy version 1.8.1","bug"
|
|
"6013","4983","datetime64 arrays don t support buffer protocol.
|
|
.","bug"
|
|
"6013","499","Importing numpy causes the loading of numpy.distutils and its hotpatches to distutils Trac 15).
|
|
Original ticket http projects.scipy.org numpy ticket 15 on 2006 02 28 by rkern, assigned to pearu.
|
|
Neither import numpy from numpy import get numpy include should trigger the importing of numpy.distutils .
|
|
Other setup.py scripts may be importing numpy to test for its existence and getting the include directory.
|
|
We should not futz around inside the distutils package unless numpy.distutils is explicitly imported.","bug"
|
|
"6013","4990","sync percentile behavior with R?.
|
|
probably the last thing we may want to do before releasing 1.9 1.9 percentile now has a few options for interpolation `linear`, `lower`, `higher`, `nearest` and `midpoint`.
|
|
R has 9 types, see http stat.ethz.ch R manual R patched library stats html quantile.html as far as I can tell only one of them that matches is type 7 in R which is `linear` in our percentile.
|
|
The question is do we want to match up some types.
|
|
E.g.
|
|
our `nearest` is not the same as R type 3 as theirs uses even order statistics.
|
|
I though their type 5 would be the same as our `midpoint` but it does not give the same result.
|
|
And the recommended type 8 is not implemented in numpy at all.
|
|
Its not really a blocker was we can always add new types that match R later should it be required, but something that should be discussed.","enhancement"
|
|
"6013","4992","[ipo] compilation error.
|
|
Hi, I am trying to compile numpy 1.8.1 with Intel c compiler icc) with an ipo switch which does inlining as its main feature.
|
|
But I am facing the below build issue.
|
|
I am using Python 2.6.6 version.
|
|
<snip> removing configtest.c configtest.o C compiler icc m64 fPIC O3 ipo fp model fast fomit frame pointer openmp xhost DMKL ILP64 compile options Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I usr include python2.6 c icc configtest.c removing configtest.c configtest.o Traceback most recent call last) File setup.py , line 239, in <module> setup package ) File setup.py , line 231, in setup package setup metadata) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils core.py , line 169, in setup return old setup new attr) File usr lib64 python2.6 distutils core.py , line 152, in setup dist.run commands ) File usr lib64 python2.6 distutils dist.py , line 975, in run commands self.run command cmd) File usr lib64 python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils command build clib.py , line 63, in run self.run command build src ) File usr lib64 python2.6 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib64 python2.6 distutils dist.py , line 995, in run command cmd obj.run ) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils command build src.py , line 153, in run self.build sources ) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils command build src.py , line 170, in build sources self.build extension sources ext) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils command build src.py , line 329, in build extension sources sources = self.generate sources sources, ext) File home indiatce reddy numpy 15.0 numpy 1.8.1 numpy distutils command build src.py , line 386, in generate sources source = func extension, build dir) File numpy core setup.py , line 463, in generate config h rep = check long double representation config cmd) File numpy core setup common.py , line 168, in check long double representation type = long double representation pyod object)) File numpy core setup common.py , line 308, in long double representation raise ValueError Could not lock sequences s) saw) ValueError Could not lock sequences None) Note W o ipo switch build works fine.
|
|
Could you please let me know if there is any work around for this issue ?
|
|
Thanks, Reddy","bug"
|
|
"6013","4993","unexpected modification of an numpy.array content.
|
|
In the little code below I did not expect that M would be modified.
|
|
Furthermore, even taking a slice of M pM = np.real M[ , ]) ) does not seem to solve the issue.
|
|
The only manner to prevent M from beeing modified is to explicitly copy it pM = np.real M.copy )) ).
|
|
This issue does not happen when e.g.
|
|
another numpy function is used pM = np.sin M) ).
|
|
The modification of the original array M is quite unexpected and not consistent.
|
|
I running on Ubuntu 14.04 LTS x64 8< .
|
|
8< .","bug"
|
|
"6013","4994","UnicodeDecodeError.
|
|
I get an error when I try to init a masked numpy string array.
|
|
Should be possible isnt it?
|
|
I use it, because I dont know the datatype in advance .
|
|
It doenst happen in python 3.2.3 It does happen in python 3.3.2 3.4.1","bug"
|
|
"6013","4995","Wierd behaviour by numpy.where ) and numpy.any ).
|
|
.
|
|
Please have a look on the above output.
|
|
The above array doesn t contain any negative values then why does where ) method says that the last element is negative?
|
|
Same applies to any ) .
|
|
Here are my system details Python 3.4.0 default, Aug 12 2014, 11 20 12) numpy, Version 1.8.1 OS Details NAME=openSUSE VERSION= 13.1 Bottle) VERSION ID= 13.1 PRETTY NAME= openSUSE 13.1 Bottle) x86 64) ID=opensuse ANSI COLOR= 0;32 CPE NAME= cpe o opensuse opensuse 13.1 Thanks.","bug"
|
|
"6013","4999","excessive computation of mode from name.
|
|
I have an inner loop that calls numpy.convolve quite a lot def inner loop g z,c i,g zs,lambda array,segments) pg zs=g zs[ ,0] for r in range 1,segments) pg zs=g z c i[0,segments r 1] numpy.convolve pg zs,lambda array, same )) g zs[ ,r]=pg zs both pg zs and lambda array are fairly short 100,3) but segments is big 1000).
|
|
Profiling revealed that substantial time is spent in [ mode from name] https github.com numpy numpy blob v1.8.1 numpy core numeric.py L814) checking types and calling the lower method of strings.
|
|
I tried to optimize the problem a bit by using 1 instead, but it didn t help much.
|
|
A pure dictionary approach is a bit faster though 198000 calls mode from name ) convolve ) same .315 s 2.888 s 1 .181 s 2.816 s dict 2.476 s I propose changing any code that executes mode= mode from name mode) to mode= mode from name dict[mode] where the dictionary is defined mode from name dict = {0 0, 1 1, 2 2, 0 0, 1 1, 2 2, v 0, s 1, f 2, V 0, S 1, F 2, valid 0, same 1, full 2, Valid 0, Same 1, Full 2, VALID 0, SAME 1, FULL 2} Although it doesn t help too much for my case, it would provide a noticeable boost for even smaller convolutions.
|
|
The only risk is that strings such as Vavoiubaowa3824 will no longer be valid mode inputs.
|
|
If you really want to capture that behavior too, KeyErrors can be caught with try, and then mode from name mode) can be called for edge cases.","bug"
|
|
"6013","500","pickling and unpickling a transposed array is not working correctly Trac 16).
|
|
Original ticket http projects.scipy.org numpy ticket 16 on 2006 03 05 by tovrstra, assigned to unknown.
|
|
Hi, I ve found a problem with numpy 0.9.5 with python 2.4.2).
|
|
Unpickeling a pickled transposed array, returns a messed up array.
|
|
Here is an example import numpy, pickle a = numpy.transpose numpy.array [[2, 9], [7, 0], [3, 8]] )) print a f = file test , w ) pickle.dump a, f) f.close ) f = file test , r ) print pickle.load f) f.close ) The output is [[2 7 3] [9 0 8]] [[2 9 7] [0 3 8]]","bug"
|
|
"6013","5000","ENH function to merge two sorted arrays.
|
|
Problem I didn t find the function to merge two sorted arrays, something like this <pre> >>>np.merge [0,2,4,9], [1,5,7,9]) array [0, 1, 2, 4, 5, 7, 9, 9]) < pre> However, such operations are needed in statistics when you want to compare different distributions.
|
|
For instance, compute Kolmogorov Smirnov stats or Cramer von Mises similarity.
|
|
You can sort the values for each distribution and then use searchsorted, but it is not linear time.
|
|
Suggestion I think, the following interface would be brilliant for such purposes <pre> >>>merged, indices1, indices2 = np.merge [0,2,4,9], [1,5,7,9], order= first ) >>>merged array [0, 1, 2, 4, 5, 7, 9, 9]) >>>indices1 array [0, 2, 3, 6]) >>>indices2 array [1, 4, 5, 7]) < pre> `indices1` and `indices2` represent the indices of initial arrays in merged array order if first , elements of first array are placed before equal elements of second as in example above) if second opposite <pre> >>>indices1 array [0, 2, 3, 7]) >>>indices2 array [1, 4, 5, 6]) < pre> if None, then indices returned are indices of the leftmost equal element <pre> >>>indices1 array [0, 2, 3, 6]) >>>indices2 array [1, 4, 5, 6]) < pre> Thanks for numpy!","bug"
|
|
"6013","5008","numpy.lib.recfunctions has no reference in the documentation.
|
|
A collection of record array functions are available under numpy.lib.recfunctions.
|
|
The documentation of the functions look OK, however they are completely missing from the documentation no index entries).
|
|
I actually discovered the existence of this module by chance... Is this intentional?
|
|
I also have some additional source code developed for myself to manipulate record arrays field cut copies from one array to another, SQL like joining of arrays) which I am happy to add possibly to this module if they are welcome.
|
|
Feel free to comment.","bug"
|
|
"6013","5009","Fix or silence deprecation warnings for bad index calls.
|
|
See gh 4991 for details.
|
|
This is about things like arr[1<<64]` which gives misleading deprecation warnings even though it is an error anyway.","bug"
|
|
"6013","501","nonzero on masked arrays fails Trac 17).
|
|
Original ticket http projects.scipy.org numpy ticket 17 on 2006 03 05 by trac user strawman, assigned to trac user sasha.
|
|
nonzero ) fails with an exception that is clearly a bug when called on a masked array astraw hdmg python Starting python version 2.3.5 Python 2.3.5 2, Nov 3 2005, 02 44 38) [GCC 3.3.5 Debian 1 3.3.5 13)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> jj= numpy.ma.masked array [0,1,2,3,0,4,5,6],mask=[0,0,0,1,1,1,0,0]) >>> numpy.ma.nonzero jj) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File home astraw py2.3 linux x86 64 lib python2.3 site packages numpy 0.9.6.2195 py2.3 linux x86 64.egg numpy core ma.py , line 320, in call m = mask or getmaskarray a), getmaskarray b)) NameError global name b is not defined >>>","bug"
|
|
"6013","5010","removed changed paths in 1.9.0rc1.
|
|
I m seeing the following errors from setup.py .
|
|
It would be nice if f2py.1 was installed in usr share man man1 f2py.1.","bug"
|
|
"6013","5011","Record comparison fails for structured arrays.
|
|
x = np.array [ 0,0)], dtype= i,i ) x[0] < x[0] result = True x[0] > x[0] result = True tuple x[0]) < tuple x[0]) the fix result = False NumPy 1.8.1 Python 2.7.6","bug"
|
|
"6013","5016","Comparison operators do not broadcast on object arrays containing arrays.
|
|
This came up [on StackOverflow] http stackoverflow.com questions 25559027 numpy where command on arrays with dtype object 25559892) and struck me as pretty strange.
|
|
Here is a short demo of the issue Python 2.x) .
|
|
As expected, a boolean array over elements.
|
|
Now make the list a numpy array .
|
|
I would expect this to raise a ValueError similar to .
|
|
Not broadcasting the equality comparison in ``foo == 3`` is very unexpected.
|
|
Is this intended behavior?","bug"
|
|
"6013","5017","ncu.log1p np.nan) produces a RuntimeWarning.
|
|
The below is produced with numpy 1.9.0rc1 on an OpenBSD i386 system.
|
|
.","bug"
|
|
"6013","5018","Behavior of the absolute value function.
|
|
I have just spent quite some time debugging my code because the absolute value function might return negative values.
|
|
See the following working example import numpy as np np.array [ 32768, 3000],dtype=np.dtype int16 )) np.abs ) While the second value is converted to 3000 as expected, the second value is left untouched.
|
|
I understand that 32768 is out of range for this datatype but I am interested whether returning 32768 is a concious design choice as I was expecting positive values from this function no matter what.","bug"
|
|
"6013","5019","FAIL test umath.test spacing on OpenBSD amd64 and numpy 1.9.0rc1.
|
|
On OpenBSD amd64 I get one regress failure.
|
|
The full test suite output is at the bottom.
|
|
I simplified the test to .
|
|
The value of the last line is False.
|
|
But strangely, if you remove the inf = np.longdouble np.inf) line, the value of the last line is True.
|
|
.","bug"
|
|
"6013","502","code cleanup for ma.py Trac 18).
|
|
Original ticket http projects.scipy.org numpy ticket 18 on 2006 03 06 by timleslie, assigned to trac user sasha.
|
|
pylint showed up minor issues in core ma.py wrong indentation missing spaces after commas unused imports Diff attached.","bug"
|
|
"6013","5023","np.random.multinomial is not generalised for any given shape.. Not able to generate multinomial samples for any given input probability shape.
|
|
Expecting same kind of behavior as binomial.
|
|
Please refer for detail http stackoverflow.com questions 25590694 multinomial sample generation in python 25591084 25591084 Thanks","bug"
|
|
"6013","5026","FAIL test umath.TestComplexFunctions.test loss of precision longcomplex on OpenBSD sparc64 and numpy 1.9.0rc1.
|
|
Here s the actual test failure .
|
|
I simplified the test case to the following .
|
|
Here s the output of the simplified code on my OpenBSD sparc64 machine .","bug"
|
|
"6013","5027","FAIL test datetime as string test datetime.TestDateTime) on OpenBSD macppc and numpy 1.9.0rc1.
|
|
There are a few regress test failures on OpenBSD macppc.
|
|
I suspect if the simplified code below is fixed the rest will be fixed too.
|
|
.
|
|
Full regress test details below.
|
|
.","bug"
|
|
"6013","5028","numpy.arange.
|
|
Wrong last element in some cases.
|
|
python 2.7.8, numpy 1.8.2. x64.
|
|
For this script .
|
|
I received the following answer .
|
|
Arange should return values from the interval [start, stop), but in the second case it included and the value stop2.","bug"
|
|
"6013","5029","f2py3 bug error in traslation Fortran to C array become functions.
|
|
Hi, I have a machine with this configuration Ubuntu 14.04, f2py3 version 2, Python 3.4.0.
|
|
I have a fortran module with a subroutine that take in input a 3D matrix and its dimension contained in two array ub ),lb ) that specify the range.
|
|
When I use f2py for create compile module happens a strange thing, gcc return error called object xe2x80x98lbxe2x80x99 is not a function or function pointer .
|
|
but If I see the module.c that f2py have generated ub and lb are a pointer .
|
|
The problem is that f2py translator replace square bracket with bracket.
|
|
If I replace in module.c the line with this .
|
|
the error disappears.","bug"
|
|
"6013","503","reference of unused variable core ma.py 320 Trac 19).
|
|
Original ticket http projects.scipy.org numpy ticket 19 on 2006 03 06 by timleslie, assigned to trac user sasha.
|
|
line 320 accesses local variable b which is not defined in this scope.
|
|
I m not sure what the correct way to fix this is.
|
|
def call self, a, args, kwargs) Execute the call behavior.
|
|
numeric tries to return scalars rather than arrays when given scalars.
|
|
m = getmask a) d1 = filled a, self.fill) if self.domain is not None m = mask or m, self.domain d1)) result = self.f d1, args, kwargs) if m is not nomask try shape = result.shape except AttributeError pass else if m.shape != shape m = mask or getmaskarray a), getmaskarray b)) return masked array result, m)","bug"
|
|
"6013","5030","array scalar numpy scalar is slow..
|
|
Some timings .
|
|
Maybe a fastpath it needed.","bug"
|
|
"6013","5032","min max of empty arrays .
|
|
A similar issue was raised in 2670, however the edge case of 0, ) shape is still not covered.
|
|
I can go around it with checking for non zero length in an ``if`` statement, but it would be more practical if ``np.min`` could handle this, too.
|
|
.","bug"
|
|
"6013","5033","ravel A) should return a view if ravel squeeze A)) does.
|
|
With NumPy 1.8.2 I get .
|
|
This is probably due to the fact that .
|
|
in contrast to .","bug"
|
|
"6013","5036","Ctrl C after running tests causes segfaults.
|
|
Maybe it is just my setup, but if I run through the tests and then press Ctrl C in the ipython session or python session), or also during the tests if early enough at least.
|
|
I am seeing regular segfaults...","bug"
|
|
"6013","5037","numpy repeat dtype.
|
|
I wanted to use numpy.repeat to get an array of repeated integers with a specific dtype, but it does not accept a dtype argument.
|
|
I can work around this by first converting the list of python integers to be repeated into a numpy array with the dtype that I want, but I d rather be able to specify a dtype in the numpy.repeat function.","enhancement"
|
|
"6013","5039","Inconsistent behavior with numpy.unique on array of tuples between 1.8.0 and 1.8.1.
|
|
I use numpy.unique to generate a unique list of tuples.
|
|
On version 1.8.0, it gives the correct results.
|
|
For example .
|
|
However, on version 1.8.1, the exact code gives .
|
|
The 2,1) tuples gets flipped to 1,2).
|
|
I can probably find a way around this.
|
|
I m not sure if it is just a fluke that it works correctly on 1.8.0 and not 1.8.1.
|
|
If this sorting is actually undefined for tuples, it probably should raise an error instead of returning inconsistent results.
|
|
I can probably find a way around this for my particular application, but I won t be depending on the specific behavior of numpy.unique in the future.","bug"
|
|
"6013","504","code cleanup for arrayprint.py Trac 20).
|
|
Original ticket http projects.scipy.org numpy ticket 20 on 2006 03 06 by timleslie, assigned to unknown.
|
|
patch attached to fix typo in variable name tabs vs spaces 2 lines) spaces after commas","bug"
|
|
"6013","5041","return NaN in npy ObjectMax ) and npy ObjectMin ) if it s an argument.
|
|
This is a patch I posted to 4903, but it must have fallen through the cracks.
|
|
It brings the code inline with the documentation in that it always returns the NaN argument first, if both).
|
|
.","bug"
|
|
"6013","5044","Incorrect sprintf ) result checks in fortranobject.c.
|
|
In f2py src fortranobject.c, there are several calls to `sprintf` which follow the pattern .
|
|
From http linux.die.net man 3 sprintf > Return value > > Upon successful return, these functions return the number of characters printed excluding the null byte used to end output to strings).
|
|
> The functions snprintf ) and vsnprintf ) do not write more than size bytes including the terminating null byte \0 )).
|
|
If the output was truncated due to this limit then the return value is the number of characters excluding the terminating null byte) which would have been written to the final string if enough space had been available.
|
|
Thus, a return value of size or more means that the output was truncated.
|
|
See also below under NOTES.)
|
|
> > If an output error is encountered, a negative value is returned.
|
|
These `sprintf` calls should check for a negative value, rather than 0.
|
|
As an aside, would it be better to use e.g.
|
|
`snprintf )`?
|
|
Especially since there doesn t seem to be any other check whether the string will overflow the allocated region.
|
|
According to the man page, `snprintf` is C99 whereas `sprintf` is C89 and later; does f2py require a C99 compiler?","bug"
|
|
"6013","5046","np.pmt causing unnecessary warnings.
|
|
Sometimes when using the np.pmt I enter 0.0 as an interest rate like so .
|
|
According to the documentation at http docs.scipy.org doc numpy reference generated numpy.pmt.html there is no restriction against a rate of 0 and the notes section in the docs even address the use case where rate = 0.
|
|
Yet every time I run the function with rate set to 0.0 I get the following warning .
|
|
I looked at the source and understand why its throwing the warning, but it really shouldn t since 0 is a valid rate and should handle that case differently.","bug"
|
|
"6013","5049","Unicode scalar problem .
|
|
Creating an array from a list of unicode scalars doesn t work when the scalars contain non ascii characters.
|
|
I m using python 2.7 32 bit) on windows with numpy 1.8.1 .
|
|
Example In [1] import numpy as np In [2] np.array u xc3xa5xc3xa4xc3xb6 ) Out[2] array u \xe5\xe4\xf6 , dtype= <U3 ) In [3] [np.array u xc3xa5xc3xa4xc3xb6 )] Out[3] [array u \xe5\xe4\xf6 , dtype= <U3 )] In [4] np.array [np.array [u xc3xa5xc3xa4xc3xb6 ])]) Out[4] array [[u \xe5\xe4\xf6 ]], dtype= <U3 ) In [5] np.array [np.array u xc3xa5xc3xa4xc3xb6 )]) UnicodeEncodeError Traceback most recent call last) <ipython input 4 918a52a6946b> in <module> ) > 1 np.array [np.array u xc3xa5xc3xa4xc3xb6 )]) c \python27\lib\site packages\numpy\core\numeric.pyc in array str a, max line width, precision, suppress small) 1613 1614 > 1615 return array2string a, max line width, precision, suppress small, , , str) 1616 1617 def set string function f, repr=True) c \python27\lib\site packages\numpy\core\arrayprint.pyc in array2string a, max line width, precision, suppress small, separator, prefix, style, formatter) 446 if isinstance x, tuple) 447 x = convert arrays x) > 448 lst = style x) 449 elif reduce product, a.shape) == 0 450 treat as a null array if any of shape elements == 0 UnicodeEncodeError ascii codec can t encode characters in position 0 2 ordinal not in range 128)","bug"
|
|
"6013","505","code cleanup for records.py Trac 21).
|
|
Original ticket http projects.scipy.org numpy ticket 21 on 2006 03 06 by timleslie, assigned to unknown.
|
|
Patch attached to fix unused imports spaces after commas","bug"
|
|
"6013","5052","numpy.percentile typo in Docs.
|
|
There s a typo at the end of the Notes paragraph in [numpy.percentile] http docs.scipy.org doc numpy reference generated numpy.percentile.html numpy.percentile) minimum if q=0``and the same as the maximum if ``q=100.
|
|
Also there s a mention of an interpolation paramete without any explanation of what that is.
|
|
Perhaps a bit more detail would be good?","bug"
|
|
"6013","5058","`bmat` does not document the arguments `ldict` and `gdict`.
|
|
See http docs.scipy.org doc numpy reference generated numpy.bmat.html https github.com numpy numpy blob master numpy matrixlib defmatrix.py L1155","bug"
|
|
"6013","506","code cleanup for numeric.py Trac 22).
|
|
Original ticket http projects.scipy.org numpy ticket 22 on 2006 03 06 by timleslie, assigned to unknown.
|
|
Patch attached to fix incorrect indentation wrong of spaces) spaces around = , == and after commas","bug"
|
|
"6013","5060","multiarray import error in 32 bit mode.
|
|
I got the following error when I simply `import numpy` in 32 bit mode.
|
|
Numpy is up to date 1.8.2).
|
|
<pre><code> arch i386 python Python 2.7.8 default, Aug 25 2014, 11 29 48) [GCC 4.2.1 Compatible Apple LLVM 5.1 clang 503.0.40)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File usr local lib python2.7 site packages numpy init .py , line 153, in <module> from .
|
|
import add newdocs File usr local lib python2.7 site packages numpy add newdocs.py , line 13, in <module> from numpy.lib import add newdoc File usr local lib python2.7 site packages numpy lib init .py , line 8, in <module> from .type check import File usr local lib python2.7 site packages numpy lib type check.py , line 11, in <module> import numpy.core.numeric as nx File usr local lib python2.7 site packages numpy core init .py , line 6, in <module> from .
|
|
import multiarray ImportError dlopen usr local lib python2.7 site packages numpy core multiarray.so, 2) no suitable image found.
|
|
Did find usr local lib python2.7 site packages numpy core multiarray.so mach o, but wrong architecture >>> < code>< pre> Below is my machine stats Machine Details Platform ID Darwin 13.3.0 x86 64 i386 64bit Processor i386 Python Implementation CPython Executable usr local opt python bin python2.7 Version 2.7.8 Compiler GCC 4.2.1 Compatible Apple LLVM 5.1 clang 503.0.40) Bits 64bit Build Aug 25 2014 11 29 48 default) Unicode UCS2","bug"
|
|
"6013","5063","Reference count error from subscript operation.
|
|
Running the following code with numpy 1.8.1 .
|
|
I get .
|
|
With 1.9.0, I don t get a reference error, but the result is still wrong .
|
|
I expect `1`.)","bug"
|
|
"6013","5064","u.resize should return array.
|
|
u.resize returns None.
|
|
If it returned the new array, this would be useful for chaining operations.","bug"
|
|
"6013","5065","np.inner segfaults with SciPy s sparse matrices.
|
|
The function `np.inner` segfaults with some of SciPy s sparse matrices.
|
|
I suspect that it will segfault for other matrix like things.
|
|
Here is a minimal working example .
|
|
with output .
|
|
I also observed the same crash with numpy `1.7.0`.","bug"
|
|
"6013","5068","round do not respect ndarray subclass.
|
|
round do not respect ndarray subclass in the following case .
|
|
the issue is similar with around.
|
|
My subclass is almost a copy of http docs.scipy.org doc numpy user basics.subclassing.html slightly more realistic example attribute added to existing array and adding \ array wrap doesn t change anything as the method is not call.","bug"
|
|
"6013","5069","numpy core tests test blasdot.py test dot 2args fails caching error?.
|
|
When running nosetests on numpy 1.9.0 numpy core tests test blasdot.py test dot 2args 44) fails.
|
|
Inserted the following line into numpy core blasdot dotblas.c DOUBLE dot 102) .
|
|
reveals for the failed test .
|
|
Huh?
|
|
.
|
|
Just to make sure, ddot works correctly, I inserted the following line right before the ddot call .
|
|
This time the test succeeds, but suprisingly when the test gets invoked a 2nd time, it fails with .
|
|
So it seems, that numpy doesn t determine the number of elements in the vector correctly.
|
|
I guess, this has something to do with the `def test vecself` in the test class, but I m not a python expert, so ... Env is Python 2.7.6 on Solaris 11.2","bug"
|
|
"6013","507","Waxc5x82 napxc4x99dowy Trac 23).
|
|
Original ticket http projects.scipy.org numpy ticket 23 on 2006 03 06 by timleslie, assigned to unknown.
|
|
Patch attached to fix use of tabs, not spaces spaces after commas","bug"
|
|
"6013","5071","cholesky should work even with an empty matrix.
|
|
I have a program where I compute a cholesky decomposition of a correlation matrix.
|
|
Sometimes, the correlation matrix is empty when there is no data).
|
|
In this case, np.linalg.cholesky reports an error.
|
|
It would be better if it returned an empty matrix.
|
|
This has been tested on numpy 1.6.2 on debian 7).","bug"
|
|
"6013","5073","np.atleast 2d of sparse matrix returns a 1x1 ndarray containing the sparse matrix object.
|
|
I assume this falls under numpy doesn t know about your custom matrix like objects so they are not our problem but I thought I d mention this anyway because it was obscuring a bug in another project.
|
|
Feel free to close as invalid.
|
|
.
|
|
.","bug"
|
|
"6013","5074","BUG numpy ufunc does not properly deal with NotImplemented.
|
|
In continuing to try to get astropy s `Quantity` to behave with ` numpy ufunc `, I found what probably is another bug.
|
|
One of our tests checks that an integer can be used as an ` index ` to, e.g., multiply a list, as in .
|
|
which yields `[ a , b , a , b ]`).
|
|
But if I make a trivial class that implements ` numpy ufunc `, the `NotImplemented` does not seem to be passed on correctly .
|
|
yields .
|
|
p.s.
|
|
This may be related to 4815","bug"
|
|
"6013","5079","any timedelta should be smaller than inf.
|
|
One might argue I shouldn t be doing this, but this behaviour is odd In [325] timedelta64 0) < float inf ) Out[325] False Use case I wanted a value that would compare larger than any timedelta64.
|
|
There is `timedelta64 NaT )`, there doesn t appear to be an equivalent for inf inf, but comparing consistently to those two seems to make sense.","bug"
|
|
"6013","508","code cleanup for defmatrix.py Trac 24).
|
|
Original ticket http projects.scipy.org numpy ticket 24 on 2006 03 06 by timleslie, assigned to unknown.
|
|
patch attached to fix unused imports spaces after commas","bug"
|
|
"6013","5080","numpy.corrcoef RuntimeWarning and NaN wrong output).
|
|
I have found a weird behaviour for numpy.corrcoef .
|
|
I reproduce with debian s squeeze python 2.6, on a compiled 2.7 python and in anaconda s 2.7 and 3.3 pythons on MacOSX.
|
|
The bug is shown in the following snippet !
|
|
usr bin env python import numpy as np wk=np.ones 400,)) 0.00282490517428 print \nCorrect output for values of , wk[0]) print corrcoef= ,np.corrcoef wk,wk)[0,1]) wk2=wk 1.e13 print \nCorrect output for values of , wk2[0]) print corrcoef= ,np.corrcoef wk2,wk2)[0,1]) wk2=wk 1.e14 print \nIncorrect output for values of , wk2[0]) print corrcoef= ,np.corrcoef wk2,wk2)[0,1]) The result is Correct output for values of 0.00282490517428 corrcoef= 1.0 Correct output for values of 28249051742.8 corrcoef= 1.0 Incorrect output for values of 282490517428.0 Users nino anaconda envs py3 lib python3.3 site packages numpy lib function base.py 1823 RuntimeWarning invalid value encountered in true divide return c sqrt multiply.outer d, d)) corrcoef= nan","bug"
|
|
"6013","5081","BUG np.broadcast arrays does not work on void arrays with object fields.
|
|
In the process of working on 4622, it was realised that `np.broadcas arrays` do not work for void arrays with object fields, because in the process of changing new strides and shapes in `as strided`, the `dtype` gets lost, and this cannot just be reset when object fields are present.
|
|
While obviously this is somewhat of a corner case, it might still be good to see if it can be fixed easily, perhaps by doing something that preserves the dtype, or by bypassing the test for object fields.
|
|
.","bug"
|
|
"6013","5082","Segmentation fault with argmin ) on unicode arrays.
|
|
`np.zeros 6031, dtype= <U9 ).argmin )` crashes `master` and at least numpy 1.8.
|
|
Some smaller arrays work fine.
|
|
Note, there is nothing particularly special about using `zeros )` here; it s just the shortest example that demonstrates the problem.","bug"
|
|
"6013","5083","assert warns does not clear emitted warning registry.
|
|
[assert warns] https github.com numpy numpy blob master numpy testing utils.py L1568) should clear the emitted warning registry, to ensure that tested warnings are emitted regardless of the order in which tests are run.
|
|
For instance, if test A emits a warning e.g.
|
|
silenced under catch warnings) and test B tries to test it using `assert warns`, whether test B passes or not depends on whether it gets run before or after test A which is not necessarily deterministic across platforms and Python Nose versions e.g.
|
|
it may pass locally and fail under travis ci, or vice versa, as happened in some cases in Scipy).
|
|
Controlling such interactions between tests is a hassle and should not be necessary.
|
|
The assumption when using `assert warns` probably is that all warnings are going to be raised, regardless of whether or not they were raised previously.
|
|
Python stores the status indicating whether a warning was previously emitted or not in a global ` warningregistry ` variable.
|
|
Python s own warning test suite manipulates the warnings state to ensure the tests are reproducible https hg.python.org cpython file c0b0dda16009 Lib test test warnings.py l18 See also https stackoverflow.com questions 2390766 how do i disable and then re enable a warning I m not fully clear currently on which module s global variable gets which warning, however...","bug"
|
|
"6013","5084","Hope genfromtxt function add nrows keyword.
|
|
numpy function `genfromtxt` can skip header and footer lines.
|
|
However sometimes we might want to control the number of rows to read directly.
|
|
For example, a txt file might contain several blocks with different number of columns, we can t read them all in once.
|
|
Specify `nrow` to read would be convenient.
|
|
Is it possible add something like `genfromtxt file,skip header=1,nrows=2)` ?","enhancement"
|
|
"6013","5085","Broken stride calculation on master.
|
|
I m running into a problem on Numpy master that I don t see on 1.9.0.
|
|
This is enough to trigger it .
|
|
On Python 2 .
|
|
On Python 3 .
|
|
9223372036854775807 is of course 1, so perhaps this is a deliberate marker for the outermost stride?
|
|
Not sure).
|
|
However, Python 3 doesn t seem to like that in the buffer interface.
|
|
See astropy astropy 2960","bug"
|
|
"6013","509","Runtime Error using Python Optimization OO) Trac 25).
|
|
Original ticket http projects.scipy.org numpy ticket 25 on 2006 03 06 by trac user jay.painter ..., assigned to unknown.
|
|
There s a little error in the file numpy init .py which prevents it from running with python optimization turned on i.e.
|
|
python OO).
|
|
Extending the module s doc attribute with = does not work, it seems, because doc is None when run with optimization.
|
|
I just removed the code extending the doc string and it works fine.","bug"
|
|
"6013","5091","UPDATEIFCOPY between F CONTIGUOUS and C CONTIGUOUS array.
|
|
I am converting the argument of my routine arg2 a numpy array C CONTIGUOUS) to a F CONTIGUOUS array with PyArray FROM OTF .
|
|
After some computation in B, I do a .
|
|
and return to the caller.
|
|
In the caller I can see that the content of the array arg2 would be correct for an F CONTIGUOUS array but not for a C CONTIGUOUS array.
|
|
However according to the array flags the array is C CONTIGUOUS.
|
|
It looks like when it does Py DECREF B) it just reuse the memory of the array data from the B array in arg2 but without updating the flags.
|
|
Thus the result is actually the transpose of B.","bug"
|
|
"6013","5094","Broken link in HOWTO DOCUMENT.
|
|
In the following file https github.com numpy numpy blob master doc HOWTO DOCUMENT.rst.txt there is a link to http docs.scipy.org numpy Questions Answers which is broken what should this link now be?","bug"
|
|
"6013","5095","BUG np.where half initializes subclass of output.
|
|
In numpy 1.9, the behaviour of `np.where condition, first, second)` changed, in that the output no longer always is an `ndarray` instance but an instance of the subclass of `first`.
|
|
However, ` array finalize self, obj)` is called with `obj=None`, and hence the subclass can not properly initialize itself according to the documentation [1], `None` is reserved for the case where a new instance is created via ` new `).
|
|
This change in behaviour has caused some problems e.g., https github.com astropy astropy pull 2958, also discussion in https github.com astropy astropy issues 1274).
|
|
I should add that I do not necessarily believe the old behaviour was preferable, but it would be good if the current behaviour is documented.
|
|
Going forward, an option which I think would be ideal though not necessarily practical...), is if `np.where` would do the equivalent of .
|
|
In this way, subclasses can check if the second argument is consistent with the first e.g., for our `Quantity` class, check the units and possibly convert).
|
|
This might be done most easily if, somehow, `np.where` could be turned into a `ufunc` and passed through the ` numpy ufunc ` mechanism.
|
|
[1] http docs.scipy.org doc numpy user basics.subclassing.html","bug"
|
|
"6013","5096","numpy.poly gives different answers for vector and diagonal matrix inputs.
|
|
The following code demonstrates this bug .
|
|
The output is not a zero vector.
|
|
See http nbviewer.ipython.org gist ketch 6be81962868fa9e4fb4f for more.","bug"
|
|
"6013","5097","should logaddexp really give a warning for inf?.
|
|
.
|
|
Maybe it could be argued that it should give a warning because log 0) gives a warning, but on the other hand it could be special cased.
|
|
https github.com scipy scipy issues 4014","bug"
|
|
"6013","5098","F2py wrapped routines shouldn t dump messages to stderr.
|
|
https github.com numpy numpy blob master numpy f2py src fortranobject.c L777 In fortranobject.c, function check and fix dimensions calls fprintf on error.
|
|
This dumps a message to stderr which should probably be included in the error message passed up as a python exception instead.","bug"
|
|
"6013","510","code cleanup for defchararray.py Trac 26).
|
|
Original ticket http projects.scipy.org numpy ticket 26 on 2006 03 07 by timleslie, assigned to unknown.
|
|
patch attached to fix tabs instead of spaces unused import","bug"
|
|
"6013","5100","NumPy arrays of objects.
|
|
Since the recent release 1.9 of NumPy, the numpy.array function no longer infer the type of class instances as object if the class defines a ` getitem ` method.
|
|
See the following notebook for an example http nbviewer.ipython.org gist jrjohansson 5c977ba22dc952ce2675 This example is a bit artificial but we first encountered this issue in the QuTiP package qutip qutip 228.
|
|
In earlier versions of NumPy it used to work without explicitly having to specify dtype=object.
|
|
I assume that this happens because the numpy.array function iterates over all elements in the list that is passed to it, and in addition also tries to iterate over all values that ` getitem ` item returns.
|
|
But at least for the use cases I often encounter this is not a very intuitite behaviour, since it doesn t make much change to unravel custom defined classes into their numerical values just because they define a ` getitem ` function.
|
|
Is it an intentional change?
|
|
Or is the way I have use the ` getitem ` function not suitable for using NumPy arrays?","bug"
|
|
"6013","5108","ma.power should not take masked values into account.
|
|
Method numpy.ma.power computes the result taking into account the masked values Example a = np.ma.array [1.8446744e 19, 1, 2, 3], mask=[True, False, False, False], fill value=1.8446744e 19, dtype=np.float32) np.ma.power a, 2) usr lib python2.7 dist packages numpy ma core.py 6012 RuntimeWarning overflow encountered in power result = np.where m, fa, umath.power fa, fb)).view basetype)","bug"
|
|
"6013","511","Bug in complex pow optimization Trac 27).
|
|
Original ticket http projects.scipy.org numpy ticket 27 on 2006 03 09 by unknown, assigned to unknown.
|
|
I just noticed that there s a bug in the complex pow optimizations that David Cooke just checked in.
|
|
There s a bi < 0) where there should be a br < 0) , which means that inverse integral powers will be wrong.
|
|
I can fix this myself, however numpy is also not compiling for me at the moment using VC7, so I ll be working on that first.","bug"
|
|
"6013","5110","argmin argmax NaN behaviour.
|
|
Be A any array like containing `np.nan`, e.g.
|
|
`A = [1,2,0, 2,np.nan, 222,1000,np.pi]` While `max A)` and `min A)` expectedly yields 1000 and 222 both `np.argmin A)` and `np.argmin A)` yields `4` which is the index of the NaN value.
|
|
I`d expect argmin and argmax to disregard the NaN entry.
|
|
However is there any good reason for this behaviour which escapes me?","bug"
|
|
"6013","5112","BUG subclass ignored for inplace operations.
|
|
Normally, ` array prepare ` and ` array wrap ` can be used to prepare output given the subclass.
|
|
However, for inplace operations where the output is an `ndarray`, these are ignored independent of ` array priority `) .
|
|
This is not an issue with ` numpy ufunc `)","bug"
|
|
"6013","5114","nanmin doesn t appear to work with Pandas Series.
|
|
Using NumPy s nanmin on a pandas Series worked as expected, now however the Series is returned with each element set to the minimum, rather than just getting the minimum.
|
|
Not sure if this is a Numpy of Pandas issue.
|
|
np.nanmin pds.Series [1,2,3,4])) Out[14] 0 1 1 1 2 1 3 1 dtype int64 np.min pds.Series [1,2,3,4])) Out[15] 1","bug"
|
|
"6013","512","bug in subtraction of two NP.poly1d s with unequal number of coefficients Trac 28).
|
|
Original ticket http projects.scipy.org numpy ticket 28 on 2006 03 09 by trac user GerardVermeulen, assigned to unknown.
|
|
The bug can be triggered by >>> NP.
|
|
version 0.9.6.2210 >>> NP.poly1d [1]) NP.poly1d [1,0]) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File home packer usr lib python2.4 site packages numpy lib polynomial.py , line 476, in sub return poly1d polysub self.coeffs, other.coeffs)) File home packer usr lib python2.4 site packages numpy lib polynomial.py , line 268, in polysub zr = NX.zeros diff, a1) TypeError data type not understood >>> and the fix is numpy numpy lib polynomial.py.dtype 2006 03 03 09 34 40.000000000 0100 numpy numpy lib polynomial.py 2006 03 03 09 36 29.000000000 0100 265,10 265,10 if diff == 0 return a1 a2 elif diff > 0 zr = NX.zeros diff, a1) zr = NX.zeros diff, a1.dtype) val = NX.concatenate zr, a1)) a2 else zr = NX.zeros abs diff), a2) zr = NX.zeros abs diff), a2.dtype) val = a1 NX.concatenate zr, a2)) if truepoly val = poly1d val) Gerard","bug"
|
|
"6013","5121","timedelta64 formatting error failing conversion to int).
|
|
import numpy format numpy.timedelta64 10, m ), ) >>> Traceback most recent call last) >>> File , line 3, in <module> >>> TypeError don t know how to convert scalar number to int This is because of the code in [multiarray scalartypes] http github.com numpy numpy blob 04f6c0fe numpy core src multiarray scalartypes.c.src L413).
|
|
A `timedelta64` with a unit cannot be directly converted to int, so is not formattable in this way.
|
|
For more conversation on the issue, see [this Stack Overflow question] http stackoverflow.com questions 26043994 printing numpy timedelta64 with format 26044487 26044487).","bug"
|
|
"6013","5123","Inconsistent keywords between histogram functions?.
|
|
Should the functions [`histogram2d`] http docs.scipy.org doc numpy reference generated numpy.histogram2d.html) and [`histogramdd`] http docs.scipy.org doc numpy reference generated numpy.histogramdd.html) have the keyword `density` added and `normed` deprecated, to match [`histogram`] http docs.scipy.org doc numpy reference generated numpy.histogram.html)?","bug"
|
|
"6013","513","reshape inconsistency between method and function Trac 29).
|
|
Original ticket http projects.scipy.org numpy ticket 29 on 2006 03 09 by trac user cjw ..., assigned to unknown.
|
|
The function reshape, when applied to an instance of a sub class, returns an array instance.
|
|
The method reshape returns and instance of the sub class.
|
|
It seems desirable that both be treated in the same way.
|
|
Colin W.","bug"
|
|
"6013","5134","Add ndarray.keys ) and ndarray.values ) methods.
|
|
It would be useful for duck typing if ndarray had keys ) and values ) methods.
|
|
Here s how I would implement them in Python def keys self) return self.dtype.names def values self) return self.flat The way I ve shown, keys ) would return a tuple and values ) would return an iterator as dict does in Python 3).
|
|
This would make it easier to write generic code which handles structured arrays, dicts, and other mapping types transparently, for example np.array [ 1, 3), 4, 5)], dtype=[ a , int), b , int)]) { a 1,3), b 4,5) } Without this, I currently have ugly code such as def hasColumn data, name) columns = data.dtype.names if isinstance data, np.ndarray) else data.keys ) return name in columns As far as I know, these methods would actually be implemented in C, specifically by adding to `array methods` in multiarray methods.c and using `flat` from multiarray getset.c).","enhancement"
|
|
"6013","5136","ImportError No module named thread.
|
|
When compiling and using numpy 1.9.0 with a single threaded python, an import numpy will lead to an ImportError saying that there is no module named thread.
|
|
The exception trace looks like follows File ... lib python2.7 site packages numpy init .py , line 189, in <module> from .
|
|
import random File ... lib python2.7 site packages numpy random init .py , line 99, in <module> from .mtrand import File mtrand.pyx , line 130, in init mtrand numpy random mtrand mtrand.c 27399) File ... lib python2.7 threading.py , line 6, in <module> import thread The issue was presumably introduced with commit 94172e1bbaf48e121f90d0252e33dc9f433b1534 of ENH replace GIL of random module with a per state lock where an from threading import Lock was added to numpy random mtrand mtrand.pyx.","bug"
|
|
"6013","5138","strange behaviour of np.nanmedian in np 1.9.0.
|
|
This looks good .
|
|
This looks ungood .
|
|
But this looks good .","bug"
|
|
"6013","514","Problem creating a numpy object array Trac 30).
|
|
Original ticket http projects.scipy.org numpy ticket 30 on 2006 03 10 by trac user cjw ..., assigned to unknown.
|
|
== Illustration of the problem == ugh11 To investigate an apparent object array problem import numpy as n a= n.array [[None, None, [5, 7, 8, 9]], [[1, 7], [5, 7, 8], None]], object) for i in range 2) for j in range 3) print i, j, a[i, j] if a[i, j] print len a[i, j]) z= 3 z= 2 In the above, [[None, None, [5, 7, 8, 9]], [[1, 7], [5, 7, 8], None]] is treated as a tuple, which is perhaps questionable, as it might appear to be two arguments.
|
|
However, wrapping the above in parentheses does not change the result.
|
|
Result of running the script is 0 0 None 0 1 None 0 2 [5, 7, 8, 9] 4 1 0 [[1, 7], [5, 7, 8], None] 3 1 1 [[1, 7], [5, 7, 8], None] 3 1 2 [[1, 7], [5, 7, 8], None] 3 Wrapping the expression in brackets adds an extra dimension but leaves the same incorrect result.
|
|
Above script with extra dimension added import numpy as n a= n.array [[[None, None, [5, 7, 8, 9]], [[1, 7], [5, 7, 8], None]]], object) for i in range 2) for j in range 3) print i, j, a[0, i, j] if a[0, i, j] print len a[0, i, j]) z= 3 z= 2 Result from the second script [Dbg]>>> 0 0 None 0 1 None 0 2 [5, 7, 8, 9] 4 1 0 [[1, 7], [5, 7, 8], None] 3 1 1 [[1, 7], [5, 7, 8], None] 3 1 2 [[1, 7], [5, 7, 8], None] 3 >>>","bug"
|
|
"6013","5141","ndarray.std returns strange value with small log.
|
|
When I search scikit learn issue, I found strange behaviour of ndarray.std.
|
|
Althought I understand this argument is too small value, I report it just in case.
|
|
I use Numpy 1.9.0, Python 3.4.1, Linux.
|
|
.","bug"
|
|
"6013","5142","Unintuitive behavior of np.clip on inconsistent amin and amax.
|
|
If `a min > a max`, the results can get a bit strange .
|
|
Now this can arguably be blamed on faulty inputs, and a precise reading of the docs it returns An array with the elements of `a`, but where values < `a min` are replaced with `a min`, and those > `a max` with `a max`. )
|
|
explains that behavior once we realize that the two tests have to be made one after the other), but throwing a `ValueError` possibly silenceable activatable by a keyword switch) could be helpful.","bug"
|
|
"6013","5143","khatri rao product.
|
|
They are asking for this on stackoverflow, and I ve also had a need for this myself.
|
|
It s an I want to save a few keystrokes argument.
|
|
http stackoverflow.com questions 26153147 special kind of row by row multiplication of 2 sparse matrices in python http en.wikipedia.org wiki Kronecker product Khatri Rao product In particular the row wise or column wise Kronecker product.","bug"
|
|
"6013","5145","np.unique no longer works with fixed float and string arrays in Python 3.
|
|
In Python 2, the following code used to work, but no longer in Python 3 .
|
|
Is this intended behavior, and if so, are there any suggested alternatives in this situation?","bug"
|
|
"6013","5147","Einsum double contraction in particular order causes ValueError.
|
|
If I try to do .
|
|
I get a `ValueError` >>> ValueError dimensions in operand 0 for collapsing index j don t match 0 != 3) However, these alternatives work .
|
|
>>> True And these agree with the naive version up to the last decimal place .
|
|
>>> True","bug"
|
|
"6013","5148","Numpy crashes with OO.
|
|
Hello, I ran ` python OO c import numpy ` on a fresh miniconda install after ` conda install numpy`) and I got the following output .
|
|
Running ` python` gives me the following output I am running OS X Mavericks on a Dec 2010 MacBook Pro) .
|
|
Finally, running `python c import numpy ` works normally.
|
|
I hope this is enough information for you all, but please let me know if you need more.
|
|
It s the second bug report I submit in my life, and the first one was about a typo in documentation!)
|
|
Thanks a lot for making numpy available to us, it is a very, very useful tool that works greatly for me.","bug"
|
|
"6013","515","Wrong values in numpy.typeNA Trac 31).
|
|
Original ticket http projects.scipy.org numpy ticket 31 on 2006 03 11 by trac user faltet ..., assigned to unknown.
|
|
Hi, There are wrong values assigned in the typeNA dictionary In [5] numpy.typeNA[numpy.int64] Out[5] Int32 In [6] numpy.typeNA[numpy.uint64] Out[6] UInt32 The next patch solves the problem numpy core numerictypes.py revision 2217) numpy core numerictypes.py working copy) 207,10 207,10 UIntname = UInt d bits Intname = Int d bits uval = typeinfo[ U ctype] typeobj = val[ 1] utypeobj = uval[ 1] if intname not in allTypes.keys ) uintname = uint d bits typeobj = val[ 1] utypeobj = uval[ 1] allTypes[intname] = typeobj allTypes[uintname] = utypeobj typeDict[intname] = typeobj 219,15 219,10 typeDict[UIntname] = utypeobj typeNA[Intname] = typeobj typeNA[UIntname] = utypeobj typeNA[typeobj] = Intname typeNA[utypeobj] = UIntname typeNA[val[0]] = Intname typeNA[uval[0]] = UIntname else typeNA[typeobj] = Intname typeNA[utypeobj] = UIntname typeNA[val[0]] = Intname typeNA[uval[0]] = UIntname typeNA[typeobj] = Intname typeNA[utypeobj] = UIntname typeNA[val[0]] = Intname typeNA[uval[0]] = UIntname add integer aliases ) We use these later Cheers","bug"
|
|
"6013","5150","integer and should probably raise divide by zero errors unconditionally, instead of just flagging divide by zero errstate.
|
|
Problem For floats, our default on divide by zero is to print a warning and return `nan`.
|
|
This is a pretty good default.
|
|
But we ve also carried this over to integers, where there is no `nan`.
|
|
Instead we pick a random value zero, apparently) and return that .
|
|
This violates the rule that we shouldn t just make up random nonsense when the user requests an impossible operation.
|
|
With a `nan` return, it s obvious that something has gone wrong; even if you don t handle the error immediately, your overall computation won t silently go wrong.
|
|
In the integer case, right now, unless you are specifically watching the console and notice the warning, there s no way to know that anything has gone wrong at all, and probably people are getting incorrect answers right now without realizing it.
|
|
My feeling is that we should after a deprecation period) make integer divide by zero a hard error.
|
|
Affected ufuncs `divide` python 2 only), `floor divide`, `remainder` `remainder` is particularly terrible `np.remainder 5, 0)` returns 0 as well, so we don t even satisfy the invariant that `num = num denom) denom num denom)`.)
|
|
Relevant mailing list thread http mail.scipy.org pipermail numpy discussion 2014 October 071306.html Not sure if the mailing list discussion has converged yet or not there is [one objection from Robert Kern] http mail.scipy.org pipermail numpy discussion 2014 October 071330.html) which [I think is unconvincing] http mail.scipy.org pipermail numpy discussion 2014 October 071331.html), but that s just me; I could be missing something.","bug"
|
|
"6013","5155","loadtxt comments=None) considers the string None as a comment symbol.
|
|
Numpy 1.8.2 `numpy.loadtxt comments=None)` considers the string ` None ` as a comment symbol.
|
|
Expected behaviour `comments=None` should indicate that there are no comments symbols.
|
|
Certainly not cast `None` to a string, what is this, Javascript?","bug"
|
|
"6013","5157","memmap docstring doesn t say how a new mapping is filled.
|
|
Is `np.memmap` like `np.zeros`, or like `np.empty` wrt.
|
|
the initial values in the array?","bug"
|
|
"6013","5159","memory leak in ndarray.flatten )?.
|
|
Unless I m missing something, in numpy 1.9.0 I haven t checked any other versions) calls to ``ndarray.flatten )`` appear to leak memory.
|
|
In example below, note how ``wset`` grows by approximately 400mb with every call to ``a.flatten )``.
|
|
Note also that calls to ``gc.collect )`` don t seem to help.
|
|
.
|
|
Perhaps related to https github.com numpy numpy issues 1601?","bug"
|
|
"6013","516","record turnaround fails Trac 32).
|
|
Original ticket http projects.scipy.org numpy ticket 32 on 2006 03 11 by trac user cjw ..., assigned to unknown.
|
|
Example borrowed from testing ar= n.recarray [ 5.1000000000000005, 61, abcde ), 5.2000000000000002, 62, fghij ), 5.3000000000000007, 63, kl )], dtype= n.record,17)) gives response File C \Documents and Settings\cjw\Documents\Py\Rec\tRecords.py , line 43, in check recarray fromfile dtype= n.record,17)) TypeError new ) got an unexpected keyword argument dtype","bug"
|
|
"6013","5163","Segmentation fault on import numpy on redhat5.
|
|
Environment GCC 4.1.2 Linux Python 2.7 How to reproduce .
|
|
Traceback of the Python interpreter .","bug"
|
|
"6013","5164","ENH Weight support for np.var.
|
|
Is there any reason why there is no `weights` keyword for `np.var` similar to `np.average`?
|
|
We sometimes need to custom code for this in sklearn sometimes, for example, [Here] https github.com scikit learn scikit learn pull 3474 files diff 383546d17769864242e4668d2338b449R136) If this is agreed upon, I can submit a Pull Request.","enhancement"
|
|
"6013","5165","Slicing huge matrix larger 1.5k,80000k)) overwrites parts of matrix with 0.
|
|
Hi all, I encountered a problem slicing huge matrices.
|
|
My matrix has approximately the following dimensions M = 1500,8000000) 1.5k, 8 million).
|
|
I tried to remove some columns of this matrix.
|
|
However, After the slicing operation is done the new matrix contains the correct dimensions but the tail of the matrix only contains zeros.
|
|
My machine has 512GB of memory.
|
|
So this shouldn t be a problem.
|
|
Here is a small example with random numbers <code> In [1] import numpy as np < code> <code> In [2] m = np.random.randn 1500,8000000) < code> <code> In [3] ind = np.arange 0,7000000) < code> <code> In [4] new m = m[ ,ind] < code> <code> In [5] new m[ , 1] < code> <code> Out[6] array [ 0., 0., 0., ..., 0., 0., 0.])
|
|
< code> The matrix new m contains now many columns with zeros.
|
|
<code> In [7] np.where new m.sum axis=0)==0)[0].shape < code> <code> Out [8] 5726623,) < code> Is this a problem someone else experienced?
|
|
Is there a slicing limit?
|
|
Best wishes, Dominik","bug"
|
|
"6013","5166","matrix power fails overflows for large exponents..
|
|
I noticed that there seems to be an overflow happening with `numpy.linalg.matrix power` when using large exponents.
|
|
In my opinion, this should at least raise an `OverflowError`.
|
|
np.version.git revision Out[8] 68538b74483009c2c2d1644ef00397014f95a696 np.version.full version == 1.6.1 In [10] A Out[10] array [[1, 2], [1, 0]]) In [11] np.linalg.matrix power A, 100) Out[11] array [[ 6148914691236517205, 6148914691236517206], [ 6148914691236517205, 6148914691236517206]])","bug"
|
|
"6013","517","Unable to import multiarray Trac 33).
|
|
Original ticket http projects.scipy.org numpy ticket 33 on 2006 03 11 by unknown, assigned to unknown.
|
|
Using numpy 0.9.5 First I tried to import numeric, this gave a process crash.
|
|
Then I tried the following, which again crashed.
|
|
C \Python24\Lib\site packages\numpy\core>python Python 2.4.2 67, Sep 28 2005, 12 41 11) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import multiarray C \Python24\Lib\site packages\numpy\core> I could copy the Windows XP response.
|
|
Colin W.","bug"
|
|
"6013","5170","numpy npz files with v1 and v2 formats are not interoperable between python 2 and python 3..
|
|
The header is a python literal, unfortunately this leaves the numpy file unreadable when it is an array and the literal for array dimensions are Long integers.
|
|
e.g.
|
|
{ descr <i2 , fortran order False, shape 100L, 200L), } \n When loading with an array in the npz file saved with python 2.x and opening with python 3.x we get the following error.
|
|
Exception SyntaxError invalid syntax , <unknown> , 1, 55, { descr <i2 , fortran order False, shape 100L, 200L), } \n The 100L and 200L are invalid python 3 literals.
|
|
A fix is to run a substitution on them, such as this one, for old files.
|
|
header = re.sub b \d )L , b \\1 , header)","bug"
|
|
"6013","5171","Numpy npz files need to not use python literals as header descriptors..
|
|
In issue 5170 we have an incompatibility between npz files from python 2.x opening in python 3.x This is caused by using a evaluated python literal storing the dictionary used for the header.
|
|
This header should be properly serialized.
|
|
There are to main options Pickle JSON Please update the npz file format accordingly, evaluated python literals are not a reliable method to store dictionaries.
|
|
See file numpy lib format.py where the use of a python pretty printed representation is documented.","bug"
|
|
"6013","5172","nanmean numpy 1.9.0 Windows 32 bit.
|
|
In some circumstances, the first time nanmean is called in my code, the result is NaN even when there are no NaNs in the array.
|
|
On the second call, the correct mean ignoring NaNs) is reported.
|
|
This behavior is not seen in 64 bit Windows numpy 1.9.0.
|
|
Tragically, I cannot reproduce it outside of a place buried deep in my code.","bug"
|
|
"6013","5173","Enhancement Request Alternative to `random.shuffle`, with an `axis` argument..
|
|
It would be nice to have an alternative to `numpy.random.shuffle` that accepts an `axis` argument, and that independently shuffles the one dimensional slices.
|
|
Here s an implementation that I ll call `disarrange`.
|
|
It works, but it would be nice to have a more efficient C implementation.
|
|
.
|
|
Example .
|
|
This request was motivated by this question on stackoverflow http stackoverflow.com questions 26310346 quickly calculate randomized 3d numpy array from 2d numpy array","enhancement"
|
|
"6013","5174","Changing warning to Exception changes FutureWarning to DeprecationWarning.
|
|
Using In [24] np.
|
|
version Out[24] 1.9.0.dev 59d3bbb .
|
|
If we try to catch the FutureWarning, the comparison returns True .
|
|
If we raise DeprecationWarnings only, the comparison returns [True] .
|
|
But if we try to raise both DeprecationWarning and FutureWarning, we get a DeprecationWarning exception .","bug"
|
|
"6013","5179","n array versions of set operations, especially intersect1d.
|
|
The stock Python set operations support multiple inputs, while the Numpy equivalents only support two arrays.
|
|
A version of intersect1d supporting n arrays was posted to numpy discussion back in 2009 http numpy discussion.10968.n7.nabble.com intersect1d for N input arrays td25956.html The posted multi array version works well for me, it would be great to see it or a variant) in Numpy.","enhancement"
|
|
"6013","518","import error in numpy.core.chararray Trac 34).
|
|
Original ticket http projects.scipy.org numpy ticket 34 on 2006 03 12 by trac user GerardVermeulen, assigned to unknown.
|
|
The bug can be triggered by >>> import numpy; numpy.
|
|
version 0.9.6.2221 >>> import numpy.core.chararray Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File home packer usr lib python2.4 site packages numpy core chararray.py , line 1, in ?
|
|
from numerictypes import character, string, unicode , \ ImportError cannot import name obj2dtype >>> Gerard","bug"
|
|
"6013","5180","np.vectorize handling of timedelta64 arrays.
|
|
It seems that `np.vectorize` casts unboxes `timedelta64` values and does so kind of inconsistently depending on the unit.
|
|
I am not sure that this was intentional, but makes it difficult to work with such arrays; in particular, it is not possible to vectorize a function which takes `timedelta64` value; >>> np.
|
|
version 1.9.0 >>> f = lambda x print type x), x) >>> xs = np.arange 1, 3).astype timedelta64[h] ) >>> xs array [1, 2], dtype= timedelta64[h] ) with `[h]` and `[s]` unit >>> = np.vectorize f) xs) <class numpy.timedelta64 > 1 hours <class datetime.timedelta > 1 00 00 <class datetime.timedelta > 2 00 00 >>> = np.vectorize f) xs.astype timedelta64[s] )) <class numpy.timedelta64 > 3600 seconds <class datetime.timedelta > 1 00 00 <class datetime.timedelta > 2 00 00 >>> = np.vectorize f, otypes= O ) xs.astype timedelta64[s] )) <class datetime.timedelta > 1 00 00 <class datetime.timedelta > 2 00 00 with `[ns]` unit >>> = np.vectorize f) xs.astype timedelta64[ns] )) <class numpy.timedelta64 > 3600000000000 nanoseconds <class int > 3600000000000 <class int > 7200000000000 >>> = np.vectorize f, otypes= O ) xs.astype timedelta64[ns] )) <class int > 3600000000000 <class int > 7200000000000","bug"
|
|
"6013","5181","Memory leak in pyfragments.swg.
|
|
I think I found a memory leak in pyfragments.swg.
|
|
In both `SWIG AsVal dec` fragments, `PyArray DescrNewFromType` is called but its return value may not get deallocated.
|
|
I moved the calls to `PyArray DescrNewFromType` to right before their value is used the calls to `PyArray CastScalarToCtype`) and added a call to `Py DECREF` and that seems to have fixed my memory leak.
|
|
I discovered this in numpy 1.4.1, but I checked 1.9.0 and I think it s still present there.","bug"
|
|
"6013","5184","numpy.gradient ) doesn t compute boundary values correctly in version 1.9.0.
|
|
It seems that numpy.gradient ) computes the boundary values incorrectly in version 1.9.0 but correctly in e.g.
|
|
version 1.8.1) To reproduce Alternative 1 >>> x = np.array [1, 2, 4], dtype=np.float) >>> np.gradient x) array [ 0.5, 1.5, 2.5]) although the resulting array should be [ 2 1) 1, 4 1) 2, 4 2) 1] = [1, 1.5, 2] Alternative 2 Just run the first example in the documentation for numpy.gradient ) >>> x = np.array [1, 2, 4, 7, 11, 16], dtype=np.float) >>> np.gradient x) The result is array [ 0.5, 1.5, 2.5, 3.5, 4.5, 5.5]) although the documentation says that it should be array [ 1. , 1.5, 2.5, 3.5, 4.5, 5. ])
|
|
the latter also seems right, while the former seems wrong.
|
|
see also http stackoverflow.com questions 26361007 numpy gradient seems to produce erroneous boundary values using first differe)","bug"
|
|
"6013","5185","MaskedArray.dot leaks masked values.
|
|
.
|
|
Note that this is a regression from oldnumeric .","bug"
|
|
"6013","5188","Uncatchable FutureWarning.
|
|
`np.array equal np.array [nan], dtype= O ), np.array [nan], dtype= O ))` issues a FutureWarning.
|
|
But how do you catch it as an exception?
|
|
.
|
|
The workaround [discussed here] https github.com pydata pandas issues 8537) does not work in this case .","bug"
|
|
"6013","519","Not able to add a title or label to a dtype Trac 35).
|
|
Original ticket http projects.scipy.org numpy ticket 35 on 2006 03 12 by trac user cjw ..., assigned to unknown.
|
|
[Dbg]>>> n.dtype [ name , >i4 , 3)]) dtype [ name , >i4 , 3)]) [Dbg]>>> n.dtype [ name , label ), >i4 , 3)]) Traceback most recent call last) File <interactive input> , line 1, in ?
|
|
TypeError data type not understood [Dbg]>>> The above is handled OK for the field name or identifier but the title, label or heading fails.
|
|
Colin W","enhancement"
|
|
"6013","5193","numpy apply along axis drops values after casting incorrectly.
|
|
Hello, I have found that when I make the following call using apply along axis and return an int rather than a float, the field with the incorrect type is cast to the right type and put into the array properly, but the value of the subsequent field is set to zero .
|
|
.
|
|
discovered with asna1005","bug"
|
|
"6013","5194","mingw w64 builds of numpy.. Hi everyone!
|
|
I m maintainer of MSYS2 project.
|
|
We provide builds of different open source software for Windows using mingw w64 toolchains.
|
|
What we provide https github.com Alexpux MINGW packages We also provide our own Python builds builded with mingw w64) and numpy for both 32 and 64 bit platforms.
|
|
Many users successfully use it.
|
|
I m read your wiki https github.com numpy numpy wiki Mingw static toolchain https github.com numpy numpy wiki Mingw w64 faq As we build our own Python we don t have any limitations with needing statically linked extensions.
|
|
To build numpy I m write small patch that enable 64 bit building and remove functionality for creating import library from MS Python DLL.
|
|
Running numpy testsuite for our 64 bit builds are good.
|
|
For example, 64 bit python 3.4.2 .
|
|
But we have some failures with numpy testsuite for 32 bit.
|
|
.
|
|
Can someone help to find real problem with it?
|
|
Two days ago we have 13 failures but we fix 9 of them in mingw w64 runtime.
|
|
Regards, Alexey.","bug"
|
|
"6013","5195","hard crash when using flat on non contiguous size 0 array.
|
|
.
|
|
Running this test script, I get .
|
|
I m using conda on OS X, with Python 2.7 and numpy 1.9.0.
|
|
Yes, this is quite an edge case, but I really did encounter it today ).","bug"
|
|
"6013","5197","Add an axis argument to generalized ufuncs?.
|
|
I would like to write generalized ufuncs, [using numba] https github.com numba numba issues 825), to create fast functions such as `nanmean` signature ` n) > ) `) or `rolling mean` signature ` n), ) > n) `) that take the axis along which to aggregate as a keyword argument, e.g., `nanmean x, axis=0)` or `rolling mean x, window=5, axis=0)`.
|
|
Of course, I have the option of writing my own wrapper function to do something like `x.swapaxes axis, 1)` before passing the array off to the gufunc.
|
|
This is not so terrible.
|
|
However, it occurs to me that this sort of dimension reordering might be handled more cleanly and efficiently) within the gufunc machinery itself, somewhat analogously to `numpy.ufunc.reduce`, except the `axis` argument if supplied) would need to be an integer or tuple with length equal to the number of core dimensions on the first gufunc argument.
|
|
Thoughts?","enhancement"
|
|
"6013","5198","numpy.in1d crashs.
|
|
using numpy 1.9 with python 2.7.6 , pythonw.exe multiarray.pyd) crashs with that code import numpy a=numpy.array [8, 8, 7, 2, 3, 4, 2, 4, 5, 2, 8, 7, 2, 1, 5, 1, 8, 6, 3, 0, 9, 3, 8, 0, 3, 0, 2, 2, 6, 2, 4, 3, 5, 0, 4, 0, 3, 7, 2, 7, 1, 3, 9, 2, 3, 9, 1, 2, 7, 4, 4, 0, 2, 8, 4, 8, 5, 6, 1, 8, 7, 6, 9, 6, 7, 5, 4, 7, 5, 8, 5, 1, 1, 8, 8, 2, 2, 9, 4, 9, 7, 5, 2, 9, 6, 2, 6, 3, 7, 4, 4, 3, 0, 4, 0, 6, 8, 2, 4, 1, 3, 4, 5, 2, 2, 2, 4, 8, 4, 8, 5, 6, 4, 1, 8, 7, 8, 1, 7, 7, 5, 3, 1, 2, 1, 0, 5, 6, 1, 9, 7, 0, 5, 6, 8, 2, 9, 9, 5, 8, 0, 4, 3, 2, 0, 3, 7, 3, 3, 4, 8, 3, 5, 8, 6, 6, 4, 2, 1, 8, 4, 0, 5, 2, 5, 5, 7, 5, 0, 6, 8, 9, 0, 5, 7, 9, 1, 9, 3, 6, 9, 8, 4, 6, 0, 0, 5, 8, 3, 6, 1, 7, 1, 2, 2, 9, 7, 7, 9, 9]) b=numpy.array [9, 0, 0, 6, 6, 7, 5, 6, 8, 6, 2, 7, 6, 8, 4, 5, 6, 6, 3, 6, 0, 1, 7, 5, 1, 8, 6, 6, 0, 3, 2, 5, 7, 2, 3, 3, 2, 2, 4, 2, 9, 1, 7, 1, 1, 3, 8, 5, 1, 2, 4, 1, 2, 9, 2, 2, 4, 9, 8, 2, 5, 8, 5, 8, 5, 0, 4, 0, 4, 5, 5, 8, 5, 1, 8, 5, 8, 0, 7, 6, 6, 0, 6, 8, 5, 8, 2, 9, 4, 3, 9, 1, 2, 8, 2, 2, 5, 6, 8, 8, 8, 1, 6, 8, 9, 7, 0, 2, 0, 8, 5, 4, 1, 1, 7, 4, 2, 7, 0, 6, 5, 9, 4, 9, 6, 8, 1, 8, 7, 3, 6, 5, 0, 7, 5, 3, 3, 3, 8, 8, 8, 6, 3, 8, 0, 6, 9, 4, 8, 4, 7, 8, 7, 3, 9, 9, 0, 8, 5, 9, 7, 6, 8, 1, 3, 5, 9, 1, 9, 6, 9, 1, 9, 2, 7, 1, 9, 5, 4, 9, 1, 7, 8, 5, 4, 2, 8, 8, 4, 5, 8, 9, 2, 6, 6, 4, 7, 2, 7, 2]) temp= numpy.in1d a,b)","bug"
|
|
"6013","5199","Using datetime converters in numpy.genfromtxt ) results in crashes.
|
|
Suppose I have a CSV file of floats, except one column e.g.
|
|
2nd column) which is a datetime column with dates in the following format 2014 04 31 01 16 11.745534 .
|
|
I want to use a converter in order to transform it into epoch values.
|
|
The datetime may or may not contain milliseconds it mostly does).
|
|
My converter is .
|
|
If I do not specify the dtype parameter in genfromtxt ) , it is considered float by default.
|
|
But then npyio.py crashes at line 1763 with this message TypeError object of type float has no len ) What happens genfromtxt ) gets smart and, instead of reading the date as a string into data , it identifies that the second column is a date and converts it into an epoch value, which is a float.
|
|
Unfortunately it discards the milliseconds in the process, by the way.
|
|
Then len row[i]) crashes with the error mentioned above because floats do not have lengths.
|
|
If I specify dtype = None , npyio.py crashes at line 1705 with the same message TypeError object of type float has no len ) .
|
|
genfromtxt ) got smart again and pre converted the datetime to epoch, which is a float.
|
|
And again, it lost the milliseconds on the way.
|
|
I think genfromtxt ) should read the columns that are to be converted as strings and let the user convert them as they like.
|
|
This is how I used genfromtxt ) .
|
|
or .
|
|
One more mention, just in case defining str2epoch as a proper function instead of a lambda function does not solve the problem.","bug"
|
|
"6013","520","several patches for linalg Trac 36).
|
|
Original ticket http projects.scipy.org numpy ticket 36 on 2006 03 14 by trac user Norbert ..., assigned to unknown.
|
|
attached a few patches for linalg against revision 2237 numpy 1 cleanup.diff just a bit of cosmetics to make the other patches cleaner numpy 2 rename functions.diff numpy.linalg used to contain a number of aliases like inv == inverse and others) this is pointless and more confusing then helpful.
|
|
Unless there is a really good reason, a library should offer one name only.
|
|
I chose the short version of each name as the official name, because I think all the abbreviations are well known and easy to understand for anyone handling numerical linear algebra.
|
|
Currently all the definition that would be needed for backwards compatibility are deactivated in a if False block.
|
|
If people think they should be deprecated slowly, this block could easily be re activated.
|
|
numpy 3 svd compute uv.diff numpy.linalg.svd now also has the compute uv option that scipy.linalg.svd already had.
|
|
Default behavior is the same as before.
|
|
numpy 4 svd bug workaround.diff the dgesdd function of the lapack library installed on my system seems to contain a strange bug.
|
|
Should probably be investigated and fixed.
|
|
For the moment I just included a workaround.
|
|
numpy 5 norm copy from scipy.diff copied and the norm function from scipy.linalg and adjusted internals to work in the new environment numpy 6 norm change default.diff the scipy.linalg.norm function was sub optimal for matrices, the frobenius norm is faster then the max svd )) by about an order of magnitude for all the test cases that I came up with.
|
|
It also is invariant under orthogonal unitary transformations, which makes it the best candidate to be the default in any case that I could come up with.
|
|
The computation of the Frobenius norm can be done with the same line of code used for the vector square norm and can even be generalized to arrays of arbitrary rank, always giving a reasonable norm for both real and complex numbers.
|
|
The most straightforward way to clean up a number of inefficiencies in the definition of norm was to introduce a norm with ord=None which always calculates sqrt sum conjugate x) x).ravel ))) which is generally the most efficient implementation and does not even need a check of the rank of the array.
|
|
I also made this choice the default of the norm function.
|
|
I know that this may cause hard to find errors in existing code if people explicitely want the svd norm for a matrix and relied on the old default setting.
|
|
Personally, I don t believe this is a real danger, because I don t believe there are many algorithms that depend on using the svd norm and break if frobenius is used.
|
|
If people are strictly against changing the default behavior, the patch should still be accepted and only the function header changed back.
|
|
numpy 7 castCopyAndTranspose.diff simplyfied this function for purely aesthetic reasons I believe a function that can be written cleanly in two lines should not be extended unnessecarily to 6 lines... numpy 8 dual.diff added norm , eigh and eigvalsh to the list of dual functions.","enhancement"
|
|
"6013","5200","assert almost equal ) for floats and assert array almost equal ) have different semantics.
|
|
When using `assert almost equal` for floats I get different results when using arrays.
|
|
This can be seen by this code .
|
|
Looking at `utils.py` the behavior can be explained Float When comparing using simple float values, this is the `if` that checks if values are equal `if round abs desired actual), decimal) != 0 ` https github.com numpy numpy blob master numpy testing utils.py L489 Array But when comparing using arrays `return around z, decimal) <= 10.0 decimal)` https github.com numpy numpy blob master numpy testing utils.py L838 Making the same assert behave in different ways.","bug"
|
|
"6013","5202","numpy.average demotes dtype of weights.. .
|
|
When x is float32 and y is float64, the average effectively converts y into float32.
|
|
For example .
|
|
Now a, b are both float64 .
|
|
This behaviour is undocumented, and unexpected.
|
|
It probably also is in contradiction to the general type conversion rule in numpy one would expect numpy to promote the float32 to float64 in this case.
|
|
Possible resolutions 1.
|
|
Document this.
|
|
still surprises but can be understood) 2.
|
|
Always cast the intermediate result to the bigger type in x and y. no document changes, for then average follows usual numpy conventions.)
|
|
Affects at least version 1.8.0","bug"
|
|
"6013","5208","fromfile tofile on subprocess pipe broken in Python 3.
|
|
Piping a numpy array as binary data to another process works in Python2, but in Python3 the following snippet results in OSError first argument must be a string or open file .
|
|
I tested with Python 3.4.0. .","bug"
|
|
"6013","5209","datetime tests fail on Solaris.
|
|
The datetime tests in numpy.test ) fail on Solaris SPARC.
|
|
This problem does not occur on Solaris x86.
|
|
The problem is present in Numpy 1.7.1 and in Numpy 1.9.0.
|
|
The problem is in the conversion of datetime64 to a string.
|
|
The type itself is working, but the str method generates the year as 0000 .
|
|
Example On Solaris 10 SPARC) Python 2.7.8 default, Sep 23 2014, 12 47 57) [GCC 4.9.0] on sunos5 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 1.9.0 >>> d = numpy.datetime64 2013 12 31 ) >>> d numpy.datetime64 0000 12 31 ) >>> d. str ) 0000 12 31 >>> d2 = d 1 >>> d2.
|
|
str ) 0000 01 01 >>> d.tolist ) datetime.date 2013, 12, 31) >>> d2.tolist ) datetime.date 2014, 1, 1) >>> str d2) 0000 01 01 >>>","bug"
|
|
"6013","521","scipy.signal.fftconvolve lacks real, fftn, ifftn Trac 37).
|
|
Original ticket http projects.scipy.org numpy ticket 37 on 2006 03 16 by trac user Stijn de Waele, assigned to unknown.
|
|
Solution In scipy\signal\signaltools.py add from scipy import real from scipy.fftpack import fftn, ifftn","bug"
|
|
"6013","5211","strange behavior of np.where with masked arrays in numpy 1.9.0. .
|
|
An empty list is returned here since ` 1` is masked .
|
|
But unexpectedly, this returns a non empty list .
|
|
Further weirdness .","bug"
|
|
"6013","5218","my norm is slow.
|
|
.
|
|
For `M = np.random.randn 4000, 4000)` the less vectorized implementation is a bit faster on my machine 3.9 seconds vs. 4.6 seconds for `timeit ..., number=100)`).
|
|
This seems weird to me.
|
|
Is there a faster way to do what I want?","bug"
|
|
"6013","522","strides accepted as an argument to records.array Trac 38).
|
|
Original ticket http projects.scipy.org numpy ticket 38 on 2006 03 19 by unknown, assigned to unknown.
|
|
strides should be rejected when object is a string, it is not passed to fromstring.","enhancement"
|
|
"6013","5220","BUG TZ offset of datetime64 t, ) depends on sign of t. It appears that `datetime64 t, unit)` with `unit` smaller than ` D ` yields an object with the system time zone offset when `t >= 0`, but in UTC when `t < 0`.
|
|
Given the opacity of the system time zone, it s probably better to avoid using it completely unless it s explicitly requested.
|
|
.","bug"
|
|
"6013","5221","BUG arange dtype= M8[ ] ) has wrong number of elements.
|
|
It looks like `arange` is misinterpreting `stop` as `len` when the requested dtype is `datetime64` .","bug"
|
|
"6013","5222","BUG nan)?
|
|
arg)?
|
|
max|min) handling of NaT inconsistent.
|
|
I would expect the handling of `NaT` to parallel the handling of `NaN` the `max` or `min` of an array containing `NaT` should be `NaT`, and the `argmax` or `argmin` of that array should be the index of the first `NaT`.
|
|
Instead, `max` and `min` ignore `NaT` as `nanmax` and `nanmin` could be expected to do, while `argmax` and `argmin` treat `NaT` as a minimal element.
|
|
This inconsistency breaks the invariant that `arr[argmin arr)] == min arr)` with the additional interpretation that `NaN == NaN` for this purpose).
|
|
To compound this, `nanmax` and `nanmin` each raise when applied to `m8` or `M8` arrays, and `nanargmax` and `nanargmin` have the same behavior as `argmax` and `argmin`.
|
|
.","bug"
|
|
"6013","5224","dtype aligned flag not respected by recarray.. As per my [stackoverflow question] http stackoverflow.com questions 26497656 numpy bug how do i align a numpy record array recarray 26514944) import numpy as np a = np.zeros 4, dtype=np.dtype [ t , <f8 ), d , [ a , <i4 ), b , <f8 )], 100,))], align=True)) b = np.require a, requirements=[ ALIGNED ]) print [x.flags[ ALIGNED ] for x in [a, b]]) prints `[False, False]` !
|
|
charris writes that The `recarray` class accepts an aligned parameter, but looks to lose it in `format parser` .
|
|
This bug breaks usage of PyOpenGL as per my [other stackoverflow question] http stackoverflow.com questions 26484665 opengl says from param received a non contiguous array).
|
|
As an aside why does c contiguous imply aligned for recarray?","bug"
|
|
"6013","5226","Installation fails.
|
|
An attempt to upgrade matplotlib results in an automatic upgrade of numpy but fails.
|
|
I have also tried installing numpy 1.8 and 1.7, but both failed with different errors).
|
|
Output log is here http pastebin.com pRaUXjDK Any idea what s wrong?
|
|
Max OSX 10.6) Thanks!","bug"
|
|
"6013","5227","MaskedArray ignores array priority .
|
|
Given a class like this .
|
|
Adding to a numpy array works as expected .
|
|
However, when I add `X )` to a masked array, `X. radd ` is not called .","bug"
|
|
"6013","5228","Feature request Add to numpy simple functions for transform coordinate systems.
|
|
It would be convenient to have these functions as a part of numpy [mathematical routines] http docs.scipy.org doc numpy reference routines.math.html).
|
|
cart2pol Transform Cartesian to polar coordinates .
|
|
pol2cart Transform polar to Cartesian coordinates .
|
|
cart2sph Transform Cartesian to spherical coordinates .
|
|
sph2cart Transform spherical to Cartesian coordinates .","bug"
|
|
"6013","5229","numpy 1.9.0 test suite failure on darwin.
|
|
The numpy 1.9.0 release when built against python 2.7.8 on OS X 10.10, produces a single test suite error of the form... Fortran f77 compiler sw bin gfortran Wall g ffixed form fno second underscore fPIC O3 funroll loops Fortran f90 compiler sw bin gfortran Wall g fno second underscore fPIC O3 funroll loops Fortran fix compiler sw bin gfortran Wall g ffixed form fno second underscore Wall g fno second underscore fPIC O3 funroll loops creating private tmp tmpf6LDcN private tmp tmpaq0PuV compile options I private tmp tmpf6LDcN src.macosx 10.10 x86 64 2.7 I sw src fink.build root numpy py27 1.9.0 1 sw lib python2.7 site packages numpy core include I sw include python2.7 c gfortran f77 private tmp tmpaq0PuV tmpJpFlbH.f sw bin gfortran Wall g L sw lib private tmp tmpf6LDcN private tmp tmpf6LDcN src.macosx 10.10 x86 64 2.7 test ext module 5403module.o private tmp tmpf6LDcN private tmp tmpf6LDcN src.macosx 10.10 x86 64 2.7 fortranobject.o private tmp tmpf6LDcN private tmp tmpaq0PuV tmpJpFlbH.o L sw lib gcc4.9 lib gcc x86 64 apple darwin14.0.0 4.9.2 L sw lib lgfortran o .
|
|
test ext module 5403.so Undefined symbols for architecture x86 64 PyArg ParseTupleAndKeywords , referenced from f2py rout test ext module 5403 t in test ext module 5403module.o f2py rout test ext module 5403 func in test ext module 5403module.o f2py rout test ext module 5403 func0 in test ext module 5403module.o f2py rout test ext module 5403 t2 in test ext module 5403module.o f2py rout test ext module 5403 string callback in test ext module 5403module.o This is when building against cython 0.21.1.
|
|
It would appear that the problem linkage... error Command sw bin gfortran Wall g L sw lib private tmp tmpVbXjEF private tmp tmpVbXjEF src.macosx 10.10 x86 64 2.7 test ext module 5403module.o private tmp tmpVbXjEF private tmp tmpVbXjEF src.macosx 10.10 x86 64 2.7 fortranobject.o private tmp tmpVbXjEF private tmp tmpX78lsh tmpp73F P.o L sw lib gcc4.9 lib gcc x86 64 apple darwin14.0.0 4.9.2 L sw lib lgfortran o .
|
|
test ext module 5403.so failed with exit status 1 is missing the required flags to created a shared module on darwin.
|
|
Is there a patch already in numpy git to fix this?","bug"
|
|
"6013","523","Build issues with numpy 0.9.6 and the Intel v9 Fortran compiler Trac 39).
|
|
Original ticket http projects.scipy.org numpy ticket 39 on 2006 03 22 by trac user marquardt.christian ..., assigned to unknown.
|
|
Hi, when installing numpy 0.9.6 and also 0.9.5) using the Intel v9 Fortran compiler, I ran into the following issues 1) I believe that the current v9.0) Intel compiler isn t configured correctly for Pentium 4 s and M s; also, no use is made from sse and sse2 architectures.
|
|
Instead, these processors were recognised as plain Pentiums, and a xM was added to the compiler options in order to support MMX.
|
|
However, the xM switch isn t recognised by the compiler any more... which also applies to xi which was used in another place for the Intel configuration.
|
|
I have fixed the files numpy distutils cpuinfo.py numpy distutils fcompilers intel.py accordingly.
|
|
2) If these options are actually used with the Intel compiler this applies particularly to the x<?> switches), the compiler will use functions from additional optimised libraries, for example from its vector math library libsvml.so), which in turn requires other shared libraries to be available at runtime in this case libimf.so).
|
|
However, the shared Fortran libraries build with the original version of f2py didn t know this, so the additional libs weren t loaded by ld.so, causing unresolved external errors.
|
|
A workaround is to specify the same optimisation switches to the compiler when creating the shared libraries.
|
|
I ll attach a patch.
|
|
Kind regards, Christian marquardt DOT christian AT gmail DOT com","bug"
|
|
"6013","5231","Error while trying to pip install numpy 1.9.0 .
|
|
Installing the current numpy Python 3.4.2, pyvenv virtual environment, gcc 4.8.2, Linux Ubuntu) .
|
|
ends with an error .
|
|
Going up the logs, I can spot the first compilation error to be .","bug"
|
|
"6013","5232","Linalg SVD fails if not computing UV.
|
|
I have a matrix which raises a type error when I run linalg.svd on it and set compute uv to false.
|
|
However, the SVD works fine if I also compute U and V. I have a pickle file with the particular matrix here, and a simple script to illustrate https www.wuala.com markwmuller Documents NumpyLinalgSVD problem ?key=GG0oL349Ak7J I m using numpy version 1.8.1, on python 2.7.6.
|
|
This works <code>u, s, v = numpy.linalg.svd mat)< code> But this does not <code>s2 = numpy.linalg.svd mat, compute uv=False)< code> The message reads <code> File svdProb.py , line 9, in <module> s2 = numpy.linalg.svd mat, compute uv=False) File C \Python27\lib\site packages\numpy\linalg\linalg.py , line 1339, in svd s = gufunc a, signature=signature, extobj=extobj) TypeError No loop matching the specified signature was found for ufunc svd m < code>","bug"
|
|
"6013","5236","Reshape of size=0 array fails.
|
|
.
|
|
I would expect this to set the shape to 5, 0).","bug"
|
|
"6013","5239","DOC python as glue doc file is out of date.
|
|
See the mailing list thread started at http mail.scipy.org pipermail numpy discussion 2014 October 071522.html.
|
|
The doc file is https github.com numpy numpy blob maintenance 1.9.x doc source user c info.python as glue.rst.
|
|
That mailing list thread makes no sense to me because it involves non technical drama that I don t understand, but in any case the relevant numpy doc file has not been significantly updated in six years while the methods of using Python as glue have been changing substantially.
|
|
Edit I see there is an existing issue https github.com numpy numpy issues 4492 for adding Cython instructions specifically.","bug"
|
|
"6013","524","reduce does not handle dtype=None correctly Trac 40).
|
|
Original ticket http projects.scipy.org numpy ticket 40 on 2006 03 23 by trac user sasha, assigned to unknown.
|
|
>>> numpy.add.reduce [1.,.5],dtype=None) 1","bug"
|
|
"6013","5240","BUG np.std with output given stores variance in output.
|
|
If output is assigned in `np.std`, the variance is stored in that output while the standard deviation is returned).
|
|
Possibly related is that for both `np.var` and `np.std`, what is returned is not the same as what is stored in the output array .","bug"
|
|
"6013","5241","Bug Adding a transposed array to itself.
|
|
Something very strange happens when you try adding a transposed array to itself .
|
|
`arr = arr arr.T` works fine of course.
|
|
It looks like because arr.T is a view of arr, while reading values out of arr.T it reads elements of arr that have already been written to.
|
|
This could lead to some really sneaky bugs.
|
|
Even just detecting the situation and raising an exception would be an acceptable solution.","bug"
|
|
"6013","5243","TypeError savez ) got multiple values for keyword argument file .
|
|
savez consumes the file parameter itself, thus the user can not save an array items named file .
|
|
This shall probably be documented.","bug"
|
|
"6013","5244","deprecate pkgload.
|
|
This seems to be obsolete, but I m not sure the best way to mark it as deprecated.
|
|
I d like to just delete it but I guess it must go through a deprecation period.
|
|
See https github.com numpy numpy pull 4977.","bug"
|
|
"6013","5246","BUG?
|
|
priority inversion when comparing scalar to non ndarray casts it to rank 0 array.
|
|
Here s how it can be reproduced .
|
|
Comparison type == `numpy.ndarray`) .
|
|
Addition type is `numpy.int64`) .
|
|
This happens on current master .
|
|
Is this a bug?","bug"
|
|
"6013","5247","Class arraymethod in ma.core is a descriptor, yet mutable..
|
|
This leads to interesting behaviour .","bug"
|
|
"6013","5248","The behavior of numpy.apply along axis function is not what expected..
|
|
When I use python set as elelments of a numpy array, the behavior of numpy.apply along axis function is not what expected.
|
|
For example, I want to get the union set for each column using the following codes import numpy import operator d = numpy.array [ [ set [1, 11]), set [2, 22]), set [3, 33]) ], [ set [4, 44]), set [5, 55]), set [6, 66]) ] ]) numpy.apply along axis lambda a reduce operator.
|
|
or , a), 0, d) i.e.
|
|
I expect to get a result like array [ set 1, 11, 4, 44), set 2, 22, 5, 55), set 3, 33, 6, 66) ] however, the result is array [ [ set 1, 11, 4, 44), set 2, 22, 5, 55), set 3, 33, 6, 66) ], [ set 1, 11, 4, 44), set 2, 22, 5, 55), set 3, 33, 6, 66) ], [ set 1, 11, 4, 44), set 2, 22, 5, 55), set 3, 33, 6, 66) ], [ set 1, 11, 4, 44), set 2, 22, 5, 55), set 3, 33, 6, 66) ] ]) The expected result was repeated 4 times, which is determined by the first element, actually it equals to the length of the first element.
|
|
Can this be viewed as a bug?
|
|
Anyway it is different with the way of the apply function in R language.
|
|
York","bug"
|
|
"6013","525","bug in multidimensional argmax Trac 41).
|
|
Original ticket http projects.scipy.org numpy ticket 41 on 2006 03 31 by trac user tim.hochberg ..., assigned to unknown.
|
|
argmax and I imagine some other stuff) produces incorrect results when operating on multidimensional matrices.
|
|
c = arange 9).reshape [3,3]) print c print argmax c) print c.argmax 0) ==> [[0 1 2] [3 4 5] [6 7 8]] [2 2 2] [2 2 2]","bug"
|
|
"6013","5250","bitwise and.identity is incorrect.
|
|
In numpy 1.9.0 I get the following .
|
|
But the correct identity should be a pattern of all 1 s for whatever the operation type is).
|
|
This leads to bizarre results like the following .","bug"
|
|
"6013","5252","Bizarre linalg errors with IRR.
|
|
The following code throws an error .
|
|
Bizarrely, if you change the `e 313` to `e 300` or `e 330`, the problem goes away on its own .
|
|
I don t have enough numerics fu to know what s going on here, unfortunately.
|
|
For now I m happy to round the array to 10 decimal places, which works around it, but thought I d file a bug since the behavior is pretty weird.","bug"
|
|
"6013","5254","Incorrect behaviour of numpy.random.binomial.
|
|
When I run the code below after about 29000 iterations the binomial function returns the negative value 2 31.
|
|
All the previous answers are corrects.
|
|
while True k = k 1 x = binomial 1e10, .5, 1) print x assert x >= 0, k","bug"
|
|
"6013","5259","Indexing a view with a view yields a non view?.
|
|
Apologies for the title; I don t quite know how to better phrase it.
|
|
The issue is easily demonstrated If I create a boolean mask, index with it, then create another boolean mask based off of the masked region, I can t nest the masking... better just to see it in action .
|
|
I expected `x[y][z]` to change value, since `x[y]` does.
|
|
Is this a bug, or do I harbor a deep misunderstanding of indexing and views?","bug"
|
|
"6013","526","src fortranobject.
|
|
files not being installed to the right location Trac 42).
|
|
Original ticket http projects.scipy.org numpy ticket 42 on 2006 03 31 by rkern, assigned to pearu.
|
|
These files are added to the configuration with a relative path config.add data files src fortranobject.c , src fortranobject.h , f2py.1 ) However, with a recent change, they end up being installed directly in the location of the package that adds them, not in the relative path that was specified.
|
|
This breaks f2py and consequently breaks the build of scipy.
|
|
f2py options [] adding build src fortranobject.c to sources.
|
|
adding build src to include dirs.
|
|
error file usr local lib python2.4 site packages numpy f2py src fortranobject.c does not exist rkern rkernx2 opt svn scipy ls usr local lib python2.4 site packages numpy f2py auxfuncs.py cb rules.py common rules.py diagnose.py f2py2e.py f90mod rules.py func2subr.py init .py setup.py use rules.py auxfuncs.pyc cb rules.pyc common rules.pyc diagnose.pyc f2py2e.pyc f90mod rules.pyc func2subr.pyc init .pyc setup.pyc use rules.pyc capi maps.py cfuncs.py crackfortran.py docs f2py testing.py fortranobject.c info.py rules.py svn version .py version .py capi maps.pyc cfuncs.pyc crackfortran.pyc f2py.1 f2py testing.pyc fortranobject.h info.pyc rules.pyc svn version .pyc version .pyc","bug"
|
|
"6013","5260","np.swapaxes doesn t always return a view.
|
|
Related somehow to the discussion in 5258.
|
|
The docs for `np.swapaxes` state that it returns a view if the first argument is an ndarray, but it has a fast path for `axis1 == axis2`, where it actually returns the original array incref ed .
|
|
I prefer that it always returned a view, even if it is of the whole array unchanged, as the docs say, and as e.g.
|
|
`np.transpose` does, even if the array is unchanged, and will be more than happy to submit a PR.
|
|
If there is any good reason why this should not be the case, then at least the docs should be updated, and the check for `axis1 == axis2` be moved to after they have both been checked for in boundness.","bug"
|
|
"6013","5261","Wrong definition of weights in numpy.polyfit.
|
|
The documentation below for numpy.polyfit is incorrect misleading regarding the definition of the optional input weights vector <b>w< b> http docs.scipy.org doc numpy reference generated numpy.polyfit.html In least squares fitting one generally defines the weights vector in such a way that the fit minimizes the squared error in Numpy notation) .
|
|
In common situation where the 1 sigma; errors sigma are known one has that the weights are the reciprocal of the variance .
|
|
see e.g.
|
|
http en.wikipedia.org wiki Least squares Weighted least squares or http www.itl.nist.gov div898 handbook pmd section4 pmd432.htm However the numpy.polyfit documentation defines the weight as weights to apply to the y coordinates .
|
|
This definition is not correct.
|
|
The weights apply to =multiply) the fit residuals, not only to the y coordinates.
|
|
More importantly, looking at the math in the Numpy v1.9.1) code, the resulting definition of squared residuals adopted by polyfit is the following, with the optional input weights vector <b>w< b> inside the parenthesis, contrary to standard practice .
|
|
in such a way that the relation between <b>w< b> and the 1 sigma; errors is .
|
|
which is different from what everybody will expect.
|
|
The confusion in the documentation likely arises from the fact that the Numpy code solves the linear problem below in the last squares sense, where the <b>w< b> vector does multiply the y coordinate .
|
|
And solving the above array expression in the least squares sense is equivalent to minimizing the expression below with <b>w< b> inside the parenthesis .
|
|
A non optimal solution, to maintain compatibility, would be to change the documentation and clearly define the weight <b>w< b> by including it in the equation for the squared error <i>E< i> given in the Notes.
|
|
One should also make clear that the adopted definition differs from standard practice by giving the relation between weights and error w=1 sigma; Even better would be to define a new optional keyword <b>weights< b>, which follows standard practice and satisfies <b>weights< b> = 1 sigma 2.
|
|
In this case, in the code one should simply calculate w=np.sqrt weight) of the input weights and the rest of the code applies unmodified.","bug"
|
|
"6013","5262","structured arrays should raise Exception if missing fields are requested in the form of a list.
|
|
If we have an `ndarray` with a structured dtype, and I index it with a single non existent field, an exception `ValueError`) is raised, as expected.
|
|
However, if I index it with a list of fields, of which at least one is non existent, no such exception is raised .
|
|
I think it would be more consistent if an error would be raised in both cases.","bug"
|
|
"6013","5268","Wrong result in matrix inverse function.
|
|
In Anaconda3 numpy 1.90).
|
|
I think there are something wrong to the results of the matrix inverse function below.
|
|
import numpy as np a = np.mat 1,2,3;4,5,6;7,8,9 ) print a.I.I) >>> G \Anaconda3\python.exe C Users 4078 test chap3.22.py [[ 2.38095238 3.0952381 3.80952381] [ 2.85714286 3.21428571 3.57142857] [ 3.33333333 3.33333333 3.33333333]] And if you change a = np.mat 1,2,3;4,5,6;7,8,9 ) to a = np.mat 1,2,3;4,5,6;7,8,9.1 ) or any other number not Singular matrix)the result would be correct.","bug"
|
|
"6013","5269","Flatten in Matrix returns 2D array numpy.matrixlib.defmatrix.matrix).
|
|
Hi, I got a nasty bug feature that I finally solve.
|
|
I was using the flatten ) function that normally returns a 1D array.
|
|
In some piece of code not relevant) this function returned 2D array!
|
|
This was the problem) The solution I was assuming that I was working with numpy.array but it turns out that I was actually using numpy.matrixlib.defmatrix.matrix.
|
|
The flatten ) function in that case returns 2D array of shape 1,n) with n being the actual number of elements in the numpy.matrixlib.defmatrix.matrix.
|
|
Casting everything to numpy.array solves the problem.
|
|
The numpy version used is 1.6.2 The OS version is Debian 3.2.60 1 deb7u3 x86 64 GNU Linux Cheers, Julien","bug"
|
|
"6013","527","zeros and ones do not take order as keyword parameter.
|
|
Trac 43).
|
|
Original ticket http projects.scipy.org numpy ticket 43 on 2006 03 31 by trac user tim.hochberg ..., assigned to unknown.
|
|
>>> numpy.zeros [3], int, C ) array [0, 0, 0]) >>> numpy.zeros [3], order= C ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError order is an invalid keyword argument for this function nuff said.","bug"
|
|
"6013","5272","Data type precision problems?.
|
|
Hi folks I ve just been investigating some type conversion and came across what appears unusual behaviour to me I would have expected these statements to more faithfully print the values at their respective precisions .
|
|
I haven t investigated playing around with .","bug"
|
|
"6013","5273","ImportError in runtests.py due to cyclic import?.
|
|
I get this error when trying to run the tests .
|
|
numpy.testing tries to import numpy.utils, which then tries to import numpy.testing but numpy.testing is not fully imported.","bug"
|
|
"6013","5276","Truthiness of object arrays.
|
|
Initially discussed on numpy discussion.
|
|
I am puzzled by the following numpy 1.9.0, python 3.4.2) .
|
|
I thought that `if x` simply calls `bool`, but apparently this is not even the case...","bug"
|
|
"6013","5278","Possible inconsistency between behaviour of integer arrays and numpy behaviour.
|
|
Hi!
|
|
I m preparing materials for my students about numpy, and found following bug inconsistency between behavioue and docs.
|
|
Docs for np.divide state that > When both x1 and x2 are of an integer type, divide will return integers and throw away the fractional part.
|
|
Moreover, division by zero always yields zero in integer arithmetic.
|
|
However when I execute examples from the documentation in my notebook I get behaviour as if floats were used np.divide 2, 4) returns 0.5 and np.divide np.array [0, 1], dtype=int), np.array [0, 0], dtype=int)) returns array [ nan, inf]).
|
|
I use numpy 1.9.1 d44b9c61499f8bc5a9fc94286cd52f05e15e003f) on Python 3.4.1.","bug"
|
|
"6013","528","Change project name from SciPy Core to NumPy Trac 44).
|
|
Original ticket http projects.scipy.org numpy ticket 44 on 2006 04 01 by rkern, assigned to rkern.
|
|
I m leery of doing this without understanding exactly what the consequences are.
|
|
I think everything will be fine if I just edit the name in trac.ini, but I would welcome any comments from Tracxperts.","bug"
|
|
"6013","5280","np.multiply is inconsistent between sparse and dense types.
|
|
The ufunc `np.multiply )` performs different operations depending on whether the arguments are sparse or dense types.
|
|
In particular, the spec for np.multiply http docs.scipy.org doc numpy reference generated numpy.multiply.html) specifies that it will Multiply arguments element wise.
|
|
But with sparse types, it appears to perform matrix multiplication .","bug"
|
|
"6013","5285","assert equal gives KeyError 0 for some Python objects..
|
|
In NumPy 1.9, `numpy.testing.assert equal` can raise a `KeyError` in some cases.
|
|
An example is below.
|
|
It s not as minimal as I would like I was unable to easily remove the dependence on the Traits library.
|
|
.
|
|
The above code fails for me under NumPy 1.9 but not NumPy 1.8), with the following output .
|
|
I have no idea why the `array` constructor would be trying to access item `0` of this object, but I can imagine that there might be valid reasons for doing so.
|
|
If so, it might be useful to turn the `except ValueError` for the `use complex` section of the `assert equal` code into `except KeyError, ValueError)`.
|
|
Code [here] https github.com numpy numpy blob master numpy testing utils.py L281 L284)).","bug"
|
|
"6013","5286","return type inconsistent between code and docs for PyArray ITEMSIZE etc.
|
|
In doc source reference c api.array.rst .
|
|
In numpy core include numpy ndarraytypes.h .
|
|
I personally think that both of these are the wrong type and that it should be `size t` or some other unsigned type.
|
|
However, in any case the inconsistency is wrong and should be corrected in one or both places.
|
|
I discovered this because `printf ... d... , PyArray ITEMSIZE arr))` is giving compiler warnings in fortranobject.c now that 5270 has been merged.
|
|
The same problem also occurs for `PyArray STRIDE` and `PyArray DIM` note the singular i.e.
|
|
the subscripted functions).","bug"
|
|
"6013","5289","numpy arithmetic error clobbers sys.
|
|
stderr .
|
|
The following three lines of code are sufficient to demonstrate the problem.
|
|
Python 3.3.2 and numpy 1.7.2rcl, 64 bit Windows 8.)
|
|
>>> import numpy >>> 1 numpy.float64 0) Traceback most recent call last) File <pyshell 84> , line 1, in <module> 1 numpy.float64 0) File C \Python33\lib\idlelib\PyShell.py , line 64, in idle showwarning file.write warnings.formatwarning message, category, filename, AttributeError NoneType object has no attribute write >>> 1 numpy.float64 0) inf The first time such an arithmetic error occurs also happens with overflows, such as `2.5 numpy.float64 775)`), an error is thrown, as above.
|
|
The second time though, no error is thrown.
|
|
This would seem to me to be undesirable behavior.
|
|
Note that I am using IDLE and a runtime error occurs instead when running such code from a command line.","bug"
|
|
"6013","529","=, etc are too liberal in what they ll accept Trac 45).
|
|
Original ticket http projects.scipy.org numpy ticket 45 on 2006 04 04 by trac user tim.hochberg ..., assigned to unknown.
|
|
=, etc will accept things they shouldn t and silently truncate them into things they can digest.
|
|
To wit, >>> a array [0, 1, 2, 3]) >>> a = 1j >>> a array [0, 0, 0, 0]) >>> a = 0.5 >>> a array [0, 0, 0, 0]) Both of those should almost certainly be type or value errors.","bug"
|
|
"6013","5290","savetxt should allow multi line headers.
|
|
I think it would be a useful addition if `savetxt` s `header` kwarg could optionally be a list, in which case multiple lines are added as header.
|
|
The same could go for footer.
|
|
What do you think?
|
|
Useful?
|
|
Should I submit a PR?","enhancement"
|
|
"6013","5293","align is not working when using only strings.
|
|
import numpy as np a = np.zeros 4, dtype=np.dtype |S4 , align=True)) print a.flags[ ALIGNED ] should be True Works with 1.8 but not with 1.9","bug"
|
|
"6013","5295","NpyIter documentation needs love.
|
|
I just found out that there are two iteration APIs in NumPy the `PyArray ITER` family and the `NpyIter` ones.
|
|
The latter are apparently the preferred API nowadays, but they re not really advertised as such.
|
|
The [array API docs] http docs.scipy.org doc numpy reference c api.array.html) in `doc source reference c api.array.rst` link to the new API, but then list the old API.
|
|
The link goes to [`doc source reference c api.iterator.rst`] http docs.scipy.org doc numpy reference c api.iterator.html), which calls the old API the existing API, as if the new one were merely proposed and bothers the reader with some development details; see gh 5294).
|
|
The documentation starts off with a conversion table, instead of encouraging new code to use it.
|
|
I m not sure how to fix this.
|
|
The first thing to do, I think, is move the conversion table down the page and start with an example of the new API.
|
|
However, that leaves the problem that the old API is still in the main docs.
|
|
We can... move the old API to the iterator API docs and put it at the bottom, merge the new API docs into the main array docs, or do both so we have a doc chapter about the old iterator API and the new API in the array docs.
|
|
Thoughts?","bug"
|
|
"6013","5297","Inconsistent casting behavior in array scalar vs array array multiplication.
|
|
Say we have an array of single precision floats.
|
|
.
|
|
If we multiply this array with a 64 bit int scalar, the result still consists of single precision floats.
|
|
.
|
|
This is also true of we multiply with a 0 d numpy array containing a 64 bit int.
|
|
.
|
|
However, if we multiply with an 1 d or n d) array of 64 bit ints, we now get double precision floats.
|
|
.
|
|
I can see the argument for either behavior since the int has 64 bits perhaps the result should too) but it seems like it should be consistent either way.
|
|
I stumbled across this when I discovered that `np.meshgrid` changes the dtype of single precision float arrays perhaps a separate issue?)
|
|
because it multiplies by an array of `np.ones` unless `copy=False`.
|
|
.","bug"
|
|
"6013","5299","using `random.choice` to sample integers in a large range.
|
|
So I wanted to generate some unique integers with the following code .
|
|
I wanted these numbers as deterministic seeds for some simulations.
|
|
When I do this almost 16 GB of memory are filled.
|
|
I looked into the code for choice and in this case it essentially generates a permutation, similar to shuffling a `np.arange max int)`, in order to then take a small slice from that array.
|
|
This seems like a bad strategy since providing only an `int` to `random.choice` makes it clear that numbers in that range should be sampled.
|
|
At least when the integer argument `a` is much larger than the argument `size`.
|
|
So for now I will just generate a set of numbers but I was wondering if there is a more general way.","bug"
|
|
"6013","530","spam Trac 46).
|
|
Original ticket http projects.scipy.org numpy ticket 46 on 2006 04 05 by trac user spammer, assigned to trac user spammer.
|
|
spam","bug"
|
|
"6013","5300","ascontiguousarray makes scalars into arrays.
|
|
Not sure if this is intended behavior or a bug, but I had been under the assumption that the result of `np.ascontiguousarray` would be the same shape as the input.
|
|
This is true except in the case of scalars.
|
|
.
|
|
If this is what should be expected, it would be helpful to document this in its docstring.
|
|
Thanks!","bug"
|
|
"6013","5303","the error message for accidentally irregular arrays is confusing.
|
|
Someone using my code reported this error message, and I would have been able to track down the problem more quickly if the message had been more informative, maybe like `ValueError setting an irregularly shaped array with a non object dtype` instead of `ValueError setting an array element with a sequence.` .","bug"
|
|
"6013","5305","Numpy Ironpython.
|
|
Hello my name is starsky.
|
|
I want to install numpy ni ironpython.
|
|
but i can t find any instructions online.
|
|
Can somebody help me","bug"
|
|
"6013","5306","`array` function s docstring is not clear on default value for `order` parameter.
|
|
In the docstring for `numpy.array`, it says that the default value for the `order` parameter is ` C ` i.e., the output will be C contiguous).
|
|
This is true if `copy=True`, but if `copy=False` then `array` acts like `asarray` not `ascontiguousarray`) and the output is not guaranteed to be contiguous.
|
|
Explicitly setting `order= C ` forces a copy if the input is not C contiguous.
|
|
The docstring needs to be changed to reflect this difference, so that it does not say that ` C ` is the default value for `order`.","bug"
|
|
"6013","531","Non native byteorder arrays do not sort correctly Trac 47).
|
|
Original ticket http projects.scipy.org numpy ticket 47 on 2006 04 05 by rkern, assigned to unknown.
|
|
It looks like the .sort ) method is using the raw values interpreted as if they were in the native byteorder.
|
|
In [57] a = linspace 0, 10, 11) In [58] a Out[58] array [ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9., 10.])
|
|
In [59] a.dtype Out[59] dtype >f8 ) In [60] c = a.astype dtype <f8 )) In [61] c Out[61] array [ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9., 10.])
|
|
In [62] c.dtype Out[62] dtype <f8 ) In [63] c.sort ) In [64] c Out[64] array [ 0., 2., 3., 4., 5., 6., 7., 8., 9., 10., 1.])","bug"
|
|
"6013","5310","Docstring for np.linalg.svd still wrong?.
|
|
The docstring of the SVD function in `np.linalg` says Factors the matrix `a` as ``u np.diag s) v``, where `u` and `v` are unitary and `s` is a 1 d array of `a` s singular values.
|
|
However, this is wrong for non square matrices `a`.
|
|
This has been discussed on the mailing list a few years back [link] http mail.scipy.org pipermail numpy discussion 2011 January 054722.html).
|
|
Nevertheless this is still wrong in my version, i.e.
|
|
1.9.0.","bug"
|
|
"6013","5311","Numpy gets slow after creating some scipy matrices and delete the reference to it.
|
|
After creating a csr matrix from scipy, assigning to a variable, and setting that variable to None, the performance of subsequent operations are affected badly.
|
|
from numpy import from scipy.sparse import import time import sys test slow = len sys.argv)>1 If this is True, test sparse is slow, otherwise it s fast if test slow for i in range int sys.argv[1])) A = random.random 1000,1000)) A = csr matrix A) If only this line is commented out, scipy completes in 0.192s A = None If only this line is commented out, scipy completes in 3s C = ndarray shape= 1000,1000)) D = ndarray shape= 1000,1000)) for i in range 20000) C[random.randint 0,1000),random.randint 0,1000)] = random.rand ) D[random.randint 0,1000),random.randint 0,1000)] = random.rand ) C1 = csr matrix C) D1 = csr matrix D) print Numpy dot product sparse start = time.time ) result = dot C,D) print Done in .3fs time.time ) start) print Scipy dot product sparse start = time.time ) result = C1.dot D1) print Done in .3fs time.time ) start) Running the file without argument causing `test slow` to be False) >Numpy dot product sparse Done in 0.470s Scipy dot product sparse Done in 0.012s That is the expected performance of numpy and scipy.
|
|
However, as further tests showed, that performance is affected by the three lines inside `test slow`.
|
|
Running the file with `1` as argument doing the seemingly useless assignment and unassignment to `A` once) >Numpy dot product sparse Done in 0.472s Scipy dot product sparse Done in 0.195s So numpy seems still unaffected, but scipy gets slower.
|
|
Running the file with `2` as argument doing the seemingly useless assignment and unassignment to `A` twice) >Numpy dot product sparse Done in 11.071s Scipy dot product sparse Done in 69.477s Both numpy and scipy get extremely slow!
|
|
This problem doesn t occur if the matrix creation the `C`, `D`, `C1`, and `D1`) happens before the seemingly useless assignment and unassignment.
|
|
If the line `A = csr matrix A)` is commented out, only scipy dot product is affected >Numpy dot product sparse Done in 0.473s Scipy dot product sparse Done in 0.194s Any explanation on this?
|
|
It seems very bizarre, as the assignment and unassignment should not affect anything.
|
|
It s not due to `random`, since only calling `A = random.random 1000,1000))` before the creation of the array doesn t affect the performance.
|
|
I m not sure whether to put this issue in scipy or numpy, but since it affects numpy also, I guess the slow in scipy is due to the slow in numpy, so I post the issue here.
|
|
My system configuration .","bug"
|
|
"6013","5312","Use an aligned allocator for NumPy?.
|
|
Regarding the f2py regression in NumPy 1.9 with failures on 32 bit Windows, the question is whether NumPy should start to use an allocator which gives guaranteed alignment.
|
|
https github.com scipy scipy issues 4168","enhancement"
|
|
"6013","5313","PyArray AsCArray causes segmentation fault with 3D arrays.
|
|
I tried to use numpy arrays with the C API and found some strange behavior I m trying to create a C array like view onto an numpy 2d and 3d) array using this function c function print requested item from 2D or 3D array static PyObject test PyObject self, PyObject args) { PyArrayObject symbol series obj; npy intp i=0, j=0, k=0; if !PyArg ParseTuple args, O!ll|l , PyArray Type, symbol series obj, i, j, k) ) return NULL; if NULL == symbol series obj) return NULL; get number of dimensions npy intp num dims = PyArray NDIM symbol series obj); help vars int typenum = NPY DOUBLE; PyArray Descr descr = PyArray DescrFromType typenum); npy intp dims[4]; incref, as PyArray AsCArray steals reference Py INCREF symbol series obj); is 2D array if num dims == 2){ double symbol series; if PyArray AsCArray PyObject ) symbol series obj, void ) symbol series, dims, 2, descr) < 0){ PyErr SetString PyExc TypeError, error converting to c array ); return NULL;} printf 2D array[ i][ i] .2f\n , int) i, int) j, symbol series[i][j]); free C like array PyArray Free PyObject ) symbol series obj, void ) symbol series); } is 3D array else if num dims == 3){ double symbol series; if PyArray AsCArray PyObject ) symbol series obj, void ) symbol series, dims, 3, descr) < 0){ PyErr SetString PyExc TypeError, error converting to c array ); return NULL;} printf 3D array[ i][ i][ i] .2f\n , int) i, int) j, int) k, symbol series[i][j][k]); free C like array PyArray Free PyObject ) symbol series obj, void ) symbol series); } else { printf error not 2D or 3D ); return NULL; } Py RETURN NONE; } python test case create array and pick random items import numpy as np from my module import test def get random item arr) selects a random item from given array with numpy and c ext get coord cords = tuple [np.random.randint dim) for dim in arr.shape]) str cords = .join [{}] .format cord) for cord in cords)) numpy print numpy arr{} = {} .format str cords, arr[cords])) c version test arr, cords) if name == main ARR2D = np.arange 0, 0.27, 0.01).reshape 3, 9) for in range 10) get random item ARR2D) ARR3D = np.arange 0, 0.27, 0.01).reshape 3, 3, 3) for in range 10) get random item ARR3D) issue python segfaults randomly when getting random items from an 3D array not always during runtime at the latest on exit) If I run the test case only with 2D arrays, everything works fine and no segfault occurs.","bug"
|
|
"6013","5315","Error detecting custom gfortran compiler.
|
|
This issue was discovered in the numpy 1.9.1 release.
|
|
I have my own custom built gcc 4.8.3 on a CentOs 5 machine.
|
|
This compiler was built using the with pkgversion option which affects the output of the gfortran version command.
|
|
For example if you build your gcc 4.8.3 using ` with pkgversion=AB2`, then `gfortran version` will output GNU Fortran AB2) 4.8.3 Copyright ....
|
|
The last regex used by GnuFCompiler.gnu version match ) in numpy distutils fcompiler gnu.py chokes on this.
|
|
It takes the 2 from AB2 as the compiler version and returns g77 .
|
|
The result is that the numpy setup.py was unable to correctly find and use my gfortran compiler.
|
|
For my site I just created a patch to hack the regex to work but I don t think my fix would be a general solution.","bug"
|
|
"6013","532","python setup.py config fc help fcompiler does nothing Trac 48).
|
|
Original ticket http projects.scipy.org numpy ticket 48 on 2006 04 07 by rkern, assigned to pearu.
|
|
As of r2332, the help fcompiler flag on config fc does not seem to do anything.","bug"
|
|
"6013","5320","Latest developer version does not import.
|
|
I tried installing the latest git version of Numpy on MacOS X 10.8 with Pytohn 2.7, and I get the following error when importing Numpy .","bug"
|
|
"6013","5329","API BUG np.dtype equality comparisons versus string like is inconsistent.
|
|
xref https github.com pydata pandas issues 8814 using ``numpy 1.9.1 macosx`` ``np.dtype`` equality checking versus string likes should return a boolean rather than raising in dtype comparisons.
|
|
This is currently inconsistent when presented with a valid dtype parse of the string, but will raise if its not valid.
|
|
e.g .","bug"
|
|
"6013","533","Indexing rank >=2 arrays with arrays that have negative numbers alters the index array Trac 49).
|
|
Original ticket http projects.scipy.org numpy ticket 49 on 2006 04 09 by rkern, assigned to teoliphant.
|
|
With r2336 [ ] python Python 2.4.1 2, Mar 31 2005, 00 05 10) [GCC 3.3 20030304 Apple Computer, Inc. build 1666)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> a = arange 25).reshape 5,5)) >>> idx = array [ 1, 0, 1]) >>> idx array [ 1, 0, 1]) >>> a[idx] array [[20, 21, 22, 23, 24], [ 0, 1, 2, 3, 4], [ 5, 6, 7, 8, 9]]) >>> idx array [4, 0, 1]) >>> idx = array [ 1, 0, 1]) >>> b = arange 5) >>> b[idx] array [4, 0, 1]) >>> idx array [ 1, 0, 1]) >>> idx = array [ 1, 0, 1]) >>> c = arange 125).reshape 5,5,5)) >>> c[idx] array [[[100, 101, 102, 103, 104], [105, 106, 107, 108, 109], [110, 111, 112, 113, 114], [115, 116, 117, 118, 119], [120, 121, 122, 123, 124]], [[ 0, 1, 2, 3, 4], [ 5, 6, 7, 8, 9], [ 10, 11, 12, 13, 14], [ 15, 16, 17, 18, 19], [ 20, 21, 22, 23, 24]], [[ 25, 26, 27, 28, 29], [ 30, 31, 32, 33, 34], [ 35, 36, 37, 38, 39], [ 40, 41, 42, 43, 44], [ 45, 46, 47, 48, 49]]]) >>> idx array [4, 0, 1]) >>> Rank 1 arrays are special cased while arrays of higher rank go through `PyArray GetMap` and friends in source trunk numpy core src arrayobject.c .
|
|
The problem appears to be around line 7794 of that file in `PyArray MapIterBind`.
|
|
Every other code path that corrects negative indices does so on a copy of the value rather than a reference.","bug"
|
|
"6013","5331","Array in place update ordering not respected.
|
|
When using = and = on arrays with a shifted copy as an argument, the ordering is not respected.
|
|
Test case import numpy as np a = np.array [1,2,3,4,5]) a[n] = a[n] a[n 1] a[1 5] = a[1 5] a[0 4] b = np.array [1,2,3,4,5]) updates in the wrong order, updating n 1 values before use expect incorrect result) b[1 5] = b[0 4] c = np.array [1,2,3,4,5]) update in the reverse order so as to not clobber the n 1 values expect correct result) c[4 0 1] = c[3 1] print a) print b) print c) assert a != b).any ) OK assert a == c).all ) fails b == c)","bug"
|
|
"6013","5336","Saving large array fails with savez compressed ) but works with save ).
|
|
I got this strange problem on a CentOS box, Python 2.6.6, Numpy 1.9.1 .","bug"
|
|
"6013","5337","Should polyfit ) accept a ddof parameter?.
|
|
Both `cov )` and `corrcoef )` accept a `ddof` parameter that allows the caller to tune the bias term.
|
|
Would it make sense for `polyfit )` to do the same, when returning the covariance matrix?
|
|
I imagine this would be implemented with an optional `ddof` parameter that would imply `cov=True` if passed.
|
|
Or for the most minimal possible change to the function signature, it could be implemented by checking `cov` for a numeric value; if `cov` is numeric, use the number passed instead of the current fixed default of 2.0.
|
|
Inspired by this question at Stack Overflow http stackoverflow.com q 27230285 577088","enhancement"
|
|
"6013","5338","show more info in numpy nosetester?.
|
|
In particular I d be curious to see https github.com numpy numpy blob maintenance 1.9.x numpy testing nosetester.py L244 report NPY RELAXED STRIDES CHECKING maybe using the method described in http docs.scipy.org doc numpy release.html npy relaxed strides checking?
|
|
On the other hand, this might be too spammy to be of general interest.","bug"
|
|
"6013","5339","cache dtype.
|
|
hash .
|
|
It would be nice to cache the hash value of dtypes, since computation can be non trivial for complex e.g.
|
|
structured) dtypes.
|
|
I would like to contribute this but I had some trouble running the test suite, and now I m not able to build Numpy anymore .","bug"
|
|
"6013","534","Loading of dumped chararrays Trac 50).
|
|
Original ticket http projects.scipy.org numpy ticket 50 on 2006 04 11 by trac user pgmdevlist ..., assigned to unknown.
|
|
Dumping a char.array works fine, but the dumped file cannot be reloaded.
|
|
Version Numpy 0.9.7.2340 char.array arange 1000,1010),itemsize=4).dump testdump ) load testdump ) gives exceptions.ValueError Traceback most recent call last) numpy core numeric.py in load file) 358 if isinstance file, type )) 359 file = file file, rb ) > 360 return cload file) 361 362 These are all essentially abbreviations numpy core internal.py in reconstruct subtype, shape, dtype) 102 103 def reconstruct subtype, shape, dtype) > 104 return ndarray.
|
|
new subtype, shape, dtype) 105 106 numpy core defchararray.py in array finalize self, obj) 38 def array finalize self, obj) 39 if not globalvar and self.dtype.char not in SU > 40 raise ValueError, Can only create a chararray from string data.
|
|
41 42 ValueError <exceptions.ValueError instance at 0x2aaaaf10d050>, <function reconstruct at 0x2aaaace129b0>, <class numpy.core.defchararray.chararray >, 0,), b ))","bug"
|
|
"6013","5341","Meshgrid shuffles the grid for 3 D or more.
|
|
Hi!
|
|
Me and pedrohasselmann found a strange issue when using numpy.meshgrid for 3 dimensions and higher.
|
|
The axis of each grid are shuffled and are not returned in the same order in which they were given.
|
|
Example .
|
|
The axis of the returned grid is ` 5, 10, 2)` when we believe that it should be ` 10, 5, 2)`, keeping the same order as the ones given.
|
|
We pinpoint that the following issue is containing in this [block] https github.com numpy numpy blob v1.9.1 numpy lib function base.py L3407 L3411).
|
|
If we removed it, we could reproduce the result expected by us.
|
|
Is this a mistake or it was designed?","bug"
|
|
"6013","5344","Should numpy.random.choice accept narray with more than one dimension ?.
|
|
It is needed to generate two or three dimensional data from list like structure in many scientific applications.
|
|
I noticed that in numpy.random.choice, the code checks whether input is one dimensional Format and Verify input a = np.array a, copy=False) if a.ndim == 0 try index must return an integer by python rules.
|
|
pop size = operator.index a.item )) except TypeError raise ValueError a must be 1 dimensional or an integer ) if pop size <= 0 raise ValueError a must be greater than 0 ) elif a.ndim != 1 raise ValueError a must be 1 dimensional ) else pop size = a.shape[0] if pop size is 0 raise ValueError a must be non empty ) It is not convenient to choose two dimensional data from structure like array [[1,1],[2,2],[3,3]]) What is the reason numpy needs to fix the dimension of input to be one?
|
|
If needed, I can rewrite choice to let it accept arrays with more than one dimension.","bug"
|
|
"6013","5345","dtype richcompare doesn t respect protocol.
|
|
When a type s `` eq `` or ``tp richcompare`` in C) doesn t understand the other type, it should return NotImplemented so that the other type s `` eq `` has a chance to be called.
|
|
Unfortunately, NumPy s ArrayDesrc doesn t respect that convention and raises TypeError instead, making comparisons fragile .
|
|
It is generally considered bad practice to raise TypeError in a comparison method, and can prevent advanced uses such as defining a custom type comparable with ArrayDescr, or simply mixing ArrayDescr with other types in e.g.
|
|
a dictionary.
|
|
Indeed let s say I use the standard `lru cache`, then hash collisions can produce unwanted errors because of the dictionary used internally comparing the keys .","bug"
|
|
"6013","5346","BUG x.view S1 ) raises for 0 dimensional stirng arrays.
|
|
This works .
|
|
This doesn t .","bug"
|
|
"6013","5348","Use GPU on Raspberry Pi for FFT.
|
|
Please, use GPU on Raspberri Pi for FFT.
|
|
Implement it in such way that calling FFT functionality explicitely has to request processing on GPU via a parameter.
|
|
The default will remain processing on CPU.
|
|
For more information, see http www.aholme.co.uk GPU FFT Main.htm and for source code, see https github.com raspberrypi userland tree master host applications linux apps hello pi hello fft","enhancement"
|
|
"6013","535","xe5x9cxa8xe8xbfx99xe9x87x8cxe7x99xbbledxe6x98xbexe7xa4xbaxe5xb1x8fxe5xb9xbfxe5x91x8axe8x81x94xe7xb3xbb020 33396536,QQ 789266 Trac 51).
|
|
Original ticket http projects.scipy.org numpy ticket 51 on 2006 04 12 by unknown, assigned to trac user led.
|
|
spam","bug"
|
|
"6013","5350","datetime64 arrays should fill in descr in array struct .
|
|
The structure exposed by ` array struct ` is currently defined as follows .
|
|
The descr field is currently only filled by record arrays, but without it, datetime64 or timedelta64) array description is incomplete.
|
|
There is no other field from which the time unit can be deduced.
|
|
See also gh 4983.","bug"
|
|
"6013","5352","ndarray astype docstring is missing new in version hints.
|
|
http docs.scipy.org doc numpy reference generated numpy.ndarray.astype.html It would have been nice to have known that `astype` doesn t do keyword arguments in numpy 1.5.1 before writing https github.com scipy scipy pull 4238 without having to go exploring in http docs.scipy.org doc numpy 1.5.x reference generated numpy.ndarray.astype.html or digging through github history.
|
|
As always pull requests are welcome!","bug"
|
|
"6013","5353","object array construction should require explicitly specifying dtype=object.
|
|
This possibility was raised by https github.com numpy numpy issues 5303 issuecomment 63904939 so I m moving this topic to its own github issue.
|
|
Problems like https github.com scipy scipy issues 4239 among others could be softened by making this change although it could bring backwards compatibility issues.","bug"
|
|
"6013","5354","clipping complex array crashes python.
|
|
This crashes my python kernel .","bug"
|
|
"6013","5356","dtype equality operator does not check endianness of fields.
|
|
For simple types, two dtype objects with different endianness are unequal .
|
|
However, for fields, the endianness is currently ignored when doing a comparison .
|
|
This causes problems with other functions.
|
|
For example, .
|
|
I m guessing that astype thought the output dtype was the same, and hence did a memcpy instead of performing the endian conversion.","bug"
|
|
"6013","5357","histogram progressive rounding error with float32 weights.
|
|
When histogramming float32 data with weights, the counts in each bin become progressively more quantized i.e.
|
|
grouped, rounded, ...) with increasing bin number.
|
|
Below is an example with random data.
|
|
This might be related to 4823.
|
|
I m not sure whether this is a bug or expected behavior, but if the latter I suggest that it at least be noted in the function docstring.
|
|
.","bug"
|
|
"6013","536","Vectorize segfaults on large arrays Trac 52).
|
|
Original ticket http projects.scipy.org numpy ticket 52 on 2006 04 12 by stefanv, assigned to unknown.
|
|
Tested with numpy revision 0.9.7.2328 import numpy as N x = N.linspace 3,2,10000) y = N.vectorize lambda x x) y x)","bug"
|
|
"6013","5362","numpy exits if lhs in lstsq lhs, rhs) contains nan.
|
|
.","bug"
|
|
"6013","5363","Incorrect histogramdd behavior with masked arrays.
|
|
Summing over the histogram should always give the number of non masked records when no weighting is used.
|
|
As shown below, the sum is reduced, but not by the correct amount.
|
|
Calling the compressed ) method and reshaping before running the histogram produces correct results.
|
|
The following uses numpy 1.8.1. .","bug"
|
|
"6013","5366","Einsum gets slow with multiple arguments.
|
|
With multiple arguments, `einsum` blows up fast .
|
|
Is there some fundamental limitation involved, or is this a bug?","bug"
|
|
"6013","537","Configuration.add data dir ) does not handle globbed directories correctly Trac 53).
|
|
Original ticket http projects.scipy.org numpy ticket 53 on 2006 04 12 by rkern, assigned to pearu.
|
|
The [http svn.enthought.com svn enthought branches pearu lib numpy numpy branch of the Enthought tool Suite] places data files in wrong places in several different ways.
|
|
For example, enthought.chaco2 is supposed to have several data directories which are added like so !python config.add data dir chaco2 ) All of the files chaco2 actually end up in the chaco2 directory directly and no subdirectories are created.
|
|
Another example is enthought.gotcha which uses the following !python config.add data dir gotcha images ) All of the image files are placed in gotcha images images","bug"
|
|
"6013","5370","searchsorted with a scalar of different type is slow.
|
|
In [4] np.
|
|
version Out[4] 1.10.0.dev 3ef77ee In [5] n, val = int 1e7), 563412 In [6] a = np.arange n, dtype= int64 ) In [7] timeit a.searchsorted val) 1000000 loops, best of 3 1.58 xc2xb5s per loop In [8] b = a.astype int32 ) In [9] timeit b.searchsorted val) 10 loops, best of 3 25.9 ms per loop In [10] fval = float val) In [11] timeit a.searchsorted fval) 10 loops, best of 3 34.8 ms per loop In [12] x = a.astype float ) In [13] timeit x.searchsorted fval) 1000000 loops, best of 3 1.38 xc2xb5s per loop In [14] timeit a == val).nonzero )[0] how much time it takes with linear search 10 loops, best of 3 20.1 ms per loop even though types do not match, because needle is a scalar it is still possible to search `O log n))`.","enhancement"
|
|
"6013","5373","f2py AssertionError with parameter declaration.
|
|
In my fortran code there is a parameter declaration .
|
|
This throws an error > File usr lib python2.7 dist packages numpy f2py crackfortran.py , line 676, in markoutercomma > assert not f, repr f, line, l, cc)) > AssertionError 1, lmax 1) lmax 1 , lmax 1) lmax 1 , ) ) However, if I change the declaration to .
|
|
It works.","bug"
|
|
"6013","5375","Strange addition with = operator.
|
|
I found something real weird when using addition in numpy arrays from numpy import x = array [[0]]) x = 0.1 2 print x > [[0]] x = x 0.1 2 print x >[[0.2]] My numpy version is 1.8.2","bug"
|
|
"6013","5376","`np.ravel` should return the same array type it was given.
|
|
If given a matrix, then `np.ravel` should return a matrix.
|
|
And similarly for a masked array.
|
|
Currently, it always returns an ndarray.
|
|
In PR 5358, `np.diagonal` was modified to return whatever array type it took in.
|
|
Also, `np.cumsum` and `np.clip` return the same array type.
|
|
So, `np.ravel` s behavior is surprising.","bug"
|
|
"6013","5377","numpy.packbits doesn t accept a bool array.
|
|
Example of problem data = numpy.array [True]) byte values = numpy.packbits data) This results with an exception TypeError Expected an input array of integer data type.
|
|
This seems to go against the documentation of the function, which says it expects a binary valued array .
|
|
To make this work, you need to first convert the data type of the array to int, which seems redundant.
|
|
eg.
|
|
byte values = numpy.packbits numpy.array data, dtype=int)) A better use case of where I encountered the problem.
|
|
data = numpy.random.sample 1024) truth values = data >= threshold bits = numpy.packbits truth values)","bug"
|
|
"6013","5378","Simple math error.
|
|
try the code import numpy as np u=np.array [170],dtype=np.uint8) v=u[0] 100000000 print v I probably naively) expected that python would recognize that the uint8 object was being multiplied to make a Long int product and yield the Longint answer 170000000000.
|
|
Instead, the result of the calculation defaults presumably to the 32 bit product, in this case a nonsensical negative number.
|
|
This is fixable by typing the initial array as 64 bits, or by explicitly typing the scalar 100000000 as 64 bit.
|
|
But this is not a standard requirement for other calculations in python.
|
|
Apologies if I have repeated an earlier bug or simply exhibited my own ignorance.
|
|
Took me a full afternoon to figure this out.","bug"
|
|
"6013","538","No thread based error handling Trac 54).
|
|
Original ticket http projects.scipy.org numpy ticket 54 on 2006 04 12 by teoliphant, assigned to unknown.
|
|
The ufunc error handling needs to work on a per thread basis.
|
|
The three dictionary look up scheme needs to be replaced with a single dictionary look up using ThreadDict.","bug"
|
|
"6013","5381","Implement new style buffer protocol for scalars on Py2.
|
|
Numpy scalars have an old style buffer interface .
|
|
but not a new style one.
|
|
.
|
|
4864 makes it sound like this inconsistency would eventually be addressed.
|
|
[PyOpenCL] pyopencl pyopencl) and [PyCUDA] inducer pycuda) are [affected] inducer pycuda pull 60) by this.","bug"
|
|
"6013","5384","numpy.ma wants to mask out my infs.
|
|
.
|
|
.","bug"
|
|
"6013","539","Scalar math Trac 55).
|
|
Original ticket http projects.scipy.org numpy ticket 55 on 2006 04 12 by teoliphant, assigned to unknown.
|
|
The scalar math module needs to be fleshed out so that array scalars have their own math tables.","bug"
|
|
"6013","5394","BUG?
|
|
np.ravel does not flatten out certain arrays.
|
|
For arrays in which the number of elements are not the same along the inner axes, `np.ravel` and `np.flatten` return the array as it is.
|
|
I m not sure if this is intended behavior or not.
|
|
In [23] a = np.ravel [[1, 2], [3, 4, 5]]) In [24] a Out[24] array [[1, 2], [3, 4, 5]], dtype=object) I thought of giving this as a shot, but I m not sure what part of the source I should modify, the ravel function just says `asarray a).ravel )` and I was not able to figure out quickly) where the ravel methods for the ndarray is.","bug"
|
|
"6013","5395","matrix nuclear norm.
|
|
Numpy already has [many norms] http docs.scipy.org doc numpy reference generated numpy.linalg.norm.html) but in my opinion it needs one more `linalg.norm x, ord= nuc )`!
|
|
The matrix nuclear norm also known as the trace class norm and the Ky Fan r norm) is one of three familiar [Schatten norms] http en.wikipedia.org wiki Matrix norm Schatten norms), two of which are already available in numpy.norm as the Frobenius norm Schatten 2 norm) and the spectral norm induced 2 norm, Schatten inf norm).
|
|
The nuclear norm is the Schatten 1 norm, and can be implemented as simply the sum of singular values of a matrix.","bug"
|
|
"6013","5399","np.equal not implemented for string arrays?.
|
|
https github.com scikit learn scikit learn issues 4014 .
|
|
.","bug"
|
|
"6013","540","String array comparisons Trac 56).
|
|
Original ticket http projects.scipy.org numpy ticket 56 on 2006 04 12 by teoliphant, assigned to unknown.
|
|
The rich compare function needs to be enhanced to handle comparisons of string arrays.
|
|
Comparison of unicode arrays needs to be thought about.
|
|
Python just compares the 32 bit unicode code point values.
|
|
We could start there.","enhancement"
|
|
"6013","5401","shape too large to be a matrix.
|
|
Does the following session represent the desired behavior of matrix multiplication of a numpy matrix and a numpy ndarray?
|
|
This question is motivated by the generalization to sparse matrices.
|
|
.","bug"
|
|
"6013","5404","np.random.multinomial gives negative samples with large n. I haven t had a chance to track down the offending code yet, but this had me scratching my head a few times.
|
|
When the number of samples is large, `np.random.multinomial` can return a forbidden) negative number Code to produce the issue .
|
|
my numpy version .","bug"
|
|
"6013","5406","masked array docs example.
|
|
In following [example] http docs.scipy.org doc numpy reference maskedarray.generic.html numerical operations) in the docs > Numerical operations can be easily performed without worrying about missing values, dividing by zero, square roots of negative numbers, etc.
|
|
.
|
|
I get a warning with a development version of numpy .
|
|
I am not suggesting that the behavior of `numpy.ma` should be changed, but maybe the example could be modified to avoid the warning or the warning could be included in the example output?","bug"
|
|
"6013","5407","Make a.diagonal ) writeable..
|
|
The ndarray.diagonal method currently returns a view, but it is not writeable.
|
|
It is documented to return a writeable view in 1.10.","bug"
|
|
"6013","5408","inconsistency in masked equal method docs and implementation.
|
|
In the masked equal method, the docstring doesn t suggest that the ``fill value`` gets updated by the ``value`` parameter of the function, but this line sets the ``fill value`` as ``value``.
|
|
output.fill value = value https github.com numpy numpy blob master numpy ma core.py L1978 Also this line https github.com numpy numpy blob master numpy ma core.py L2176 suggests that masked equal shouldn t change the ``fill value``, while the masked values does change the ``fill value`` x = np.arange 5) x array [0, 1, 2, 3, 4]) Expected ma.masked equal x, 2) masked array data = [0 1 3 4], mask = [False False True False False], fill value=999999) In the development version ma.masked equal x, 2) masked array data = [0 1 3 4], mask = [False False True False False], fill value=2)","bug"
|
|
"6013","541","ufunc methods need improved BUFFER loop Trac 57).
|
|
Original ticket http projects.scipy.org numpy ticket 57 on 2006 04 12 by teoliphant, assigned to unknown.
|
|
The BUFFER loop may need to be re written to be similar to the NOBUFFER case but with a different inner loop that uses the buffers.
|
|
This is similar to the change made for the generic function case which was found to be much faster.","enhancement"
|
|
"6013","5411","genfromtxt with names=True doesn t work with python3.
|
|
Using python 3.4.2 on Arch Linux with numpy 1.9.1, the following code doesn t work !
|
|
usr bin env python import sys import numpy as np data = np.genfromtxt sys.stdin, names=True) When running with a simple input file, it gives the following error Traceback most recent call last) File .
|
|
test.py , line 6, in <module> data = np.genfromtxt sys.stdin, names=True) File usr lib python3.4 site packages numpy lib npyio.py , line 1395, in genfromtxt if comments in first line TypeError in <string> requires string as left operand, not bytes The input file looks like this x y z 2.4 2.3 0.1 3.5 5.6 0.2 There is no problem with python2.","bug"
|
|
"6013","5412","Data corruption initializing array from a PIL style buffer .
|
|
I encountered this error in my own extension module, but it can be easily demonstrated using the ` tesbuffer` module that is included with Python 3.4 .
|
|
It looks like numpy fails to implement proper pointer dereferencing in the presence of suboffsets.
|
|
In this case .","bug"
|
|
"6013","5414","bitwise and.reduce ) only bitwise ands bit 1.
|
|
All other bits come out 0 .
|
|
Surely this is because .
|
|
is 1, when it should be 1 i.e.
|
|
all bits set).","bug"
|
|
"6013","5419","svd requires positive size matrices.
|
|
If I m using svd to help compute the norm of a matrix by looking at its singular values, then I want a size 0 matrix to have a length 0 sequence of singular values whose aggregate function e.g.
|
|
max , sum) will be 0.
|
|
But numpy explicitly forbids empty inputs with ` assertNoEmpty2d a)`.","bug"
|
|
"6013","542","setting a shape of a non contiguous view of an array Trac 58).
|
|
Original ticket http projects.scipy.org numpy ticket 58 on 2006 04 13 by rc, assigned to teoliphant.
|
|
Note below the fill method is used, but the bug works the same also for b[ ] = 3... == Example == In [43] a = nm.zeros 10,5) ) In [44] b = a[ ,2] In [47] b.fill 3 ) In [48] a Out[48] array [[0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0], [0, 0, 3, 0, 0]]) ok In [49] b.fill 0 ) In [50] a Out[50] array [[0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0]]) In [51] b.shape = 5,2) In [52] b Out[52] array [[0, 0], [0, 0], [0, 0], [0, 0], [0, 0]]) In [53] b.fill 3 ) In [54] a Out[54] array [[0, 0, 3, 3, 3], [3, 3, 3, 3, 3], [3, 3, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0]]) wrong == Expected behaviour?
|
|
== either raise an exception setting a shape of a non contiguous array view) or let the b shape change, but always work with the original column data of a?","bug"
|
|
"6013","5420","The norm of a size zero array should be zero.. Just as the empty sum is 0 and the empty product is 1, the empty norm should be 0 in my opinion.
|
|
In numpy, this convention has been implemented and is unit tested for the default vector and matrix norms .
|
|
Some of the non default empty norms raise exceptions.
|
|
The induced matrix 2 norm raises exceptions for two reasons first, attempting to compute the singular values of a size 0 matrix raises an exception, and second, even if the the empty array of singular values were to be returned, attempting to compute the max of an empty sequence also raises an exception.
|
|
The empty induced matrix 1 norm and np.inf norm each also raise an exception because they try to compute an empty max.","bug"
|
|
"6013","5422","Vectors containing zero get incorrect norm with negative order.
|
|
The negative order norm for vectors are failing when some of the elements of the vector are zeros.
|
|
The documentation says that negative norm, strictly speaking is not math , but they are used in general purpose.
|
|
[here] https github.com numpy numpy blob v1.9.1rc1 numpy linalg linalg.py L1953) But some of the examples provided in the documentation fails at finite negative order >>> a = np.arange 9) 4 >>> a array [ 4, 3, 2, 1, 0, 1, 2, 3, 4]) >>>la.norm a, 1) 4.6566128774142013e 010 >>> la.norm a, 2) nan >>> la.norm a, 3) nan The results on the dev version are >>>la.norm a, 1) home maniteja FOSS numpy numpy linalg linalg.py 2118 RuntimeWarning divide by zero encountered in reciprocal absx = ord 0.0 >>> la.norm a, 2) home maniteja FOSS numpy numpy linalg linalg.py 2118 RuntimeWarning divide by zero encountered in power absx = ord 0.0 >>> la.norm a, 3) 0.0 The raising of 0 to negative power is causing an error, 0.0 1 Traceback most recent call last) File <stdin> , line 1, in <module> ZeroDivisionError 0.0 cannot be raised to a negative power But the raising of a ``ndarray`` with zero to negative power is causing these exceptions.
|
|
a = np.arange 9) >>> a 1 main 1 RuntimeWarning divide by zero encountered in power main 1 RuntimeWarning invalid value encountered in power array [ 2147483648, 1, 0, 0, 0, 0, 0, 0, 0]) Suppose the initial array was not consisting of zeros, it was working fine, as in following example.
|
|
>>> a = np.arange 9) 1 >>> la.norm a, 1) 0.35348576237901524 >>> la.norm a, 2) 0.80588373958852921 >>> la.norm a, 3) 0.94194431453393834","bug"
|
|
"6013","5424","ma.median cannot perform reduce with flexible type.
|
|
Consider the following snippet .
|
|
The last line throws a .
|
|
using version 1.9.1, but works with versions prior to 1.9.0 i didn t try 1.9.0).
|
|
Is this expected behavior?
|
|
Many thanks","bug"
|
|
"6013","5426","fromiter does not work with dtype=numpy.long.
|
|
Running `numpy.long 2 127)` works fine, but numpy.fromiter 2 127 for v in xrange 10)), dtype= numpy.long, count=10) fails with `OverflowError long too big to convert`.
|
|
I assume the long is converted to an int and then back to a long, and that fails.
|
|
Explicitly casting it to `numpy.long` does not help either numpy.fromiter numpy.long 2 127) for v in xrange 10)), dtype=numpy.long, count=10) Using `dtype=numpy.float` works, though.","bug"
|
|
"6013","5428","os.environ gives different results.
|
|
I am compiling numpy 1.9.1 with Intel C on Ubuntu 14.04.
|
|
The `PATH` in my system is ` opt intel composer xe 2015.1.133 bin intel64 opt intel composer xe 2015.1.133 debugger gdb intel64 mic bin usr local cuda bin usr local sbin usr local bin usr sbin usr bin sbin bin usr games usr local games`.
|
|
This is also the same string if I run `os.environ.get PATH )` in the python interpreter.
|
|
To prove the `icc` can be found icc version icc ICC) 15.0.1 20141023 Copyright C) 1985 2014 Intel Corporation.
|
|
All rights reserved.
|
|
But when I run `python setup.py config compiler=intelem build clib compiler=intelem build ext compiler=intelem install` according to the [Intel official instruction][1], it complains `icc` cannot be found, and the `os.environ.get PATH )` in the function `find executable` of `numpy distutils exec command.py` becomes `.
|
|
usr local sbin usr local bin usr sbin usr bin sbin bin`.
|
|
The prepending happens in `numpy distutils misc util.py` and the function `get config cmd`.
|
|
There, the `old path` is exactly ` usr local sbin usr local bin usr sbin usr bin sbin bin`, which fails the installation.
|
|
UPDATE I used `sudo python setup.py install` so the `os.environ` is different.
|
|
[1] https software.intel.com en us articles numpyscipy with intel mkl","bug"
|
|
"6013","5429","Custom ufuncs can t take typed timedeltas and datetimes.
|
|
When creating a custom ufunc e.g.
|
|
with Numba) that uses the generic resolver, timedeltas and datetimes with a unit are forbidden, which kinds of make timedelta and datetime support useless.
|
|
Here s a simple identity function, compiled with Numba .
|
|
It can be used on unit less timedeltas .
|
|
but it fails when a unit is specified .","bug"
|
|
"6013","543","Segfault when indexing on second or higher dimension with a list or tuple Trac 59).
|
|
Original ticket http projects.scipy.org numpy ticket 59 on 2006 04 13 by alberts, assigned to teoliphant.
|
|
The following code segfaults with NumPy 0.9.7.2348, tested with Python 2.4.1 on Fedora Core 4.
|
|
This bug also existed in NumPy 0.9.6.2168, tested with Python 2.4.2 on Windows Server 2003. import numpy as N F = N.zeros 1,1)) F[ ,[0]] = 0 The following also doesn t work F[ , 0,)] = 0 Please add test cases when committing the fix for this bug.
|
|
Thanks!","bug"
|
|
"6013","5430","array2string prints 0 D arrays incorrectly.
|
|
The behavior is easy to reproduce .
|
|
Why is a seemingly arbitrary floating point value being printed?
|
|
The function [array2string] https github.com numpy numpy blob master numpy core arrayprint.py L344) prints 0 D arrays by computing .
|
|
According to the array2string docstring, style ) should be passed the array `a`, but instead it is passed the value of a.item ).
|
|
.
|
|
The call to item ) returns garbage, so garbage is printed instead of something meaningful.
|
|
I ve filed a separate issue for item )).
|
|
It seems like array2string, when called with a 0 D array, should either throw an exception or return something meaningful, perhaps even the empty string.
|
|
At the very least, the docstring needs to be corrected or `a` needs to be passed to the `style` function to be consistent with the docstring.","bug"
|
|
"6013","5431","ndarray.item ) reveals contents of memory for 0 D arrays.
|
|
When ndarray.item ) is called on 0 D arrays, the contents of memory are unnecessarily exposed.
|
|
This causes unpredictable and non deterministic behavior in places where item ) returns values of uninitialized regions of memory.
|
|
This also looks like it could be a security problem.
|
|
To reproduce .
|
|
It seems like item ) should raise an exception when called on a 0 D array.","bug"
|
|
"6013","5433","a bug in zeros array in numpy.
|
|
import numpy as np a=np.zeros 10,dtype=np.int8) a[0]=64 2 print a) you will find the the entry is 128","bug"
|
|
"6013","5434","problems with numpy.delete.
|
|
Just noticed this In [122] np.delete np.arange 1, 100, 1.0), np.arange 1, 100, 1.0)) Out[122] array [ 1.])
|
|
In [123] np.delete np.arange 1, 100, 1.0), np.arange 0, 100, 1.0)) Out[123] array [], dtype=float64) In [124] np.delete np.arange 0, 100, 1.0), np.arange 1, 100, 1.0)) Out[124] array [ 0.])
|
|
not sure if this is intended behavior, but probably worth looking into!","bug"
|
|
"6013","5435","Possible overflow for long integer in matrix multiplication.
|
|
I try to play with the Fibonacci number and find this issue.
|
|
The operation is relatively easy .
|
|
in which the latter matrix contains the 91st and 92nd Fibonacci number.
|
|
The correct result should be .
|
|
But the program returns .
|
|
I am not sure if this is the bug of the program, or I should do something different with this kind of extreme long integers.
|
|
Thanks in advance!","bug"
|
|
"6013","5437","linspace does not properly handle denormal values.
|
|
This is pretty low priority since AFAIK no user has actually run into this, but, since it [was pointed out on python ideas] http news.gmane.org gmane.comp.python.ideas) .
|
|
The problem is that `linspace` does `step = stop start) samples 1)`, and here that becomes .
|
|
The correct output would be something like `[0, 0, 5e 324, 5e 324, ...]`.","bug"
|
|
"6013","544","bool leaks memory Trac 60).
|
|
Original ticket http projects.scipy.org numpy ticket 60 on 2006 04 15 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code leaks memory from numpy import bool bool 1) Valgrind says ==32531== 82 80 direct, 2 indirect) bytes in 2 blocks are definitely lost in loss record 7 of 25 ==32531== at 0x400444E malloc vg replace malloc.c 149) ==32531== by 0x45442E8 array alloc arrayobject.c 5330) ==32531== by 0x454F18D PyArray NewFromDescr arrayobject.c 4153) ==32531== by 0x4551844 Array FromScalar arrayobject.c 5768) ==32531== by 0x45602B7 PyArray FromAny arrayobject.c 6630) ==32531== by 0x4570065 bool arrtype new scalartypes.inc 2855) ==32531== by 0x2FBF6E within usr lib libpython2.4.so.1.0) ==32531== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) The second leak that Valgrind reports is from this code in ma.py MaskType = bool nomask = MaskType 0) Tested with !NumPy 0.9.7.2358.","bug"
|
|
"6013","5441","BUG Wrong sorting of byte swapped complex arrays.
|
|
A minimal not)working example .
|
|
The problem is in ` new argsortlike` and ` new sortlike` in `item selection.c` when the array is byteswapped, it is copied and swapped into a temporary buffer, but rather than using the `dtype` s `copyswapn` function, this is implemented with some generic code that calls ` unaligned strided byte copy` and ` strided byte swap` directly based on the `dtype` s `elsize`.
|
|
For complex arrays, byte swapping therefore results in reversing both the real and imaginary part as a whole, i.e.
|
|
it fixes the byte ordering, but swaps the real and imaginary parts.
|
|
This is very easy to fix, as mentioned above, by calling the `dtype` s `copyswapn` function directly, thus letting it handle the internal structure of the dtype.
|
|
I found this out while working on an ambitious refactor of the sorting functionality, trying to get rid of the ugly global storage used in generic comparisons.
|
|
If I cannot wrap the whole thing up in a couple of weeks, I ll submit a PR fixing this only, opened the issue mostly as a note to self.","bug"
|
|
"6013","5443","DOC Update Pushing Changes to Main Repo Dev Workflow.
|
|
In a recent updated to the Numpy Development Workflow documentation 5332), some issues were brought up regarding the Pushing Changes to Main Repo section, but a consensus on the correct procedure was not reached.
|
|
See the previous pull request for some of the discussions on potential changes.
|
|
Specifically, a series of comments charris, juliantaylor, and tacaswell starting from a line comment by rgommers, as well as a comment at the end of the thread by charris.","bug"
|
|
"6013","5444","raising of NotImplementedError could be more helpful.
|
|
numpy 1.9.0 numpy ma core.py line 5462 says raise NotImplementedError Not implemented yet, sorry... ) This is in MaskedArray.tofile ).
|
|
I came across this when answering a user support question where my user asserted I am not writing a masked array to the file .
|
|
There are a few places in this file that give ambiguous errors like that.
|
|
It would be more helpful if the exception message were to say what it is that is not implemented.
|
|
For example, in this case raise NotImplementedError cannot write masked array to file ) There are other places in ms core.py where there are similarly ambiguous messages, so I conjecture there may be other ambiguous messages in other places.
|
|
I suggest a sweep of the full numpy source code for NotImplementedError and clarify all the messages.
|
|
I think this would be easy for a numpy developer.
|
|
I am not very familiar with numpy myself.
|
|
Since this is a trivial mostly) change, I could possibly make a pull request, if I have some indication that it would be accepted without me making a major commitment to testing or learning the numpy developer work flow.
|
|
Unfortunately, I don t have time to make a serious commitment to becoming a numpy developer.)","enhancement"
|
|
"6013","5445","BUG Sorting of unorderable object types in Python3.
|
|
In Python3 not all objects are comparable .
|
|
Numpy does not acknowledge this possibility, so when trying to sort an object array, even though the error is raised, the sorting goes on as if all unorderable objects compared equal .","bug"
|
|
"6013","5448","Agregation functions don t use full machine int magnitude on 64 bit Windows.
|
|
Linux 64 bit .
|
|
Linux 32 bit .
|
|
Windows 64 bit .
|
|
As you can see, Win64 behaves like 32 bit Linux here the operands are extended to 32 bits rather than 64 bits.","bug"
|
|
"6013","545","Valgrind warning when creating MaskedArray with mask Trac 61).
|
|
Original ticket http projects.scipy.org numpy ticket 61 on 2006 04 15 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code performs a conditional jump or move that depends on uninitialised values from numpy.core.ma import array array 1,mask=[1]) Valgrind says ==356== Conditional jump or move depends on uninitialised value s) ==356== at 0x4683F1F PyArray Newshape multiarraymodule.c 525) ==356== by 0x4684784 PyArray Reshape multiarraymodule.c 369) ==356== by 0x46847C1 array shape set arrayobject.c 4640) ==356== by 0x2CFAC3 within usr lib libpython2.4.so.1.0) ==356== by 0x2EDD2C PyObject GenericSetAttr in usr lib libpython2.4.so.1.0) ==356== by 0x2ED624 PyObject SetAttr in usr lib libpython2.4.so.1.0) ==356== by 0x31FFA4 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==356== by 0x324EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==356== by 0x2D9BE0 within usr lib libpython2.4.so.1.0) ==356== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) ==356== by 0x2CCC75 within usr lib libpython2.4.so.1.0) ==356== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) Tested with !NumPy 0.9.7.2358.","bug"
|
|
"6013","5450","attribute optimize XXX ))) is only available on gcc > 4.4.
|
|
When compiling numpy with gcc 4.2.1 I see a lot of these warnings .
|
|
This is on the following system .
|
|
This is because the optimize attribute was only introduced in [gcc 4.4] https gcc.gnu.org gcc 4.4 changes.html).
|
|
But on OpenBSD, gcc 4.2.1 is used because it s the last GPLv2 release.
|
|
As a work around I ve used a patch similar to the below, but not sure if this is the best way .","bug"
|
|
"6013","5451","remove the funroll loops switch.
|
|
On OpenBSD macppc, using the funroll loops option causes a segfault with scipy.
|
|
Proposed patch below just removes the optimization flag across the board though.
|
|
In the OpenBSD ports tree, this flag is sometimes removed due to buggy results.
|
|
For what it s worth gentoo also recommends against it http wiki.gentoo.org wiki GCC optimization But I get better performance with funroll loops fomg optimize.21 .","bug"
|
|
"6013","5452","datetime64 chokes on sub nanosecond input string.
|
|
When a datetime64 is created from a string with more than 9 decimal places in the second, the result is corrupt.
|
|
.
|
|
The first instance with 9 decimal places in the seconds is correct.
|
|
The second, with 10 decimal places is incorrect.
|
|
Versions numpy 1.9.1; OSX 10.9.5; Python 2.7.9; Python and numpy installed using macports.
|
|
Replicated on RHEL Python 3.2.5 numpy 1.8.0","bug"
|
|
"6013","5456","BUG?
|
|
Inconsistency in slicing of object arrays.
|
|
I am not sure again if this is the expected behavior or not.
|
|
In [16] x1 = [[1], [2, 3]] In [17] b = np.array x1, dtype=object) In [18] x = np.random.rand 5) In [19] x[b[1]] Out[19] array [ 0.65138485, 0.52464837]) In [20] x2 = [[1], [2]] In [21] b = np.array x2, dtype=object) In [22] x[b[0]] IndexError Traceback most recent call last) <ipython input 22 89c92d26bcc0> in <module> ) > 1 x[b[0]] IndexError arrays used as indices must be of integer or boolean) type Should this be a list?
|
|
In [23] b[0] Out[23] array [1], dtype=object)","bug"
|
|
"6013","546","Valgrind warning in MaskedArray code Trac 62).
|
|
Original ticket http projects.scipy.org numpy ticket 62 on 2006 04 15 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code performs a conditional jump or move that depends on uninitialised values from numpy.core.ma import zeros, masked where, MaskType atest = zeros 1,1)) btest = zeros atest.shape, MaskType) ctest = masked where btest,atest) atest ctest Valgrind says ==590== Conditional jump or move depends on uninitialised value s) ==590== at 0x455ECF4 PyArray MapIterReset arrayobject.c 7786) ==590== by 0x456D0A7 array ass sub arrayobject.c 1812) ==590== by 0x2C2BE0 PyObject SetItem in usr lib libpython2.4.so.1.0) ==590== by 0x321014 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==590== by 0x324EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==590== by 0x32369B PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==590== by 0x324EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==590== by 0x32369B PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==590== by 0x324EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==590== by 0x2D9BE0 within usr lib libpython2.4.so.1.0) ==590== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) ==590== by 0x2CCC75 within usr lib libpython2.4.so.1.0) Tested with !NumPy 0.9.7.2358.","bug"
|
|
"6013","5460","public macros missing in C API documentation.
|
|
Noticed after trying to look up ``NPY LIKELY`` used in another PR.
|
|
These seem to be missing NPY LIKELY NPY UNLIKELY NPY UNUSED NPY INTERRUPT H NPY SIGSETJMP NPY SIGLONGJMP NPY SIGJMP BUF NPY SIGINT ON NPY SIGINT OFF NPY HAVE ENDIAN H There isn t yet an obvious place to add these in the reference guide.
|
|
Maybe a new sub section under Numpy core libraries except for ``HAVE ENDIAN`` which belongs in System configuration )?
|
|
Suggestion for a good name?","enhancement"
|
|
"6013","5463","array2string handles floats differently for structured array and ndarray.
|
|
This may not be a bug per se, but it is unexpected to me) >>> arr = np.array [1.10000000000003,2.1], dtype= <f8 ) >>> np.array2string arr, separator= , ) [ 1.1, 2.1] vs. >>> arr = np.array 1.10000000000003,2.1), dtype=[ x , <f8 ), y , <f8 )]) >>> np.array2string arr, separator= , ) 1.10000000000003, 2.1) this is with Numpy v1.9.1)","bug"
|
|
"6013","5465","No clear documentation of how to go about working on Numpy.
|
|
Currently there is no way to use a development Numpy without installing it, which makes interactive use and testing a pain .","bug"
|
|
"6013","5466","Creation of arrays inside an object array.
|
|
I want to declare an array of object and later to include arrays in it.
|
|
I can do it this way .
|
|
But since Numpy has iterators, I wanted to use them .
|
|
and the message is .
|
|
It seems, according to stackoverflow, it is an example which could be usefull for your discussion on creation of object array.","bug"
|
|
"6013","5467","Interrupting build leaves working copy in a broken copy.
|
|
If you try to build Numpy, for example .
|
|
and press Ctrl C before the end, the working copy ends up broken.
|
|
Any further invocation fails .
|
|
You then have to issue a non trivial invocation of git clean to start again from a blank slate.","bug"
|
|
"6013","5469","BUG partition and argpartition don t bound check kth for some types.
|
|
For dtypes that do not have a type specific implementation of the introselect algorithm, both `partition` and `argpartition` resort to fully arg)sorting the input array.
|
|
When doing so, the `kth` array of elements to partition by is not bound checked .","bug"
|
|
"6013","547","Typo in numpy core src arraytypes.inc.src Trac 63).
|
|
Original ticket http projects.scipy.org numpy ticket 63 on 2006 04 15 by alberts, assigned to unknown.
|
|
diff ur numpy core src arraytypes.inc.src numpy.new core src arraytypes.inc.src numpy core src arraytypes.inc.src 2006 04 13 14 02 59.000000000 0200 numpy.new core src arraytypes.inc.src 2006 04 15 15 03 34.000000000 0200 479,7 479,7 n = PyTuple GET SIZE names); if PyTuple GET SIZE op) != n) { PyErr SetString PyExc ValueError, size of tuple must match \ size of tuple must match \ number of fields.
|
|
); return 1; } Diff against !NumPy 0.9.7.2358.","bug"
|
|
"6013","5474","fails to use gcc4.8.1 to compile on AIX6.1.
|
|
running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building py modules sources building library npymath sources error don t know how to compile C C code on platform posix with gcc compiler","bug"
|
|
"6013","5477","documentation on sliced arrays leaves out critical detail.
|
|
I m a Python beginner, coming from Matlab.
|
|
So was puzzled that taking an array, say, A = arange 35).reshape 7,5) the trying to get a slice, like A[ ,0 3] gives array [[ 0, 1, 2], [ 5, 6, 7], [10, 11, 12], [15, 16, 17], [20, 21, 22], [25, 26, 27], [30, 31, 32]]) that is, 3 array elements from the 2nd dimension instead of 4 which was specified)...
|
|
Deep down somewhere else, a web page on Numpy says subscripting an array with a b extracts elements from a to b NOT INCLUDING b.
|
|
This is non intuitive and needs to be highlighted.
|
|
or else, made compatible with something rational like Matlab).
|
|
The fewer differences people need to keep track of so they can do multiple languages, the better","bug"
|
|
"6013","5479","Windows wheel package .whl) on Pypi.
|
|
Please make Windows wheel packages and put them on Pypi.
|
|
Currently it is possible to download Windows wheel packages for numpy here http www.lfd.uci.edu gohlke pythonlibs numpy It would be great if the wheels were directly available on the Pypi server https pypi.python.org pypi so that they can be installed with pip.","bug"
|
|
"6013","548","Valgrind warning when indexing heterogeneous nested array Trac 64).
|
|
Original ticket http projects.scipy.org numpy ticket 64 on 2006 04 15 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code performs a conditional jump or move that depends on uninitialised values from numpy import array descr = [ x , [ y , [ z , c16 , 2,)),]),]),] buffer = [6j,4j],),),) h = array buffer, dtype=descr) h[ x ][ y ][ z ] Valgrind says ==899== Conditional jump or move depends on uninitialised value s) ==899== at 0x454EE42 PyArray NewFromDescr arrayobject.c 4117) ==899== by 0x4550909 PyArray GetField arraymethods.c 265) ==899== by 0x45509AE array getfield arraymethods.c 291) ==899== by 0x2EB1CA PyCFunction Call in usr lib libpython2.4.so.1.0) ==899== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) ==899== by 0x31E1BF PyEval CallObjectWithKeywords in usr lib libpython2.4.so.1.0) ==899== by 0x2C537A PyObject CallObject in usr lib libpython2.4.so.1.0) ==899== by 0x4559D89 gentype generic method scalartypes.inc 354) ==899== by 0x455A53E voidtype getfield scalartypes.inc 1819) ==899== by 0x455A65C voidtype subscript scalartypes.inc 2160) ==899== by 0x2C2A21 PyObject GetItem in usr lib libpython2.4.so.1.0) ==899== by 0x321CBE PyEval EvalFrame in usr lib libpython2.4.so.1.0) Tested with !NumPy 0.9.7.2358.","bug"
|
|
"6013","5484","numpy.mean returns nan when indices are equal.
|
|
I was working on finding the mean value over a set of array elements where sometimes the bounds are equal.
|
|
numpy.mean returns nan in such a case.
|
|
I m guessing that it subtracts the bounds and gets a divide by zero.
|
|
aa = numpy.zeros 5,5),dtype=numpy.float) aa = 5 print numpy.mean aa) print numpy.mean aa[3,3]) print numpy.mean aa[3 4,3 4]) print numpy.mean aa[3 3,3]) print numpy.mean aa[3,3 3]) print numpy.mean aa[3 3,3 3]) 5.0 5.0 5.0 nan nan nan I can include a test for the equality of the bounds, but it seems like it should work with equal bounds.
|
|
Thanks Mike","bug"
|
|
"6013","5485","dot passes overlapping memory to blas.
|
|
forgot if that is allowed or not but openblas up to git head) has issues with it also not sure if it is a regression, I haven t seen the issue in openblas earlier, but its a while since I ran a valgrind check.
|
|
.","bug"
|
|
"6013","5486","Make numpy.distutils recognize Fortran 2003 and Fortran 2008 files.
|
|
NumPy build utils does not know what to do with .
|
|
and .
|
|
files.
|
|
This makes it difficult to compile a library written in Fortran 2003 or 2008.
|
|
Fortran 2003 is e.g.
|
|
useful if we want to make a C callable library without depending on an ABI wrapper.
|
|
A typical usecase is calling Fortran from Cython, which is more flexible and efficient than using f2py.
|
|
f2py does not parse Fortran 2003.
|
|
But that is not a problem if we just want to use ., and not .
|
|
which will invoke f2py).","bug"
|
|
"6013","549","Another Valgrind warning when indexing heterogeneous nested array Trac 65).
|
|
Original ticket http projects.scipy.org numpy ticket 65 on 2006 04 15 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code performs a conditional jump or move that depends on uninitialised values from numpy import array descr = [ x , i4 , 2,))] buffer = [3,2],) h = array buffer, dtype=descr) h[ x ] Valgrind says ==992== Conditional jump or move depends on uninitialised value s) ==992== at 0x454EE42 PyArray NewFromDescr arrayobject.c 4117) ==992== by 0x4550909 PyArray GetField arraymethods.c 265) ==992== by 0x456C074 array subscript arrayobject.c 2010) ==992== by 0x456D626 array subscript nice arrayobject.c 2250) ==992== by 0x2C2A21 PyObject GetItem in usr lib libpython2.4.so.1.0) ==992== by 0x321CBE PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==992== by 0x324EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==992== by 0x325227 PyEval EvalCode in usr lib libpython2.4.so.1.0) ==992== by 0x341559 within usr lib libpython2.4.so.1.0) ==992== by 0x3427D1 PyRun SimpleFileExFlags in usr lib libpython2.4.so.1.0) ==992== by 0x343268 PyRun AnyFileExFlags in usr lib libpython2.4.so.1.0) ==992== by 0x34916C Py Main in usr lib libpython2.4.so.1.0) Note that this stack trace differs slightly from the one in ticket 485, but both errors might be caused by the same bug.
|
|
Tested with !NumPy 0.9.7.2358.","bug"
|
|
"6013","5491","Opening Pickle Files using ma.dump in Python 3.
|
|
When using numpy.ma.dump to save a masked array in Python 3 .
|
|
I got the following error .
|
|
This is with NumPy 1.9.1 and Python 3.3.2.
|
|
From what I found out, it is because in Python 3 pickle files has to be opened in binary mode, i.e.
|
|
L7142 should be `F = open F, wb )`.
|
|
Indeed, if I were to change it to .
|
|
the error goes away.","bug"
|
|
"6013","5497","Error in help documentation of numpy.zeros.
|
|
The help documentation for numpy zeros method `import numpy; help numpy.zeros)`) shows an example like this .
|
|
The documentation throughout `multiarray` module assumes that `numpy` is imported as `np` but the value for `dtype` keyword is passed as `numpy.int`.
|
|
While the above snippet conveys the point across, it will throw a name error if some one actually tried to execute it .
|
|
This can be fixed by consitent usage of `numpy` or `np`.","bug"
|
|
"6013","5499","Python3 arraypad.py s normalize shape not accepting all types of ints.
|
|
This is Numpy V1.9.1.
|
|
I found the bug while testing a working py2 code in py3 .
|
|
is working fine, but .
|
|
is crashing with .
|
|
it s related to the fact that .
|
|
returns .
|
|
in Python3 but .
|
|
in Python2.
|
|
Replacing .
|
|
with .
|
|
in arraypad.py at line 1024 should do the trick.
|
|
I could do the fix but I never commited a thing before....","bug"
|
|
"6013","550","More descriptive docstring for sqrt Trac 66).
|
|
Original ticket http projects.scipy.org numpy ticket 66 on 2006 04 18 by stefanv, assigned to unknown.
|
|
The .
|
|
docstring specified in .)
|
|
states y = sqrt x) square root elementwise.
|
|
A docstring that explains the differences in output for different input types would be useful.
|
|
Something along the lines of y = sqrt x) square root elementwise.
|
|
For real x, the domain is restricted to x>=0.","enhancement"
|
|
"6013","5506","f2py compilation error with array argument in Fortran subroutine.
|
|
When I constructed an array named ``list`` in fortran, like code shown below .
|
|
I encountered the following error, which seems to treat exponentiation as a pointer.
|
|
Any ideas how to get rid of it?
|
|
Thanks!
|
|
.","bug"
|
|
"6013","5507","Performance of numpy average and numpy.mean function.
|
|
I need a weightened average function on a VERY large Dataset some 1e8 numbers or more).
|
|
The numpy functions mean and average serve me well and fast, but I discovered, that numpy.average is slower than builing the weightened average myself with two numpy.mean functions, as shown by the example https gist.github.com skuschel 2d148a37a2ce17925fb0 `np.average a,weights=b)` takes 0.32 sec on my computer, but `np.mean a b) np.mean b)` takes 0.23 sec for the equally sized dataset, yielding the same result.
|
|
How does that make sense?","enhancement"
|
|
"6013","551","Complex array comparison may leak memory.
|
|
Trac 67).
|
|
Original ticket http projects.scipy.org numpy ticket 67 on 2006 04 18 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1, the following code leaks memory from numpy import array from numpy.testing import assert almost equal x = array [1 2j]) assert almost equal x 1), [1 1 2j)]) Valgrind says ==18687== 56 40 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 7 of 26 ==18687== at 0x400444E malloc vg replace malloc.c 149) ==18687== by 0x466D2A8 array alloc arrayobject.c 5331) ==18687== by 0x467814D PyArray NewFromDescr arrayobject.c 4153) ==18687== by 0x47D3454 construct loop ufuncobject.c 1000) ==18687== by 0x47D3E2D PyUFunc GenericFunction ufuncobject.c 1402) ==18687== by 0x47D4FBF ufunc generic call ufuncobject.c 2655) ==18687== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) ==18687== by 0x2C547B PyObject CallFunction in usr lib libpython2.4.so.1.0) ==18687== by 0x466B40B PyArray GenericBinaryFunction arrayobject.c 2593) ==18687== by 0x2C3299 within usr lib libpython2.4.so.1.0) ==18687== by 0x2C3CFF PyNumber Multiply in usr lib libpython2.4.so.1.0) ==18687== by 0x4696790 PyArray Round multiarraymodule.c 268) ==18687== by 0x4696C19 array round arraymethods.c 1494) ==18687== by 0x2EB1CA PyCFunction Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with !NumPy 0.9.7.2360.","bug"
|
|
"6013","5511","System complex functions not detected used on Fedora 21.. Current state on Fedora 21. .
|
|
Looks like library function is not used.
|
|
gcc 4.9.2 gives `1.000000000000000000 0.000000000000000000i`, where denormals look to be flushed to zero, probably a side affect of how the function compiles.
|
|
The code for that is .","bug"
|
|
"6013","5514","RandomState.shuffle has a terrible overhead for NumPy arrays.
|
|
The shuffle function for .
|
|
is implemented like this .
|
|
This implementation is close to optimal for Python lists, as it would just swap .
|
|
pointers.
|
|
We could probably do a little better by declaring .
|
|
list, but not much.
|
|
However for NumPy arrays this implementation is awful.
|
|
First, the indexing is implemented as Python function calls, as Cython does not know that .
|
|
is an ndarray.
|
|
Along with it comes all the overhead like bounds checking, etc.
|
|
Second, the swapping statement .
|
|
continously creates and destroys dtype scalars Python objects).
|
|
In the multidimensional case it creates temporary ndarrays instead.
|
|
All in all, this is as bad as it gets.
|
|
In the single dimensional case we should have the shuffling loop .
|
|
as a C function, with specialisations depending on the dtype e.g.
|
|
a switch statement on the typenum).
|
|
Another possibility to avoid C is to use fusedtypes in Cython to code this generically.
|
|
In the multidimensional case we could shuffle an array of void pointers to the subarrays, and use this to fill in the output array.
|
|
This will kill enormous amount of Python overhead in Monte Carlo simulations where we need to shuffle, e.g.
|
|
in permutation tests.","bug"
|
|
"6013","552","PyArray ToFile ignores return value from fwrite Trac 68).
|
|
Original ticket http projects.scipy.org numpy ticket 68 on 2006 04 18 by alberts, assigned to unknown.
|
|
The `PyArray ToFile` function in `numpy core src arrayobject.c`, around line 1202 calls the `fwrite` function, but ignores the return value.
|
|
The return value should be checked to ensure that all the requested bytes were written to the file and if not, an error should be raised.
|
|
This problem came to light as a result of a GCC 4 compiler warning numpy core src arrayobject.c In function PyArray ToFile numpy core src arrayobject.c 1202 warning ignoring return value of fwrite, declared with attribute warn unused result Confirmed in !NumPy 0.9.7.2360.","bug"
|
|
"6013","5520","how can i read the numpy.where ) in the sources.
|
|
i just use .
|
|
only , i want to know where i could read the function","enhancement"
|
|
"6013","5521","np.lib.test ) FAILED.
|
|
ubuntu 14.04 pip freeze backports.ssl match hostname==3.4.0.2 certifi==14.5.14 Cython==0.21.2 docutils==0.12 h5py==2.4.0 ipython==2.3.1 Jinja2==2.7.3 MarkupSafe==0.23 matplotlib==1.4.2 mock==1.0.1 nose==1.3.4 numexpr==2.4 numpy==1.9.1 numpydoc==0.5 pandas==0.15.2 Pillow==2.7.0 Pygments==2.0.2 pyparsing==2.0.3 python dateutil==2.4.0 pytz==2014.10 pyzmq==14.5.0 scikit image==0.10.1 scikit learn==0.15.2 scipy==0.15.1 six==1.9.0 Sphinx==1.2.3 tables==3.1.1 tornado==4.0.2 In [3] np.lib.test ) Running unit tests for numpy.lib NumPy version 1.9.1 NumPy is installed in usr local lib python2.7 dist packages numpy Python version 2.7.6 default, Mar 22 2014, 22 59 56) [GCC 4.8.2] nose version 1.3.4 ...................................................................................................................................................... ...................................................................................................................................................... ...................................................................................................................................................... ...................................................................................................................................................... ...................................................................................................................................................... ...................................................................................................................................................... ............................................F......................................................................................................... ...................................................................................................................................................... ...................................................................................................................................................... ====================================================================== FAIL test allnans test nanfunctions.TestNanFunctions Median) Traceback most recent call last) File usr local lib python2.7 dist packages numpy lib tests test nanfunctions.py , line 613, in test allnans assert len w) == 3) File usr local lib python2.7 dist packages numpy testing utils.py , line 53, in assert raise AssertionError smsg) AssertionError Ran 1350 tests in 5.710s FAILED failures=1) Out[3] <nose.result.TextTestResult run=1350 errors=0 failures=1>","bug"
|
|
"6013","5524","argpartition fails on non ndarray array likes.
|
|
While `partition` works, as advertised, array likeinputs like, e.g.
|
|
lists .
|
|
`argpartition` raises an error if its first argument is not an `ndarray` .","bug"
|
|
"6013","5526","Newstyle buffers with suboffsets are probably wrongly used instead of rejected?.
|
|
Just don t have a clear idea how to easy test this, but it seems to me that we will happily interpet a newstyle buffer with suboffsets as one without suboffsets.
|
|
In principle, we should likely even check if all suboffsets may be negative in which case suboffsets are effectively not being used) and we can handle it anyway.
|
|
it only tangentially touches this, but see also http bugs.python.org issue23352)","bug"
|
|
"6013","5528","Polyfit better error message.
|
|
.
|
|
The following calls all result in the above error message.
|
|
.
|
|
Yes, these are all invalid cases for polynomial, but may arrise when input is computed itself and a better error message will help, something like Not enough distinct points to construct the polynomial .
|
|
One learns the former error message and the latter are loosely related but the latter is more explicit and does not leak the implementation details.
|
|
This is numpy version 1.8.2","bug"
|
|
"6013","5529","ImportError cannot import name multiarray.
|
|
nils linux 5fmr git numpy> python setup.py install prefix= HOME local Running from numpy source directory.
|
|
Traceback most recent call last) File setup.py , line 251, in <module> setup package ) File setup.py , line 235, in setup package from numpy.distutils.core import setup File home nils git numpy numpy distutils init .py , line 37, in <module> from numpy.testing import Tester File home nils git numpy numpy testing init .py , line 14, in <module> from .utils import File home nils git numpy numpy testing utils.py , line 17, in <module> from numpy.core import float32, empty, arange, array repr, ndarray File home nils git numpy numpy core init .py , line 14, in <module> from .
|
|
import multiarray ImportError cannot import name multiarray","bug"
|
|
"6013","553","Why numpy headers are installed using add data dir and not add headers?
|
|
Trac 69).
|
|
Original ticket http projects.scipy.org numpy ticket 69 on 2006 04 18 by pearu, assigned to unknown.
|
|
Need to answer this question in scipy FAQ.","bug"
|
|
"6013","5530","Issue when using loadtxt with string data in Python 3.
|
|
Let s say I use a data file ``data.txt`` containing .
|
|
If I do in Python 3.4) .
|
|
I get .
|
|
Note that the ``b`` is actually part of the string, so there is something wrong here.
|
|
The same happens if I use ``dtype= U ``.
|
|
I am using Numpy dev as of a02f5c8fd384b5aeffd1942747b4b6c4efdcd3a4.","bug"
|
|
"6013","5532","Patch for unraised exceptions in cygwin64.
|
|
On cygwin64, some floating point exceptions are not raised correctly .
|
|
The reason is that a specific `fenv.h` provided by numpy probably for cygwin32) is used in `ieee754.c.src`, instead of the system wide version.
|
|
I guess this dates back to times when there was no `fenv.h` on cygwin, but it is now available on both cygwin32 and cygwin64.
|
|
A simple patch follows, enabling the system wide `fenv.h` in `ieee754.c.src` and removing the references to the local `fenv.h` and `fenv.c` in `numpy core setup.py`.
|
|
I don t know if these files can still be useful on some exotic systems or if they can be deleted.
|
|
With the patch, the correct exception or warning is raised, and almost all tests pass on cygwin64.
|
|
.","bug"
|
|
"6013","5533","numpy.dot cannot handle arrays with >2 31 elements without support in the underlying BLAS.
|
|
[The bug here is that some BLAS libraries will crash if you pass in > 16 GiB arrays, because they use 32 bit indices internally.
|
|
The most obvious solution is to break the dot call into multiple calls to dgemm, though there are also [other possibilities] https github.com numpy numpy issues 5533 issuecomment 114309520).
|
|
Original report follows ] On python 2.7.8, numpy 1.9.1, on Mac OS X import numpy numpy.random.seed 1) X = numpy.random.random 50000,100)) numpy.dot X, X.T) Results in Segmentation fault 11 Segfault doesn t occur on smaller arrays e.g.
|
|
30000x100 is fine).
|
|
In case it s useful, some linkage info >>> numpy.
|
|
config .show ) atlas threads info NOT AVAILABLE blas opt info extra link args = [ Wl, framework , Wl,Accelerate ] extra compile args = [ msse3 , DAPPLE ACCELERATE SGEMV PATCH , I System Library Frameworks vecLib.framework Headers ] define macros = [ NO ATLAS INFO , 3)] atlas blas threads info NOT AVAILABLE openblas info NOT AVAILABLE lapack opt info extra link args = [ Wl, framework , Wl,Accelerate ] extra compile args = [ msse3 , DAPPLE ACCELERATE SGEMV PATCH ] define macros = [ NO ATLAS INFO , 3)] openblas lapack info NOT AVAILABLE atlas info NOT AVAILABLE lapack mkl info NOT AVAILABLE blas mkl info NOT AVAILABLE atlas blas info NOT AVAILABLE mkl info NOT AVAILABLE","bug"
|
|
"6013","5534","Document size of ndarray.view when using dtypes of different sizes.
|
|
The docs for ndarray.view explicitly allows for views with dtypes of different sizes although it warns that the result is layout dependent), but does not specify how the size of the new array is calculated.
|
|
I believe from quick experimentation that the number of dims is conserved, but haven t tried e.g.
|
|
weirdly strided arrays so I don t know the answer either.","bug"
|
|
"6013","5537","Failure in array ) with complex 1 element array in list.
|
|
Spotted on stackoverflow this fails with a somewhat mysterious error .
|
|
The expected result would be to treat the 1 element array as a scalar, similarly as in .","bug"
|
|
"6013","5538","histogram2d docstring too restrictive.
|
|
The docstring for `histogram2d )` does not mention that the `bins` parameter can be of the form [int, array] or [array, int], but this does seem to work.
|
|
It would be nice if this were documented.
|
|
Maybe the same should be done as well for similar functions, if it applies.","bug"
|
|
"6013","554","Improve copyswap and copyswapn functions Trac 70).
|
|
Original ticket http projects.scipy.org numpy ticket 70 on 2006 04 19 by teoliphant, assigned to unknown.
|
|
We need to adjust copyswap and copyswapn functions to take advantage of compiler optimizations for aligned data.
|
|
We should probably adapt them to handle strided cases as well.
|
|
This requires adjusting the interface to pass in more information like whether or not the data is aligned and any stride information).
|
|
Perhaps passing in the pointer to the array object is the best thing like the other data type functions.","enhancement"
|
|
"6013","5541","Mathematically incorrect behaviour of adding numpy vectors.
|
|
Hi!
|
|
Currently If I add an 3x1 vector to an 1x3 vector I get a matrix .
|
|
This is 1.
|
|
Mathematically not correct; either it should return an 3x1 or 1x3 vector or return an Exception that the dimensions do not match 2.
|
|
Unpractilly for programming, because dimension mismatches go on without being noticed, and therefore are hard to debug, and always have to be checked seperately.
|
|
I think the best way would do not permit such a behaviour.
|
|
At least I don t see a use for such a behaviour.
|
|
If I m wrong correct me please.
|
|
In Octave at least a warning is provided .","bug"
|
|
"6013","5543","ndarray should offer format that can adjust precision.
|
|
In many wonderful cases an ndarray can be used in place of a Python float and Just Work.
|
|
But not in one case .
|
|
The output of the above code, at least under Python 3.4, is .
|
|
It would be a great convenience if the ndarray grew a ` format )` method that understood the tiny mini language of float formatting, and used the number of digits of precision specified there to make its own call to the standard NumPy vector array formatting.
|
|
Users could control array appearance on the screen using a Python standard that many programmers already understand.","bug"
|
|
"6013","555","The Kronecker product function kron should return a matrix instead of an array) at least for matrix inputs Trac 71).
|
|
Original ticket http projects.scipy.org numpy ticket 71 on 2006 04 19 by trac user svetosch gmx net, assigned to unknown.
|
|
kron always returns a numpy array, even when working with numpy matrices exclusively.
|
|
Example >>> import numpy as n >>> n.kron n.asmatrix n.ones 1,2))), n.asmatrix n.zeros 2,2)))) array [[0, 0, 0, 0], [0, 0, 0, 0]]) There were also further suggestions on the mailing list thread Toward release 1.0 of Numpy ) to restrict inputs to 2d arrays including matrices).","bug"
|
|
"6013","5550","powerpc aligned contig cast double to longdouble Assertion `npy is aligned dst, builtin offsetof struct {char c; npy longdouble v;}, v)) failed.
|
|
Running unit tests for numpy.
|
|
Hello, with 1.9.2rc1 in Debian we got the following error running tests with python2.7 debug flavour when building the package on powerpc architecture .
|
|
full build log at https buildd.debian.org status fetch.php?pkg=python numpy arch=powerpc ver=1 3A1.9.2 rc1 1 stamp=1423426097","bug"
|
|
"6013","5552","Bug in `ndarray.fill )` in combination with indexing using a boolean array.
|
|
Sorry, if this is a duplicate but I could not find a matching issue..
|
|
I am reporting an issue that was introduced after version 1.8.2 and is present in all of 1.9.1, 1.9.2rc1 and master a4cdc5b8d314fcf18faf29d82e84c877c4ed0e3f).
|
|
I am on Python 2.7.9.
|
|
Here s a snippet to illustrate the issue .
|
|
which produces.. .
|
|
I ve tried to `git bisect` to the responsible commit, but git seems to get confused, it stops the bisect run with the following message .
|
|
Here s the `git bisect log` afterwards .","bug"
|
|
"6013","5553","Handling Fortran data types with numpy.dtype.
|
|
Hi, I m converting some Fortran code to Python, and the data is in these types .
|
|
`2I4` means two consecutive columns have the data type `I4`) I noticed that `numpy.dtype I4 )` seems to work, but `numpy.dtype F6.1 )` does not.
|
|
Any advice on how I should handle this?","bug"
|
|
"6013","5554","Options for Intel Fortran Composer XE 2015 outdated.
|
|
According to https software.intel.com en us forums topic 538120 The two Fortran compiler options Qlowercase us used in numpy scripts are already removed from Intel R) Visual Fortran 15.0.
|
|
The suggested replacement are names lowercase and assume underscore.
|
|
Under class IntelVisualFCompiler BaseIntelFCompiler) Change .
|
|
to .
|
|
This change was necessary in order to compile scipy.
|
|
I do not know what this change implies for other versions of the Fortran compiler.","bug"
|
|
"6013","5557","Get error from loading very small numbers from FORTRAN.
|
|
For example, I get solution including 3.14 114. with real kind=8).
|
|
But python cannot identify the data.
|
|
The error is as follows usr lib64 python2.7 site packages numpy lib npyio.pyc in loadtxt fname, dtype, comments, delimiter, converters, skiprows, usecols, unpack, ndmin) 825 vals = [vals[i] for i in usecols] 826 Convert each value according to its column and store > 827 items = [conv val) for conv, val) in zip converters, vals)] 828 Then pack it according to the dtype s nesting 829 items = pack items items, packing) ValueError invalid literal for float ) 5.687138413593 304","bug"
|
|
"6013","5558","Masked arrays don t properly edit views.
|
|
I tried to change the mask of `a` through a subindexed view, but was unable.
|
|
Using this setup I can reproduce this in the 1.9.1 version of NumPy.
|
|
import numpy as np a = np.arange 6).reshape 2,3) a = np.ma.masked array a, mask=np.ma.getmaskarray a), shrink=False) b = a[1 2,1 2] c = np.zeros b.shape, b.dtype) c = np.ma.masked array c, mask=np.ma.getmaskarray c), shrink=False) c[ ] = np.ma.masked This yields what one would expect for `a`, `b`, and `c` seen below).
|
|
masked array data = [[0 1 2] [3 4 5]], mask = [[False False False] [False False False]], fill value = 999999) masked array data = [[4]], mask = [[False]], fill value = 999999) masked array data = [[ ]], mask = [[ True]], fill value = 999999) Now, it would seem reasonable that to copy data into `b` from `c` one can use ` setitem ` seen below).
|
|
b[ ] = c This results in new data and mask for `b`.
|
|
masked array data = [[ ]], mask = [[ True]], fill value = 999999) This should, in turn, change `a`.
|
|
However, the mask of `a` remains unchanged seen below).
|
|
masked array data = [[0 1 2] [3 0 5]], mask = [[False False False] [False False False]], fill value = 999999)","bug"
|
|
"6013","5559","Numpy structured arrays and logical indexing.
|
|
Assigning values to Numpy structured arrays produces very odd behavior.
|
|
For example, the following does not assign correctly H = np.zeros 4, dtype=[ ind , float )]) H[ ind ] = np.arange 1,5) H[ ind ] array [ 1., 2., 3., 4.])
|
|
x = [False, True, False, True] H[ ind ][np.ix x)] array [ 2., 4.])
|
|
H[np.ix x)][ ind ] = [ 1, 1] H array [ 1.0,), 2.0,), 3.0,), 4.0,)], dtype=[ ind , <f8 )]) If I attempt to make the change recommended here https github.com numpy numpy issues 3561, it still does not even change the first entry indexed H[np.ix x)][ ind ][ ] = [ 1, 1] H array [ 1.0,), 2.0,), 3.0,), 4.0,)], dtype=[ ind , <f8 )]) Why might this be the case?","bug"
|
|
"6013","556","Scalar objects can t compare with other types of objects Trac 72).
|
|
Original ticket http projects.scipy.org numpy ticket 72 on 2006 04 19 by rkern, assigned to unknown.
|
|
In [1] from numpy import In [2] a = linspace 0, 10, 11) In [3] a[0] == auto exceptions.TypeError Traceback most recent call last) opt svn tempo numpy src lib em <ipython console> TypeError invalid string data typesin comparison","bug"
|
|
"6013","5560","MaskedArray.trace ) inconsistent with ndarray.trace ) for arrays with ndim>2.
|
|
The `trace` method for a masked array with more than 2 dimensions does not appear to respect the `axis1` and `axis2` methods.
|
|
In fact, for a 3 dimensional array, `MaskedArray.trace` does not return an array as `ndarray.trace` does, but appears to return the sum of the relevant array.
|
|
See the code sample below.
|
|
.","bug"
|
|
"6013","5561","tools swig numpy.i gives syntax error when swigged in C mode.
|
|
The ifdef cplusplus guard at the end of tools swig numpy.i doesn t seem to be honored by the latest SWIG 3.0.5, or indeed older ones back to 2.0.4).
|
|
As a result, swig imports std complex.i, and tries to setup std complex definitions, which it can t manage in C mode.
|
|
By contrast, everything works with swig c .","bug"
|
|
"6013","5562","NumPy scalars are not collections.Hashable on Python 3. xref https github.com pydata pandas issues 9276 For example .
|
|
On Python 2.7, the last output here is `True` which is the correct value)","bug"
|
|
"6013","5569","Unitialized variable in numpy core src multiarray compiled base.c.
|
|
Looks like `tp dict` in `numpy core src multiarray compiled base.c arr add docstring` can be used uninitialized.
|
|
juliantaylor git blame is not your friend ;)","bug"
|
|
"6013","557","Flesh out site.cfg.example Trac 73).
|
|
Original ticket http projects.scipy.org numpy ticket 73 on 2006 04 20 by rkern, assigned to rkern.
|
|
The current example is quite sparse.
|
|
I need to write a fuller example that demonstrates all of the features one can use e.g.
|
|
using colons to separate directories in a list, using commas to separate library names, the [DEFAULT] section, etc.).
|
|
The file should also provide examples for common configurations, like using ATLAS or the MKL or reference BLAS and LAPACK source directories.","enhancement"
|
|
"6013","5570","gentype alloc in scalartypes.c.src does not check for memory error.. As stated.
|
|
This function could possibly be simplified to use `PyObject {New, NewVar}` if the extra bits of the object don t need to be initialized to zero, the steps are otherwise identical if the error check is omitted.","bug"
|
|
"6013","5573","numpy.take ) with a ignore mode.
|
|
For me numpy.take ) is a great function but the possibility to behave on out of bounds with just ignoring these is really missing.
|
|
Example with not yet existing mode= ignore feature) .
|
|
shall result in > array [7, 34, 76]) ignoring the index 6, which is out of bounds ...
|
|
I tried to add this to numpy by my own, but unfortunately although I think it s not a big deal if you know how numpy library is build up) I failed at first with the finding of the right source code place...","bug"
|
|
"6013","5574","IndexError when extracting noncontiguous columns and rows from matrix.
|
|
Test case .
|
|
However, `A[ 0,3), 0,2,3)]` is a question with a very definite answer which can be obtained by calling .
|
|
I can t think of any situation in which this might be ambiguous or indefinite, so if in fact none exists, I propose that numpy should return this result instead of an exception.
|
|
Thank you for your work!","bug"
|
|
"6013","5575","Problems with handling AMD64 Pile Driver.
|
|
scipy errors and failures due to piledriver 1) ERROR test improvement test quadpack.TestCtypesQuad) WindowsError [Error 193] 1 is not a valid Win32 application 2) ERROR test typical test quadpack.TestCtypesQuad) WindowsError [Error 193] 1 is not a valid Win32 application 3) FAIL test interpolate.TestInterp1D.test circular refs ReferenceError Remaining reference s) to object 4) FAIL test gcutils.test assert deallocated ReferenceError Remaining reference s) to object Colin W. with thanks to Carl Kleffner","bug"
|
|
"6013","5578","wrong eigenvectors when using linalg.eig.
|
|
.
|
|
returns [[ 1.00000000e 00 1.00000000e 00] [ 0.00000000e 00 1.11022302e 16]] The second eigenvector is [0, 0], that is OK, but for the first one I was expecting [1, 0] like in [here] http www.wolframalpha.com input ?i=eigenvectors 28 5B1 2C 2 5D 2C 5B0 2C 1 5D 29) It is also easy to check that the correct solution should be [1, 0] and not [1, 1].
|
|
Is this a bug or the expected behaviour?
|
|
Am I doing something wrong?","bug"
|
|
"6013","558","Enable email notification of assigned tickets Trac 74).
|
|
Original ticket http projects.scipy.org numpy ticket 74 on 2006 04 20 by rkern, assigned to rkern.
|
|
Currently ticket creators, assignees, and the CCed do not receive email notification of their tickets.","bug"
|
|
"6013","5581","Strange issue with casting a list of strings and longs to an array.
|
|
This is for numpy version `1.6.1` .","bug"
|
|
"6013","5586","Integer arithmetic doesn t use SIMD.
|
|
See [this SO question] http stackoverflow.com q 28592415 2399799).
|
|
Addition of two integer arrays should be using SIMD, as should all other ufunc ish) integer functions.
|
|
Currently all integer arithmetic takes roughly 3 5 times longer than 32 bit float operations.
|
|
In the most extreme example, for the case of 8bit integers, this means there is probably a 12 20 times potential speedup not being used, which is a real shame.","bug"
|
|
"6013","5588","numpy.linalg.svd hangs on bad input, fails to raise LinAlg SVD did not converge ) error.
|
|
When an SVD calculation does not converge, normally we get `LinAlgError SVD did not converge`.
|
|
For some matrices, this does not occur, and numpy remains in `svd n s` forever .
|
|
Rather than raising an exception, this causes numpy to hang indefinitely.","bug"
|
|
"6013","5589","numpy.core.multiarray.datetime data should return tuple in reverse order.
|
|
`datetime data` is a great function to expose for time unit extraction.
|
|
It currently works like this .
|
|
`timedelta64` constructor takes the arguments in the reverse order from which `datetime data` returns them.
|
|
I think it would be more intuitive if they were the same.","bug"
|
|
"6013","559","Indexing into 1 D arrays is slow Trac 75).
|
|
Original ticket http projects.scipy.org numpy ticket 75 on 2006 04 21 by rkern, assigned to unknown.
|
|
From Travis Oliphant s email We are already special casing the integer select code but could special case the getitem code so that if nd==1 a faster construction is used.
|
|
I think right now a 0 dim array is being created only to get destroyed later on return.
|
|
Please add a ticket as this extremely common operation should be made as fast as possible.
|
|
This is a little tricky because array big item is called in a few places and is expected to return an array.
|
|
If it returns a scalar in those places segfaults can occur.
|
|
Either checks need to be made in each of those cases or the special casing needs to be in array big item nice.
|
|
I m not sure which I prefer....","bug"
|
|
"6013","5591","BUG numpy.ma.polyfit masks NaNs incorrectly.
|
|
The current implementation of [``numpy.ma.polyfit``] https github.com numpy numpy blob master numpy ma extras.py L1898) fails with masked arrays that mask NaNs.
|
|
For example this import numpy as np x = np.array [1, 2, np.nan, 4]) y = np.array [1, 2, 3, 4]) x masked = np.ma.masked invalid x) y masked = np.ma.masked invalid y) print np.ma.polyfit x masked, y masked, 1) yields the following error message when run using numpy 1.9.0 and the latest development version with python 2.7.9 on mac OSX 10.10 On entry to DLASCL, parameter number 4 had an illegal value On entry to DLASCL, parameter number 4 had an illegal value On entry to DLASCL, parameter number 4 had an illegal value On entry to DLASCL, parameter number 4 had an illegal value On entry to DLASCL, parameter number 5 had an illegal value On entry to DLASCL, parameter number 4 had an illegal value [ nan nan] I think this is due to the way ``numpy.ma.polyfit`` is implemented.
|
|
Currently it sets the weight of any point that is masked in the fit to zero which is fine for any finite value of the input arrays, however if one of the masked values is NaN then it will still propagate through to the result.","bug"
|
|
"6013","5593","numpy.interp always uses a binary search of entire array may be slower than necessary.. `numpy.interp x, xp, fp)` uses a binary search https github.com numpy numpy blob master numpy core src multiarray compiled base.c L495) of the entire input array to find the proper index for the linear interpolation.
|
|
For nearly monotonic target array ``x``) with spacing finer or similar to the original data, storing the last index and looking in the near vicinity of it say 1) can find the correct index much faster for many cases.
|
|
For the very common cases of a) interpolating data onto a finer grid, or b) making a small perturbation of the ``xp`` say, a constant offset or a slight expansion or rotation), storing the previous value and looking 1 can avoid binary search at almost all values.
|
|
Though binary search is efficient, a search through the entire array can be a serious bottleneck for large arrays.
|
|
Thus, storing and the index found for element N and using it as the starting search value for element N 1 may offer a significant boost in performance.","bug"
|
|
"6013","560","Keyword argument support for vectorize.
|
|
Trac 2100).
|
|
Original ticket http projects.scipy.org numpy ticket 2100 on 2012 04 06 by mforbes, assigned to unknown.
|
|
In a few place, people have requested being able to use keyword arguments with vectorize.
|
|
For example http thread.gmane.org gmane.comp.python.scientific.user 29362 focus=29371 Here is a simple patch that defines a decorator kwvectorize simply wrapping vectorize to provide simple support for kwargs.
|
|
It simply looks at the order of arguments defined in the original function, and then reorders any kwargs provided appropriately, calling vectorize with only positional arguments.","enhancement"
|
|
"6013","5600","ufunc out parameter used as out=None throws TypeError Python 2.7.9, numpy 1.9.1).
|
|
The out parameter for ufuncs cannot be used as a keyword argument with value None which is cumbersome for many applications.
|
|
For some ndarray a , the command np.add a,1,out=None) raises a TypeError return arrays must be of ArrayType.
|
|
The ufunc docstring suggests the use of out as a keyword arguments as op X, out=None) .
|
|
np.add a,1,a) works and returns a np.add a,1,None) works and returns a new object np.add a,1,out=a) works and returns a np.add a,1,out=None) DOES NOT WORK","bug"
|
|
"6013","5601","BUG npyio np.savez is limited by tmp size.
|
|
When creating the npz file, each array is written to a temporary file before adding to a zip archive https github.com numpy numpy blob master numpy lib npyio.py L601 This means that if tmp is close to full, npz file creation will fail even when the actual target device has plenty of space.
|
|
Even worse, it fails with a cryptic IOError message that doesn t point to tmp.","bug"
|
|
"6013","5602","read array header 1 0 returns None.
|
|
When trying to read from a .npy file, the np.lib.format read array header 1 0 function returns None instead of the intended triple of shape, fortran, dtype.
|
|
Seems to me that numpy lib format.py line 379 needs a return to get the tuple back from the call.
|
|
Code snippet to reproduce the error Create some junk data import numpy as np filename = mp.npy data = np.random.random 10,100)).astype np.float32) np.save filename, data) The error is caused by with open filename, rb ) as fhandle major, minor = np.lib.format.read magic fhandle) shape, fortran, dtype = np.lib.format.read array header 1 0 fhandle)","bug"
|
|
"6013","5604","may share memory returning false negative?!.
|
|
when setting write flag to false and using ` array ` >>> class Foo ... def init self, arr, write=True) ... self.arr = arr ... self.arr.setflags write=write) ... def array self) ... return self.arr ... >>> a = Foo np.arange 10)) >>> np.may share memory a, a) True >>> a = Foo np.arange 10), False) >>> np.may share memory a.arr, a.arr) True so far so good, but >>> np.may share memory a, a) False is this false negative or am i missing something?
|
|
numpy `1.9.1`)","bug"
|
|
"6013","5606","set printoptions does not work with numpy structured arrays .
|
|
I want to set the print precision for a numpy structured array, but it does not seem to work.
|
|
H = np.zeros 1, dtype=[ x , float ,2), f , float )]) H[ x ]= np.random.uniform 0,1, 1,2)) H[ f ]= np.random.uniform 0,1) np.set printoptions precision=3) print H) leaves [ [0.5928384015048476, 0.6187004067605606], 0.2981617279773586)] not [ [0.593, 0.619], 0.298)] as I expected.
|
|
Printing H[ x ] works because that s not a structured array), but I have many fields in my structured array and I don t want to have to print each one when monitoring the progress of my code.","bug"
|
|
"6013","5608","development environment .rst doc is not advertised?.
|
|
https github.com numpy numpy blob master doc source dev development environment.rst This file has useful information especially about `runtests.py` but I do not see how it is linked with the rest of the numpy documentation.","bug"
|
|
"6013","561","Document changes in casting rules for 1.7 release Trac 2101).
|
|
Original ticket http projects.scipy.org numpy ticket 2101 on 2012 04 10 by rgommers, assigned to pv.
|
|
Casting rules have undergone some changes in corner cases, due to the NA related work.
|
|
These haven t caused any problems so far, but they should be documented for the 1.7 release.
|
|
See http news.gmane.org gmane.comp.python.numeric.general which discussed the above.","bug"
|
|
"6013","5610","Add `isfinite` support for `datetime64` and `timedelta64`.
|
|
`numpy.isfinite` should support `datetime64` and `timedelta64` objects.
|
|
It should return `True` for any ordinary value, and `False` for `datetime64 NaT )` and `timedelta64 NaT )`.
|
|
As a side effect, this will make `isclose` and `allclose` usable for `datetime64` and `timedelta64` objects.
|
|
.","bug"
|
|
"6013","5611","Unexpected result when feeding empty array into atleast 2d.
|
|
I was expecting a square array `array [], shape= 0, 0), dtype=float64)`, but got `shape= 1,0)`.
|
|
Is this the intended behavior?
|
|
In [14] np.empty 0).shape Out[14] 0,) In [15] np.atleast 2d np.empty 0)) Out[15] array [], shape= 1, 0), dtype=float64) The reason is that I needed a 2D array from `hsplit` and `vhsplit`, but the dimension is not preserved when the result happens to be empty gh 4102).
|
|
I tried to work around the issue with `atleast 2d`, and found that it doesn t work because `atleast 2d` doesn t give me a square array with shape ` 0,0)`","bug"
|
|
"6013","5612","Python long raises OverflowError while same value as np.uint64 does not..
|
|
Trying to assign a value of `2 64 1` to an array of dtype `uint64` raises an `OverflowError`, but only if the value is a Python long rather than a Numpy scalar.
|
|
This seems like a bug to me because the value obviously doesn t overflow and they compare equal.
|
|
.","bug"
|
|
"6013","5613","Default fill value dtype does not match that of data for masked arrays.
|
|
The fill value type chosen for integers is always `int64` and always `999999` even when that doesn t make sense [numpy.ma.MaskedArray.fill value] http docs.scipy.org doc numpy reference maskedarray.baseclass.html numpy.ma.MaskedArray.fill value ).
|
|
For instance, take the type `uint8`.
|
|
In this case, the fill value is way outside of bounds and ends up getting converted to `63`.
|
|
If the behavior must be preserved, then for `uint8` the fill value should be `63` and it s dtype should be `uint8`.
|
|
Alternatively, better fill values could be picked that work on a range of types i.e.
|
|
`0` for integers and `nan` for floats and complex) or something similar.","bug"
|
|
"6013","5617","Consider deprecating writeable broadcast arrays.
|
|
It might make sense to deprecate broadcast arrays giving back a writeable view see also gh 5371).
|
|
This should be relatively easy by copying how the diagonal deprecation is done.
|
|
Just tagged for 1.10, but probably no need to do it for that.
|
|
We have the new broadcast to in it which is readonly though.","bug"
|
|
"6013","562","numpy.abs is 5x slower on complex64 than on complex128 under 32 bit Vista Trac 2102).
|
|
Original ticket http projects.scipy.org numpy ticket 2102 on 2012 04 11 by trac user hgomersall, assigned to unknown.
|
|
I experience the following on my Windows Vista 32 bit machine note the significant slowdown when doing abs ) on the complex64 array) In [95] a = numpy.random.randn 256,2048) 1j numpy.random.randn 256,2048) In [96] b = numpy.complex64 a) In [97] timeit numpy.abs a) for comparison 10 loops, best of 3 30 ms per loop In [98] timeit numpy.abs b) slow slow slow) 1 loops, best of 3 153 ms per loop In [99] timeit numpy.sqrt a numpy.conj a)) 10 loops, best of 3 61.9 ms per loop In [100] timeit numpy.sqrt b numpy.conj b)) 10 loops, best of 3 27.2 ms per loop The final 2 lines suggest the problem is in abs ) itself.
|
|
This bug is consistent and reliable.
|
|
It is not present on my 64 bit Ubuntu installation.","bug"
|
|
"6013","5624","wishlist Add an axis argument or two) to np.convolve np.correlate.
|
|
This would be nice to have for a variety of things.
|
|
Could probably be a simply loop, preferably in the C level.","enhancement"
|
|
"6013","5625","Have an `exp` for pure imaginary numbers.
|
|
It would be nice to have a pure imaginary `exp` implemented like `npy cexp`, but only for pure imaginary arguments maybe `npy iexp` or `npy ciexp`).
|
|
It would take a real value ` \theta ` and evaluate ` e {i\cdot\theta} `.
|
|
The hope is this would be a little bit faster in cases where it is know there is no real part to the complex number.","enhancement"
|
|
"6013","5626","numpy.linalg.norm ) broken for multidimensional integer arrays.
|
|
.
|
|
.","bug"
|
|
"6013","5628","gradient ) fuction for bi dimensional array returns first in y direction and then in x.
|
|
Seems the return of gradient ) function, for a two dimensional array, is changed compared to matlab.
|
|
The biggest problem is it s not documented.
|
|
See, in MATLAB .
|
|
Now, in NumPy .
|
|
As we can see, the return of gradient ) at NumPy is not in x direction and later in y.
|
|
It is in y direction and then x.
|
|
In my opinion that is ok this change.
|
|
However, when you are trying to port a large code from matlab to python, like me, this behavior can be a pain.
|
|
Also, the return is not well explained at docs, since there is only shown the gradient with a one dimensional array.","bug"
|
|
"6013","563","Binary incompatibility warning Trac 2103).
|
|
Original ticket http projects.scipy.org numpy ticket 2103 on 2012 04 12 by fonnesbeck, assigned to unknown.
|
|
I have built numpy from the master branch source today I make new builds every couple of weeks), and am seeing a new warning that I have not seen before In [1] import numpy Library Python 2.7 site packages numpy 1.7.0.dev 0c5f480 20120412 py2.7 macosx 10.7 x86 64.egg numpy random init .py 91 RuntimeWarning numpy.ndarray size changed, may indicate binary incompatibility from mtrand import I am using the same build script as I always do.
|
|
I am on OS X 10.7.3, and am using the following env variables export MACOSX DEPLOYMENT TARGET=10.7 export CFLAGS= arch x86 64 export ARCHFLAGS= arch x86 64 export FFLAGS= static ff2c export LDFLAGS= Wall undefined dynamic lookup bundle arch x86 64 export PYTHONPATH= Library Python 2.7 site packages export CC=gcc 4.2 export CXX=g 4.2 Again, nothing has changed here since last I built the package, which did not have this message last time.","bug"
|
|
"6013","5630","[BUG] np.allclose does not work with np.timedelta64.
|
|
http stackoverflow.com questions 28857937 trying to check data frequency with pandas series of datetime64 objects 28858228?noredirect=1 comment45982435 28858228 It works fine with `datetime64`.
|
|
Currently you have to convert to datetime in order to check allclose","bug"
|
|
"6013","5631","Cryptic SystemError when creating array with weird structured but empty dtype.
|
|
Trying to create an array with a weird structured but empty dtype results in a `SystemError` .
|
|
This should be caught somewhere and a suitable exception should be raised instead of `SystemError`.","bug"
|
|
"6013","5632","When accessing multiple fields of a structured array, existence is not checked and return dtype may be empty.
|
|
When accessing multiple fields of a structured array, numpy fails to verify that all the fields exist.
|
|
Instead, it will only return existing fields.
|
|
If none exist, it return a strange object that has a dtype that is empty.
|
|
Normally, creating an object with an empty dtype is not supposed to be possible `zeros shape= 5,), dtype=[])` results as expected in `TypeError Empty data type`), so this might actually be a dual bug.
|
|
.","bug"
|
|
"6013","5637","f2py C to fortran mapping file files under python 3.3.
|
|
Under python 3.3, f2py fails to read the file `.f2py f2cmap`, for mapping fortran types to C types.
|
|
It gives an exception `RuntimeError dictionary size changed during iteration`.
|
|
This is due to python 3 s `keys )` and `items )` members of dictionaries returning iterators instead of lists.
|
|
In `capi maps.py` there is iteration over these functions, during which the dictionary is modified.
|
|
This should be fixable by wrapping these calls with `list )`, which reproduces the old python 2 behaviour.","bug"
|
|
"6013","5639","How can I install numpy1.92?.
|
|
WIN7 64bit Winpython Python 3.4.3.1 64bit Problems as follows F \numpy 1.9.2>python setup.py install Running from numpy source directory.
|
|
D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutils\dist.p y 260 UserWarning Unknown distribution option test suite warnings.warn msg) non existing path in numpy\\distutils site.cfg non existing path in numpy\\f2py docs non existing path in numpy\\f2py f2py.1 F2PY Version 2 blas opt info blas mkl info libraries mkl,vml,guide not found in [ D \\Program Files\\WinPython 64bit 3.4.
|
|
3.1\\python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3 .1\\python 3.4.3.amd64\\libs ] NOT AVAILABLE openblas info libraries openblas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\ python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\p ython 3.4.3.amd64\\libs ] NOT AVAILABLE atlas 3 10 blas threads info Setting PTATLAS=ATLAS libraries tatlas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\py thon 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\pyt hon 3.4.3.amd64\\libs ] NOT AVAILABLE atlas 3 10 blas info libraries satlas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\py thon 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\pyt hon 3.4.3.amd64\\libs ] NOT AVAILABLE atlas blas threads info Setting PTATLAS=ATLAS libraries ptf77blas,ptcblas,atlas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 6 4bit 3.4.3.1\\python 3.4.3.amd64\\libs ] NOT AVAILABLE atlas blas info libraries f77blas,cblas,atlas not found in [ D \\Program Files\\WinPython 64bi t 3.4.3.1\\python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\python 3.4.3.amd64\\libs ] NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1603 UserWarning Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) blas info libraries blas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\pyth on 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\pytho n 3.4.3.amd64\\libs ] NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1612 UserWarning Blas http www.netlib.org blas ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [blas]) or by setting the BLAS environment variable.
|
|
warnings.warn BlasNotFoundError.
|
|
doc ) blas src info NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1615 UserWarning Blas http www.netlib.org blas ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [blas src]) or by setting the BLAS SRC environment variable.
|
|
warnings.warn BlasSrcNotFoundError.
|
|
doc ) NOT AVAILABLE non existing path in numpy\\lib benchmarks lapack opt info openblas lapack info libraries openblas not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\ python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\p ython 3.4.3.amd64\\libs ] NOT AVAILABLE lapack mkl info mkl info libraries mkl,vml,guide not found in [ D \\Program Files\\WinPython 64bit 3.4.
|
|
3.1\\python 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3 .1\\python 3.4.3.amd64\\libs ] NOT AVAILABLE NOT AVAILABLE atlas 3 10 threads info Setting PTATLAS=ATLAS libraries tatlas,tatlas not found in D \Program Files\WinPython 64bit 3.4.3.1\ python 3.4.3.amd64\lib libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\lib libraries tatlas,tatlas not found in C \ libraries lapack atlas not found in C \ libraries tatlas,tatlas not found in D \Program Files\WinPython 64bit 3.4.3.1\ python 3.4.3.amd64\libs libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\libs <class numpy.distutils.system info.atlas 3 10 threads info > NOT AVAILABLE atlas 3 10 info libraries satlas,satlas not found in D \Program Files\WinPython 64bit 3.4.3.1\ python 3.4.3.amd64\lib libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\lib libraries satlas,satlas not found in C \ libraries lapack atlas not found in C \ libraries satlas,satlas not found in D \Program Files\WinPython 64bit 3.4.3.1\ python 3.4.3.amd64\libs libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\libs <class numpy.distutils.system info.atlas 3 10 info > NOT AVAILABLE atlas threads info Setting PTATLAS=ATLAS libraries ptf77blas,ptcblas,atlas not found in D \Program Files\WinPython 64bi t 3.4.3.1\python 3.4.3.amd64\lib libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\lib libraries ptf77blas,ptcblas,atlas not found in C \ libraries lapack atlas not found in C \ libraries ptf77blas,ptcblas,atlas not found in D \Program Files\WinPython 64bi t 3.4.3.1\python 3.4.3.amd64\libs libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\libs <class numpy.distutils.system info.atlas threads info > NOT AVAILABLE atlas info libraries f77blas,cblas,atlas not found in D \Program Files\WinPython 64bit 3.
|
|
4.3.1\python 3.4.3.amd64\lib libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\lib libraries f77blas,cblas,atlas not found in C \ libraries lapack atlas not found in C \ libraries f77blas,cblas,atlas not found in D \Program Files\WinPython 64bit 3.
|
|
4.3.1\python 3.4.3.amd64\libs libraries lapack atlas not found in D \Program Files\WinPython 64bit 3.4.3.1\p ython 3.4.3.amd64\libs <class numpy.distutils.system info.atlas info > NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1505 UserWarning Atlas http math atlas.sourceforge.net ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [atlas]) or by setting the ATLAS environment variable.
|
|
warnings.warn AtlasNotFoundError.
|
|
doc ) lapack info libraries lapack not found in [ D \\Program Files\\WinPython 64bit 3.4.3.1\\py thon 3.4.3.amd64\\lib , C \\ , D \\Program Files\\WinPython 64bit 3.4.3.1\\pyt hon 3.4.3.amd64\\libs ] NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1516 UserWarning Lapack http www.netlib.org lapack ) libraries not found.
|
|
Directories to search for the libraries can be specified in the numpy distutils site.cfg file section [lapack]) or by setting the LAPACK environment variable.
|
|
warnings.warn LapackNotFoundError.
|
|
doc ) lapack src info NOT AVAILABLE F \numpy 1.9.2\numpy\distutils\system info.py 1519 UserWarning Lapack http www.netlib.org lapack ) sources not found.
|
|
Directories to search for the sources can be specified in the numpy distutils site.cfg file section [lapack src]) or by setting the LAPACK SRC environment variable.
|
|
warnings.warn LapackSrcNotFoundError.
|
|
doc ) NOT AVAILABLE D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutils\dist.p y 260 UserWarning Unknown distribution option define macros warnings.warn msg) D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutils\dist.p y 260 UserWarning Unknown distribution option test suite warnings.warn msg) running install running build running config cc unifing config cc, config, build clib, build ext, build commands compiler opti ons running config fc unifing config fc, config, build clib, build ext, build commands fcompiler opt ions running build src build src building py modules sources creating build creating build\src.win amd64 3.4 creating build\src.win amd64 3.4\numpy creating build\src.win amd64 3.4\numpy\distutils building library npymath sources Looking for python34.dll customize GnuFCompiler Could not locate executable g77 Could not locate executable f77 customize IntelVisualFCompiler Could not locate executable ifort Could not locate executable ifl customize AbsoftFCompiler Could not locate executable f90 customize CompaqVisualFCompiler Could not locate executable DF customize IntelItaniumVisualFCompiler Could not locate executable efl customize Gnu95FCompiler Found executable D \Program Files\WinPython 64bit 3.4.3.1\tools\mingw32\bin\gfor tran.exe Found executable D \Program Files\WinPython 64bit 3.4.3.1\tools\mingw32\bin\gfor tran.exe Traceback most recent call last) File setup.py , line 251, in <module> setup package ) File setup.py , line 243, in setup package setup metadata) File F \numpy 1.9.2\numpy\distutils\core.py , line 169, in setup return old setup new attr) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\core.py , line 148, in setup dist.run commands ) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\dist.py , line 955, in run commands self.run command cmd) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\dist.py , line 974, in run command cmd obj.run ) File F \numpy 1.9.2\numpy\distutils\command\install.py , line 60, in run r = old install.run self) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\command\install.py , line 539, in run self.run command build ) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\cmd.py , line 313, in run command self.distribution.run command command) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\dist.py , line 974, in run command cmd obj.run ) File F \numpy 1.9.2\numpy\distutils\command\build.py , line 39, in run old build.run self) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\command\build.py , line 126, in run self.run command cmd name) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\cmd.py , line 313, in run command self.distribution.run command command) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\dist.py , line 974, in run command cmd obj.run ) File F \numpy 1.9.2\numpy\distutils\command\build src.py , line 153, in run self.build sources ) File F \numpy 1.9.2\numpy\distutils\command\build src.py , line 164, in build sources self.build library sources libname info) File F \numpy 1.9.2\numpy\distutils\command\build src.py , line 299, in build library sources sources = self.generate sources sources, lib name, build info)) File F \numpy 1.9.2\numpy\distutils\command\build src.py , line 386, in gener ate sources source = func extension, build dir) File numpy\core\setup.py , line 684, in get mathlib info st = config cmd.try link int main void) { return 0;} ) File D \Program Files\WinPython 64bit 3.4.3.1\python 3.4.3.amd64\lib\distutil s\command\config.py , line 243, in try link self.
|
|
check compiler ) File F \numpy 1.9.2\numpy\distutils\command\config.py , line 87, in check co mpiler c compiler=self.compiler) File F \numpy 1.9.2\numpy\distutils\fcompiler\ init .py , line 842, in new fcompiler c compiler=c compiler) File F \numpy 1.9.2\numpy\distutils\fcompiler\ init .py , line 816, in get default fcompiler c compiler=c compiler) File F \numpy 1.9.2\numpy\distutils\fcompiler\ init .py , line 765, in fin d existing fcompiler c.customize dist) File F \numpy 1.9.2\numpy\distutils\fcompiler\ init .py , line 540, in cust omize self.set libraries self.get libraries )) File F \numpy 1.9.2\numpy\distutils\fcompiler\gnu.py , line 337, in get libra ries raise NotImplementedError Only MS compiler supported with gfortran on win64 ) NotImplementedError Only MS compiler supported with gfortran on win64","bug"
|
|
"6013","564","Link option missing ?
|
|
Trac 2104).
|
|
Original ticket http projects.scipy.org numpy ticket 2104 on 2012 04 12 by trac user medernac, assigned to pearu.
|
|
Hello, I tried to use f2py like this f2py c m dpolft dpolft.f and obtain this error undefined reference to `MAIN It appears that some arguments are missing when linking the dpolft.so shared object.
|
|
I have modified this file numpy distutils fcompiler gnu.py with linker so [None, g , Wall ], linker so [None, g , shared fPIC Wall ], and now it compiles.
|
|
Hope this helps.
|
|
Best regards, Emmanuel Medernach","bug"
|
|
"6013","5643","Loading object arrays from npz files fails on Python 2.6.
|
|
At 1.10.0.dev0 7ad33a2 on Python 2.6 Pythons >= 2.7 all work) .
|
|
git bisect indicates cfae0143b4 Not sure if this is worth fixing, given that it only affects Python 2.6","bug"
|
|
"6013","5645","np.dtype d,L ) > NameError name L is not defined.
|
|
the title says it all.
|
|
additional info python 2.7.9, numpy 1.9.2, both from conda.
|
|
other combinations not using L seems to work fine.","bug"
|
|
"6013","5647","test collections hashable fails on windows python 3.4. test introduced in gh 5326 isinstance np.array []), collections.Hashable) is true on windows python 3.4.3 built with msvc though what is weird that both hash x) and x. hash ) raise an exception as expected, no clue whats going on.
|
|
cournape any idea?","bug"
|
|
"6013","5648","Numpy array record field assignment is broken in numpy 1.9. .
|
|
But this works .","bug"
|
|
"6013","5649","numpy.dual.cholesky behaves differently than numpy.linalg.cholesky.
|
|
The documentation for numpy.dual states Aliases for functions which may be accelerated by Scipy.
|
|
One of the functions supported is cholesky.
|
|
One would assume I did) that the results will be the same regardless of whether scipy is installed or not; only the execution time might vary.
|
|
However, the implementation in numpy and scipy for cholesky are different.
|
|
The numpy version returns a lower triangular matrix, and the scipy version returns an upper triangular matrix.
|
|
Hence you get different behavior depending on whether you have scipy installed or not.
|
|
Below is a script that duplicates this behavior without forcing you to uninstall scipy to test).
|
|
import numpy as np import numpy.linalg as la import numpy.dual as dual P = np.cov np.diag [1., 2., 3.]))
|
|
c1 = la.cholesky P) c2 = dual.cholesky P) assert c1[1,2] == c2[2,1] assert c1[1,2] != c2[1,2] Suggested fix I dunno.
|
|
Some choices that come to mind Add a lower parameter to numpy, defaulted to true, and add a non optional parameter lower to dual.cholesky.
|
|
This will break all code using dual.cholesky, but I am not a fan of silent failures.
|
|
Remove cholesky from dual scipy s cholesky has an optional lower param that can be set to true.
|
|
If dual.py could be modified to set that to be true than numpy.linalg.cholesky and dual.cholesky will behave the same, which is presumably the behavior that the programmer wants.
|
|
I say presumably because the programmer may be used to using the scipy version and expecting the upper triangular behavior; they will be surprised by this choice.
|
|
Add a warning in the documentation that the behavior is inconsistent.
|
|
I dislike this because I found out about this problem the hard way my output looked wrong.
|
|
Sure, documentation would have helped me find it faster, but either consistent behavior or an exception would have been preferred by me.
|
|
Make cholesky in both scipy and numpy return upper triangular.
|
|
This is a terrible option IMO because it would break so much code.
|
|
The top two are the most painful to the users of the library, but they have the advantage of forcing them to think about and deal with the problem.
|
|
The others allow silent failures.","bug"
|
|
"6013","565","fix function name collisions in fftpack with Sun Studio sunperf library Trac 2105).
|
|
Original ticket http projects.scipy.org numpy ticket 2105 on 2012 04 13 by trac user rroliver, assigned to unknown.
|
|
When using Sun Studio Solaris Studio 12.2) compilers and utilising the sunperf libraries for lapack and blas you end up with collisions with the naming of the fft functions used in fftpack.
|
|
Sunperf provides cfft as single precision float, dcfft as double precision float and rfft as single precision float.
|
|
Here in NumPy s fftpack cfft and rfft are double precision float.
|
|
Patch attached basically renames functions to np d d prefix as these are double precision) to avoid these collisions.
|
|
Thoughts?","bug"
|
|
"6013","5650","numpy.linalg.cholesky performance.
|
|
Both scipy.linalg.cholesky and numpy.linalg.cholesky call DPOTRF.
|
|
However, the timings are .
|
|
Perhaps things can be improved in Numpy?","enhancement"
|
|
"6013","5652","logsumexp with sign indicator enable calculation with negative signs.
|
|
So my problem has been the following I had to calculate the logsumexp of some numbers in logspace for accuracy reasons, but some of the elements of the array have negative sign.
|
|
I implemented the following code, which you might want to integrate into the upstream logsumexp implementation of numpy.
|
|
The `sign` argument is a matrix assumed to be filled with 1 or 1 depending on the sign of the number before the log transformation.
|
|
.
|
|
I m not sure I passed on the `b` parameter correctly I never made use of it before).
|
|
Here is a test method .","bug"
|
|
"6013","5653","Creating a complex array crashes Python.
|
|
I am running Python 2.7.9 on Windows 7 SP1 64 bit with numpy 1.9.2.
|
|
Running the following code abruptly crashes the Python interpreter for me, with the typical python.exe has stopped working popup from Windows and no other error information import numpy as np x = np.array 1, dtype= F ) This is preventing me from importing scipy.stats because the decomp.py module crashes when executing line 29 .
|
|
For context, I began observing this after updating from numpy 1.7.1 to 1.9.2.","bug"
|
|
"6013","5654","allclose returns True when comparing an empty array to a scalar.
|
|
In Numpy v 1.9.1 .
|
|
returns `True`.
|
|
The value of the scalar doesn t matter.
|
|
It also returns `True` when comparing an empty array to an array with only one value, e.g.
|
|
.
|
|
also returns `True`.
|
|
It should return `False` unless comparing two empty arrays if you ask me.","bug"
|
|
"6013","5655","loadtxt fails with complex data under python 3.
|
|
If the file `complex data.txt` contains the following .
|
|
Then the code .
|
|
gives the exception `TypeError complex ) argument must be a string or a number, not bytes `.
|
|
It works perfectly fine under python 2.
|
|
This appears to be caused by `loadtxt` failing to properly decode the text file.
|
|
May be related to 2715, 4208, 4543, 5530.
|
|
These issues give me the impression that `loadtxt` is essentially broken under python 3, and needs a radical overhaul.
|
|
Are there any suggested workarounds?","bug"
|
|
"6013","5656","Fix alignment issue for 1.10.
|
|
There was quick fix for 1.9.2 that was a backport of 5316, but the proper fix for 1.10 is still open.","bug"
|
|
"6013","5657","Silent overflow in absolute ).
|
|
Python .
|
|
NumPy .
|
|
[In C this is undefined] http stackoverflow.com q 11243014 125507) because abs ) returns an int and there s no 2147483648, but NumPy should convert to int64?
|
|
and then `absolute array 2 63))` should convert to object array with longs instead of returning 9223372036854775808?
|
|
Related to https github.com numpy numpy issues 593?","bug"
|
|
"6013","566","einsum crashes with repeated zero dimension arrays Trac 2106).
|
|
Original ticket http projects.scipy.org numpy ticket 2106 on 2012 04 16 by trac user rcl33, assigned to unknown.
|
|
einsum crashes when summing over the same array more than once, when the array has zero dimension a = zeros 3,3,0,0)) b = zeros 0,)) einsum ijpr,p,r , a, b, b) hangs then Python crashes On the other hand these both work einsum ijpr,p , a, b) c = zeros 3,3,1,1)) d = zeros 1,)) einsum ijpr,p,r , c, d, d)","bug"
|
|
"6013","5660","lexsort does not sort lexicographically in both colums.
|
|
Hi!
|
|
I have the following issue Given a list of real numbers, I tried to sort it with lexsort, but the result is not lexicographically ordered, at least not in my understanding of lexicographic the second column is not in order) liste2[lexsort liste2[ ,1],liste2[ ,0]))] [[ 8.27989043 0.19634954] [ 8.27989043 2.94524311] [ 8.27989043 2.94524311] [ 8.27989043 0.9817477 ] [ 8.27989043 1.76714587] [ 8.27989043 0.9817477 ] [ 8.27989043 1.76714587] [ 8.27989043 2.55254403] .... maybe I have some misconception, but shouldn t the order be something like that [ [ 8.27989043 2.94524311] [ 8.27989043 1.76714587] [ 8.27989043 0.9817477 ] [ 8.27989043 0.19634954] [ 8.27989043 0.9817477 ] [ 8.27989043 1.76714587] [ 8.27989043 2.55254403] [ 8.27989043 2.94524311] ,,,, So only the first column is ordered the second one is ignored.
|
|
But why?
|
|
Here the code .
|
|
Python Version Python 2.7.5 Numpy version 1.8.2 OS Fedora 20 Processor Intel Xeon","bug"
|
|
"6013","5662","BUG some ufunc methods do not yet take tuple for output.
|
|
5621 introduced the nice option to allow tuples for output arrays, including a tuple holding a single array for a single output.
|
|
However, it missed also changing the outputs for some of the ufunc methods such as reduce .
|
|
cc jaimefrio","bug"
|
|
"6013","5663","tensordot handles contraction over 0 length dimensions differently than dot.
|
|
.
|
|
If tensordot is asked to contract over an axis that has length 0 it raises a ValueError.
|
|
In contrast, dot and einsum give an array of zeros.
|
|
I could see both behaviors as reasonable choices, but I feel that tensordot should be a generalization of dot and the behavior of the two should be consistent.
|
|
This as of numpy 1.9.2.","bug"
|
|
"6013","5664","timedelta64 doesn t accept int64 as input.
|
|
Seems to me that timedelta should be able to accept numpy integers as input, but it can t .
|
|
This is numpy installed via conda on kubuntu 14.04 Not sure if this is related to https github.com numpy numpy issues 2543","bug"
|
|
"6013","5665","BUG segmentation fault for np.modf.at np.arange 10.
|
|
), 1,)).
|
|
Found this while trying to thing of ufunc methods for ufuncs with two outputs see 5662).
|
|
While not something that I would expect to work, the following also shouldn t lead to a segmentation fault .","bug"
|
|
"6013","5667","einsum double diagonal fails when combined with contraction.
|
|
`einsum` has the ability to contract indices, but also to extract diagonals between indices.
|
|
A combination of the two should also be possible.
|
|
E.g.
|
|
if we have an array of shape 2,3,3) we can contract the first index, and take the diagonal between the last two as follows .
|
|
However, it fails if we take more than one diagonal.
|
|
E.g.
|
|
.
|
|
To me, this seems like a bug because the same thing works if we just change the order of the axes .
|
|
Tested on NumPy version 1.8.2 with Python version 2.7.9 on openSuse 13.1 x64 86.","bug"
|
|
"6013","5668","shift operator does not work with numpy arrays with dtype=uint64.
|
|
I have tested the following python code on 64 bit Windows 7 import sys print python version is \n , sys.version import numpy as np print numpy version is \n , np.version.version a = np.array [1,3], dtype = np.int64) print a[1] is , a[1] print type of a[1] is , type a[1]) print a[1] << 1 is , a[1] << 1 works fine for dtype = int64 u = np.array [1,3], dtype = np.uint64) print u[1] is ,u[1] print type of u[1] is , type u[1]) print u[1] << 1 is , u[1] << 1 but for dtype = uint64 there is a problem This script produces the following output python version is 2.7.6 default, Nov 10 2013, 19 24 18) [MSC v.1500 32 bit Intel)] numpy version is 1.8.1 a[1] is 3 type of a[1] is <type numpy.int64 > a[1] << 1 is 6 u[1] is 3 type of u[1] is <type numpy.uint64 > u[1] << 1 is Traceback most recent call last) File test uint64 array.py , line 12, in <module> print u[1] << 1 is , u[1] << 1 TypeError ufunc left shift not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule safe I can do my job with int64 instead of uint64.
|
|
I just want to report this problem.","bug"
|
|
"6013","5669","Call array finalize when creating a view of a different dtype.
|
|
When calling `array.view SomeArraySubclass)`, a new instance of `SomeArraySubclass` is created, and its ` array finalize ` is called per the usual array subclassing rules.
|
|
However, when calling `array.view new dtype)`, where `new dtype` is just a different dtype for which we would like to view the array, ` array finalize ` is still called.
|
|
However, the argument passed to it is the original array with the original dtype.
|
|
There is no way to find out what we are changing the dtype to in the view casting.
|
|
This is because `PyArray View` calls `PyArray NewFromDescr` which handles all the ` array finalize ` stuff where applicable, then afterwards just directly updates the `dtype` attribute.
|
|
It would be nice if instead the new dtype were elevated a bit in importance here so that it is possible to discover through ` array finalize `, though I m not sure yet how best to handle that.
|
|
Apologies of this has come up before, but I didn t see any issues in GH concerning this.)
|
|
I have an ndarray subclass wherein some of its initialization depends on its dtype, so viewing it to a different dtype results in incorrect initialization.
|
|
A workaround that seems okay is to make `dtype` a property, like .
|
|
but this is still pretty hacky, and also won t suffice if, say, I don t want the dtype attribute to be overridable.
|
|
I guess another workaround I could try would be to just override the `view` method, but I still feel like that shouldn t be necessary.
|
|
I d be happy to come up with a fix for this in Numpy if it seems like an acceptable change.","bug"
|
|
"6013","567","numpy False is not python False Trac 2107).
|
|
Original ticket http projects.scipy.org numpy ticket 2107 on 2012 04 16 by trac user alexa, assigned to unknown.
|
|
i doubt this is a bug, but i found the following behavior unexpected.
|
|
>np.False is False returns `False` >np.False == False returns `True`","bug"
|
|
"6013","5671","Function for converting datetime.time to np.timedelta64.
|
|
like `np.timedelta64`s, `datetime.time`s are a relative measure of time time since midnight).
|
|
However, there is no easy method to convert from and array of `datetime.time`s to `np.timetelta`s.
|
|
About the simplest way is .
|
|
But this doesn t work with arrays of datetimes.
|
|
The other option is to `datetime.combine` the `datetime.time`s with an arbitrary `datetime.date`, and then subtract a `datetime.datetime` with the same date at midnight from them, to get a `datetime.timedelta`, and convert that.
|
|
Convoluted as hell, and it also doesn t work with arrays.
|
|
It d be nice to just ignore `datetime.time` altogether to me it seems like a stupid datatype, when `datetime.timedelta` already exists, and has no practical difference.
|
|
But there are cases when it can t be avoided.
|
|
For example, pandas converts all date time indexes into Timeindexes, which are based on datetimes.
|
|
And when you do something like `df.groupby df.index.time).mean )` to get an average daily cycle), the resultant dataframe is indexed by `datetime.time` objects.
|
|
This is a pretty common precedure, for me at least, and it d be really good to be able to convert those times back in to something useful, like `timedelta64`s.","enhancement"
|
|
"6013","5672","Segmentation fault when creating a nargs>32 ufunc using frompyfunc in numpy 1.9.2.
|
|
Title says most of it, here s a minimal example faulter.py reads .
|
|
In 1.8.2 this is handled nicely, as when I run this with my usr bin python3 I just get a ValueError .
|
|
However, I have a virtualenv with the numpy 1.9.2. and there this happens .
|
|
If I run this in an interactive gdb session using .
|
|
I get .
|
|
I m running Ubuntu 14.10 and python 3.4.2.
|
|
The system wide installation that has numpy 1.8.2. came from the Ubuntu binary repositories where as the virtualenv one with numpy 1.9.2. was installed with pip.
|
|
Let me know if more details are needed helpful.","bug"
|
|
"6013","5674","Ufunc override should normalize the sig keyword argument.
|
|
Opening this as a reminder of unfinished work after merging 5659.
|
|
The ufunc override mechanism should make sure that, if `sig` is part of the keyword arguments, it gets normalized to `signature`.
|
|
It may also be a good idea to make `signature` the keyword argument announced in the docs, with a small footnote saying that `sig` is also acceptable for backwards compatibility reasons, rather than having them advertised as equally valid alternatives.","enhancement"
|
|
"6013","5678","argsort ..., axis=None) doesn t use flattened array for 0d inputs..
|
|
There seems to be an inconsistency in `numpy.argsort` with respect to zero dimensional arrays when `axis=None`.
|
|
The docs for the `axis` parameter say > If None, the flattened array is used.
|
|
But for a 0d array, this doesn t seem to be true I get a scalar out where I expected a 1d array .
|
|
I d expect both `Out[30]` and `Out[31]` above to give me `array [0])`.
|
|
For 2d arrays, the results are the same with and without explicit flattening, as expected .","bug"
|
|
"6013","5679","`numpy.where` and `numpy.ma.where` are inconsistent when using `None` as a scalar.
|
|
Two pull requests have been offered https github.com numpy numpy pull 5582 ) https github.com numpy numpy pull 5583 ) to address the inconsistency.
|
|
Below are four examples.
|
|
>>> import numpy >>> numpy.where True) ... array [0]),) >>> numpy.where True, None, None) ... array None, dtype=object) >>> numpy.ma.where True) ... array [0]),) >>> numpy.ma.where True, None, None) ... array [0]),) It it worth noting that it doesn t matter if the first argument the conditional) is an array or scalar.
|
|
The inconsistency regards the use of `None` as a scalar, which is currently permitted by `numpy.where`, but not by `numpy.ma.where`.","bug"
|
|
"6013","568","Datetime failures with MinGW Trac 2108).
|
|
Original ticket http projects.scipy.org numpy ticket 2108 on 2012 04 16 by rgommers, assigned to unknown.
|
|
See https github.com numpy numpy pull 214 These failures are hard to fix for MinGW 3.4.5.
|
|
Until we can do everything with MinGW 4.x including distributing binaries) we do have a problem.","bug"
|
|
"6013","5680","On Py3, import numpy adds an mtrand package to sys.modules.
|
|
Noticed this while randomly reading through the source code to `mtrand.c` as generated from `mtrand.pyx`) .
|
|
The problem is the following code inside the Cython generated module init function .
|
|
which is just... if this is python 3, then do `sys.modules[ mtrand ] = this module` .
|
|
I don t know why Cython does this.
|
|
If you look at the above REPL trace, it s clear that the `mtrand` module does end up with a properly qualified ` name `, and does end up in the proper fully qualified place in `sys.modules` as well, so... maybe this is just a bug in Cython, and the C code I quoted above should just not be generated?","bug"
|
|
"6013","5681","numpy.mean is extremely inaccurate for large float32 arrays.
|
|
I know it says in the docs it can be inaccurate, but I had no idea it was this bad.
|
|
Maybe this should be the example given?
|
|
In [37] np.mean np.random.binomial 1,.5, 100000,5000)).astype float32 )) Out[37] 0.033554431","bug"
|
|
"6013","5684","array2string documentation is missing from docs.scipy.org.
|
|
There is a docstring for [array2string] https github.com numpy numpy blob master numpy core arrayprint.py L344) but it is not online at docs.scipy.org, as far as I can tell from google and the [index] http docs.scipy.org doc numpy genindex.html).
|
|
There are references to it online, for example [here] http docs.scipy.org doc numpy reference generated numpy.array str.html numpy.array str), but `array2string` shows up without a link!","bug"
|
|
"6013","5686","Use of NameValidator in np.genfromtxt is inconsistent with the rules for naming structured array fields.
|
|
This issue has cropped up a [couple][1] of [times][2] before on StackOverflow.
|
|
`np.genfromtxt` uses a `numpy.lib.
|
|
iotools.NameValidator` which automatically modifies field names by replacing spaces and stripping out certain non alphanumeric characters etc.
|
|
.
|
|
However, these are all perfectly legal field names for structured arrays .
|
|
This behavior is not very well covered in the documentation for `np.genfromtxt` at best it s indirectly alluded to in the descriptions for some of the keyword arguments `deletechars`, `autostrip`, `replace space` etc.).
|
|
What is perhaps most confusing is the fact that `np.genfromtxt` will even mangle field names that you pass directly via the `names=` kwarg.
|
|
The rationale behind mangling field names in this way is also not clear to me.
|
|
It should either be the case that spaces and non alphanumeric characters are disallowed in field names altogether and my second example should raise an exception) , or `np.genfromtxt` should leave field names alone by default.
|
|
[1] http stackoverflow.com q 29097917 1461210 [2] http stackoverflow.com q 16020137 1461210","bug"
|
|
"6013","5687","tanh has problems with large complex numbers.
|
|
Hi, when working with complex impedance data, i encountered an issue and narrowed it down to >>> from numpy import >>> x= 616.47292227535877 53.814558958179042j) >>> tanh x) Warning from warnings module) File C \python\Test\Impedance class reduced.py , line 1 from numpy import RuntimeWarning overflow encountered in tanh Warning from warnings module) File C \python\Test\Impedance class reduced.py , line 1 from numpy import RuntimeWarning invalid value encountered in tanh When actually, it should return 1 0j), see http tinyurl.com nm4goq3 Then i tried >>> sinh x) cosh x) Warning from warnings module) File C \python\Test\Impedance class reduced.py , line 1 from numpy import RuntimeWarning overflow encountered in cdouble scalars Warning from warnings module) File C \python\Test\Impedance class reduced.py , line 1 from numpy import RuntimeWarning invalid value encountered in cdouble scalars nan nan j) However, the solution was to use numpy.divide >>> divide sinh x),cosh x)) 1 0j) This took me quite some time to figure out, so i guess it would be a good idea to include it in the next release maybe even with with coth?).
|
|
Thanks a lot!","bug"
|
|
"6013","5688","I installed numpy 1.10.0.dev0 c60f1c6, the latest in the Git Hub and found the following error when running numpy.test full ).
|
|
I installed numpy 1.10.0.dev0 c60f1c6, the latest in the Git Hub and found the following error when running numpy.test full ).
|
|
Could some please let me know if this is a serious bug that can cause problem later?
|
|
====================================================================== FAIL test from object array unicode test defchararray.TestBasic) Traceback most recent call last) File usr local lib python2.7 dist packages numpy core tests test defchararray.py , line 26, in test from object array unicode self.assertRaises ValueError, np.char.array, A,)) AssertionError ValueError not raised ====================================================================== FAIL test from unicode array test defchararray.TestBasic) Traceback most recent call last) File usr local lib python2.7 dist packages numpy core tests test defchararray.py , line 63, in test from unicode array self.assertRaises UnicodeEncodeError, fail) AssertionError UnicodeEncodeError not raised ====================================================================== FAIL Ticket 1240.
|
|
Traceback most recent call last) File usr local lib python2.7 dist packages numpy core tests test regression.py , line 1376, in test unicode to string cast self.assertRaises UnicodeEncodeError, fail) AssertionError UnicodeEncodeError not raised Ran 6059 tests in 54.762s FAILED KNOWNFAIL=3, SKIP=9, failures=3) <nose.result.TextTestResult run=6059 errors=0 failures=3>","bug"
|
|
"6013","569","mean ) with axis=None causes promotion of floating point types Trac 2109).
|
|
Original ticket http projects.scipy.org numpy ticket 2109 on 2012 04 17 by dwf, assigned to unknown.
|
|
It seems that `mean )` with no `axis` argument and no `dtype` incorrectly promotes to float64 when taking the mean of an array of dtype float32.
|
|
Here s a test that demonstrates def test mean promotion regression bug ) x = np.zeros 5, 2, 2), dtype= float32 ) np.testing.assert x.mean axis=0).dtype == x.dtype) np.testing.assert x.mean axis=1).dtype == x.dtype) np.testing.assert x.mean axis=2).dtype == x.dtype) This assertion currently fails.
|
|
np.testing.assert x.mean ).dtype == x.dtype) Also, even specifying `dtype= float32 ` in this circumstance fails def test mean promotion regression bug dtype specified ) x = np.zeros 5, 2, 2), dtype= float32 ) np.testing.assert x.mean axis=0, dtype= float32 ).dtype == x.dtype) np.testing.assert x.mean axis=1, dtype= float32 ).dtype == x.dtype) np.testing.assert x.mean axis=2, dtype= float32 ).dtype == x.dtype) This assertion currently fails.
|
|
np.testing.assert x.mean dtype= float32 ).dtype == x.dtype) In addition to not sharing this behaviour with `sum )`, docstring reads dtype data type, optional Type to use in computing the mean.
|
|
For integer inputs, the default is `float64`; for floating point inputs, it is the same as the input dtype.
|
|
So I believe this is indeed a bug.
|
|
I don t know !NumPy s reduction machinery well enough and don t have time to go digging right now, but am filing it here so it doesn t get lost.
|
|
This behaviour seems to go back as far as 1.3.x, at least, so it doesn t seem to be the result of any of the recent NA work.","bug"
|
|
"6013","5690","including numpy.i in other projects.
|
|
What is the prefered way to include numpy.i in the source distribution of other projects using numpy.i.
|
|
currently there is no license header associated with this file.
|
|
Adding the contents of LICENSE seems to be only way.
|
|
If there is any prefered way let me know","enhancement"
|
|
"6013","5691","nanmax strangely slow 3x slower than nanmin).
|
|
On my system Numpy 1.9.2 from conda, Windows 7 64b), there is almost a 3x difference between nanmin and nanmax while there is barely any difference between np.amin np.amax.
|
|
I do not see any reason for this to be the case.
|
|
I ve tracked the difference to np.fmin.reduce np.fmax.reduce but going further is beyond my abilities... .","bug"
|
|
"6013","5692","Inconsistent and incorrect datetime64 string representations when in structured array.
|
|
datetime64 string representations are inconsistent when in structured arrays.
|
|
When the value should be `NaT`, they are shows as `None` inside a structured array.
|
|
The showing of timezones is also inconsistent, as the code snippet below shows.
|
|
.","bug"
|
|
"6013","5695","Add examples of more complex NpyIter usage to the documentation..
|
|
The iterator is nice in some regards, simple things are simple, and simply iterating over all elements is straightforward.
|
|
However, how to do things like what `PyArray IterAllButAxis` does but with multiple broadcasted arrays, etc is pretty opaque.
|
|
To me at least; maybe I m just dense.)
|
|
My currently go at using nested iterators to try things out is [here] https gist.github.com ewmoore e88ee9dd84c1d9d58892), FWIW.
|
|
I m performing a reduction along a specified tuple of axes, computing both the sum and product stored in the same output array.
|
|
Its a bit of mess, but it does seem to work.
|
|
Doing this kind of thing appears to take quite a lot of setup though.","enhancement"
|
|
"6013","5696","Mean of dtype object array works sometimes.
|
|
Hi.
|
|
We ran into this at scikit learn and it is more of an integration testing issue than anything else.
|
|
It boils down to this expression .
|
|
which either produces ``4.5`` or raises an error.
|
|
I d like to know what causes the error, in which version it is, and how to robustly test for it.
|
|
Maybe I m overlooking something, but testing it on several platforms, what I get is this .
|
|
So even for the same Python and numpy version, the behavior is different between Ubuntu and OS X.
|
|
Any help would be much appreciated.
|
|
cc trevorstephens","bug"
|
|
"6013","5697","Numerical precision inconsistency in np.linalg.norm vector vs. matrix).
|
|
There is a slight numerical inconsistency in `np.linalg.norm` in evaluating the length of vectors vs. the column wise or row wise length of matrices.
|
|
Minimal example to reproduce on 1.9.1 .
|
|
Prints .
|
|
The last few digits of the output do not agree.
|
|
I understand this passes from the `allclose` perspective, but in some applications it is extremely surprising and hard to debug.
|
|
I encountered this when writing an approximate nearest neighbour lookup tree and realising points were not being assigned to the same leaf nodes depending on whether I used the one vector at a time or matrix based approach.
|
|
Numpy 1.8.1 gives the same result in both cases.","bug"
|
|
"6013","5699","invalid value encountered in double scalars.
|
|
Everything goes well, and I can get my answers.
|
|
But there is warning whenever I run the code I was running python 2 .","bug"
|
|
"6013","570","Documentation for reshape case of reshape i, j, k)?
|
|
Trac 2110).
|
|
Original ticket http projects.scipy.org numpy ticket 2110 on 2012 04 18 by lebigot, assigned to unknown.
|
|
The documentation string for `ndarray.reshape )` and `numpy.reshape )` doe not mention the `reshape i, j, k)` syntax.
|
|
On the other hand, this syntax is used in the official documentation http scipy.org Numpy Example List), and it does work.
|
|
It would be nice to have a documentation string and an example list that are consistent with each other.
|
|
PS Should the `reshape i, j, k)` syntax be avoided?","bug"
|
|
"6013","5700","On Python3, round with one argument should return an integer type..
|
|
Unlike Python2, Python3 s builtin) round function is documented to return an integer when called with a single argument.
|
|
However, numpy float dtypes do not satisfy this, returning a float instead.","bug"
|
|
"6013","5702","Occasional wildly incorrect results from numpy.linalg.lstsq.
|
|
Hello, I m new to numpy and not exactly an expert in statistics.
|
|
However, I think I ve run across a serious bug in numpy s linear regression.
|
|
I m running under Linux, Python3.4 on a 64bit processor and OS.
|
|
Please see the script below.
|
|
The graphical output speaks for itself, so I encourage you to run it.
|
|
See also the comments inline.
|
|
I suspect the bug is related to an integer floating type cast overflow or something along along those lines.
|
|
With my data in similar ranges, it seems to occur when the x coordinates are nearer to 2 32 1, though I don t have a lot of examples.
|
|
.","bug"
|
|
"6013","5705","interpolate needs fixing for short arrays of sample points.. See 5594.","bug"
|
|
"6013","571","Trac is driving me crazy ; ) Trac 2111).
|
|
Original ticket http projects.scipy.org numpy ticket 2111 on 2012 04 20 by trac user nicopinto, assigned to unknown.
|
|
Hey guys, Would you be interested in moving away from this slow trac website and try to bug the bug reports in Github Issues?
|
|
I understand that they may not provide the same features but I heard many times from various people that it was a massive PITA to submit track bug reports to numpy or scipy and possibly link them to pull requests using the current interface).
|
|
It may be slowing down progress of these awesome projects.
|
|
What do you think?
|
|
Cheers, N","enhancement"
|
|
"6013","5710","ERR Broadcast error on assignment prints dimensions in wrong order.
|
|
xref 2357 .
|
|
The first error message is confusing at best.","bug"
|
|
"6013","5711","issubdtype double, bool).
|
|
Maybe I am missing something, but is this correct?
|
|
.
|
|
At the very least, this is confusing.","bug"
|
|
"6013","5712","numpy.lib.format.read array header don t return as specified.
|
|
See [here] https github.com numpy numpy blob master numpy lib format.py L392) and [here] https github.com numpy numpy blob master numpy lib format.py L425).
|
|
Fix should be adding `return` to these lines.
|
|
Fix incoming.","bug"
|
|
"6013","5715","record dtype hashing is fragile.
|
|
When pickling and unpickling a record type, hash equality can fail sporadically tested under Python 3.4).
|
|
If you take this script .
|
|
and run it in a loop e.g.
|
|
`while python nprecbug.py; do ; done`), eventually you ll hit a failure .
|
|
As you can see, ` hash ` seems to be sensitive to the ordering of fields in the mappingproxy.
|
|
By contrast, ` eq ` does the right thing.","bug"
|
|
"6013","5716","Possible bug with multiple views and basic slicing .
|
|
We were able to reproduce this in numpy 1.8.1 and 1.9.2.
|
|
When given a slice from the beginning of the array relative to the end) to some other point also relative to the end) with a step size greater than 1, we receive a view which appears to operate as expected.
|
|
However, when this view is viewed as an `|S<n>` dtype where `<n>` is greater than 1 we receive a `ValueError` .
|
|
Changing the slice step to 1 works as expected.
|
|
Performing a copy and then creating a view will also work as expected.","bug"
|
|
"6013","5717","misleading linalg error message for junk input.
|
|
In this toy example I tried to invert a meaningless Python object, but the error message complains about a 0 dimensional array.
|
|
I was able to decipher this error message because I have some experience using numpy, but maybe the message should be changed to indicate that numpy does not recognize the input as a matrix or array type?
|
|
.","bug"
|
|
"6013","5718","Make it possible to initialize array from dictionary values in Python 3.
|
|
At the moment arrays can t be initialized from dictionary values in Python 3 .
|
|
and one has to instead explicitly convert to a list first .
|
|
It might be worth adding a special case in the initializer for the array to automatically do this conversion, since ``d.values )`` is a valid iterator?","bug"
|
|
"6013","5719","segfault while casting rational dtype.
|
|
On numpy master and numpy types master.
|
|
.
|
|
Instead, this should raise an overflow error.
|
|
This could possibly related to this [issue 3359] https github.com numpy numpy issues 3359).","bug"
|
|
"6013","572","provide option for silent install in Windows installers Trac 2112).
|
|
Original ticket http projects.scipy.org numpy ticket 2112 on 2012 04 23 by rgommers, assigned to rgommers.
|
|
NSIS installers support the S command line argument, indicating a silent install.
|
|
This requires some changes to the install script though, right now this doesn t seem to do anything.
|
|
See http nsis.sourceforge.net Docs Chapter4.html 4.12","bug"
|
|
"6013","5721","Inconsistent behavior in the flags attribute of ndarray.
|
|
I have taken a close look at the `flags` attribute of `ndarray`s, and have found several inconsistencies with the C API definitions, one of which is almost certainly a bug.
|
|
To set things up for the discussion, here s a list of the attributes of `arr.flags`, the flags that they check for, and the equivalent C function macro `contiguous` and `c contiguous` checks for `NPY ARRAY C CONTIGUOUS` C equivalent `PyArray ISCONTIGUOUS` `PyArray IS C CONTIGUOUS` `fortran` and `f contiguous` checks for `NPY ARRAY F CONTIGUOUS` C equivalent `PyArray IS F CONTIGUOUS` `updateifcopy` checks for `NPY ARRAY UPDATEIFCOPY` C equivalent ` ` `owndata` checks for `NPY ARRAY OWNDATA` C equivalent ` ` `aligned` checks for `NPY ARRAY ALIGNED` C equivalent `PyArray ISALIGNED` `writeable` checks for `NPY ARRAY WRITEABLE` C equivalent `PyArray ISWRITEABLE` `behaved` checks for `NPY ARRAY BEHAVED` C equivalent `PyArray ISBEHAVED` also checks dtype is NBO) `carray` checks for `NPY ARRAY CARRAY` C equivalent `PyArray ISCARRAY` also checks dtype is NBO) `forc` checks for `NPY ARRAY F CONTIGUOUS || NPY ARRAY C CONTIGUOUS` C equivalent `PyArray ISONESEGMENT` `fnc` checks for `NPY ARRAY F CONTIGUOUS !NPY ARRAY C CONTIGUOUS` C equivalent `PyArray ISFORTRAN` `farray` ` NPY ARRAY ALIGNED || NPY ARRAY WRITEABLE || NPY ARRAY F CONTIGUOUS) !NPY ARRAY C CONTIGUOUS` C equivalent ` ` Based on the above, here s what I would like to change The one that s obviously wrong is `farray`, which rather than checking for the array being aligned and writeable and Fortran contiguous, checks if it is aligned or writeable or Fortran contiguous.
|
|
It also checks that it is not C contiguous, a check that the C macro `PyArray ISFARRAY` does not do.
|
|
I would like to make it behave the same as `PyArray ISFARRAY`.
|
|
`fortran` is inconsistent with the C behavior.
|
|
I believe it should behave as `PyArray ISFORTRAN`, and hence be the same as `fnc`, not the same as `f contiguous`.
|
|
I would like to make that change, but understand we may be stuck with the current behavior on account of backwards compatibility.
|
|
The object returned by `arr.flags` holds a reference to `arr`, as well as a local copy of the array s flags at the moment of creation.
|
|
All of the checks are now carried out on the local copy.
|
|
This is wrong, as discussed [in this thread] http mail.scipy.org pipermail numpy discussion 2015 March 072563.html).
|
|
There are two things I would like to change here Do not use the local copy if the array is available.
|
|
Use the C functions macros directly on the array, rather than equivalent implementations.
|
|
This would create some subtle changes in behavior, e.g.
|
|
`behaved` and `carray` would start checking that the dtype of the array be in native byte order.","bug"
|
|
"6013","5722","Fancy indexing inconsistency with object dtype.
|
|
Here is a test case, where I expect the results to always be the same, but they are not.
|
|
import numpy as np dtype=object print [[0]] = array like ) A = np.zeros 2, dtype=dtype) A[[0]] = [[1,2,3,4]] print A) print [[0]] = ndarray ) B = np.zeros 2, dtype=dtype) B[[0]] = np.array [[1,2,3,4]]) print B) print [[T,F]] = array like ) C = np.zeros 2, dtype=dtype) C[np.array [True, False])] = [[1,2,3,4]] print C) print [[T,F]] = ndarray ) D = np.zeros 2, dtype=dtype) D[np.array [True, False])] = np.array [[1,2,3,4]]) print D) print [0] = array like ) E = np.zeros 2, dtype=dtype) E[0] = [1,2,3,4] print E) To be clear, I expected the result to always be `[ [1,2,3,4], 0]`.
|
|
The results in 1.8 are different, but also wrong.
|
|
A probably related bug when using `dtype=int`, then I would expect all the calls to trigger an error message, but they do not.
|
|
Tested on both 1.9.2 and current master.","bug"
|
|
"6013","5725","some numpy linalg matrix norms do not like negative axes.
|
|
.","bug"
|
|
"6013","5727","the linalg norm axis keyword argument versionadded is not documented.
|
|
According to comments it should be 1.8. http stackoverflow.com questions 20168596 linalg norm not taking axis argument http docs.scipy.org doc numpy reference generated numpy.linalg.norm.html Also in the current master branch the `keepdims` versionadded appears to be documented but possibly not using correct sphinx formatting.
|
|
I think it needs more newlines.","enhancement"
|
|
"6013","5728","Recent updates cause importing numpy failed..
|
|
!
|
|
[screenshot 1] https cloud.githubusercontent.com assets 3604312 6872279 392ced72 d4e3 11e4 8941 f8f113127780.png) python 3.4.3 numpy 1.9.2","bug"
|
|
"6013","5729","Piecewise bug with scalar domains and more than 2 conditions.
|
|
Commented in 4792) There, seems to be an issue with scalar domains and three or more conditions .
|
|
Funny enough it works fine with two conditions this is what has been tested) .
|
|
The three conditions version works fine with select .
|
|
Tested on python 3.4 with numpy 1.9.0 and python 2.7.6 with numpy 1.8.2","bug"
|
|
"6013","573","Python 3 dictionary view objects not supported Trac 2113).
|
|
Original ticket http projects.scipy.org numpy ticket 2113 on 2012 04 24 by trac user khaeru, assigned to unknown.
|
|
This seems to be related to 2612, except there is no segfault; just that code that worked under Python 2.7 no longer works under 3.2. import numpy d = {0 numpy.zeros [5]), 1 numpy.ones [5])} numpy.sum d.values )) On Ubuntu, with the packages python2.7 2.7.3 0ubuntu3 and python numpy 1 1.6.1 6ubuntu1, the result is 5.0 With python3.2 3.2.3 0ubuntu1 and python3 numpy 1 1.6.1 6ubuntu1, the result is dict values [array [ 0., 0., 0., 0., 0.
|
|
]), array [ 1., 1., 1., 1., 1.])])
|
|
NumPy appears not to recognize the new Python 3 [http docs.python.org py3k library stdtypes.html dictionary view objects dictionary view object ].
|
|
In order to get the same result, I must do numpy.sum list d.values )) This gets very cluttered when feeding data from dicts to NumPy functions.","bug"
|
|
"6013","5730","place has different casting rules for datetime64.
|
|
I can put an integer into a datetime array directly `` setitem ``), or with putmask, but it fails with a ValueError if I try with place.
|
|
.","bug"
|
|
"6013","5734","random.multivariate normal doesn t work in 1d case.
|
|
I don t think a mathematician would write a multivariate function that doesn t work in the particular case of n = 1.
|
|
I ve written code that in several places generates correlated normals with a given covariance matrix, but I now need to re write the code to handle the special case in which the incoming data is one dimensional.
|
|
I consider this a bug, because I cannot imagine why anyone would design it like that.","bug"
|
|
"6013","5735","The view method does not work with numpy scalars .
|
|
Compare the following two behaviors .
|
|
I believe `type x.view ma.MaskedArray)` should always be `ma.MaskedArray` regardless of what `x` is.","bug"
|
|
"6013","5736","many boundary issues detected in clang 3.6 with fsanitize=undefined for numpy 1.9.2.
|
|
Building numpy 1.9.2 with the clang 3.6 compilers and the fsanitize=undefined triggered many runtime error messages during the unit tests for numpy.
|
|
A sampling of these detected on x86 64 apple darwin14 are... numpy core src multiarray iterators.c 333 5 runtime error index 1 out of bounds for type npy intp [32] numpy core src multiarray iterators.c 1413 9 runtime error index 1 out of bounds for type npy intp [32] numpy core src multiarray common.h 161 35 runtime error load of misaligned address 0x7fb8ca9519f2 for type unsigned int , which requires 4 byte alignment numpy core src multiarray datetime.c 991 13 runtime error index 26 out of bounds for type NPY DATETIMEUNIT [16][4] numpy core src multiarray datetime.c 992 16 runtime error index 27 out of bounds for type NPY DATETIMEUNIT [16][4] sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value 34440 is outside the range of representable values of type short sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value 130 is outside the range of representable values of type signed char .
|
|
sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value 325 is outside the range of representable values of type unsigned char .numpy core src multiarray arraytypes.c.src 536 5 runtime error load of misaligned address 0x7fb8cac0730c for type PyObject aka struct object ), which requires 8 byte alignment Knumpy core src umath loops.c.src 1490 40 runtime error division by zero sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value inf is outside the range of representable values of type long sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value nan is outside the range of representable values of type long sw src fink.build root numpy py27 1.9.2 3 sw lib python2.7 site packages numpy core multiarray.so 0x4e15ea) runtime error value inf is outside the range of representable values of type long ...for a total of 16696 runtime errors detected by UBSan as described in... https docs.python.org devguide clang.html","bug"
|
|
"6013","5737","Piecewise regression with multidimensional arrays with extra otherwise condition.
|
|
Comments from 4792) The changes applied from this pull request don t seem to work well with multidimensional arrays that have an extra otherwise condition .
|
|
The same code produces the following result in NumPy 1.8.2 .
|
|
If you reshape the array to a single dimension, it works as intended.
|
|
Also, I noticed there aren t any multidimensional test cases, so maybe this can be one of them?
|
|
These conditions were tested with Python 2.7 NumPy 1.8.2 and Python 2.7.9 NumPy 1.9.2.","bug"
|
|
"6013","5738","ENH Add ma=False option to filled and compressed methods of MaskedArray.
|
|
Ideally, I would like to see .filled ) and .compressed ) return MaskedArray by default, but this is probably not possible for backward compatibility reasons.
|
|
Although MaskedArray is derived from ndarray, there are way too many cases where masked arrays cannot be used instead of ndarrays.)
|
|
The motivation for this feature is that when working with masked arrays, it is important that all operations are performed as masked, but if you have an expression like .
|
|
the division will be performed as ndarray division and infinities won t be masked.
|
|
This is a very common mistake and a work around is rather verbose .
|
|
With the proposed feature, one will be able to write .","enhancement"
|
|
"6013","5739","ENH Unify ma.dot and MaskedArray.dot interfaces.
|
|
I am opening a new issue to follow up on the discussion started at gh 5709.
|
|
> Why not add an out to the function also?
|
|
charris) In the past, it was considered safe to give ma methods additional arguments that were not present in ndarray prototypes.
|
|
It turned out that was not a good idea.
|
|
Consider what happened to the `numpy.dot` function.
|
|
Not being a ufunc, it did not originally have an `out` argument, so when a matching method was impemented in ma, it was considered safe to have `ma.dot a, b, strict=False)`, but when `out=None` was later added to `numpy.dot` and people started using the third argument in `numpy.dot`, it became impossible to drop in `ma.dot` in expressions like `numpy.dot a, b, out)`.
|
|
Now, no matter what we do to `ma.dot`, it is not possible to simultaneously have backward compatibility and drop in substitution.
|
|
What we can probably do is to deprecate use of strict as a positional argument.
|
|
For example, we can change `ma.dot` signature to `ma.dot a, b, args, out=None, strict=False)` and issue a deprecation warning whenever args is non empty.
|
|
After a deprecation period, we can probably start supporting `ma.dot a, b, out)`.","bug"
|
|
"6013","574","tofile ) truncation on arrays >= 2 32 on 64 bit OSX Trac 2114).
|
|
Original ticket http projects.scipy.org numpy ticket 2114 on 2012 04 24 by trac user embray, assigned to unknown.
|
|
After a fair bit of debugging we ve tracked down a bug in OSX s fwrite ) actually in an internal function that affects fwrite ), fprintf ), and other functions that write to a file handle).
|
|
This bug was originally discovered by trying to write out some large arrays with Numpy.
|
|
As far as I can tell from some Google searches) this bug isn t otherwise well known yet.
|
|
The bug is that at some point the size passed to fwrite ) is stuffed into a 32 bit register and checks if it s a multiple of 0x1000 4096) and then branches off to some separate routine for doing writes that are a multiple of one block size.
|
|
Thus, if the size is a multiple of 4096 and >= 2 32, the size gets silently truncated to `size 0xffffffff`.
|
|
The attached test program illustrates the problem.
|
|
This has been tested and been shown buggy on Leopard and Lion and so presumably the bug exists in Snow Leopard not sure about earlier OSX versions).
|
|
This is what the output looks like gcc g Wall arch x86 64 Wextra writetest.c o writetest .
|
|
writetest 0x100000000 ls l test.array size t bytes 8 array size 4294967296 array size cast as size t 4294967296 wrote 4294967296 bytes rw r r 1 embray 31 0 Apr 24 11 03 test.array As you can see, fwrite ) even returns that it wrote 4294967296 bytes , though in reality it wrote zero bytes.
|
|
Likewise .
|
|
writetest 0x100001000 ls l test.array size t bytes 8 array size 4294971392 array size cast as size t 4294971392 wrote 4294971392 bytes rw r r 1 embray 31 4096 Apr 24 11 04 test.array Further testing has shown that this holds for any multiple of 4096.
|
|
The fix that was implemented for 2256, where arrays are written in 2GB chunks, would also solve this problem.
|
|
So I think it would probably be sufficient to just enable the same chunked write code block in `PyArray ToFile )` on OSX as well.
|
|
Although the OSX bug only occurs on those 4K boundaries and only for sizes >= 2 32, for the sake of simplicity I think it s fine to just use more or less the same workaround.","bug"
|
|
"6013","5741","ValueError in numpy core with mmap.
|
|
I m using SunPy which uses AstroPy which uses numpy.
|
|
I m not sure where should I go. )
|
|
Here s the traceback .","bug"
|
|
"6013","5743","Feature Request make np.tile and np.repeat have the same syntax.
|
|
Naively, I would expect that these have the same syntax, but they don t....
|
|
It would be nice if someone could make their syntax consistent.
|
|
In [23] np.repeat vec3, 2) Out[23] array [ 1., 1., 2., 2., 3., 3.
|
|
], dtype=float32) In [24] np.tile vec3, 2) Out[24] array [ 1., 2., 3., 1., 2., 3.
|
|
], dtype=float32) In [25] np.tile vec3, 2,)) Out[25] array [ 1., 2., 3., 1., 2., 3.
|
|
], dtype=float32) In [26] np.repeat vec3, 2,)) ValueError Traceback most recent call last) <ipython input 26 ddd537e03337> in <module> ) > 1 np.repeat vec3, 2,)) u lisa local export.soft.lisa.master linux x86 64 fc14.x86 64 v3 canopy lib python2.7 site packages numpy core fromnumeric.pyc in repeat a, repeats, axis) 388 except AttributeError 389 return wrapit a, repeat , repeats, axis) > 390 return repeat repeats, axis) 391 392 ValueError a.shape[axis] != len repeats)","bug"
|
|
"6013","5744","Increase maximum number of array dimensions?.
|
|
At my lab, we re working with arrays with very many dimensions.
|
|
We ve run up against the hardcoded limit of 32 dimensions for `np.array`s.
|
|
What is the rationale for this limit, and is it possible to increase it?
|
|
Thanks!","enhancement"
|
|
"6013","5745","uint64 converted silently to float64 when adding an int.
|
|
This code .
|
|
prints .
|
|
which was a big surprise for me.
|
|
Why would adding an integer to uint64 result in a floating point value?","bug"
|
|
"6013","5746","Cannot reliably compare uint64 values to integers?.
|
|
According to this program any value between ...cde0 and ...cdef is equal to uint64 value of ...cdef which was a big surprise .
|
|
Seems like uint64 is converted to float even when it is compared against standard integers which was another big surprise for me.","bug"
|
|
"6013","575","asbytes function missed Trac 2115).
|
|
Original ticket http projects.scipy.org numpy ticket 2115 on 2012 04 24 by trac user mcilfone, assigned to rgommers.
|
|
Working with Scipy 0.9.0, stats module of Scipy, Numpy 1.4.1 and Python2.5, I get an import error on the following line from scipy import stats The error is <type exceptions.ImportError > cannot import name asbytes args = cannot import name asbytes ,) message = cannot import name asbytes in file C \Python25\lib\site packages\scipy\optimize\linesearch.py in ) 2 import numpy as np 3 from numpy.compat import asbytes 4 5 all = [ line search wolfe1 , line search wolfe2 , 6 scalar search wolfe1 , scalar search wolfe2 , numpy undefined, asbytes undefined In the source code of NumPy 1.4.1, I don t find this function.","bug"
|
|
"6013","5752","NumPy dot hangs when used with multiprocessing potentially Apple Accelerate related?).
|
|
I m having a devil of a time making a minimal test case, but this seems to be my issue http stackoverflow.com questions 23963997 python child process crashes on numpy dot if pyside is imported In my case, I have code that farms out a bunch of calculations, including matrix products to a `multiprocessing.Pool` with `pool.map`.
|
|
The computation hangs partway through, and some hacky `print` based debugging shows it hanging on a call to `np.dot` down in the guts of the program.
|
|
Replacing `pool.map` with the built in serial) `map` makes everything work.
|
|
I used not to have this issue, then something changed lunar eclipse?)
|
|
and now my computation hangs consistently whenever multiprocessing is used.
|
|
A minimal test case continues to elude me.
|
|
It s not enough to simply generate 10 random NxN arrays and dot them in a `multiprocessing` based way.)
|
|
.","bug"
|
|
"6013","5754","rfftn returns full sized fftn array instead of half 1 sized .
|
|
My `np.fft.rfftn` returns a wrongly shaped array `Enthought Canopy Python 2.7.6 | 64 bit | default, Sep 15 2014, 17 43 19) ` `[GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin` `Type help , copyright , credits or license for more information.` `>>> import numpy as np` `>>> grid = np.random.random 8,8,8))` `>>> np.fft.rfftn grid).shape` ` 8, 8, 8)` The shape should be ` 8, 8, 5)` in this case.
|
|
What could be causing this behavior?
|
|
Numpy version is 1.8.1.
|
|
I m running on Mac OS X 10.9 Mavericks).","bug"
|
|
"6013","5755","numpy.MachAr does not allow for denormals.
|
|
Using numpy 1.8.2 on python 2.7.9, the return value of ``MachAr.xmin`` does not take into consideration of denormals, getting a value of ``e 308`` see below), but much lower values exist for floats in the form of denormals.
|
|
.","bug"
|
|
"6013","5756","numpy.all may return a generator, which always evals to True.
|
|
The problem is that you can shoot yourself in the foot with the numpy version of `all`, because it is not consistent with the Python build in `all` in return value when generator is passed to it In [1] import numpy In [2] numpy.all a for a in [False]) Out[2] <generator object <genexpr> at 0x7f22623eeaa0> In [3] bool numpy.all a for a in [False])) Out[3] True In [4] builtin .all a for a in [False]) Out[4] False This is especially confusing when used with `ipython pylab` in which case the buildin .all is shadowed with the numpy one ipython pylab ... blah blah...
|
|
In [2] all.func code Out[2] <code object all at 0x7ff967d8c5b0, file usr lib64 python2.7 site packages numpy core fromnumeric.py , line 1842>","bug"
|
|
"6013","5759","Inconsistent concatenate behaviour.
|
|
Numpy under Python 2.7.9 .
|
|
I would expect the second use of concatenate to throw an exception, too.
|
|
Is this behaviour intentional?","bug"
|
|
"6013","576","MaskedArray record TypeError len ) of unsized object Trac 2116).
|
|
Original ticket http projects.scipy.org numpy ticket 2116 on 2012 04 26 by trac user mwtoews, assigned to pierregm.
|
|
I m getting a inconsistent error while trying to get the length of a record from a masked array, but only when the mask is enabled.
|
|
See the example import numpy as np my dtype = [ a , i ), b , f )] mar = np.ma.zeros 5, my dtype) Length of the first record; no error raised assert len mar[0]) == 2 Change mask for one of the fields print mar[0]) 0, 0.0) mar.mask[0][0] = True print mar[0]) , 0.0) Repeat same command as above to reveal this bug assert len mar[0]) == 2 Raises the error Traceback most recent call last) File <interactive input> , line 1, in <module> TypeError len ) of unsized object I am using !NumPy version 1.6.1 obtained from http www.lfd.uci.edu gohlke pythonlibs numpy and the version of Python is 2.5.1 [MSC v.1310 32 bit Intel)].","bug"
|
|
"6013","5760","nanpercentile issues.
|
|
`nanpercentile` crashes when an slice is filled with `nans` and multiple percentiles are requested .
|
|
It also sometimes returns an incorrectly sized array compare with `np.percentile`) .
|
|
Finally, the function also seems not mentioned at all in the official docs except in the release notes).","bug"
|
|
"6013","5761","Incorrect conversion from Python `datetime` to `np.datetime64` when timezone specified..
|
|
Here s the behavior I m seeing, both on Ubuntu 14.04 and on OS X Mavericks using Python 2.7.8 >>> from pytz import timezone >>> import numpy as np >>> np.datetime64 dt.datetime 2015, 4, 3, 16, tzinfo=timezone US Eastern ))) numpy.datetime64 2015 04 03T13 56 00.000000 0700 ) It s getting the offset right but for some reason off by 56 minutes.
|
|
This could also be an issue in `pytz` but `numpy` seems more plausible.
|
|
When I use naive timezone `tzinfo=None` implicitly), I get >>> np.datetime64 dt.datetime 2015, 4, 3, 16)) numpy.datetime64 2015 04 03T09 00 00.000000 0700 ) So, correct conversion if we interpret `tzinfo=None` to be UTC, and my machine is on PDT, which is presumably why in both cases we are seeing ` 0700` for timezone.","bug"
|
|
"6013","5762","TypeError Cannot change data type for object array.
|
|
If I try and read in the array contained in this npy file https gist.github.com astrofrog 8c2d188005f31e0bba36 raw 3065c8fa220a6eaccbff20565d0d520c07e5e7e6 test.npy then try and print out the array, so .
|
|
I get .
|
|
This is with the latest developer version of Numpy 3c5409e4e38e6034d69d0042bf2a3bc854ef2e53) and Python 3.4 on MacOS X.
|
|
The dtype can be printed, as can individual columns, but the array as a whole can t. This doesn t occur in the latest stable release so may be a regression?","bug"
|
|
"6013","5763",".H for ndarray?.
|
|
`np.matrix` has `.conj )`, `.T`, and `.H`, while `np.ndarray` has only `.conj )` and `.T` but not `.H`.
|
|
I found an extensive discussion about the idea of adding `.H` to `ndarray` a couple of years ago.
|
|
Remarks that I ve interpreted as favorable towards `.H` Alan G Isaac http mail.scipy.org pipermail numpy discussion 2013 July 067082.html Ralf Gommers http mail.scipy.org pipermail numpy discussion 2013 July 067230.html Bryan Van de Ven http mail.scipy.org pipermail numpy discussion 2013 July 067247.html Dave Hirschfeld http mail.scipy.org pipermail numpy discussion 2013 July 067249.html Pauli Virtanen http mail.scipy.org pipermail numpy discussion 2013 July 067256.html Stxc3xa9fan van der Walt http mail.scipy.org pipermail numpy discussion 2013 July 067197.html Votes against, concerns, alternatives, compromises, and devil s advocacy Nathaniel Smith http mail.scipy.org pipermail numpy discussion 2013 July 067245.html Fernando Perez `.H )`) http mail.scipy.org pipermail numpy discussion 2013 July 067250.html Jerome Kieffer `.H )`) http mail.scipy.org pipermail numpy discussion 2013 July 067257.html Sebastian Haase `.H )`) http mail.scipy.org pipermail numpy discussion 2013 July 067283.html Dag Sverre Seljebotn copy overhead) http mail.scipy.org pipermail numpy discussion 2013 July 067251.html Chris Barker use matrix) http mail.scipy.org pipermail numpy discussion 2013 July 067281.html Josef Perktold http mail.scipy.org pipermail numpy discussion 2013 July 067285.html Stxc3xa9fan van der Walt wants ufunc views) http mail.scipy.org pipermail numpy discussion 2013 July 067299.html An offer to write a NEP as of a couple years ago...) Stxc3xa9fan van der Walt http mail.scipy.org pipermail numpy discussion 2013 July 067284.html A summary Alan G Isaac http mail.scipy.org pipermail numpy discussion 2013 July 067280.html An ancient discussion about adding `.T` which has some points relevant to `.H` Bill Baxter http mail.scipy.org pipermail numpy discussion 2006 July 009124.html And an ancient voting thread that included `.T` and `.H` proposals Travis Oliphant http mail.scipy.org pipermail numpy discussion 2006 July 009203.html It seems that the strongest reasons to not do this are 1) The core ndarray interface should be kept pure without `.H`, and even `.T` was a stretch or a mistake.
|
|
2) Although `.T` is a view, `.H` would be a copy, so it would be slow bloated and inconsistent.
|
|
3) It would be cool to automatically use conjugate transpose BLAS LAPACK flags automatically with `.H` instead of explicitly constructing the conjugate transpose matrix, but numpy s machinery wouldn t be able to do this with a naive implementation of `.H`.","bug"
|
|
"6013","5766","Cannot generate random variates from noncentral chi square distribution with dof = 1.
|
|
The issue is a duplicate of scipy scipy 4720 The [document] http docs.scipy.org doc numpy reference generated numpy.random.noncentral chisquare.html numpy.random.noncentral chisquare) suggests that the parameter degree of freedom should be `df >= 1` to use `numpy.random.noncentral chisquare`.
|
|
However, taking `df = 1` raises error.
|
|
.
|
|
It makes me suspect that the random variate generation of the noncentral chi square distribution with degree of freedom `df` and noncentrality parameter `nc` is implemented in numpy using a method similar to the following.
|
|
.
|
|
However, this method is valid only for the case where `df > 1` because with `df = 1`, it needs to generate a chi square random variate with degree of freedom 0, which is not mathematically well defined.
|
|
On the other hand, noncentral chi square distribution is indeed mathematically well defined for all `df > 0`.
|
|
It simply requires a different algorithm for random variate generation if `0 < df <= 1`, such as the following one .","bug"
|
|
"6013","5767","issubdtype None, float) == issubdtype np.float64, None) == True ??.
|
|
Using Python 2.7.6 numpy 1.8.1, passing `None` and either `float` or a `numpy.float ` subtype to `issubdtype` always yields `True` >>> np.issubdtype None, float) True >>> np.issubdtype float, None) True >>> np.issubdtype np.float64, None) True >>> np.issubdtype None, np.float64) True If I understand the nature of `NoneType` correctly, it seems to me that `issubdtype` should return `False` in all cases where `None` is passed to either argument, save for when `None` is passed as both arguments.
|
|
`np.issubdtype None, int)` yields `False`, as I would expect.
|
|
Apologies if I m misunderstanding, or if this has been fixed between 1.8.1 and 1.9.2.","bug"
|
|
"6013","5768","Can t find gfortan on RHEL 6.4.
|
|
I cannot build Scipy against a current numpy on RHEL 6.4.
|
|
Running `gfortran version` or `gfortran dumpversion` both provide the same output .
|
|
This has to do with the changes made to how gnu fortran compilers are detected in 5333.
|
|
Going back to 8a2dd0680e8470aab7ad63622c0d0af703f58d42 works.
|
|
This isn t the newest RHEL, but it is still in the full support phase.","bug"
|
|
"6013","5769","MaskedArray.mean returns zero length array when self.
|
|
mask is nomask.
|
|
Should return scalar.. [Copied from this stackoverflow question] http stackoverflow.com questions 29689040 testing equivalence of means of numpy maskedarray instances raises attribute err) Edit So far it appears that this is a bug that was added somewhere between versions 1.8.1 and 1.9.2.
|
|
I have yet to track down the root cause.
|
|
In 1.8.1 `np.ma.MaskedArray.mean )` would return a scalar.
|
|
In 1.9.2 it is returning a zero dimensional MaskedArray.
|
|
This appears to only occur when `foo.mask` is set to `np.ma.nomask`.
|
|
I tried to track this issue down, but am having difficulty following the inheritance for `MaskedArray.mean )`.
|
|
Specifically, I am having difficulty with line 2711 in numpy core fromnumeric.py which appears to call back to `MaskedArray.mean )` when called with a `MaskedArray` instance.
|
|
Original post I m probably doing something funny here, but I can t seem to find the problem.
|
|
When I test the equivalence of the means of two different instances of `np.ma.MaskedArray` an `AttributeError` is raised.
|
|
Creating the arrays In [1] import numpy as np In [2] foo = np.ma.array [1,2,3,4]) In [3] bar = np.ma.array [1,2,3,4]) In [4] foo.mean ) Out[4] masked array data = 2.5, mask = False, fill value = 1e 20) In [5] bar.mean ) Out[5] masked array data = 2.5, mask = False, fill value = 1e 20) Comparison of the arrays works fine In [6] foo == bar Out[6] masked array data = [ True True True True], mask = False, fill value = True) Testing equivalence of the means fails In [7] foo.mean ) == bar.mean ) AttributeError Traceback most recent call last) <ipython input 7 3b824b0972e3> in <module> ) > 1 foo.mean ) == bar.mean ) users .local lib python2.7 site packages numpy ma core.pyc in eq self, other) 3705 mask = np.all [[f[n].all ) for n in mask.dtype.names] 3706 for f in mask], axis=axis) > 3707 check.
|
|
mask = mask 3708 return check 3709 AttributeError numpy.bool object has no attribute mask In [8] foo.mean ) != bar.mean ) AttributeError Traceback most recent call last) <ipython input 8 0947fa5da1ed> in <module> ) > 1 foo.mean ) != bar.mean ) users .local lib python2.7 site packages numpy ma core.pyc in ne self, other) 3738 mask = np.all [[f[n].all ) for n in mask.dtype.names] 3739 for f in mask], axis=axis) > 3740 check.
|
|
mask = mask 3741 return check 3742 Testing for larger or smaller of the means works fine AttributeError numpy.bool object has no attribute mask In [9] foo.mean ) >= bar.mean ) Out[9] masked array data = True, mask = False, fill value = True) In [10] foo.mean ) <= bar.mean ) Out[10] masked array data = True, mask = False, fill value = True) Python version 2.7.2 with Numpy version 1.9.2 In [11] np.
|
|
version Out[11] 1.9.2 It appears that in `np.ma.MaskedArray.
|
|
ne )` the variable `check` is created by a line that looks like check = np.ndarray.
|
|
eq foo.filled 0), bar.filled 0)).view type foo)) which I would expect to return a new masked array.
|
|
However, since `np.ndarray.
|
|
eq )` returns an instance of `np.bool `.
|
|
attempting to make a `MaskedArray` view of the `check` simply produces another `np.bool ` instance.
|
|
The error occurs when the routine attempts to assign to `check.
|
|
mask` since `np.bool ` does not have a ` mask` attribute.
|
|
Any ideas what is going on here?
|
|
A bug or stupidity on my part?","bug"
|
|
"6013","577","inconsistent comparison of object type arrays Trac 2117).
|
|
Original ticket http projects.scipy.org numpy ticket 2117 on 2012 04 26 by yarikoptic, assigned to unknown.
|
|
This issue became more visible since 1.6.x allowed to construct such heterogeneous arrays without specification of dtype=object python c import numpy as np; print np.
|
|
version ; a = np.array [np.array [0, 1]), np.array 1)]); print a.dtype; print a == a.copy ); print a == np.array [np.array [0, 1]), np.array 1)]) 1.7.0.dev 3cb783e object [ True True] False So comparing an object array to itself s copy worked just fine but comparison to identically created another one failed.","bug"
|
|
"6013","5775","Inconsistent arange behaviour with different int types.
|
|
Given this code .
|
|
Obviously step 0 is an invalid input for creating a range.
|
|
So the division error is understandable, but might be more explicit.
|
|
What surprised me, was the following behaviour .
|
|
The code is as wrong as before, but now instead of throwing any error, it just returns an empty list.
|
|
Shouldn t it throw exactly the same error as above?
|
|
An error explicitly stating, that a range with a step of 0 can t be specified?","bug"
|
|
"6013","5776","numpy.ma.allequal raises MaskError on structured arrays.
|
|
Minimal example code In [1] import numpy as np In [2] import numpy.ma as ma In [3] import numpy.version In [4] numpy.version.full version Out[4] 1.9.2 In [5] x = np.zeros 2,), dtype = [ a , int), b , int)]) In [6] y = ma.array x) In [7] ma.allequal y, y) MaskError Traceback most recent call last) <ipython input 7 5ad9932e8937> in <module> ) > 1 ma.allequal y, y) usr lib64 python3.4 site packages numpy ma core.py in allequal a, b, fill value) 6924 y = getdata b) 6925 d = umath.equal x, y) > 6926 dm = array d, mask=m, copy=False) 6927 return dm.filled True).all None) 6928 else usr lib64 python3.4 site packages numpy ma core.py in array data, dtype, copy, order, mask, fill value, keep mask, hard mask, shrink, subok, ndmin) 5881 return MaskedArray data, mask=mask, dtype=dtype, copy=copy, subok=subok, 5882 keep mask=keep mask, hard mask=hard mask, > 5883 fill value=fill value, ndmin=ndmin, shrink=shrink) 5884 array.
|
|
doc = masked array.
|
|
doc 5885 usr lib64 python3.4 site packages numpy ma core.py in new cls, data, mask, dtype, copy, subok, ndmin, fill value, keep mask, hard mask, shrink, options) 2722 msg = Mask and data not compatible data size is i, \ 2723 mask size is i.
|
|
> 2724 raise MaskError msg nd, nm)) 2725 copy = True 2726 Set the mask to the new value MaskError Mask and data not compatible data size is 1, mask size is 2.
|
|
Python 3.4.3 NumPy 1.9.2 Thanks!","bug"
|
|
"6013","5777","Discrepancy in numpy.round depending on output type.
|
|
When the exact result is representable in the output type, there shouldn t be any error such in the int32 case below .","bug"
|
|
"6013","5778","The phase of 0 is not defined, but numpy.angle 0) works.
|
|
`numpy.angle 0)` returns 0.0.
|
|
However, in mathematics, 0 does not have any angle.
|
|
I would have expected feeding 0 to `numpy.angle )` to raise an exception or maybe generate NaN??).
|
|
Now, the standard `cmath` library has a similar function, except that it is called `phase )`.
|
|
This function does return a 0 phase for 0 unlike a mathematical phase).
|
|
However, this behavior is [documented] https docs.python.org 3 library cmath.html cmath.phase).
|
|
Thus, at the very least, the behavior of `numpy.angle )` for 0 should be defined in the documentation.
|
|
Since the current behavior is not officially documented yet, and since the name of the function is different from the similar `cmath.phase )` name, maybe NumPy has the opportunity to produce a more mathematically correct result for 0, like raising an exception or maybe returning NaN?","bug"
|
|
"6013","5779","np.round ) with explicit output doesn t work with complex arrays.
|
|
The output array isn t filled in .","bug"
|
|
"6013","578","bincount2d Trac 2118).
|
|
Original ticket http projects.scipy.org numpy ticket 2118 on 2012 04 27 by trac user amueller, assigned to unknown.
|
|
I would love to have a function bincount2d that is the equivalent of histogram2d but where the bins are given by the integers and possibly weighted.","enhancement"
|
|
"6013","5780","np.round ) always returns positive zero on Windows.
|
|
On Linux .
|
|
On Windows .
|
|
Note that Python s round ) gets it right .","bug"
|
|
"6013","5781","abs ) of numpy floating point scalar fails on signed zeros.
|
|
.
|
|
Compare with .
|
|
np.abs ) seems to get it right .","bug"
|
|
"6013","5782","Wrong lexsort with negative axis.
|
|
When calling `np.lexsort` on multidimensional arrays with a negative `axis` argument, the `axis` gets changed to the one with the smallest stride of the first `keys` array .
|
|
Using `keys=[b, a]` results in all calls with negative `axis` returning the same as with `axis=0`, for this is the smallest strided axis of `b` .
|
|
This seems to be due to the initialization of iterators over each of the `keys` arrays before bounds checking the axis parameter, [here] https github.com numpy numpy blob f1f9e14e5d63aee6c8309ad425e3928192c0873b numpy core src multiarray item selection.c L1441), which for negative `axis` results in resetting `axis` to the minimum strided one, according to the [docs] https docs.scipy.org doc numpy 1.9.0 reference c api.array.html c.PyArray IterAllButAxis).","bug"
|
|
"6013","5783","Int32 overflow when creating array from large list.
|
|
When trying to create an array from a huge list, I receive the following error ValueError negative dimensions are not allowed After some debugging, I found that it seems to be a problem with some INT32 overflow in the C code of numpy.
|
|
So you can quite easily reproduce the error the following way .","bug"
|
|
"6013","579","Indexing with a boolean scalar array > segfault Trac 2119).
|
|
Original ticket http projects.scipy.org numpy ticket 2119 on 2012 04 28 by njsmith, assigned to unknown.
|
|
This has the same result in 1.5.1 as shipped by Ubuntu 11.04) and in current master 3cb783ecc96fb) >>> import numpy as np >>> a = np.zeros 2, 2)) >>> a[np.asarray True)] zsh segmentation fault python","bug"
|
|
"6013","5790","dtype f64 silently results in float32 .
|
|
I ve recently been bitten by this .
|
|
Some proposals what to change 1.
|
|
Warning for ` f64 ` input?
|
|
2.
|
|
Error for ` f64 ` input?
|
|
3.
|
|
Make f64 result in float64 I know change is difficult for numpy because of backwards compatibility concerns.
|
|
Hopefully it s possible to be more strict here ... in my code I was getting slightly incorrect results for half a year because I was using 32 bit floats where I know I needed to use 64 bit floats, but for some reason wrote f64 instead of float64 .","bug"
|
|
"6013","5794","numpy.reshape overflow of np.prod somewhere?.
|
|
I try to reshape a large array.
|
|
numpy.reshape returns ValueError total size of new array must be unchanged I couldn t find it, but is it possible that there is a overflow of np.prod in the source code?
|
|
I checked the shapes and they are certainly ok!
|
|
I read in a from a binary file and if I increase the size, the error sets in, where it was lacking at the beginning.","bug"
|
|
"6013","5795","np.average vs np.ma.average for axis tuple.
|
|
np.average allows to average over multiple axis at once but np.ma.average does not.
|
|
Compare import numpy as np a = np.random.rand 4, 4, 4) np.average a, axis= 1, 2)) np.ma.average a, axis= 1, 2))","bug"
|
|
"6013","5798","An error in odeint program.
|
|
I wrote a program to use odeint to solve a differential equation.
|
|
But it had a problem.
|
|
When I setted Cosmopara as np.array [70.0,0.3,0, 1.0,0]), it gave a warning that invalid value encountered in sqrt and invalid value encountered in double scalars in h = np.sqrt y1 2 Omega M t 3) Omega DE y2) .
|
|
But I checked that line and didn t find any error.
|
|
If Cosmopara = np.array [70.0,0.3,0.0, 1.0,0.0]), Y shouldn t change but it changed.
|
|
Besides, If I chose Cosmopara = np.array [70.0,0.3,0.1, 1.0,0.1]), this program could give a right result.
|
|
I don t know why.
|
|
So please help me.
|
|
Followwing is my program.
|
|
import numpy as np import matplotlib.pyplot as plt from scipy.integrate import odeint global CosmoPara Cosmopara = np.array [70.0,0.3,0.0, 1.0,0.0]) def derivfun Y,t) Omega M = Cosmopara[1] Sigma 0 = Cosmopara[2] omega = Cosmopara[3] delta = Cosmopara[4] Omega DE = 1 Omega M Sigma 0 2 y1 = Y[0] y2 = Y[1] h = np.sqrt y1 2 Omega M t 3) Omega DE y2) dy1dt = 3.0 y1 t delta Omega DE y2) t h) dy2dt = 3.0 1 omega) 2.0 delta y1 h) y2 t return np.array [dy1dt,dy2dt]) z = np.linspace 1,2.5,15001) time = 1.0 z Omega M = Cosmopara[1] Sigma 0 = Cosmopara[2] omega = Cosmopara[3] delta = Cosmopara[4] Omega DE = 1 Omega M Sigma 0 2 y1init = Sigma 0 y2init = 1 Yinit = np.array [y1init,y2init]) Y = odeint derivfun,Yinit,time) y1 = Y[ ,0] y2 = Y[ ,1] h = np.sqrt y1 2 Omega M time 3) Omega DE y2) plt.figure ) plt.plot z,h) plt.show )","bug"
|
|
"6013","5799","BUG numpy.arctan returns nan for large complex arguments.
|
|
The cmath.atan is able to handle large complex values >>> import numpy as np >>> import cmath >>> t = 1.10076957173e 292 0j >>> cmath.atan t) Out[37] 1.5707963267948966 0j) >>> cmath.atan 1j t) Out[38] 1.5707963267948966 9.084553440447778e 293j) while numpy.arctan does not as shown here >>> np.arctan t) Out[39] nan nan j) np.arctan 1j t) Out[40] nan nan j)","bug"
|
|
"6013","580","unravel index 0, )) should return ) Trac 2120).
|
|
Original ticket http projects.scipy.org numpy ticket 2120 on 2012 04 29 by njsmith, assigned to unknown.
|
|
Currently, this idiom works for all arrays that have ndim >= 1 for i in xrange np.prod a.shape)) idx = np.unravel index i, a.shape) a[idx] = ... Obviously there are usually better ways to do this, but anyway.)
|
|
But, it fails for 0 d arrays >>> np.unravel index 0, )) Traceback most recent call last) File <stdin> , line 1, in <module> ValueError dims must have at least one value This should just return ), which is the unique index that can be used with 0 d arrays.","enhancement"
|
|
"6013","5801","Elements of int64 array made from converting Python 2.x longs not int64 even though ndarray dtype is int64.
|
|
This is in Python 2.x only since int and long got unified in 3.x.
|
|
I tested it in numpy 1.9.2 on CPython 2.6.6 and in numpy 1.8.2 on CPython 2.7.8.
|
|
It seems Numpy has a builtin ability to convert longs to ints and then to int64, which was a bit of a surprise really.
|
|
But, it was after that that the real surprise came, which seems to be a Numpy bug.
|
|
If I make an ndarray from longs that are small enough to fit into int64 none are larger than ``2 64 ), the array s dtype is int64, but if I extract an element from it and see if it is an int64, it fails.
|
|
.
|
|
The dtypes of a and its element are both int64.
|
|
Even the type of its element is int64, yet it isn t the same one.
|
|
.","bug"
|
|
"6013","5802","segfault of linalg.cond with some nan matrices.
|
|
.
|
|
Counts to 8 and then segfaults for me.
|
|
Linux, 64bit, numpy 1.9.2.
|
|
It also crashes for other multiples of 8 for xsize like 16 and 24.
|
|
Everything else seems fine.
|
|
Just tested, it also hits `linalg.inv`","bug"
|
|
"6013","5803","numpy.distutils.misc util.appendpath issue.
|
|
I m using python 2.7, numpy 1.9.1 on OSX 10.9.1 I m trying to install yt for the dark sky simulation data access.
|
|
When I execute python setup.py install an importerror appears because of the line from numpy.distutils.misc util import appendpath saying this doesn t exist.
|
|
Any help would be appreciated, thanks!","bug"
|
|
"6013","5804","ix ) returned array has dtype=float64 when given empty range.
|
|
.
|
|
The correct dtype should be int32, since float indices are usually not accepted as indices in numpy arrays.
|
|
That dtype is returned for non empty ranges .","bug"
|
|
"6013","5806","mp.ma.masked singleton causes difficulties.
|
|
Whenever a MaskedArray method returns a scalar value but that value should be masked, all MaskedArray methods return a reference to a singleton instance `np.ma.masked` of type `MaskedConstant`.
|
|
I think the motivation was to be able to write code like if result is masked .
|
|
However, returing a singleton causes some problems that make it difficult to use np.ma.MaskedArray as a drop in replacement for ndarray.
|
|
Here is a summary of the problems I see, in the hope that they can be fixed.
|
|
1.
|
|
Many operations involving `masked` are coerced to float Probably the worst issue is that `masked` is of type `float`.
|
|
This means the return value of a method may be of a different type than the original array.
|
|
This is especially bad for boolean arrays.
|
|
For example, if arr is a boolean array, but all or any return a masked value, the following line will fail since you cannot do ` ` to a float.
|
|
>>> a = np.ma.array [True, True], mask=[True, True]) >>> a.all ) It also means that certain series of operations on masked arrays will sometimes get cast to float when they wouldn t be with ndarrays.
|
|
2.
|
|
Overwriting `masked` causes strange results in completely separate code A less serious problem arises if someone tries to assign to the return value of a MaskedArray method, which would end up assigning to the singleton.
|
|
That will then affect code anywhere that involves the masked singleton.
|
|
I came across this when one numpy unit test would modify the singleton, and then another would read it, and I would get an error depending on the order the unit tests were run.
|
|
The problem arises in case like `marr2[ ] = marr1.method )` if the method returns `masked`.
|
|
This means marr2 will get filled with arbitrary gargage, but maybe that s not a problem since those values will be masked garbage.
|
|
Although, it was a somewhat confusing bug to fix).
|
|
3.
|
|
Code acting on a return value of a MaskedArray method can fail if `masked` was returned) Consider some code of the form >>> result = arr.sum ) >>> dosomething result) This might work fine most of the time, but fail in the possibly rare) case that the sum returns the singleton.
|
|
It might be that the operation is not allowed on `np.ma.masked`, or it might be that further use of `np.ma.masked` wouldn t work as before.
|
|
Most cases of dosomething I checked seem OK, but here are some that cause problems a) What if someone decides to remove the mask on a return value?
|
|
Eg >>> result = arr.sum ) >>> result.mask = False if arr.sum happened to return the masked singleton, this would cause havoc.
|
|
rgommers suggested making `.mask` readonly which sounds like a good idea to me, although it means the code will generally run fine for most arr but will raise an error in the possibly rare case the sum is fully masked.
|
|
b) writing to a scalar Consider >>> result[ )] = 6 which would be fine for ndarrays, but raises an error for masked arrays if result is masked though it s hard to imagine a case where someone would want to index a numpy scalar this way).
|
|
c) passing `masked` as the `out` parameter of a ufunc >>> np.ma.log inputarr, out=result) I think though there are other bugs involved here) that using a variable which might be `np.ma.masked` as the out parameter to a ufunc will cause problems.","bug"
|
|
"6013","5807","numpy.polfit freezes pyotherside app on jolla.
|
|
I m not quite sure whether im right here, maybe this is pyotherside s issue or my build of numpy).
|
|
when I run numpy s polyfit function on my jolla phone in a pyotherside app, it always hangs exactly at the same place without returning or throwing any error.
|
|
It works fine when running the same code in the terminal Interpreter on the phone, or in pyotherside on the desktop, it does not matter how big the attempted fit is it does not simply run out of memory) To investigate further, running the debugger on polyfit a,b,1,w=c) showed me the last few lines before also freezing Pdb) Return > usr lib python3.4 site packages numpy core methods.py 38) any ) >False > return umr any a, axis, dtype, out, keepdims) Pdb) Return > usr lib python3.4 site packages numpy core fromnumeric.py 1848)any ) >False > return arr.any axis=axis, out=out, keepdims=keepdims) Pdb) > usr lib python3.4 site packages numpy core machar.py 233) do init ) > i = i 1 Pdb) > usr lib python3.4 site packages numpy core machar.py 234) do init ) > k = k k Pdb) > usr lib python3.4 site packages numpy core machar.py 223) do init ) > for in range max iterN) Pdb) > usr lib python3.4 site packages numpy core machar.py 224) do init ) > y = z Pdb) > usr lib python3.4 site packages numpy core machar.py 225) do init ) > z = y y Pdb) Has anyone an idea what could be causing this?
|
|
I havent seen any bug like this yet.
|
|
To reproduce, I created a quick demo [numpy bug showcase] https github.com dasimmet numpy polyfit bug on jolla ) run the installed app with sailfish numpy in terminal on the jolla and press the button, it should bring up the python debugger.
|
|
the version of numpy I use is built by installing dependencies and running pip wheel numpy on the jolla phone, then extracted into the app dir.
|
|
The architecture it is built against is armv7hl.
|
|
Any other numpy function I tried seemed to work flawlessly.
|
|
Thanks for any kind of help","bug"
|
|
"6013","5808","numpy.arange stop precision issue.
|
|
All elements of the array generated by `numpy.arange start, stop, step)` should be strictly smaller than `stop`.
|
|
On my Ubuntu 14.04 installation this happened .
|
|
Is this an issue on other installations?","bug"
|
|
"6013","5809","numpy.ndarray.shape is represented by different types under Linux and Windows.
|
|
Under Linux it returns a tuple of ints .
|
|
while under Windows it s a tuple of longs .
|
|
It causes problems, for example, with writing C extensions.
|
|
The following code works fine under Linux but under Windows NULL is returned and it s unobvious.
|
|
.
|
|
Python 2.7, numpy 1.9.2","bug"
|
|
"6013","581","Variance can return negative values Trac 2121).
|
|
Original ticket http projects.scipy.org numpy ticket 2121 on 2012 04 29 by charris, assigned to unknown.
|
|
Apropos ticket 2545, I don t think it is really fixed.
|
|
The following should probably raise an error.
|
|
In [4] x = np.array [1, 1, 2]) In [5] x.var ddof=10) Out[5] 0.095238095238095233 The nan for the std is an artifact of taking a square root.
|
|
Robert, is there any use for negative variance?
|
|
For the 1.6.2 backport of 2545 I return nan in these circumstances, but that is a dirty fix.","bug"
|
|
"6013","5810","Generator for normally distributed random numbers that reproduces MATLAB.
|
|
We ve faced the problem of reproducing MATLAB s randn ) while porting a project from MATLAB to Python see this issue https github.com scipy scipy issues 4666).
|
|
The solution that worked for us was to use MATLAB Coder for generating a C code for randn ) function and wrapping it as a C extension to Python.
|
|
MATLAB uses Ziggurat method for generating gaussian random numbers and there were a lot of proposals here to implement it.
|
|
Maybe the code generated by MATALB Coder could be used by numpy for implementing Ziggurat in numpy?
|
|
A project with a C extension for Python can be found here https github.com Shambricool py matlab randn","enhancement"
|
|
"6013","5811","numpy.random.vonmises documentation contains incorrect code example..
|
|
The numpy.random [reference] http docs.scipy.org doc numpy reference generated numpy.random.vonmises.html) for the von Mises distribution contains an incorrect code sample pasted below).
|
|
.
|
|
While this works for the example, line 5 doesn t concur with the definition of the von Mises probability density function, given above the example.
|
|
According to documentation about the [special functions] http docs.scipy.org doc scipy reference special.html module scipy.special), the modified bessel function of order real order is given by scipy.special.iv ).
|
|
Using that function instead means the sign doesn t need to be changed.
|
|
I noticed the problem when I tried to emulate the code and I found that area under the pdf didn t sum to 1 when I used the trapezium rule and would sometimes even go below the x axis, when neither the exponential function or the bessel function should ever be negative.
|
|
Compare .
|
|
.
|
|
Excuse my use of tau.
|
|
Using Python 2.7.9 Anaconda 2.2.0 64 bit), IPython 3.1.0, numpy 1.9.2","bug"
|
|
"6013","5812","f2py incorrect call to Python interpreter.
|
|
Using the following wheel from PyPI on OS X 10.9.5 .
|
|
the shebang in ` usr local bin f2py` is incorrect .","bug"
|
|
"6013","5813","doc dead links.
|
|
As mentioned in https github.com numpy numpy issues 5811 issuecomment 97124230 the page http docs.scipy.org doc numpy reference generated numpy.random.vonmises.html has some dead links.","bug"
|
|
"6013","5816","argsort bug.
|
|
I come across a wrong numpy output while using argsort ) >>> x array [ 2.43752439e 01, 2.12345154e 01, 3.00682512e 01, 8.82461778e 01, 6.54040476e 01, 9.66741454e 04, 1.41006860e 01, 3.42346428e 02, 2.20661171e 01, 1.04476288e 00, 1.82253038e 01, 5.79845517e 01, 5.15250240e 01, 1.31020432e 00, 1.77514292e 00]) >>> x.argsort ) array [ 5, 7, 6, 10, 1, 8, 0, 2, 12, 11, 4, 3, 9, 13, 14]) Apparently the number 1.04476288e 00 in x[9] should be the third largest, but it yields that the number 2.20661171e 01 in x[8] is the third largest.
|
|
It doesn t seem like an error caused by this simple code itself, so I consider it as a bug in argsort ) I am using numpy 1.6.1, python 2.7.2.","bug"
|
|
"6013","5818","`np.random.normal` with 0 scale.
|
|
Currently `np.random.normal` refuses to generate random variates with no standard deviation i.e., a stream of zeros).
|
|
While this could make sense for more featureful random libraries e.g.
|
|
scipy s, as the pdf becomes harder to define), when all we can have is a stream of numbers, the limitation seems arbitrary.
|
|
Use case Be able to generate simulated data with zero noise using the same code as I generate data with some gaussian noise, just by passing a std of 0, rather than special casing it.","enhancement"
|
|
"6013","5819","unexpected behaviors when subclassing ndarray.
|
|
I m running into a few unexpected behaviors when subclassing `ndarray`, as described here http docs.scipy.org doc numpy user basics.subclassing.html).
|
|
Here s a simple subclass that doesn t try to add any functionality .
|
|
Notice that the behavior of `.sum )` differs between an `ndarray` and `MyArray` .
|
|
I d like my subclass to have the `np.array` behavior I don t want `MyArray.sum )` to return 0 dimensional arrays.
|
|
Before I try to solve this problem, I want to point out that the difference is also there for the object `dtype`.
|
|
Here s a pretty useless object definition, that supports addition .
|
|
Notice that the behavior of `.sum )` still differs in terms of return type .
|
|
It seems I can force the behavior I want, using ` array wrap ` .
|
|
.
|
|
... but I m wondering whether this is the preferred recommended way to deal with this, and whether guidelines for this issue should be added to the documentation.
|
|
One related potential unexpected behavior is that the exceptions raised in ` array wrap ` seem to get caught at a higher level, and therefore appear to be ignored.
|
|
Even odder, raising in ` array wrap ` also solves the 0 dimensional array problem .
|
|
.","bug"
|
|
"6013","582","numpy.take gives corrupted result with inplace operation Trac 2122).
|
|
Original ticket http projects.scipy.org numpy ticket 2122 on 2012 04 29 by fengy research, assigned to unknown.
|
|
Take doesn t work as expected when the output is inplace.
|
|
When mode= clip or mode= wrap , the operation is truly in place, and the simple take algorithm currently implemented will overwrite the array elements before they are memmoved to the correct reindexed positions.
|
|
A simple test case is listed below, In [37] a = array [0, 1, 2]) In [38] ind = array [2, 1, 0]) In [39] b = a.copy ) In [40] a.take ind, out=a) Out[40] array [2, 1, 0]) In [41] a = b.copy ) In [42] a.take ind, out=a, mode= clip ) Out[42] array [2, 1, 2]) Out[40] is correct, for mode= raise secretly creates a copy of the output before copying.
|
|
But Out[42] is in correct.
|
|
No error or warning is given.
|
|
A fix is either to create a copy for the other two modes, or to actually use a permuting algorithm.
|
|
The latter calls for a full rewrite of PyArray Take but will save memory usage significantly.
|
|
A proper implementation is used in GSL gsl permute.","bug"
|
|
"6013","5820","np.allclose does not handle infinities correctly.
|
|
Running np.allclose [np.inf, np.inf], np.inf) raises the following error my machine macosx, python 2.7 numpy 1.9.2 from macports) IndexError Traceback most recent call last) <ipython input 6 318ae0ed72a3> in <module> ) 1 a = np.full 2, np.inf) > 2 np.allclose a, np.inf) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy core numeric.pyc in allclose a, b, rtol, atol) 2224 return False 2225 Check that sign of inf s in x and y is the same > 2226 if not all x[xinf] == y[xinf]) 2227 return False 2228 IndexError index 1 is out of bounds for axis 1 with size 1 Interestingly, running the same code on a linux machine pyhton 2.7 numpy 1.8.1) fails at the same line, but raises a `ValueError too many boolean indices` error.
|
|
I suspect this is because I compare an array with a single number, since `np.allclose [np.inf, np.inf], [np.inf, np.inf])` works.
|
|
However, `np.allclose [1, 1], [1])` returns `True`.
|
|
I m not sure what `np.allclose` should return when a single element is compared to an array, but there certainly is an inconsistency here.","bug"
|
|
"6013","5823","Unexpected behaviors numpy.fromfile on windows 7 Python 2.7.
|
|
I m running the following code .
|
|
To download the file test https dl.dropboxusercontent.com u 21272835 band PA2.lum) The correct tab.size is 902 But on Windows 7 Python 2.7.9 Numpy 1.9.2, the size is... 276 !!!
|
|
Why ???
|
|
On Windows 7 same computer) Python 3.4.3 Numpy 1.9.2, the size is correct, it s 902.
|
|
Correct also on OS X, Ubuntu with Python 2.7.x","bug"
|
|
"6013","5826","ma.where does not work with structured types..
|
|
I think this is an easy fix for someone familiar with the ma module.
|
|
.","bug"
|
|
"6013","583","numpy.core.records.fromrecords fails for zero length lists Trac 2123).
|
|
Original ticket http projects.scipy.org numpy ticket 2123 on 2012 05 01 by trac user ktegan, assigned to unknown.
|
|
I have a patch that lets fromrecords handle zero length lists.
|
|
If the list is length zero and dtype and formats are both None then a ValueError is raised.","bug"
|
|
"6013","5832","distutils makes it too hard to install from source with mingw32.
|
|
One can build from source with python setup.py build compiler=mingw32 but then installing fails.
|
|
Doing python setup.py install compiler=mingw32 results in ... error option compiler not recognized And doing python setup.py install results in this obscure error ... running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building py modules sources building library npymath sources No module named numpy.distutils.msvccompiler in numpy.distutils; trying from distutils Unable to find productdir in registry Checking environ VS100COMNTOOLS error Unable to find vcvarsall.bat This was reported at http numpy discussion.10968.n7.nabble.com Compiling NumPy on Windows for Python 3 3 td8653.html, can be found on various StackOverflow issues and is related to http bugs.python.org issue2943.
|
|
Most SO answers which all recommend installing MSVC) and Python issue 2943 are not the whole story though, because if you add a file ``distutils.cfg`` in the ``distutils`` directory a local ``setup.cfg`` or a `` HOME .pydistutils.cfg`` should also work) containing [build] compiler=mingw32 then ``python setup.py install`` works fine.
|
|
Because this will probably never be fixed in distutils bug report open for 7 years...), maybe it can be fixed in ``numpy.distutils``.
|
|
And otherwise at least this issue serves as documentation....","bug"
|
|
"6013","5833","broken links at numpy.org.
|
|
Everything linked to wiki.scipy.org is broken.","bug"
|
|
"6013","5834","f2py CLI configuration problem.
|
|
I tried today to compile a small fortran code for my python modules and encountered the following problem Python 2.7.9, OS X 10.10.3, Numpy 1.9.2) .
|
|
This happened both with virtualenv numpy and system one.
|
|
`python c from numpy import f2py; f2py.test ) ` worked fine.
|
|
I had to edit the ` Users b Repositories BBN pyBBN env bin f2py` to replace the weird shabang to be ` !
|
|
usr bin env python`.
|
|
Now f2py executes fine","bug"
|
|
"6013","5835","numpy.cov ) and numpy.var ) default bias are inconsistent numpy 1.9.2).
|
|
cov ) uses a bias of 1 by default.
|
|
var ) uses a bias of 0 by default.
|
|
Such that .
|
|
will only print the second line.","bug"
|
|
"6013","5837","numpy.where s result is read only for multi dimensional arrays.
|
|
See http stackoverflow.com questions 28663142 why is np wheres result read only for multi dimensional arrays","bug"
|
|
"6013","5839","broadcast to should have versionadded in docstring.
|
|
Is the policy to label all of the new functions with the version in which they were introduced?","bug"
|
|
"6013","584","genfromtxt and unicode strings Trac 2124).
|
|
Original ticket http projects.scipy.org numpy ticket 2124 on 2012 05 02 by trac user anntzer, assigned to unknown.
|
|
With bytes in Python 3 speak) fields, genfromtxt dtype=None) sets the sizes of the fields to the largest number of chars npyio.py line 1596), but it doesn t do the same for unicode fields, which is a pity.
|
|
See example import io, numpy as np s = io.BytesIO ) s.write b abc 1\ndef 2 ) s.seek 0) t = np.genfromtxt s, dtype=None) or converters={0 bytes}) print t, t.dtype) > [ b a , 1) b b , 2)] [ f0 , |S1 ), f1 , <i8 )] s.seek 0) t = np.genfromtxt s, dtype=None, converters={0 lambda s s.decode utf 8 )}) print t, t.dtype) > [ , 1) , 2)] [ f0 , <U0 ), f1 , <i8 )] I tried to change npyio.py around line 1600 to add that but it didn t work; from my limited understanding the problem comes earlier, in the way StringBuilder is defined ?
|
|
).","bug"
|
|
"6013","5842","MANT Raise an error for O4 and O8 dtype descriptors.
|
|
Complete the deprecations started in 5840 and 5841.","bug"
|
|
"6013","5844","Decide on what the resolution rules for op rop numpy ufunc actually are.
|
|
There is a complex set of questions around how to handle method resolution in the presence of ` numpy ufunc `.
|
|
Currently in master is an extremely complicated set of rules that isn t documented and that I don t actually understand see 5748 for the latest set of changes to this), so it s kinda hard to know whether they are correct, but I suspect not.
|
|
And this is a blocker for 1.10, b c whatever we release in 1.10 will be set in stone forever.
|
|
I strongly feel that we cannot include ` numpy ufunc ` in a release without at least having a document somewhere describing what the actual dispatch rules are.
|
|
I hope that doesn t mean we have to defer ` numpy ufunc ` for another release, but if it does then it does.
|
|
AFAICT this is how `a op b` dispatch works for ndarrays, BEFORE ` numpy ufunc ` i.e., this is how 1.9 works) First Python uses the subclass rule to decide whether to invoke `a. op b)` or `b. rop a)`.
|
|
So in the case where one of these objects is a proper subclass of the other, that object always gets to do absolutely anything, so that s fine.
|
|
The interesting cases are the ones where neither is a proper subclass of the other either because it s like, `matrix masked array`, or because it s like `ndarray scipy.sparse`).
|
|
So without loss of generality, let s focus on the case where Python calls `a. op b)`, and `a` is either an instance of `ndarray` or else an instance of a subclass of `ndarray` which has not overridden ` op `, i.e.
|
|
we re getting `ndarray.
|
|
op a, b)`.
|
|
`ndarray.
|
|
op ` has the following logic see `PyArray GenericBinaryFunction` in `number.c`) If `b` is not an `ndarray` at all even a subclass), and `b` has a higher ` array priority ` than `a`, then we return `NotImplemented` and let control pass to `b. rop a)`.
|
|
Otherwise, we call `np.op a, b)` and let the ufunc machinery take over.
|
|
`np.op a, b)` does the following see `PyUFunc GenericFunction`, `PyUFunc GeneralizedFunction`, in `ufunc object.c`, and also `ufunc generic call` which converts ` 2` return values from the previous into `NotImplemented` so you have to audit their whole call stack) If `b` is not an `ndarray`, and calling `np.array b)` returns an object array presumably because coercion failed... though I guess this could also be hit if `b. array )` return an object array or something), AND `b` has a higher ` array priority ` than `a`, and `b` has an ` rop ` method, then return `NotImplemented`.
|
|
If any of our arrays contain structured dtypes or strings, and there are no special struct ufunc loops registered, but not if any of our arrays contain objects, then return `NotImplemented`.
|
|
This is buried in `get ufunc arguments`, search for `return 2`.)
|
|
Otherwise we return the actual ufunc result.
|
|
Now, my suggestion is that the way we would EVENTUALLY like this to look is First, Python uses the subclass rule to decide whether to invoke `a. op b)` or `b. rop a)`.
|
|
As above, let s assume that it invokes `ndarray.
|
|
op a, b)`.
|
|
`ndarray.
|
|
op a, b)` calls `np.op a, b)` which in turn invokes all the standard ufunc stuff, including ` numpy ufunc ` resolution).
|
|
There is no step 3.
|
|
I submit that it is obvious that IF we can make this work, then it is obviously the ideal outcome, because it is the simplest possible solution.
|
|
But is it too simple?
|
|
To determine this we have to answer two questions 1) Will it adequately address all the relevant use cases?
|
|
2) Can we get there from here?
|
|
So let s compare the current rules to my dream rules.
|
|
First, we observe that everything that currently happens inside the ufunc machinery looks like it s totally wrong.
|
|
The first check can only be triggered if `b` is a non `ndarray` that has a higher ` array priority ` among other things), but if we look above, we see that those conditions are sufficient to trigger the check in `ndarray.
|
|
op `, so checking again at the ufunc level is redundant at best.
|
|
And the second check is just incoherent nonsense AFAICT.
|
|
The only reason to return `NotImplemented` is b c you want to pass control to another ` r)op ` method, and there s no reason arrays containing structured dtypes in particular should somehow magically have different ` r)op ` methods available than other arrays.
|
|
So we can just get rid of all the ufunc stuff immediately, great.
|
|
That leaves the ` array priority ` stuff.
|
|
We have two problems here we can t just drop this immediately b c of backcompat issues, and we need to have some way to continue to support all the use cases that this currently supports.
|
|
The first problem is just a matter of having a deprecation period.
|
|
For the second, observe that a class which defines a ` numpy ufunc ` method gets complete control over what any ufunc call does, so it has almost as much power as a class that currently sets ` array priority `.
|
|
The only additional power that ` array priority ` currently gives you is that it lets you distinguish between e.g.
|
|
a call to `ndarray.
|
|
add a, b)` versus a call to `np.add a, b)`.
|
|
So the only code that really loses out from my proposed change is code which wants `a b` and `add a, b)` to do different things.
|
|
AFAIK in the entire history of numpy there is only one situation where this power has been used on purpose the definition of matrix classes where `a b` is matmul, but `np.multiply a, b)` is elmul.
|
|
And we ve all agreed that such classes should be deprecated and eventually phased out cite PEP 465).
|
|
So, I conclude that EVENTUALLY my dream rules should work great.
|
|
The only problem is that we need some temporary compromises to get us from here to there.
|
|
Therefore, I propose we use the following dispatch rules in numpy 1.10, with the goal of moving to my dream rules in some future version First, Python uses the subclass rule to decide whether to invoke `a. op b)` or `b. rop a)`.
|
|
As above, let s assume that it invokes `ndarray.
|
|
op a, b)`.
|
|
`ndarray.
|
|
op a, b)` does the following If `b` does not define ` numpy ufunc ` and is not an `ndarray` at all even a subclass), and `b` has a higher ` array priority ` than `a`, then we issue a deprecation warning and return `NotImplemented` and let control pass to `b. rop a)`.
|
|
bolded parts are changes compared to the current behaviour) If ` op ` is ` mul ` and `b >tp class >tp name.startswith scipy.sparse.
|
|
)`, then return `NotImplemented`.
|
|
This rule is necessary in addition to the above, because `scipy.sparse` has already made a release containing ` numpy ufunc ` methods, so the exception above doesn t apply.)
|
|
Otherwise, we call `np.op a, b)` and let the ufunc machinery take over.
|
|
I believe that this is adequate to covers all practical use cases for the current dispatch machinery, and gives us a clean path to better dispatch machinery in the future.
|
|
The main alternative proposal is Pauli s, which involves a very complicated check I won t try to summarize here, see [this comment and following code] https github.com numpy numpy blob eecb2e3c07f29c0ac991d364a846a2f8293a432a numpy core src multiarray number.c L103)).
|
|
The goal of that approach is to continue supporting classes where `a b` and `add a, b)` do different things.
|
|
I don t think that keeping substantial additional complexity around indefinitely is worth it in order to support functionality that no one has ever found a use for except in one very specific case overriding ` mul `), and where we generally agree that that one specific case should be phased out as possible.
|
|
I would very much appreciate feedback from scipy.sparse and astropy in particular on whether the above covers all their concerns.
|
|
Partial) History 4815, 5748 CC pv, cowlicks, mhvk","bug"
|
|
"6013","5846","Inconsistency in setdiff1d return value dtype.
|
|
I noticed this discrepancy in the dtype of the returned array because it lead to an Implicitly casting between incompatible kinds warning down the line .
|
|
It looks like this could be easily fixed in https github.com numpy numpy blob 50e4eb87370b77792f1f6d21a5edd6ff20abee68 numpy lib arraysetops.py L478 by returning `ar1` instead of `aux`, is there another reason for returning `aux` instead?","bug"
|
|
"6013","5849","PendingDeprecationWarning in Python 3.4.
|
|
The `imp` module will be deprecated in favor of importlib.
|
|
It is used in setupegg.py runtests.py numpy import tools.py numpy testing utils.py numpy distutils misc util.py numpy core setup.py","bug"
|
|
"6013","585","Memory leak in numpy?
|
|
Trac 2125).
|
|
Original ticket http projects.scipy.org numpy ticket 2125 on 2012 05 02 by trac user eschlafly, assigned to unknown.
|
|
Under numpy 1.6.0, the following code leaks about a hundred megs of memory for me import numpy dtype = [ var1 , 100f8 )] for i in xrange 100000) arr = numpy.zeros 1, dtype=[ var1 holder , dtype)]) arr[ var1 holder ] = numpy.zeros 1, dtype=dtype)[0] Admittedly, this code should probably trigger some kind of exception anyway.
|
|
Rewriting the code slightly as below gets rid of the leak arr[ var1 holder ][0] = numpy.zeros 1, dtype=dtype) Still, I would much rather have had an exception thrown than a silent memory leak in otherwise well behaving code.","bug"
|
|
"6013","5850","UserWarning in tests.. ` usr local lib python3.4 dist packages numpy distutils cpuinfo.py 120 UserWarning [Errno 2] No such file or directory proc cpuinfo ` This may be due to the way Travis CI runs tests.","bug"
|
|
"6013","5851","Bug in np.random.dirichlet for small alpha parameters.
|
|
Hi, I encountered a bug when using np.random.dirichlet with small alpha parameters.
|
|
Call and traceback are below.
|
|
.
|
|
I am using numpy 1.9.1.
|
|
I believe this is a floating point issue, the distribution has almost all of its mass very close to either 1, 0, 0) or 0, 0, 1) .
|
|
The float division error already occurs for larger values<1, e.g.
|
|
0.001.
|
|
It is likely that this occurs because of the Dirichlet distribution is usually sampled via the Gamma distribution followed by normalization.
|
|
If all values returned from Gamma sampling are zero than a float division error occurs.
|
|
In addition .
|
|
produces nan most of the time, while it should be alternating almost always between 1, 0) and 0, 1) .
|
|
It might not be able to fix that in the current algorithmic framework but maybe it is possible to discourage prevent users from supplying too small parameters.
|
|
Wow, this wasn t supposed to become such a long post.
|
|
Thanks to anyone reading considering this issue.","bug"
|
|
"6013","5853","Equality specifier for numpy.unique.
|
|
I would like to suggest a way to specify the equality function used in `np.unique`, e.g.
|
|
with a `equality` argument typical use case `np.isclose` for arrays of floats).
|
|
True, `isclose` is not transitive but it s probably still fine.
|
|
Implementation wise, it s just a matter of replacing `aux[1 ] != aux[ 1]` l. 198) by the new equality function.","enhancement"
|
|
"6013","5856","floating point errors can deadlock when using multiple embedded sub interpreters.
|
|
If an application is using multiple embedded sub interpreters that use numpy, there is a possibility that floating point errors can cause the thread to deadlock.
|
|
I traced this down to ufunc object s error handler ) method where it does NPY ALLOW C API and that freezes on PyGILStateEnsure ).
|
|
Interestingly, the PyGILState API is allegedly unsafe for applications using sub interpreters, but it works fine as long as long as each sub interpreter is on its own thread.
|
|
There is a workaround to call numpy.seterr all= ignore ) as that will not even enter the method.
|
|
Ideally the print option would also not get the GIL as the GIL should not be needed for that.
|
|
In the meantime I added a warning to Jep when it detects an interpreter on the same thread in the hopes of deterring the conditions that can cause this.
|
|
I have a test case that illustrates the problem at https github.com mrj0 jep blob v3.3.0rc src jep test numpy TestNumpyGILFreeze.java This is somewhat related to gh 3837.
|
|
If you have problems running the test case, you may need to configure the environment variables PATH, LD LIBRARY PATH, or LD PRELOAD depending on your system.","bug"
|
|
"6013","5857","Tiny memory leak when using numpy in embedded sub interpreters.
|
|
There appears to be a tiny memory leak somewhere in numpy that shows up when using numpy in embedded sub interpreters.
|
|
I ve seen this on RHEL 5 and RHEL 6 in long running server processes, and it appears to be there in all versions of numpy though it s significantly smaller slower in newer numpy releases).
|
|
It s rather challenging to spot but I ve written a test case that will illustrate the memory climbing higher continuously.
|
|
See https github.com mrj0 jep blob v3.3.0rc src jep test numpy TestNumpyMemoryLeak.java The workaround is to eventually restart the process.
|
|
If you have trouble running the test case, you may need to configure your environment variables PATH, LD LIBRARY PATH, or LD PRELOAD depending on your system.","bug"
|
|
"6013","5859","np.dot deprecations.
|
|
[This isn t actually a 1.10 blocker, but hopefully it will be resolved trivially before that so sticking the tag on so it doesn t get lost.]
|
|
http thread.gmane.org gmane.comp.python.numeric.general 60520","bug"
|
|
"6013","586","Possibly unwanted behaviour of numpy.median when the array contains numpy.nan Trac 2126).
|
|
Original ticket http projects.scipy.org numpy ticket 2126 on 2012 05 04 by trac user koji, assigned to unknown.
|
|
Because the median function is dependent on the sort function, which places the nan entries at the end, the median function may overestimate median in an unfortunate situation.
|
|
I got really surprised to see that line 38 returned 11.5.
|
|
I was expecting either np.nan or 11.0.
|
|
Perhaps an explicit handling of np.nan either take away from the sorting to begin with) would be better, or make it return np.nan when there s one ore more nan s in the array.
|
|
It makes me wonder if anybody had tripped over this without realising it.
|
|
In [33] np.sort np.array [np.nan, 10])) Out[33] array [ 10., nan]) In [34] np.sort np.array [np.nan, 10, 11])) Out[34] array [ 10., 11., nan]) In [35] np.sort np.array [np.nan, 10, 11, 12])) Out[35] array [ 10., 11., 12., nan]) In [36] np.median np.array [np.nan, 10])) Out[36] nan In [37] np.median np.array [np.nan, 10, 11])) Out[37] 11.0 In [38] np.median np.array [np.nan, 10, 11, 12])) Out[38] 11.5 In [39] np.
|
|
version Out[39] 1.5.1 Python 2.7.1 |EPD 7.0 1 32 bit)| r271 86832, Dec 3 2010, 15 41 32)","bug"
|
|
"6013","5861","creating an array from a Pillow 1 bit image object results in junk data.. Python 3.4.2 Windows 64 bit NumPy 1.9.2 pillow 2.8.1 When `numpy.array` is called on an 1 bit Pillow image, the resulting array is filled with junk data and does not correctly represent the image object it should contain.
|
|
I ve tested this with a small image and a large image.
|
|
Small image 3x3) !
|
|
[test] https cloud.githubusercontent.com assets 4407285 7554999 853ef4c2 f73d 11e4 9747 2e9c79190344.png) Pillow represents it like this .
|
|
while the resulting numpy array varies in how it looks .
|
|
I tried this out on a bigger image represented below), and I saw some interesting patterns in the output.
|
|
The code I tested with .
|
|
Input image !
|
|
[2015 05 10 17 50 18] https cloud.githubusercontent.com assets 4407285 7554994 1031077e f73d 11e4 9923 f07bf9117c4e.png) Output image !
|
|
[2015 05 10 17 48 37] https cloud.githubusercontent.com assets 4407285 7554995 121907da f73d 11e4 8e1f f574fba57567.png) Notice the row of miniaturized versions of the input along the top.","bug"
|
|
"6013","5862","broadcast shape is broken when 32 < nargs.
|
|
xref https github.com numpy numpy pull 5371 .
|
|
In [this line] https github.com numpy numpy blob 05b5335ecf25e59477956b4f85b9a8edbdf71bcc numpy lib stride tricks.py L123) `b` is ultimately converted to an array of shape `1 x 1`; so it will broadcast with the rest of args regardless of their shape.","bug"
|
|
"6013","5867","np.roll improvements.
|
|
Two quick suggestions add an `inplace={False|True}` or perhaps just `out`, like for the ufuncs).
|
|
support shifting along multiple axes simultaneously either allow `shift` to be an array like of length equal to the array s ndim, or allow passing two sequences, `shift` and `axis`, of the same length or allow both).","enhancement"
|
|
"6013","5868","vectorize fails for zero dimensioned arrays.
|
|
The following used to work and return an empty 5 by 0 array.
|
|
.
|
|
Now it stack traces .
|
|
Seems to me that there should first be a check for empty arrays before attempting to see the return types of the function?","bug"
|
|
"6013","587","numpy.savetxt See also section needs small change Trac 2127).
|
|
Original ticket http projects.scipy.org numpy ticket 2127 on 2012 05 06 by trac user iosonofabio, assigned to pv.
|
|
In the documentation for the function numpy.savetxt, in the See also section, we currently find savez Save several arrays into a .npz compressed archive This is wrong since savez does not compressed.
|
|
I suggest to change compressed to uncompressed and add a reference to savez compressed.","bug"
|
|
"6013","5871","Inconsistency for indexing multidimensional array.
|
|
I noticed the implicit transpose when index multidimensional array.
|
|
>>> a = numpy.zeros 5,6,7,8)) >>> a[1, ,numpy.arange 3)].shape 3, 6, 8) >>> a[1][ ,numpy.arange 3)].shape 6, 3, 8) >>> a[1, , ,numpy.arange 3)].shape 3, 6, 7) >>> a[1][ , ,numpy.arange 3)].shape 6, 7, 3) But h5py and pytable do not imply this rule and I feel that s a natural way to indexing subarray).
|
|
What s the rules of the transpose for subarray?
|
|
This problem is in numpy 1.6, 1.8, and I didn t check numpy 1.9.","bug"
|
|
"6013","5875","np.testing.assert array equal prints [repr failed] on some matrices.
|
|
With numpy 1.9.2 on Ubuntu 14.04 with Python 2.7 .
|
|
It seems that 1x1 matrices work, and changing `np.matrix` to `np.array` works.
|
|
Also, manually calling `repr` on the matrix works.","bug"
|
|
"6013","5876","f2py fix extended common block.
|
|
I try to include a common f77 block from legacy code.
|
|
There once can continue a common block with a new common statement with the same common block name.
|
|
In crackfortran.py there is statement that wrongly) explicitly catches this case.
|
|
That should be removed reverted.
|
|
Example common namecom namezone nzonei nzonec nzoneb), names nsubz) common namecom unitzone nzonei nzonec nzoneb) [...hundreds more such lines...] doing this all with just line continuation is not an option, there is too many of them hence compilers have overflowed in the past beyond some 20 lines).
|
|
The patch, based on numpy 1.9.2, is .
|
|
Can you please include this patch in future versions of f2py?","bug"
|
|
"6013","5877","f2py bug in calcualtion of array dimensions.
|
|
I try to include a common f77 block from legacy code.
|
|
some calculations in the conversion go wrong.
|
|
I have in my common block lines .
|
|
resulting an a comlilation error .
|
|
it seems a bad line is generated and f2py does not resolve multiple occurrences nvar) of the same parameter variable in an expression?
|
|
at least inside sufficiently complex expressions?
|
|
If I define a parameter with nvar nvar 1) and use that new parameter, things work, but maybe if you get a chance to fix this fluke as well, that would be nice.
|
|
or let me me know in what function for crackfortran even to look...)","bug"
|
|
"6013","588","Documentation typo PyArrayDescr Check should be PyArray DescrCheck Trac 2128).
|
|
Original ticket http projects.scipy.org numpy ticket 2128 on 2012 05 07 by trac user komnomnomnom, assigned to pv.
|
|
On the c api array api page `PyArray DescrCheck` is incorrectly referenced as `PyArrayDescr Check`.
|
|
[http docs.scipy.org doc numpy reference c api.array.html?highlight=pyarraydescr check PyArrayDescr Check] .
|
|
I don t think it s mentioned anywhere else in the documentation.","bug"
|
|
"6013","5881","BUG np.broadcast does not properly handle np.broadcast objects.
|
|
This bit us with a bug in https github.com numpy numpy issues 5862 Instead they are treated as scalar arrays .
|
|
It would be nice to fix or at least document this.
|
|
cc mhvk","bug"
|
|
"6013","5883","Numpy datetime array pickling unpickling fails for array constructed from byte string .
|
|
I m working with Apache Spark to communicate arrays of various datatypes to Pyspark, amongst which timestamps.
|
|
The timestamps are communicated from Java to python as a byte strings from the raw 64 bit longs representing milliseconds since the unix epoch.
|
|
These are loaded as numpy arrays with ..
|
|
The initial communication between Java and Python works just fine.
|
|
The issues is when these numpy arrays are communicated to another Python process.
|
|
I think I ve managed to isolate the issue to pickling by numpy of the array.
|
|
I can successfully pickle and unpickle a datetime array .
|
|
I think the construction of the array from a byte string with .
|
|
isn t really the issue .
|
|
However pickling and then unpickling again doesn t .
|
|
Inspecting the `dtype` of the array shows that the type maybe wrong .
|
|
The dissected pickle looks like .
|
|
I m no pickle expert, but I do see .
|
|
indicating the byte order and .
|
|
indicating the milliseconds interpretation of the longs.
|
|
Also the string seems to be exactly with the byte string with which the array was created via ..
|
|
I m running 64 bit Fedora 21.
|
|
I m using python 2.7.8 for this and numpy version 1.9.2 git revision 762c6f15bf57506e21239a4dd68192e4da6014f7).","bug"
|
|
"6013","5885","BUG np.cross should ignore axisc if return is scalar.
|
|
When applying `cross` to two 2D vectors, the return is a scalar, so the `axisc` parameter is meaningless and should be ignored, or an error raised if it is specified.
|
|
Instead, it can alter the shape of the return .python) >>> a = np.random.rand 3, 4, 2) >>> b = np.random.rand 3, 4, 2) >>> np.cross a, b).shape 3, 4) >>> np.cross a, b, axisc=0).shape 4, 3)","bug"
|
|
"6013","5887","BLAS issues on OS X 10.10.3.
|
|
It occurs at using scikit TSNE algorthm on the following line TSNE n components=3, verbose=1).fit transform X) Break bus error Generating stack trace... [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8ee1b8ea in cblas dscal 0x2f from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8ee6418d in dtrScale 0x47 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8edf50b6 in APL dsyrk 0xa5 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8eef3696 in cblas dgemm 0x478 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x000000010a216775 in dotblas matrixproduct 0x1665 from usr local lib python3.4 site packages numpy core dotblas.so [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fb869 in PyEval EvalFrameEx 0x51fd from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099feb38 in fast function 0x141 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fb6fb in PyEval EvalFrameEx 0x508f from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x000000010998422f in function call 0x174 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109963e2a in PyObject Call 0x67 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fbdaf in PyEval EvalFrameEx 0x5743 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x000000010998422f in function call 0x174 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109963e2a in PyObject Call 0x67 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fbdaf in PyEval EvalFrameEx 0x5743 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099feb38 in fast function 0x141 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fb6fb in PyEval EvalFrameEx 0x508f from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099feb38 in fast function 0x141 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fb6fb in PyEval EvalFrameEx 0x508f from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099feb38 in fast function 0x141 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099fb6fb in PyEval EvalFrameEx 0x508f from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f64b5 in PyEval EvalCodeEx 0x656 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001099f5e59 in PyEval EvalCode 0x3f from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109a1a185 in run mod 0x3a from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109a1a424 in PyRun FileExFlags 0xb2 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109a19c31 in PyRun SimpleFileExFlags 0x386 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109a2cfcf in Py Main 0xc0b from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python 0x0000000109954e27 in <unknown> from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Resources Python.app Contents MacOS Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8739b5c9 in start 0x1 from usr lib system libdyld.dylib Break segmentation violation Generating stack trace... [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00000001144d5a2c in ZN5TCint12ResetGlobalsEv 0x34 from usr local lib root libCore.so [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000114481c25 in ZN5TROOT20EndOfProcessCleanupsEb 0x33 from usr local lib root libCore.so [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x000000011450924c in ZN11TUnixSystem4ExitEib 0x40 from usr local lib root libCore.so [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000114507aa3 in ZN11TUnixSystem15DispatchSignalsE8ESignals 0xe7 from usr local lib root libCore.so [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8a511f1a in sigtramp 0x1a from usr lib system libsystem platform.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x0000000109a0a2a1 in vgetargskeywords 0x2f0 from usr local Cellar python3 3.4.3 Frameworks Python.framework Versions 3.4 Python [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8ee1bb57 in catlas dset 0x2c from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib 0x00007fff8ee1b8ea in cblas dscal 0x2f from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8ee6418d in dtrScale 0x47 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8edf50b6 in APL dsyrk 0xa5 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x00007fff8eef3696 in cblas dgemm 0x478 from System Library Frameworks Accelerate.framework Versions A Frameworks vecLib.framework Versions A libBLAS.dylib [invalid usage] unrecognized option d Usage atos [ p pid] [ o executable] [ f file] [ s slide | l loadAddress] [ arch architecture] [ printHeader] [address ...] 0x000000010a216775 in dotblas matrixproduct 0x1665 from usr local lib python3.4 site packages numpy core dotblas.so <Here I stopped process>","bug"
|
|
"6013","5888","NumPy ABI does not successfully maintain forward compatibility should it?.
|
|
Not sure whether this is something we consider a bug, but flagging for potential discussion and so it doesn t get lost I hadn t realized until today that NumPy in practice does not provide a forward compatible ABI, i.e., if you build with numpy 1.9 then try to run against 1.8, this may not work.
|
|
Apparently packages that care about this like sklearn are actively working around this by carefully installing old versions of numpy before building wheels.
|
|
In particular, we have several times added extra fields to the `dtype` struct.
|
|
In practice this is normally fine b c no one actually accesses these fields, but Cython in particular does struct size checking.
|
|
For backwards compat build against 1.8 and then run against 1.9 the struct appears to get larger, and Cython merely issues a warning which we suppress).
|
|
For forward compat build against 1.9 and then run against 1.8 the struct appears to get smaller, and in this case Cython [issues a hard error] https github.com cython cython blob master Cython Utility ImportExport.c L403).
|
|
We could work around this by simply exposing a truncated struct to user code, so that Cython sees a small struct when doing `sizeof`, and the actual object is always larger then this, meaning that we always hit the warning path rather than the error path.
|
|
I don t know if this is the only problem we would have to fix in order to achieve forward compatibility, e.g.
|
|
I haven t checked the C API import code to see what import multiarray or import umath do when they find themselves running against an older version of numpy.
|
|
If we want to take ABI compatibility seriously I guess we should probably also start requiring C API users to explicitly state which version of the ABI they expect, and enforce that they don t get access to anything newer than that.
|
|
This would at least give us the option then in the future to provide different versions of the same function to old users and new users.","bug"
|
|
"6013","589","Segfault on umath.cpython 32mu.so Trac 2129).
|
|
Original ticket http projects.scipy.org numpy ticket 2129 on 2012 05 08 by trac user Sachiel, assigned to unknown.
|
|
On a code involving Numpy, Matplotlib and Netcdf4, I m getting a Segmentation Fault from numpy core umath.cpython 32mu.so during the execution of the pyplot.show ) funcion.
|
|
AT first I thought it was a problem from Matplotlib, but given that the segfaulted library it s from Numpy, I decided to report it here.
|
|
System Python 3.2.3 Numpy 1.6.1 Matplotlib 1.1.0 Python NetCDF4 0.9.9","bug"
|
|
"6013","5891","stdout suppression in tests.
|
|
At least one of the tests closes `sys.stdout` https github.com numpy numpy blob maintenance 1.9.x numpy lib tests test regression.py L171 but according to https github.com numpy numpy commit 3fd6b62c1db915 closing `sys.stdout` won t close it.
|
|
I m not sure what is the best way to fix it.
|
|
This issue arose in review of a scipy PR https github.com scipy scipy pull 4805 which had copied this pattern from numpy.","bug"
|
|
"6013","5894","numpy.maximum numpy.minimum fail quietly on sparse matrices.
|
|
As I understand, the numpy.maximum and numpy.minimum functions do not support sparse arrays.
|
|
However, they fail quietly , producing erroneous output when given a scipy.sparse.dok matrix .
|
|
When given a scipy.sparse.csr matrix, these functions fail with an exception, but the error is not clear .
|
|
I think an exception should be raised if the data type is not supported.","bug"
|
|
"6013","5895","float32 has priority over float64 for scalar integer ufunc inputs with other arrays.
|
|
Consider this .
|
|
Apparently, the ufunc loop selector prefers to cast integers to float32 in this case, which is somewhat surprising.
|
|
This may have to do with the fact that given an integer array as one of the first two arguments, there are still two possible loop choices, and deciding which to use would need considering appropriate scalar casting rules for the last two arguments.
|
|
Note that if one of the last arguments is an array instead of a scalar, the dtype decision is as expected, so maybe this is an issue with scalar special cases.","bug"
|
|
"6013","5896","Segfault with boolean indexing and complex arrays casting to float.
|
|
I am experiencing a tricky segfault right now.
|
|
To trigger the segfault, assign a `complex128` array to a large zeroed `float64` array using a boolean index array.
|
|
Here is a program the reproduces the behavior .
|
|
Running gives .
|
|
Assigning a float with the boolean index array, and assigning a complex number with fancy indexing works although you get a `ComplexWarning` because of casting).
|
|
However, assigning the complex array to the float array using the boolean indexing array causes a segfault.
|
|
Here is where it gets weird.
|
|
If I set the `shape` parameter to anything less than ` 8, 63)` that is, if I make 8 smaller or 63 smaller), the segfault doesn t occur.
|
|
So it seems to me that the shape of the boolean and zero arrays must have at least 8 rows and at least 63 columns for some reason.
|
|
I have no idea why this is the case, but it may be a clue for what the problem is.
|
|
The size of the complex array doesn t matter, as long as it fits in the float array, of course.","bug"
|
|
"6013","5897","Is there a list of functions which do don t release the GIL?.","bug"
|
|
"6013","5898","Numpy inverse of very large matrix returns all zero matrix without error.
|
|
I have been experimenting with large matrix inversions have to inverse the whole matrix is my specific case) to check the runtime.
|
|
It all works well until I tried to inverse a 50000 by 50000 matrix .
|
|
The run time of 6 seconds is very small.
|
|
But more importantly I should not have the inv ) function to return an all zero matrix without any error!
|
|
Am I missing something here?
|
|
Is it a result I should be expecting?
|
|
About my environment Right before running the snippet above, I made sure I had enough RAM .
|
|
Running Conda 3.12.0 with numpy 1.9.2 py27 p0 [mkl] OS Linux seftrtools1 3.16.0 34 generic 47 14.04.1 Ubuntu SMP Fri Apr 10 17 49 16 UTC 2015 x86 64 x86 64 x86 64 GNU Linux .
|
|
.","bug"
|
|
"6013","590","array ) with subok=True broken with subclass in nested sequence Trac 2130).
|
|
Original ticket http projects.scipy.org numpy ticket 2130 on 2012 05 11 by dopplershift, assigned to unknown.
|
|
The following works fine import numpy as np a = np.ma.array [1,2,3]) print a masked array data = [1 2 3], mask = False, fill value = 999999) print np.array a, subok=True) masked array data = [1 2 3], mask = False, fill value = 999999) This is broken print np.array [a], subok=True) array [[1, 2, 3]])","bug"
|
|
"6013","5900","numpy.diff docstring does not match operation for nth order differences.
|
|
An [nth order difference] http en.wikipedia.org wiki Finite difference Higher order differences) refers to a difference scheme which converges to the continuous difference as the nth power of the discrete mesh size used.
|
|
For n greater than 1, what numpy.diff calculates is not the nth order finite difference, but rather a first order forward approximation to the nth difference of the series.
|
|
<code>import numpy as np a = np.array [3,2,1,0]) np.diff a, n=2) Out[65] array [0, 0])< code> Expected <code>Out[65] array [1,1])< code> I don t suggest changing the code to match the docstring, rather, the docstring should probably be changed.
|
|
It currently reads Calculate the n th order discrete difference along given axis.
|
|
The first order difference is given by out[n] = a[n 1] a[n] along the given axis, higher order differences are calculated by using diff recursively.
|
|
Parameters a array like Input array n int, optional The number of times values are differenced.
|
|
axis int, optional The axis along which the difference is taken, default is the last axis.
|
|
Returns diff ndarray The n order differences.
|
|
The shape of the output is the same as a except along axis where the dimension is smaller by n. It should probably read Calculate the n th discrete difference along given axis.
|
|
The first difference is given by out[n] = a[n 1] a[n] along the given axis, higher differences are calculated by using diff recursively.
|
|
Parameters a array like Input array n int, optional The number of times values are differenced.
|
|
axis int, optional The axis along which the difference is taken, default is the last axis.
|
|
Returns diff ndarray The n th differences.
|
|
The shape of the output is the same as a except along axis where the dimension is smaller by n. By removing all instances of the word order , the docstring would no longer misleadingly use nth order , which traditionally refers to the rate of convergence of a differencing scheme, and not which difference is being calculated.","bug"
|
|
"6013","5901","Mergesort not sorting stable.
|
|
Example .
|
|
Result array [ 1, Anton ), 1, Paul ), 3, Herbert ), 3, Otto ), 4, Karl ), 5, Bernd ), 8, Alfred )], dtype=[ number , <i8 ), name , S25 )]) Shouldn t 3, Otto ) placed before 3, Herbert ) in this example?
|
|
Or where am I wrong?","bug"
|
|
"6013","5902","savez needs a lot more memory than save.
|
|
Hello numpy 1.7.1, Python 2.7.2 default, Jun 12 2011, 14 24 46) [MSC v.1500 64 bit AMD64)] on win32.
|
|
Definitely running 64 bit as ctypes.c size t==ctypes.c ulonglong.
|
|
.
|
|
By my reckoning an array of 30,000 x 30,000 float64 should take 7.2GB, and the task manager confirms this.
|
|
Before saving I have 21GB free, so there shouldn t be a problem compressing this data.
|
|
`numpy.save foo ,a)` works, though `numpy.savez` fails.","bug"
|
|
"6013","5903","round incorrect result on int64.
|
|
Using python 3.4.3 anaconda 2.1.0 64 bit on Ubuntu 14.04 NumPy version 1.9.2 Python version 3.4.3 |Anaconda 2.1.0 64 bit)| default, Mar 6 2015, 12 03 53) [GCC 4.4.7 20120313 Red Hat 4.4.7 1)] correct >>> numpy.round 1432116610123450000, 7) 1432116610120000000 incorrect >>> numpy.round 1432116610123450000, 6) 1432116610123000064","bug"
|
|
"6013","5904","DOC Return type of eig may be real .
|
|
The documentation for [linalg.eig] http docs.scipy.org doc numpy reference generated numpy.linalg.eig.html numpy linalg eig) claims that The resulting array will be always be of complex type.
|
|
This is not true .
|
|
In fact, the result is always cast to real whenever imaginary part is exactly zero.","bug"
|
|
"6013","5905","ENH matrix rank should support stacks of matrices.
|
|
Several of the linear algebra routines listed above are able to compute results for several matrices at once, if they are stacked into the same array.
|
|
[Linalg documentation] http docs.scipy.org doc numpy reference routines.linalg.html linear algebra on several matrices at once)) However, `matrix rank` is limited to a single matrix even though it is implemented using `linalg.svd` which supports stacks of matrices.","enhancement"
|
|
"6013","5906","lapack lite should use 64 bit integer indices.
|
|
Currently lapack lite uses `int` for indices, so large arrays cause overflow and things crash e.g.
|
|
5898).
|
|
There s a separate need to handle this issue when using real BLAS LAPACK which may or may not handle arrays with >2 31 elements, depending on the vendor), but when falling back on our built in code there s no reason we shouldn t get this right.","bug"
|
|
"6013","5907","Einsum produces incorrect result when casting to non contiguous out array.
|
|
Like so .
|
|
This seems to be an iterator problem.
|
|
It works correctly in the non casting case.","bug"
|
|
"6013","5908","documentation of limitations of random integer sampling.
|
|
According to https github.com numpy numpy issues 4085 numpy cannot sample integers larger than 32 bits on systems where `long` is 32 bits 64 bit Windows but not 64 bit Linux or Mac), causing complications like https github.com scipy scipy issues 4552 which attempts to sample the sparsity structure of a large sparse matrix.
|
|
This limitation could be documented more clearly on pages like http docs.scipy.org doc numpy reference routines.random.html or http docs.scipy.org doc numpy reference generated numpy.random.randint.html.","bug"
|
|
"6013","591","PyArray TakeFrom does not release GIL Trac 2131).
|
|
Original ticket http projects.scipy.org numpy ticket 2131 on 2012 05 11 by fengy research, assigned to unknown.
|
|
PyArray TakeFrom does not release the GIL in the lengthy memmove.
|
|
It shall probably release the GIL after all buffers are copied, and re obtain the GIL before raising the exception in raise mode.","bug"
|
|
"6013","5915","ImportError dynamic module does not define module export function PyInit .multiarray) with Python 3.5 Beta.
|
|
There seems to be an issue with running Numpy latest, 1.9.2) using the just released Python 3.5 beta 1 https www.python.org downloads release python 350b1 ).
|
|
Below is the issue I seem to be running into .
|
|
I m running Arch Linux 64 bit on kernel 4.0.1, and have built a fresh Python 3.5 manually build log at http sprunge.us gUab).
|
|
I have attempted to install numpy both through pip and manually, but both run into the same error.
|
|
The manual build log is located at http sprunge.us UZBV.
|
|
A friend, also running Arch Linux 64 bit, has confirmed he receives the same error when using Python 3.5, and I have reproduced the issue on a Gentoo Linux 64 bit box as well.
|
|
If it helps, the relevant .so file exists .","bug"
|
|
"6013","5917","np.can cast 127, np.int8) is False??.
|
|
On a 32 bit Linux with Python 2.7 and numpy 1.9.2, `np.can cast 127, np.int8)` unexpectedly returns False.
|
|
Puzzlingly, it s the only value that fails `np.can cast 126, np.int8)` and `np.can cast 128, np.int8)` return True) and it only happens on 32 bit.","bug"
|
|
"6013","5918","F CONTIGUOUS loss during array column slicing.
|
|
Hello, my current version of numpy is 1.9.2 .
|
|
It is not clear to me why I lose F CONTIGUOUS in this case.
|
|
Because of that later In my script I have a problem .
|
|
when I do something like this .
|
|
It seems that it is quite natural that F CONTIGUOUS should be preserved in case I slice by column.
|
|
And that is the case when one slices more than one column .
|
|
I found workaround for this behaviour with reshape .
|
|
And also in my case `y` should be a view of an array not a copy.
|
|
I would highly appreciate any clarifications about this slicing behaviour.","bug"
|
|
"6013","5919","numpy float multiplied by a string does not raise a TypeError.
|
|
This illustrates the issue .
|
|
The same construction using a regular python float raises a `TypeError`, as I would expect .
|
|
I haven t looked very closely at the implementation of the numpy float class.
|
|
Please feel free to close this issue if this is working as expected.
|
|
This is on numpy 1.9.2 and python 2.7.9","bug"
|
|
"6013","592","Offline documentation search triggers JS error in Chrome Trac 2132).
|
|
Original ticket http projects.scipy.org numpy ticket 2132 on 2012 05 15 by trac user talcite, assigned to unknown.
|
|
When entering a search query the offline documentation in chrome, a JS error is triggers in the developer console and the search does not complete.
|
|
Actual behavior waits forever) The error in the developer console is XMLHttpRequest cannot load file home <scrubbed> Downloads numpy html searchindex.js.
|
|
Origin null is not allowed by Access Control Allow Origin.
|
|
There is no problem in Firefox with this JS error.","bug"
|
|
"6013","5922","ufunc.at perfomance >10x too slow.
|
|
I have created a Matlab like [accumarray] http uk.mathworks.com help matlab ref accumarray.html?refresh=true) function which tries to squeeze as much performance form numpy as possible for a specific list of functions `sum any all max min mean...`etc.
|
|
The functions is available as a gist [here] https gist.github.com d1manson 5f78561c0f52d3073fe8).
|
|
There is another accumarray implementation available on github [here] https github.com ml31415 accumarray), which was originally created by ml31415, but over the last couple of days has had many of my suggestions incorporated I have also updated my own gist in line with my recent suggestions).
|
|
The original purpose of this other implementation was to use `scipy.weave` to get massively improved performance over what was assumed to be the best possible raw numpy version.
|
|
However it appears that most of the functions can be fairly heavily optimized without `scipy.weave`.
|
|
[It s not really important here, but for reference, the remaining functions which are difficult to optimize are `min max` and `prod`.]
|
|
The main point , however, is that it should be simple to optimize by just using `ufunc.at` for the relevant `ufunc` as this is exactly what `ufunc.at` is intended for), yet `ufunc.at` gets miserable performance....about 15x slower than `scipy.weave` and 10 25x slower than a bunch of carefully written alternative numpy algorithms where such optimized algorithms exist).
|
|
Surely `ufunc.at` could be improved?
|
|
Also, and on a separate note, would there be any interest in including accumarray itself in numpy?
|
|
Here are some benchmarking stats produced by my function testing and benchmarking code is inclued at the bottom of the [gist] https gist.github.com d1manson 5f78561c0f52d3073fe8)).
|
|
Note that the `baseline` times are obtained by sorting spliting and looping, using the named numpy function for each group; whereas the optimised functions do some kind of handcrafted vectorised operation in most cases, except `max min` and `prod` which use `ufunc.at`.
|
|
Note also that the actual observed speedup depends on a variety of properties of the input.
|
|
Here we are using `100 000` indices uniformly picked from the interval `[0, 1000)`.
|
|
Specifically, about 25 of the values are `0` for use with the bool tests), the remainder are uniformly distributed on `[ 50,25)`.
|
|
.","enhancement"
|
|
"6013","5924","Array contiguous change during np.nan to num.
|
|
Hello, my current version of numpy is 1.9.2 .
|
|
It is not clear to me why CONTIGUOUS flags of the array `y` are different from ones of `x`.
|
|
I looked at the code in type check.py on the master branch, and it is different from type check.py on my machine.
|
|
So probably it had been already fixed.
|
|
But still, could anyone point me to a good explanation what the `subok` argument in array creation function is?
|
|
I would highly appreciate any clarifications about this nan to num behaviour.","bug"
|
|
"6013","5926","Indicial notation wrapper for numpy.einsum?.
|
|
Would it be interesting to have a wrapper class that let you write indicial notation expressions?
|
|
It looks doable.
|
|
I made this code work a = WrappedTensor np.r [ 4].reshape 2,2).astype float)) b = WrappedTensor [[1,0.1],[0.11,1.5]]) i,j,k,l= ijkl Note .tensor returns the ndarray assert a[i,i].tensor == np.trace a.tensor) assert np.all a[j,i].tensor == a.tensor.T) Regular dot product assert np.all np.dot a.tensor, b.tensor) == a[i,j] b[j,k]).tensor) assert np.all a[i,j] a[j,i]) 2.0).tensor == a.tensor a.tensor.T) 2.0) assert np.all a[i,j] KroneckerDelta 2)[j,k] b[k,l]).tensor == a[i,j] b[j,k]).tensor) x = WrappedTensor [1.0,0,0]) y = WrappedTensor [0,1.0,0]) z = WrappedTensor [0,0,1.0]) assert np.all x[i] y[j] LeviCivita[i,j,k]).tensor == [0,0, 1]) assert np.all x[k] y[j] LeviCivita[i,j,k]).tensor == [0,0, 1]) assert np.all x[i] x[j] y[i] y[j] z[i] z[j]).tensor == np.eye 3)) I d be happy to share my code.","bug"
|
|
"6013","5927","Incomplete documentation for sort ) and friends order argument.
|
|
In the documentation of `numpy.sort )` and friends `numpy.recarray.sort )`, `numpy.ndarray.sort )`, at least), the `order` parameter is lacking two pieces of information The current documentation mentions that it is a list , but the example is with a string record name) the documentation should state that a string is also accepted.
|
|
The current documentation does not describe what is the default behavior of `order` when it is not specified.
|
|
I guess that the lexicographic order starting from the first field) on the records is used?
|
|
or maybe only the first record field is sorted?
|
|
It would be useful to document this so that user know what they are getting, in this important default) case.","bug"
|
|
"6013","593","overflow not caught on operators with int32 array Trac 2133).
|
|
Original ticket http projects.scipy.org numpy ticket 2133 on 2012 05 16 by trac user mwtoews, assigned to unknown.
|
|
I see good overflow warnings operations on int32 scalars, but not for int32 arrays import numpy as np Case 1 Good, proper data types are used to avoid overflow np.array [1], dtype=np.long) np.int32 2 31 1) array [2147483648], dtype=int64) Case 2 Bad, overflow happens and no warning raised np.array [1], dtype=np.int32) np.int32 2 31 1) array [ 2147483648]) Case 3 Similar bad, array vs. array np.array [1], dtype=np.int32) np.array [2 31 1], dtype=np.int32) array [ 2147483648]) Case 4 Better, a warning is raised np.int32 1) np.int32 2 31 1) main 1 RuntimeWarning overflow encountered in long scalars 2147483648 Here is what I can determine between operator op between arrays or scalars 1. array long op scalar 32 = array long, all good 2. array 32 op scalar 32 = array 32, no overflow warning!
|
|
3. array 32 op array 32 = array 32, no overflow warning!
|
|
4. scalar 32 op scalar 32 = scalar 32, helpful overflow warning I would expect an overflow runtime warning for cases 2 and 3, similar to case 4.
|
|
These results are with NumPy 1.6.1 Python 2.5.1 on Windows 32 bit, which I installed from `numpy unoptimized 1.6.1.win32 py2.5.exe` from [http www.lfd.uci.edu gohlke pythonlibs numpy gohlke], but are reproducible on 64 bit Linux.","bug"
|
|
"6013","5932","numpy.allclose does not work as expected for unsigned dtypes.
|
|
Here is a brief example .
|
|
This is using version 1.9.1","bug"
|
|
"6013","5933","Enh Object array creation function.
|
|
As discussed in issue 5303, currently it is not possible to create arrays of object dtype containing equal length sequences, since the sequence is automatically read in as array elements.
|
|
There is a suggestion to only do this for lists, but this would be a major backwards compatibility break and would require a long deprecation period.
|
|
Another approach would be to have a function explicitly for creating arrays with an object dtype.
|
|
Perhaps this could be called objectarray .
|
|
The default for this function would be to take in a sequence, and consider each element of the sequence as an element in a 1D object array.
|
|
The function, however, could have an optional ndim or depth argument, that could be used to specify how many levels of the sequence should be considered part of the array.
|
|
This would default to 0 only the outermost level is considered).
|
|
This would raise an exception if the dimensions don t match.
|
|
Note that this approach is not mutually exclusive with the alternative, but has the advantage that it wouldn t break backwards compatibility.
|
|
So for example >>> arr = objectarray [ 1, 2, 3), 4, 5, 6)), 7, 8, 9), 10, 11, 12))]) >>> arr array [ 1, 2, 3), 4, 5, 6)), 7, 8, 9), 10, 11, 12))], dtype=object) >>> arr.shape 2,) >>> arr = objectarray [ 1, 2, 3), 4, 5, 6)), 7, 8, 9), 10, 11, 12))], depth=1) >>> arr array [[ 1, 2, 3), 4, 5, 6)], [ 7, 8, 9), 10, 11, 12)]], dtype=object) >>> arr.shape 2, 2) >>> arr = objectarray [ 1, 2, 3), 4, 5, 6)), 7, 8, 9), 10, 11, 12))], depth=2) >>> arr array [[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]], dtype=object) >>> arr.shape 2, 2, 3)","bug"
|
|
"6013","5934","ENH rec.fromrecords accept dict.
|
|
It is currently relatively complicated to create a structured or record array from a `dict`.
|
|
This simplest approach I can find, `numpy.rec.fromrecords`, still requires a fair amount of boilerplate data={1 [1,2,3], 2 [4,5,6]} np.rec.fromrecords list data.values )), names=list data.keys ))) I suggest allowing `numpy.rec.fromrecords` to allow using a dict directly data={1 [1,2,3], 2 [4,5,6]} np.rec.fromrecords data) Alternatively, there could be separate convenience functions to construct record arrays and structured arrays from dicts or sequences of key value pairs.","enhancement"
|
|
"6013","5935","ndarray does not accept memoryview for buffer= under python2.
|
|
Hello up there.
|
|
I ve discovered that `ndarray buffer=memoryview)` fails on python 2 .
|
|
Creating ndarray backed by memoryview fails .
|
|
though it succeeds if we pass memoryview to array ) .
|
|
[7] works on python 3, of course.
|
|
Thanks beforehand, Kirill","enhancement"
|
|
"6013","5936","AttributeError when splitting empty Pandas DataFrame using numpy.array split.
|
|
I use numpy.array split to split [Pandas DataFrame] http pandas.pydata.org pandas docs version 0.16.1 dsintro.html?highlight=dataframe dataframe) into list of smaller dataframes.
|
|
I use Pandas version.
|
|
When I use numpy version 1.8.2, everything seems OK.
|
|
But when I use numpy version 1.9.2, I got the following error .
|
|
Here is code to reproduce that error .
|
|
I use Ubuntu 14.04.","bug"
|
|
"6013","5937","linspace should give a Warning or Exception when num is negative.
|
|
I just finished tracking down a bug that would have been much easier to find if `linspace` did not just give back an empty array whan `num` is negative.
|
|
An exception or at least a warning to be backwards compatible) would have been nice to have.","enhancement"
|
|
"6013","5938","Bug when filling an ndarray valued field of a Structured Array.","bug"
|
|
"6013","5939","Filling an array valued field of a Structured Array.
|
|
After several days of debugging an application, I determined that the source of my problem was related to an unexpected behavior bug?)
|
|
when assigning ndarray values to structured arrays.
|
|
Depending on how the assignment is expressed, only the first entry in the array gets updated.
|
|
I am using numpy version 1.9.2.
|
|
Here is a snippet to demonstrate the issue.
|
|
.
|
|
Here is the output.
|
|
Notice the last assignment `x[3][ a ] = b[3]` did not fill in all three components.
|
|
.","bug"
|
|
"6013","594","slicing and write Trac 2134).
|
|
Original ticket http projects.scipy.org numpy ticket 2134 on 2012 05 16 by trac user vascot, assigned to unknown.
|
|
Writing sliced arrays using Pythons buildin open function does not work as expected.
|
|
The wrong slice is written.
|
|
Example code import numpy as np foo = np.fromfunction lambda x,y,z x 10 y 100 z, 3,3,3)) print foo[ , ,2] f = open tmp , wb ) f.write foo[ , ,2]) f.close ) print np.fromfile tmp ) Gives output [[ 200.
|
|
210.
|
|
220.]
|
|
[ 201.
|
|
211.
|
|
221.]
|
|
[ 202.
|
|
212.
|
|
222.]]
|
|
[ 200.
|
|
10.
|
|
110.
|
|
210.
|
|
20.
|
|
120.
|
|
220.
|
|
1.
|
|
101.]
|
|
The second array should contain the same items as the first.
|
|
A working example is first copying the slice import numpy as np foo = np.fromfunction lambda x,y,z x 10 y 100 z, 3,3,3)) print foo[ , ,2] f = open tmp , wb ) f.write foo[ , ,2].copy )) f.close ) print np.fromfile tmp ) Output [[ 200.
|
|
210.
|
|
220.]
|
|
[ 201.
|
|
211.
|
|
221.]
|
|
[ 202.
|
|
212.
|
|
222.]]
|
|
[ 200.
|
|
210.
|
|
220.
|
|
201.
|
|
211.
|
|
221.
|
|
202.
|
|
212.
|
|
222.]
|
|
As expected.","bug"
|
|
"6013","5942","`numpy.bool True) is True` evaluates to `False`.
|
|
Even though `numpy.bool True) == True` evaluates to `True`.
|
|
Doesn t this violate the [PEP8] http legacy.python.org dev peps pep 0008 other recommendations) recommendation that ``Comparisons to singletons like None should always be done with is or is not, never the equality operators.`?
|
|
Sorry, if this is a dupe.
|
|
I searched for bool comarison and pep8 and could not find an relevant issue.
|
|
`numpy.
|
|
version ` 1.9.0.dev 297f54b","bug"
|
|
"6013","5945","Warning message for one function is used by several, and contains extraneous information.
|
|
In [this Stackoverflow question] http stackoverflow.com questions 30679192 how to remove multiple values from an array at once ) the inquirer points out the following discrepancy in an error message In [5] np.delete x, x == a) usr lib python2.7 dist packages numpy lib function base.py 3254 FutureWarning in the future insert will treat boolean arrays and array likes as boolean index instead of casting it to integer of casting it to integer , FutureWarning) Since the warning is produced by a call to `delete )` it would be better if the message didn t refer to a specific function.
|
|
Also, why does the string ` of casting it to integer ` appear at the end of the warning message?","bug"
|
|
"6013","5949","When np.arange is called with certain parameters the output array is unexpected!.
|
|
The problem is reproducable on two different machines x) Windows 7, Python 3.4.3, numpy 1.9.2 x) Centos 6, Python 2.6.6, numpy 1.4.1 When np.arange is called with start parameter 0.5 and stop parameter 1.1 the returned array includes one more value than expected.
|
|
Any other combinations of start and stop parameters I tried so far deliver the expected output.
|
|
Here are two examples of the concrete problem, which are also shown in the screenshot attached.
|
|
np.arange 0.5, 1.1, 0.1) returns array [ 0.5, 0.6, 0.7, 0.8, 0.9, 1. , 1.1]) np.arange 0.5, 1.1, 0.2) returns array [ 0.5, 0.7, 0.9, 1.1]) !
|
|
[python numpy arange bug] https cloud.githubusercontent.com assets 4442327 8023256 ed7cd11e 0d02 11e5 90ff 87d9bb882ae4.png)","bug"
|
|
"6013","595","nditer memory leak with negative strides and UPDATEIFCOPY Trac 2135).
|
|
Original ticket http projects.scipy.org numpy ticket 2135 on 2012 05 18 by mwiebe, assigned to unknown.
|
|
Numpy uses its base attribute in two different ways, both to manage the memory of an external data buffer, and to track an array for UPDATEIFCOPY purposes.
|
|
The nditer was creating a temporary view with negative strides to match the strides of input arrays, but then also wanted to use UPDATEIFCOPY.
|
|
Unfortunately this choice to overload base for both purposes prevents that from working, and the result was a memory leak.
|
|
This bug was found from reviewing Nathaniel s pull request https github.com numpy numpy pull 280.","bug"
|
|
"6013","5951","Support zip input in savetxt in Python 3.
|
|
I and I think a number of other people, judging from code I ve seen) have often made use of the following way of calling savetxt in Python 2 .
|
|
but in Python 3 this now raises an error because ``zip`` is now an iterator .
|
|
In order to minimize pain when upgrading from Python 2 to Python 3, I would like to suggest that we explicitly check whether the data argument is a ``zip`` instance or even better an iterable object).
|
|
I don t think this would be ambiguous, and it would be useful in any case to continue supporting the above syntax.
|
|
If there is support for this, I can open a pull request.","enhancement"
|
|
"6013","5954","How to limit cross correlation window width in Numpy?.
|
|
I am learning numpy scipy, coming from a MATLAB background.
|
|
The xcorr function in Matlab has an optional argument maxlag that limits the lag range from xe2x80x93maxlag to maxlag.
|
|
This is very useful if you are looking at the cross correlation between two very long time series but are only interested in the correlation within a certain time range.
|
|
The performance increases are enormous considering that cross correlation is incredibly expensive to compute.
|
|
What is troubling me is that numpy.correlate does not have a maxlag feature.
|
|
This means that even if I only want to see correlations between two time series with lags between 100 and 100 ms, for example, it will still calculate the correlation for every lag between 20000 and 20000 ms which is the length of the time series).
|
|
This theoretically) gives a 200x performance hit!
|
|
Is it possible that I could contribute this feature?
|
|
I have introduced this question as a [scipy issue] https github.com scipy scipy issues 4940) and on the [scipy dev list] http mail.scipy.org pipermail scipy dev 2015 June 020757.html).
|
|
It seems the best place to start is with numpy.correlate, so that is what I am requesting.
|
|
I have done a [simple implementation] https gist.github.com bringingheavendown b4ce18aa007118e4e084) which gives 50x speedup under [my conditions] https github.com scipy scipy issues 4940 issuecomment 110187847).
|
|
This is my first experience with contributing to open source software, so any pointers are appreciated.
|
|
Other issues related to correlate functions include [ENH Fold fftconvolve into convolve correlate functions as a parameter 2651] https github.com scipy scipy issues 2651), [Use FFT in np.correlate convolve?
|
|
Trac 1260) 1858] https github.com numpy numpy issues 1858), and [normalized cross correlation Trac 1714) 2310] https github.com numpy numpy issues 2310).","enhancement"
|
|
"6013","5957","Cannot subscribe to mailing list read direct archives.
|
|
The [links for Archives and Subscribe] http www.scipy.org scipylib mailing lists.html) for the numpy discussion list yield cannot connect to server .","bug"
|
|
"6013","596","PyArray Diagonal view transition for 1.9 Trac 2136).
|
|
Original ticket http projects.scipy.org numpy ticket 2136 on 2012 05 19 by njsmith, assigned to unknown.
|
|
Originally this was scheduled to happen in 1.8, but 1.7 was released 2013 02 10, and then we re accelerating the release schedule so that 1.8 will be only a few months later which seems too soon to actually take the next step in this deprecation plan.
|
|
Current plan make the changes below in whichever release comes on or after ??
|
|
to be determined on mailing list) Starting in 1.<?>, PyArray Diagonal is supposed to start returning a non writeable view.
|
|
To do Make the trivial) change at the bottom of PyArray Diagonal Update the numpy.diagonal documentation Make a note in the release notes Optionally, remove NPY ARRAY WARN ON WRITE flag.
|
|
If so, see the below pull request to find the code to change.
|
|
Even if we decide to keep it around for a rainy day, the message in array might be written should have the references to diagonals removed.
|
|
File a new ticket to make the returned array writeable, at some future date.
|
|
Reference https github.com numpy numpy pull 280","bug"
|
|
"6013","5961","BUG object dtype quicksort can segfault.
|
|
Running the example from 3879 with current master segfaults .
|
|
This happens because of the changes introduced in 5670.
|
|
We now use the same quicksort algorithm for objects as for other numeric types, instead of the stdlib s `qsort`.
|
|
Our algorithm sorts the first, last and middle items of the array, both to choose the middle one as the pivot, and to have the first and last items as guards against run away pointers, which lets us avoid extra checks in the innermost loop.
|
|
Because on failure `PyObject Compare` returns 1, and we do no checking if an error was raised in `OBJECT compare`, these end guards do not serve their purpose anymore, hence the runaway pointer and the segfault.
|
|
I see two potential solutions 1.
|
|
Modify `npy quicksort` s inner loop to do explicit inbounds checks before dereferencing a pointer.
|
|
2.
|
|
Modify `OBJECT compare` to check if `PyErr Occurred`, and return 0 rather than 1.
|
|
This would unify the behavior with Python 3, where `PyObject RichCompareBool` is called with `Py LT`, then with `Py GT`, and if none succeeds it returns 0 without explicitly checking with `Py EQ`.
|
|
I prefer option 2.","bug"
|
|
"6013","5969","np.ma.median returns masked array; doesn t match example in docs.
|
|
From the examples .
|
|
On CentOS 6 64 with a conda python and numpy 1.9.2 .
|
|
In addition to not matching the Example, I think this is also inconsistent with the return type text and with the behavior of other masked aggregate functions np.ma.mean) which seems to just return a np.float64 on the same `x`.","bug"
|
|
"6013","597","cov fails with complex data Trac 2137).
|
|
Original ticket http projects.scipy.org numpy ticket 2137 on 2012 05 20 by trac user mph, assigned to unknown.
|
|
The cov function fails with complex data due to line 1990 in function base.py X = array m, ndmin=2, dtype=float) For example, with a = np.array 1, 2, 3), 1j, 2j, 3j))) np.cov a) gives array [[ 1., 0.
|
|
], [ 0., 0.]])
|
|
whereas the correct answer is array [[ 1., 1.j], [ 1.j, 1.]])","bug"
|
|
"6013","5972","numpy.full does not fill all types of data.
|
|
doesnt work v= numpy.full shape, [], dtype= object ) works v=numpy.empty shape, dtype= object ) v.fill [])","bug"
|
|
"6013","5974","Unable to set fill value for c type.
|
|
Unable to set fill value for string instantiated with c char, but no problem with string instantiated with S1 .
|
|
Applies to version 1.9.2 and the current code in the repository.
|
|
To reproduce, run the following lines of code.
|
|
The first passes set fill value passes while the second fails.
|
|
.
|
|
.
|
|
The problem originates in lines 424 to 426 in numpy ma core.py copied below .
|
|
When a 1 length string dtype is instantiated with c , the char is c .
|
|
When a 1 length string type is instantiated with S1 , the char is S .
|
|
Otherwise the dtypes are the same.
|
|
The char value however is not correctly identified as a string on line 424.
|
|
If this makes sense, I recommend adding c to the OSVU or using the kind property of dtype.","bug"
|
|
"6013","5975","BUG numpy.asarray does not work with large lists.
|
|
numpy.asarray throws an exception with lists with length greater than numpy.iinfo np.int32).max.
|
|
Try the following, but mind that much memory is needed .
|
|
The output on my system .
|
|
Below are some version info Python 3.4.3 NumPy 1.9.2 OS linux, 64 Bit","bug"
|
|
"6013","5976","NpzFile and memmap arrays.
|
|
Hi, `numpy.load` with memory mapped file support is great, however the argument `mmap mode` is silenty ignored when the file being loaded is anything else other than a `.npy` file.
|
|
I need to combine `numpy.savez compressed` with loading directly to mem mapped files.
|
|
`numpy.load` returns an `NpzFile` when `.npz` files are loaded.
|
|
`NpzFile` is hard coded to load data directly to numpy arrays.
|
|
Is it a possibility to pass the `mmap mode` argument to the initialization of the `NpzFile` and have it restore arrays to a memory mapped file?","enhancement"
|
|
"6013","5979","seg fault in concatenate.
|
|
I m running numpy 1.9.1 on a X86 64 machine tried more than one with slightly different builds) The following code causes a segfault import numpy as np x=np.arange 1,100) np.concatenate x,100) I know that the last line is wrong but it should bitch at me rather than segfaulting!!
|
|
!","bug"
|
|
"6013","598","numpy.interp does not release the GIL Trac 2138).
|
|
Original ticket http projects.scipy.org numpy ticket 2138 on 2012 05 22 by fengy research, assigned to unknown.
|
|
numpy.interp does not release the GIL.
|
|
The function is implemented in ` compiled base.c` This is seems to be related to http projects.scipy.org numpy ticket 80, which is closed 6 years ago before interp gets into numpy?)
|
|
Anyhow, it appears to me it is safe to release the GIL at line 535 before the slope calculation starts and re grab the GIL at line 553 before the destruction of all allocated objects.","bug"
|
|
"6013","5984","Fix relaxed stride checking faux stride for 1.10 release..
|
|
Stride in mainline currently ridiculous to smoke out problems.","bug"
|
|
"6013","5986","Drop index from numpy ufunc signature?
|
|
out always a tuple?.
|
|
This is a copy of https github.com numpy numpy pull 4753 issuecomment 113632983 by njsmoth in 4753, which I think has broader scope than the PR there OK, going to say something unpopular.
|
|
Certainly I m annoyed at myself for saying it, but nonetheless...) I ve been debating this with myself for a while, and I ve come to the conclusion that we should bite the bullet and drop the index argument from numpy ufunc s signature.
|
|
This will require some care to avoid breaking already released versions of scipy.sparse lesson learned we should never have released scipy.sparse with numpy ufunc before it came out in numpy), but that s better than leaving it forever.
|
|
Rationale the index argument is never useful.
|
|
The only time it s useful at all is for classes that want to cast self to ndarray and recurse, and even in those cases they can just as well do a loop and handle all objects of their type at once.
|
|
It ll be faster too.
|
|
None of the example classes we ve written during the binop discussion have used it at all.
|
|
the impossibility of having a useful value for output arguments makes there index even more useless.
|
|
Sure we can set it to n > nin, but what s the use of this?
|
|
Writing code to check for this and do something sensible is more complicated than just finding the array by hand.)
|
|
there are almost certainly other arguments we should dispatch on, like where=.
|
|
What do we set the index to in this case?
|
|
Sorry.
|
|
I m the one who suggested adding it in the first place, but it was a bad idea Also while we re at it, we should make the out= argument unconditionally a tuple, rather than sometimes a single ndarray and sometimes a tuple of them.
|
|
This would simplify argument processing see all the times we wrote if not isinstance out, tuple) out = out,) on the example binop implementations).","enhancement"
|
|
"6013","599","Named dtype array Difference between a[0][ name ] and a[ name ][0] Trac 2139).
|
|
Original ticket http projects.scipy.org numpy ticket 2139 on 2012 05 22 by trac user bmu, assigned to unknown.
|
|
I came across a question on stackoverflow http stackoverflow.com q 9470604) and I am wondering if this is a bug import numpy as np dt = np.dtype [ tuple , int, 2))]) a = np.zeros 3, dt) type a[ tuple ][0]) ndarray type a[0][ tuple ]) ndarray a[ tuple ][0] = 1,2) ok a[0][ tuple ] = 1,2) ValueError shape mismatch on array construction Travis wrote that this something like a design bug in a post at numpy discussion http permalink.gmane.org gmane.comp.python.numeric.general 50058","bug"
|
|
"6013","5993","diagonal views need to be made writable.. Was scheduled for 110.","enhancement"
|
|
"6013","5994","Implement structure subset views.
|
|
Was scheduled for 1.9, but there was a potential problem with changes in filler.
|
|
If done, can probably go straight to making them writable.","enhancement"
|
|
"6013","5995","Documentation error for numpy.loadtxt Example 1.
|
|
[Line 699] https github.com numpy numpy blob v1.9.1 numpy lib npyio.py L699) should be from io import StringIO","bug"
|
|
"6013","5997","DEP Python 3.4 deprecates the imp module.. Should have a workaround applied before then.","bug"
|
|
"6013","600","Error in deleting an empty row with numpy.delete Trac 2140).
|
|
Original ticket http projects.scipy.org numpy ticket 2140 on 2012 05 22 by trac user jansb, assigned to unknown.
|
|
Hello, I get an error when I try to delete nothing from an array with a shape of one axis equal zero.
|
|
What works In [1] z=np.zeros 0,)) In [2] z, z.shape Out[2] array [], dtype=float64), 0,)) In [3] numpy.delete z,[],0) Out[3] array [], dtype=float64) What gives an index error In [4] z=np.zeros 0,2)) In [5] z, z.shape Out[5] array [], shape= 0, 2), dtype=float64), 0, 2)) In [6] numpy.delete z,[],0) IndexError Traceback most recent call last) home jansb .local share Trash files estimbuilder 3 tests <ipython input 6 f9a27bc2cbeb> in <module> ) > 1 numpy.delete z,[],0) usr lib python2.7 dist packages numpy lib function base.pyc in delete arr, obj, axis) 3383 obj = setdiff1d all, obj) 3384 slobj[axis] = obj > 3385 new = arr[slobj] 3386 if wrap 3387 return wrap new) IndexError invalid index What works In [8] z=np.zeros 0,2)) In [9] z, z.shape Out[9] array [], shape= 0, 2), dtype=float64), 0, 2)) In [10] numpy.delete z,[],1) Out[10] array [], shape= 0, 2), dtype=float64) Is this a bug or wanted behaviour?
|
|
Best regards, jansb","bug"
|
|
"6013","6002","numpy.linalg.eigh Eigenvalues do not converge for some matrices in numpy 1.9.1.
|
|
Dear numpy developers, There seems to be a bug in the numpy.linalg.eigh function in version 1.9.1 of numpy.
|
|
For some particular matrices the eigenvalues do not converge.
|
|
You will find an example below.
|
|
The problem does not appear for this example using numpy 1.8.2 or the scipy 0.14.0 and 0.15.1 version of the function.
|
|
Also, if I take the real part of the matrix which has imaginary part equal to zero), the diagonalization goes without any problem.
|
|
The python version is 2.7.3.
|
|
```python import numpy as np dm=[[ 0.08538110465924602 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j], [0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j)], [0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j)], [ 0.020017637891624775 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j], [0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j)], [0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j)], [ 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j], [0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j)], [0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j)], [ 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j], [0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j)], [0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j)], [ 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j], [0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j)], [0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j)], [ 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j], [0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j)], [0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j)], [ 0.0020512906310953673 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j], [0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j)], [0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.006079106448764422 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j)], [ 0.002310558750212695 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0j, 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j], [0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0j, 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j)], [0j, 0j, 0.002310558750212695 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.005264877091694114 0j), 0j, 0j, 0.020017637891624775 0j), 0j, 0j, 0.0020512906310953673 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.006079106448764422 0j), 0j, 0j, 0.08538110465924602 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.007640366523797612 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0j, 0j, 0.012250910597278966 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.0018230013446288607 0j), 0j, 0.0001696647814559791 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0.00018153178596027909 0j), 0j, 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j), 0j, 0.0018230013446288607 0j), 0.0001696647814559791 0j), 0j, 0.00018153178596027909 0j), 0.005017379541362799 0j)], [ 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.19189041737312648 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j], [ 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0j, 0.19189041737312654 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0037183102798007384 0j), 0j], [ 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.19189041737312654 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0037183102798007384 0j)], [ 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.04916072807548792 0j), 0j, 0j, 0.19189041737312654 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j], [ 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0j, 0.002980602600651591 0j), 0j, 0j, 0.19189041737312648 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0037183102798007384 0j), 0j], [ 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.19189041737312654 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.10523317949793976 0j), 0j, 0j, 0.0019946497883291323 0j), 0j, 0j, 0.0006468726435089934 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.015018579696547865 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.0037183102798007384 0j)], [ 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002594791231222671 0j), 0.008598907906774166 0j), 0.008598907906774166 0j), 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.19189041737312654 0j), 0j, 0j, 0.04916072807548792 0j), 0j, 0j, 0.003142991873797905 0j), 0j, 0j, 0.003204172711472622 0j), 0j, 0j, 0.002980602600651591 0j), 0j, 0j, 0.003918829551120404 0j), 0j, 0j, 0.0035624379400979677 0j), 0j, 0j, 0.0037183102798007384 0j), 0j, 0j, 0.018764189993089025 0j), 0j, 0j, 0.015018579696547","bug"
|
|
"6013","6008","numpy.linalg.eig eigenvectors not enough precision.
|
|
I ve been taking a Coursera class and came across an issue with Python 3.4 using numpy 1.8.1 on 32 bit Windows Vista when solving the following problem using eigenvalues and eigenvectors import numpy as np import matplotlib.pyplot as plt M = np.matrix [[ .75,0,.75,.75,0],[0, .75,0,.75,.75],[.75,0, .75,0,.75],[.75,.75,0, .75,0],[0,.75,.75,0, .75]]) DOESN T RETURN CORRECT PRECISION eigM=np.linalg.eig M) W = np.matrix [[.6,.1,.1,.1,.1],[.1,.6,.1,.1,.1],[.1,.1,.6,.1,.1],[.1,.1,.1,.6,.1],[.1,.1,.1,.1,.6]]) u = np.matrix [[.6],[.5],[.6],[.2],[.1]]) h = W u vss = 0 for i in range 0,len eigM[0])) vss = np.dot np.squeeze np.asarray h)),np.squeeze np.asarray eigM[1][ ,i]))) 1 eigM[0][i])) eigM[1][ ,i] The problem is the result for vss doesn t have the right precision.
|
|
When I solve the equation numerically I get the correct answer with the right precision v = np.matrix [[1.0],[1.0],[1.0],[1.0],[1.0]]) tau = .1 dt = .001 for i in range 1,100000) v = v 1 tau v h M v) dt I also wrote equivalent code in MATLAB to confirm that the eig ) function there is working correctly and it is.
|
|
In looking at the results, it seemed like the eigenvalues matched closely, but the eigenvectors did not.
|
|
Hopefully you can see what is going on here better than I can, I think numpy is awesome!
|
|
Thanks for your time and efforts!","bug"
|
|
"6013","601","numpy.round ) doesn t work on zero dimensional vectors or floats Trac 2141).
|
|
Original ticket http projects.scipy.org numpy ticket 2141 on 2012 05 23 by pbrod, assigned to unknown.
|
|
as shown in the following example In [12] np.
|
|
version Out[12] 1.6.1 In [13] np.round 3.434523, decimals=3) Out[13] 3.4350000000000001 In [14] np.round np.array 3.434523), decimals=3) Out[14] 3.4350000000000001 However, correct result is obtained when the value is converted to a one dimensional array, as shown here In [15] np.round [3.434523], decimals=3) Out[15] array [ 3.435]) In [16] np.round np.array [3.434523]), decimals=3) Out[16] array [ 3.435]) A related issue is Ticket 2244","bug"
|
|
"6013","6010","numpy.ma.masked equal seems to leak.
|
|
.
|
|
leads to .","bug"
|
|
"6013","6011","ndarray should have a numpy ufunc method.
|
|
As per https github.com numpy numpy issues 5986 issuecomment 114552976 CC mhvk","bug"
|
|
"6013","6012","rfft and irfft ignore float32 and complex64 dtypes, and always return float64 or complex128 arrays.
|
|
When i use a float32 dtype array as the argument to numpy.fft.rfft, I get a complex128 dtype array as the output instead of a complex64 dtype array.
|
|
Similarly, when I use a complex64 dtype array as the argument to numpy.fft.irfft, I get a float64 dtype array instead of a float32 dtype array.
|
|
Here is a code snippet In [4] a = np.array [1., 2., 3.
|
|
], dtype=np.float32) In [5] a Out[5] array [ 1., 2., 3.
|
|
], dtype=float32) In [6] b = np.fft.rfft a) In [7] b Out[7] array [ 6.0 0.j , 1.5 0.8660254j]) In [8] b.dtype Out[8] dtype complex128 ) In [9] b.astype np.complex64) Out[9] array [ 6.0 0.j , 1.5 0.86602539j], dtype=complex64) In [10] c = np.fft.irfft b) In [11] c Out[11] array [ 2.25, 3.75]) In [12] c.dtype Out[12] dtype float64 )","bug"
|
|
"6013","6014","regression in np.concatenate with object dtypes in 1.10.
|
|
This was working as of this commit 1.10.0.dev0 95b2c24 and in 1.9.2), so just broken very recently .
|
|
.","bug"
|
|
"6013","6019","Masked array repr fails for structured array with multi dimensional column.
|
|
As first noted in https github.com astropy astropy issues 3877, there is a problem printing out a masked structured array with a multi dimensional column where at least one of the mask values is True .","bug"
|
|
"6013","602",".max 0) on reshaped array returns inconsistent results.
|
|
Trac 2144).
|
|
Original ticket http projects.scipy.org numpy ticket 2144 on 2012 05 25 by thouis, assigned to unknown.
|
|
The following code fails on the current master 7a254bd) import numpy as np b = np.array [0, 1, 2, 3, 4, 5], np.int64) a = b.reshape 3, 2) while True np.testing.assert array equal np.atleast 1d np.array a.max 0), np.float)), np.atleast 1d np.array a.max 0), np.float))) This is on OSX 10.6.8, numpy compiled with gcc 4.0.1 Python 2.7 r27 82508, Jul 3 2010, 21 12 11) [GCC 4.0.1 Apple Inc. build 5493)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.version.version 1.7.0.dev 7a254bd","bug"
|
|
"6013","6020","In place operators in 1.9 respected array priority ; in current master they don t. In 1.9.2 .
|
|
But in current master .
|
|
This is because .
|
|
I m pretty sure this is due to 5964 5864, i.e., my fault, doh.
|
|
Though I m not sure how, exactly; e.g.
|
|
pv [caught that there was potentially an issue here] https github.com numpy numpy pull 5864 issuecomment 101191802) but then decided it [actually only affected object arrays] https github.com numpy numpy pull 5864 issuecomment 101246043), so right now I m not even 100 sure that this is the responsible change or what exactly happened.
|
|
Some careful analysis is needed.
|
|
I think the behaviour in master is probably correct, but it could break people s code.
|
|
So options Leave it alone unless someone complains Fix it to issue a `DeprecationWarning` in 1.10, error in 1.11.
|
|
And if we go with the latter option, then we have to figure out exactly which cases should return `NotImplemented` and issue the warning.
|
|
This doesn t need to block 1.10 alpha, but it does need some sort of decision before 1.10 final, so I ll tag it with the milestone.","bug"
|
|
"6013","6021","KeyError when constructing an array with a dict like class instance.. Numpy version 1.9.2, Python version 2.7.3 See test case below.
|
|
I would have expected the same behaviour as when using a `dict`.
|
|
Removing the ` iter ` method prevents the error.
|
|
.
|
|
.","bug"
|
|
"6013","6022","Invalid offset in buffer format string generation, when trying to use PyOpenCl with Numpy.
|
|
I am porting over one of my projects from cpu to gpu and in the process I simply copied some of the code from http documen.tician.de pyopencl howto.html However I keep getting a numpy error.
|
|
Code import pyopencl as cl import pyopencl.tools import pyopencl.array import numpy as np def gpuTest ) context = cl.create some context interactive=False) queue = cl.CommandQueue context) my struct = np.dtype [ field1 , np.int32), field2 , np.float32)]) my struct, my struct c decl = cl.tools.match dtype to c struct context.devices[0], my struct , my struct) my struct = cl.tools.get or register dtype my struct , my struct) print my struct c decl) ary host = np.empty 20, my struct) ary host[ field1 ].fill 217) ary host[ field2 ].fill 1000) ary host[13][ field2 ] = 12 print ary host) ary = cl.array.to device queue, ary host) prg = cl.Program context, my struct c decl ... kernel void set to 1 global my struct a) ... { ... a[get global id 0)].field1 = 1; ... } ... ).build ) evt = prg.set to 1 queue, ary.shape, None, ary.data) print ary) gpuTest ) input ) Here is the console output typedef struct { int field1; float field2; } my struct; [ 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 12) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000) 217, 1000)] Traceback most recent call last) File C \Users\\Desktop\Programming\Python Programming\Pygpu\main.py , line 36, in <module> gpuTest ) File C \Users\\Desktop\Programming\Python Programming\Pygpu\main.py , line 22, in gpuTest ary = cl.array.to device queue, ary host) File C \Python34\lib\site packages\pyopencl\array.py , line 1680, in to device result.set ary, async=async) File C \Python34\lib\site packages\pyopencl\array.py , line 670, in set is blocking=not async) File C \Python34\lib\site packages\pyopencl\ init .py , line 1051, in enqueue copy return cl.
|
|
enqueue write buffer queue, dest, src, kwargs) RuntimeError This should never happen Invalid offset in buffer format string generation.
|
|
Please report a bug to the Numpy developers.
|
|
Press any key to continue .
|
|
.
|
|
.
|
|
I cant seem to find anyone else who has had this kind of problem, is there something I could have done wrong when installing?
|
|
Thank you in advance","bug"
|
|
"6013","6023","`np.dtype string )` only works on Python 2.
|
|
In Python 2, there s no problem git numpy2 python2 Python 2.7.6 default, Jun 22 2015, 17 58 13) [GCC 4.8.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.10.0.dev0 38d6f09 >>> np.dtype string ) dtype S ) In Python 3, this throw an exception git numpy python Python 3.4.0 default, Jun 19 2015, 14 20 21) [GCC 4.8.2] on linux Type help , copyright , credits or license for more information.
|
|
>>> import numpy as np >>> np.
|
|
version 1.10.0.dev0 38d6f09 >>> np.dtype string ) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError data type string not understood","bug"
|
|
"6013","6024","Creating array from sympy matrix with dtype.
|
|
I try to convert an sympy matrix to a numpy array.
|
|
Then this strange behaviour pops up Does not work .
|
|
Works .
|
|
numpy.
|
|
version = 1.9.2","bug"
|
|
"6013","6025","Document versioning of C API.
|
|
Looking at http docs.scipy.org doc numpy reference c api.html it doesn t seem to be documented which macros and or magic values to check for the availability of certain APIs say, for example, I want to access the datetime C APIs, but also provide some compatibility with Numpy 1.6).
|
|
It would be nice to document that somewhere.","bug"
|
|
"6013","6026","numpy.pad breaks if constant values are float32.
|
|
First of all, thank you so much to all the numpy devs.
|
|
It s an awesome beast!
|
|
Here s a minimal code example to illustrate the issue This works .
|
|
But, if we use `np.float32` numbers for `constant values`, it fails .
|
|
This also fails if `arr` has `dtype=np.float32`.
|
|
My proposal would either be 1. change the definition of `np.pad` so that it accepts `constant values` with any float dtype.
|
|
Or 2. modify the docs and the `ValueError` text to tell the user that only `np.float64` numbers are allowed for `constant values`.
|
|
update This refers to numpy version 1.9.2 on Ubuntu Linux 15.04.","bug"
|
|
"6013","6027","Handling ndarrays with singular matrices in linalg functions.
|
|
I am trying to solve several independent systems of equations at the same time using `numpy.linalg.solve`, e.g., `a` has shape N, M, M).
|
|
The challenge I m running into is how to deal with the case when, for some values of N, the last two dimensions comprise a singular matrix.
|
|
Consider this example .
|
|
On numpy 1.9.2, this will raise a `LinAlgError`, but I still would like to know the solutions for all the full rank matrices.
|
|
I would be okay with getting back NaNs for the singular cases.)
|
|
I m aware I can compute the singular values and use fancy indexing to slice the array, but in my algorithm I have to do several slicing filtering steps and I would prefer not to lose the alignment of `a` with `b` my real world `a` typically has 6 7 dimensions and all the indexing arrays make the code hard to follow.
|
|
My ideal solution seems to involve masked arrays since I end up doing several other filtering steps, but it appears that most all?)
|
|
functions in the `linalg` family ignore array masks.
|
|
Is there an alternative solution already available, or perhaps a suggestion on how I could contribute a solution that I could package as a PR?","bug"
|
|
"6013","6028","initializing ndarray from empty array loses type.
|
|
empty arrays should keep their type, but they always come back as `double` .
|
|
i m using python 2.7.9, numpy 1.9.2","bug"
|
|
"6013","603","Numpy does not compile with Python 3.3.0a3 Trac 2145).
|
|
Original ticket http projects.scipy.org numpy ticket 2145 on 2012 05 28 by astrofrog, assigned to unknown.
|
|
I have tried compiling Numpy 1.5.1 and 1.6.2) with Python 3.3.0a3 on MacOS 10.7 with XCode 4.3.2 and using g95 as the fortran compiler, and am getting failures whereas Python 3.2 works fine).
|
|
I have attached the install log for both Python 3.2 which works) and Python 3.3.0a3 which doesn t) with Numpy 1.6.2 to help diagnose the issue.
|
|
Maybe this is an issue with Python 3.3.0a3 rather than with Numpy?","bug"
|
|
"6013","6030","min ) and argmin ) inconsistent with NaTs.
|
|
Either NaT is ignored or it isn t when looking for a minimum, but min ) and argmin ) disagree .","bug"
|
|
"6013","6032","argmin ) and argmax ) buggy with NULL objects.
|
|
The source code for argmin ) and argmax ) purports to handle NULL objects for object arrays), but the code handling them is buggy .
|
|
min ) and max ) take another approach they interpret the NULL pointers as references to None.","bug"
|
|
"6013","6034","document numpy bylaws.
|
|
I am a researcher interested in Numpy.
|
|
I am trying to find information about numpy s governance structure.
|
|
Many OSS project have their process documented somehow.
|
|
There is a lot of interesting material about this in this mailing list discussion http mail.scipy.org pipermail numpy discussion 2012 February 060452.html e.g.
|
|
these comments by John Hunter http mail.scipy.org pipermail numpy discussion 2012 February 060640.html Documenting this process might help resolve conflicts of the kind suggested by the discussion here https github.com numpy numpy issues 5844 issuecomment 117746511","bug"
|
|
"6013","6037","polyfit ) fails on vectors with NaN.
|
|
I have two Pandas DataFrames .
|
|
Then I sort these series by increasing order of df a [0, 2, 1, 3, 4, 5, 13, 6, 7, 8, 9, 10, 11, 12].
|
|
Apply polynomial fitting tried for several degrees) .
|
|
And getting an error .
|
|
.
|
|
The warning is OK, but I still expect function to fit the polynomial.
|
|
You can check that all the points are located near the imaginary line.
|
|
The problem is NaN handling, I suppose.
|
|
Also, I m not sure what does the Intel MKL error mean.
|
|
SW versions python 3.4.2) numpy 0.15.1) cython 0.21.2) pandas 0.15.2) All libraries are taken from http www.lfd.uci.edu gohlke pythonlibs .
|
|
Platform Windows 7 x64.","bug"
|
|
"6013","6038","intp on 32bit architectures not same as int32.
|
|
np version 1.9.2 but bug in previous versions as well) python version 3.4.3 but bug with previous versions as well) The following demonstrates the problem On 64Bit architecture .
|
|
On 32Bit architecture .","bug"
|
|
"6013","604","Bus error seg fault when using flat on sliced, memmap d array Trac 2148).
|
|
Original ticket http projects.scipy.org numpy ticket 2148 on 2012 05 31 by trac user farrowch, assigned to unknown.
|
|
The following code crashes on numpy 1.6.1, on OSX 10.6. import numpy as np shape = 8, 8) dtype = np.dtype np.uint8) image = np.random.randint 0, 256, shape).astype dtype) image.tofile test image.bin ) image = np.memmap test image.bin , dtype=dtype, shape=shape, mode= r ) arr = image[ 2, 2] np.sum arr.flat) Here s what I ve discovered so far about this the bus error only occurs with mode r the dimensionality of the array appears to be irrelevant if the array slice does not change the strides, the bus error does not occur no arr.flat , no bus error Other aggregating functions e.g.
|
|
fmin.reduce) will induce the error Iterating over arr.flat will not cause a bus error Based on this, I suspect the issue is with the C facing side of the flat iterator.","bug"
|
|
"6013","6040","numpy histogram precision bug.
|
|
When dealing with weights with a large dynamic range, the result is problematic.
|
|
The problem can be reproduced with this code in numpy 1.9.1 and 1.9.2 .
|
|
This should print two identical numbers which is the case if, for example, we use `xmin= 1.`) but instead it produces .
|
|
This is not a datatype issue because every variable is `float64`.
|
|
Replacing the bin number parameter with double precision bins does not help.","bug"
|
|
"6013","6041","DEV adding unbias option to the correlate function.
|
|
The correlate function currently does not support the unbias option.
|
|
Please see the issue discussed here http nbviewer.ipython.org gist jankoslavic d4298dae2fbe9d598db1 I also have a working proposal for the implementation https github.com jankoslavic numpy commit 3fc57351e0dacec5cdef6b0923d8fc80f52d37a0 Here is another implementation http statsmodels.sourceforge.net 0.6.0 modules statsmodels tsa stattools.html ccf","enhancement"
|
|
"6013","6044","cumsum and accumulation operators generally?)
|
|
should have an option to leave the identity in the first spot.
|
|
I just found myself writing .
|
|
and it occured to me that this is definitely not the first time I ve found myself doing the awkward concatenate a zero dance in order to get a cumsum array whose shape matches what I want.
|
|
Maybe we should have a convenience argument to `cumsum` and to `ufunc.accumulate` in general I guess) to insert the operation s identity as the first argument.
|
|
No brain to figure out the details of how this would work right now, but filing a bug to remind me us...","bug"
|
|
"6013","6045","Possible bug in numpy ma core.py.
|
|
pyflakes notes the ` sharedmask` is assigned but never used.
|
|
I suspect this may be a bug.
|
|
Note the the `FIXME` message is not yet in master.
|
|
.","bug"
|
|
"6013","605","interp nan, ...) should yield nan Trac 2149).
|
|
Original ticket http projects.scipy.org numpy ticket 2149 on 2012 05 31 by trac user ehiggs, assigned to unknown.
|
|
Passing nan to the first argument of interp appears to return the first element of fp.
|
|
This isn t a meaningful value and can silently corrupt data since nan went in and undetectable garbage came out.
|
|
>>> import numpy as np >>> np.interp 1, [ 10, 10], [ 2, 2]) 0.20000000000000018 >>> np.interp np.nan, [ 10, 10], [ 2, 2]) 2.0 >>> np.interp [np.nan, 1], [ 10, 10], [ 2, 2]) array [ 2. , 0.2]) This 2.0 is not a meaningful answer.
|
|
The interpolation of a nan value should probably also be nan.
|
|
Thanks","bug"
|
|
"6013","6050","argsort sort with O n) algorithms counting sort, radix sort.
|
|
You can already sort with [couting sort] https en.wikipedia.org wiki Counting sort) using numpy as stated in an [SO answer] http stackoverflow.com a 18502321 2399799), it is simply .
|
|
But there is no way as far as I can tell) to do an argsort.
|
|
I think it would be fairly easy to implement and slot into the argsort sort documentation.
|
|
The benefit of counting sort is that it s O n), although it only works for not too large non negative integers...however those are actually quite common in the real world!
|
|
Pandas uses it behind the scenes [apparently] http wesmckinney.com blog tag pandas ).","enhancement"
|
|
"6013","6051","Numpy functions return a scalar when the input is a 0 dimensional array.
|
|
This may be a duplicate issue, but so far I ve only found few somewhat related issues, e.g.
|
|
3161 , 2500 and perhaps 1774 and 3311.)
|
|
When creating an array of a single non list) element, the result is a 0 dimensional array >>> a = np.array 1) >>> type a), a.shape, a.ndim <class numpy.ndarray >, ), 0) When applying a numpy function to `a`, the result is a numpy scalar instead >>> b = np.cos a) >>> type b), b.shape, b.ndim <class numpy.float64 >, ), 0) or one can simply multiply by an int >>> b = 100 a >>> type b), b.shape, b.ndim <class numpy.int64 >, ), 0) ) This effectively changes the type, and options available to the result.
|
|
For example, one can assign to an indexed 0 dimensional array, but not to an indexed numpy scalar >>> a[True] = 5 >>> b[True] = 5 Traceback most recent call last) File <stdin> , line 1, in <module> TypeError numpy.float64 object does not support item assignment Of course, indexing a scalar is awkward, but it is consistent if one doesn t known whether the input is a Python) float or int to which as)array has been applied), a list, a one element array `[1]` or a 1 or higher dimensional array code written would not need to distinguish between a scalar and an array.
|
|
My workaround for now is to use `asarray` >>> b = np.asarray b) >>> b[True] = 5 but I wonder if there is a better way.
|
|
Moreover, how consistent should the return type of numpy functions be with the input type?
|
|
Apologies if this topic has come up before, and may have been discussed to dead, but there was relatively little I could find that appears to address exactly this issue.","bug"
|
|
"6013","6056","Using C int default leads to overflow on 32 bit platforms.
|
|
I did not expect this behavior .
|
|
I assume this is because .
|
|
Is there a good reason not to default to int64 int casting on 32 bit platforms?
|
|
Is the speed difference large enough to justify this risk?
|
|
Or is it the memory consumption?","bug"
|
|
"6013","6059","polybase.py the way to declear abstract class works only for Python 2.
|
|
The Python 2 way to specify abstract class .
|
|
Python 3 way .","bug"
|
|
"6013","606","distutils should support Debian multi arch fully Trac 2150).
|
|
Original ticket http projects.scipy.org numpy ticket 2150 on 2012 06 04 by rgommers, assigned to cournape.
|
|
See http thread.gmane.org gmane.comp.python.numeric.general 49957 Patch is available in link [2] of the first message of the linked thread.","bug"
|
|
"6013","6060","Proposal Add a todense method to matrixes..
|
|
The method would be an identity map for numpy matrixes.
|
|
It is needed to match the interface offered by scipy.
|
|
The particular situation which tripped me up was the following .
|
|
I find the case to add the method compelling and do not see any obvious downsides.
|
|
If it is acceptable to add it, then should it return the same object or should it return a copy ?
|
|
With this information, I should be able to work on a pull request.","bug"
|
|
"6013","6061","Calling searchsorted on an int haystack with str needle converts haystack to str.
|
|
With `numpy 1.9.2` I see quite counter intuitive behavior of `searchsorted` on integer arrays with a `str` needle .
|
|
The main reason why I believe this behavior is broken is because when one converts an integer array to strings, it s not sorted anymore.
|
|
I believe that the least surprising result would be an exception, especially because this is what happens if the array contains floats and not integers.
|
|
P.S.
|
|
I initially asked about this on [StackOverflow] http stackoverflow.com q 31325001 1258041), and jaimefrio suggested creating a bug report in the comments.","bug"
|
|
"6013","6062","BUG ix modifies input arrays.
|
|
Regression that was introduced in gh 5805.
|
|
Discovered when running Scipy tests against numpy master ====================================================================== ERROR test qhull.TestUtilities.test more barycentric transforms Traceback most recent call last) File usr local lib python2.7 site packages nose case.py , line 197, in runTest self.test self.arg) File Users rgommers Code bldscipy scipy spatial tests test qhull.py , line 310, in test more barycentric transforms grid = np.c [list map np.ravel, np.broadcast arrays np.ix [x] ndim)))))].T File Users rgommers Code numpy numpy lib index tricks.py , line 77, in ix raise ValueError Cross index must be 1 dimensional ) ValueError Cross index must be 1 dimensional A simple way to reproduce the issue >>> x = np.linspace 0, 1, num=5) >>> np.ix [x] 4)) Traceback most recent call last) File <ipython input 6 b0c751ea5a84> , line 1, in <module> np.ix [x] 4)) File Users rgommers Code numpy numpy lib index tricks.py , line 77, in ix raise ValueError Cross index must be 1 dimensional ) ValueError Cross index must be 1 dimensional >>> x.shape 5, 1, 1, 1) Solution is simply to make a copy when constructing output arrays.","bug"
|
|
"6013","6063","FAIL test umath.TestComplexFunctions.test branch cuts.
|
|
Got following error when running runtests.pyt on linux 64 bit.
|
|
.","bug"
|
|
"6013","6068","3 out of 5 numpy imports do not work.
|
|
I have Python installed in at least) 5 places in my system.
|
|
All but two do not accept a numpy installation.
|
|
1. windows 64 bit system) registered Python27 works D \cmd>python c import numpy; print numpy.get include ) e \Python27\lib\site packages\numpy\core\include D \cmd> 2.
|
|
CYGWIN64 python 2.7.10 numpy 1.7 OR 1.9 does not work python c import numpy; print numpy.get include ) Traceback most recent call last) File <string> , line 1, in <module> File usr lib python2.7 site packages numpy init .py , line 137, in <module> import add newdocs File usr lib python2.7 site packages numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File usr lib python2.7 site packages numpy lib init .py , line 13, in <module> from polynomial import File usr lib python2.7 site packages numpy lib polynomial.py , line 17, in <module> from numpy.linalg import eigvals, lstsq, inv File usr lib python2.7 site packages numpy linalg init .py , line 48, in <module> from linalg import File usr lib python2.7 site packages numpy linalg linalg.py , line 23, in <module> from numpy.linalg import lapack lite ImportError No such file or directory 3.
|
|
MSYS2 does not work greg Homerw7 MSYS python c import numpy Traceback most recent call last) File <string> , line 1, in <module> File usr lib python2.7 site packages numpy init .py , line 170, in <module> from .
|
|
import add newdocs File usr lib python2.7 site packages numpy add newdocs.py , line 13, in <module> from numpy.lib import add newdoc File usr lib python2.7 site packages numpy lib init .py , line 8, in <module> from .type check import File usr lib python2.7 site packages numpy lib type check.py , line 11, in <module> import numpy.core.numeric as nx File usr lib python2.7 site packages numpy core init .py , line 6, in <module> from .
|
|
import multiarray ImportError No such file or directory .
|
|
4.
|
|
MSYS2 Mingw32 python 2.7.9 does not work .
|
|
5.
|
|
MSYS2 Mingw64 python 2.7.9 works greg Homerw7 MINGW64 python c import numpy; print numpy.get include ) C msys64 mingw64 lib python2.7 site packages numpy core include C msys64 mingw64 lib python2.7 site packages numpy core init .py 6 Warning Numpy built with MINGW W64 on Windows 64 bits is experimental, and only available for testing.
|
|
You are advised not to use it for production.
|
|
CRASHES ARE TO BE EXPECTED PLEASE REPORT THEM TO NUMPY DEVELOPERS from .
|
|
import multiarray greg Homerw7 MINGW64 python version Python 2.7.9","bug"
|
|
"6013","6069","non contiguous indices, fancy indexing corner case.
|
|
To boggle everyones minds, this should not tbe the same .","bug"
|
|
"6013","607","flatiter does not work correctly with non writeable arrays Trac 2151).
|
|
Original ticket http projects.scipy.org numpy ticket 2151 on 2012 06 06 by trac user tux, assigned to unknown.
|
|
I found out that numpy.flatiter objects do not retain the base object s writeable flag.
|
|
While this only leads to unexpected behavior for normal non writeable arrays, this causes a segfault if the array s buffer is a read only memory area e.g.
|
|
mmap).
|
|
crashme1.py shows one part of the problem.
|
|
The array b is created with the mmaped array a as its base.
|
|
The current code creates a copy and sets the UPDATEIFCOPY flag if a is not C CONTIGUOUS I used Fortran order here).
|
|
Now when b is destroyed it tries to write to its base UPDATEIFCOPY) which causes a segfault because the buffer is a read only mmap.
|
|
I tried to fix this problem by only setting updateifcopy only if the original array is writeable.
|
|
If not, the flatiter should be non writeable as well in my opinion to prevent people from trying to change the original array with it which would not work.
|
|
I m not perfectly sure that I got the semantics of iter array iterators.c) right.
|
|
Somebody with more insight into numpy should judge.
|
|
At least the comment of iter array should also be updated to reflect the new situation.
|
|
After I implemented this, stumbled on another bug.
|
|
Even if the flatiter is non writeable, the function array flat set getset.c) does not actually check the flag before writing to the array.
|
|
This is demonstrated in the example write the unwriteable.py.
|
|
If the non writeable array is a read only memory area, we get a segfault again see crashme2.py).
|
|
To fix this I added a check for the writeable flag at the beginning of array flat set which throws an appropriate exception the same that direct assignment would cause).
|
|
I tested all of the above on Debian Sid Python 2.7.3 and numpy 1.6.2), Ubuntu Precise Python 2.7.3 and numpy 1.6.1) and a self compiled version of Python 2.7.1 and numpy 1.6.1.
|
|
I could also reproduce it with python 3.2.
|
|
I attached my patches against the current git HEAD to this report.","bug"
|
|
"6013","6070","Please Deprecate creation of numpy arrays for arbitrary objects..
|
|
NOTE as discussed at SciPy2015 just thought I d get an issue in.
|
|
numpy currently will default to making an dtype object array for anything it can t make other sense of.
|
|
This is hardly ever what someone would want.
|
|
For example, I just banged my head against a bug due to this In [5] np.array object) Out[5] array <type object >, dtype=object) Granted, due to a stupid typo but why would anyone want this?
|
|
When a user does want the object dtype, numpy almost never can get it right by default anyway, so it s not much burden to have to specify the object dtype if you really want that.","bug"
|
|
"6013","6074","Inconsistency in assignment to object array..
|
|
Consider the following code >>> import numpy as np >>> x = np.zeros 1,1), dtype=object) >>> y = np.zeros 1,1), dtype=object) >>> z = np.zeros 1,1), dtype=object) >>> a = np.array None) >>> x[0] = a >>> y[0,0,...] = a >>> z[0,0] = a >>> x array [[None]], dtype=object) >>> y array [[None]], dtype=object) >>> z This is inconsistent!!!!
|
|
array [[array None, dtype=object)]], dtype=object) As it can be seen, assignment of a zero rank array to an object array causes nesting of array inside an array.
|
|
I believe that the third case assignment to `z`) should behave like the other cases, otherwise one can quickly generate arrays inside arrays inside arrays and necessitate extra code to treat special cases.
|
|
This was tested in numpy 1.9.2 under python 3.4.2.","bug"
|
|
"6013","6076","numpy.fromstring behaviour.
|
|
I encountered strange behaviour of numpy.fromstring .
|
|
Is this desired?
|
|
To me it feels like this should raise an error.","bug"
|
|
"6013","6078","Optional output of full fancy index for argmin, argmax, argort?.
|
|
In 6075, I wondered whether it might be possible to > to have some simple way to use the output of `np.arg[min|max|sort]` as a proper index to avoid hackery as in https github.com astropy astropy blob master astropy time core.py L896 if there is a simpler way already, let me know!)
|
|
and jaimefrio suggested that it might be simpler to have > a `make me an indexing tuple=False` new keyword argument for the `argxxx` functions?
|
|
That indeed seems better, so I m raising this as a separate issue.
|
|
The code linked to above could be used as a template on how to generate the index tuple.","enhancement"
|
|
"6013","6079","support masked arrays for np.insert.
|
|
Currently np.insert do not follow the masked rules of a np.ma.Maskedarray.
|
|
How to Reproduce .
|
|
outputs .
|
|
So the insert function doens t take the mask in account even if the input is masked.
|
|
The same behaviour occurs when input is not masked, so the operation causes a lost of the masking information it would be desirable that even in this case the output would be at least not masked and the operation would preserve the mask information).
|
|
I believed that the mask information would be passthrough to the new array.","enhancement"
|
|
"6013","608","np.inner is no inner product conjugate symmetry) Trac 2152).
|
|
Original ticket http projects.scipy.org numpy ticket 2152 on 2012 06 06 by trac user nschloe, assigned to unknown.
|
|
The documentation says it Ordinary inner product of vectors for 1 D arrays without complex conjugation).
|
|
This means that, for complex vectors, np.inner other than its name suggests does not define an inner product as the symmetry isn t conjugate.
|
|
For the sake of consistency, either this function should be renamed or np.inner should be made compliant with actual inner products.
|
|
This of course would break backward compatibility, but for the sake of consistency with mathematics, this seems justified.
|
|
Is there, at this point, actually any difference between np.inner and np.dot?
|
|
).","bug"
|
|
"6013","6080","Missing keepdims in documentation for ndarray.min max.
|
|
ndarray.min ) and ndarray.max ) accept the keepdims argument, but this is missing from the documentation.","bug"
|
|
"6013","6081","BUG arccos inaccurate for small complex numbers.
|
|
Consider the following examples The first few terms of taylor series for arccos is given by >>> import numpy as np >>> def arccos 3 x) ... return np.pi 2 x x 3 6 3 x 5 40 Then we compare the taylor series with numpy.arccos for small complex number >>> x = 1je 14 >>> v1 = np.arccos x); v1 1.5707963267948966 9.9920072216263584e 15j) >>> v2 = arccos 3 x); v2 1.5707963267948966 1e 14j) >>> v2 v1).imag v2.imag 0.00079927783736416112 Lets compute the complex step derivative of arcos >>> y = 0 >>> h = 1e 14 >>> np.arccos y 1j h).imag h 0.99920072216263589 The exact value is 1 >>> arccos 3 1j h).imag h 1.0 In both the examples the relative error of the complex part of numpy.arccos function is around 1e 3.
|
|
It should be relatively easy to fix this as the complex part of arccos function is proportional to x when x is small.","bug"
|
|
"6013","6082","BUG arcsin inaccurate for small complex numbers.
|
|
Consider the following examples The first few terms of taylor series for arcsin is given by >>> import numpy as np >>> def arcsin 3 x) ... return x x 3 6 3 x 5 40 Then we compare the taylor series with numpy.arcsin for small complex number >>> x = 0.01 1je 14 >>> v1 = np.arcsin x); v1 0.010000166674167112 1.0103029524088975e 14j) >>> v2 = arcsin 3 x); v2 0.010000166674166668 1.00005000375e 14j) >>> v2 v1).imag v2.imag 0.010252435998650904 Lets compute the complex step derivative of arcsin >>> y = 0.01 >>> h = 1e 14 >>> np.arcsin y 1j h).imag h 1.0103029524088976 The exact value is 1. np.sqrt 1 0.01 2)=1.0000500037503126 >>> arcsin 3 1j h).imag h 1.00005000375 In both the examples the relative error of the complex part of numpy.arcsin function is around 1e 2.
|
|
It should be relatively easy to fix this as the complex part of arcsin function is proportional to x when x is small.","bug"
|
|
"6013","6083","BUG arctan inaccurate for small complex numbers.
|
|
Consider the following examples The first few terms of taylor series for arctan is given by >>> import numpy as np >>> def arctan 3 x) ... return x x 3 3 x 5 5 Then we compare the taylor series with numpy.arctan for small complex number >>> x = 0.01 1je 14 >>> v1 = np.arctan x); v1 0.0099996666866652394 9.9920072216263095e 15j) >>> v2 = arctan 3 x); v2 0.009999666686666667 9.9990001e 15j) >>> v2 v1).imag v2.imag 0.0006993577661521059 Lets compute the complex step derivative of arctan >>> y = 0.01 >>> h = 1e 14 >>> np.arctan y 1j h).imag h 0.99920072216263089 The exact value is >>>1.
|
|
1 y 2) 0.9999000099990001 >>> arctan 3 1j h).imag h 0.9999000100000001 In both the examples the relative error of the complex part of numpy.arcsin function is around 1e 4.
|
|
It should be relatively easy to fix this as the complex part of arctan function is proportional to x when x is small.","bug"
|
|
"6013","6084","Document our backcompat guarantees for np.random.
|
|
I met someone at SciPy this year who had implemented their own PRNG routine from scratch, because they needed to guarantee it was deterministic and weren t sure if `np.random` results were stable between versions and machines.
|
|
Given how much effort we put into guaranteeing exactly this stability, we should probably let people know we re doing it ) Specifically, the guarantee as I understand it is so long as you use the same seed, the same sequence of calls to `np.random` functions `np.random.RandomState` methods will always return the same sequence of results, modulo small a few units in the last place) differences in floating point values due to unavoidable differences in rounding errors between compilers.
|
|
Anything else is a bug.)","bug"
|
|
"6013","6085","Broadcasting not working with record array with a single record.
|
|
This code .
|
|
Throws this exception .
|
|
However this code works fine .
|
|
` p.fill 1)` works in both cases.
|
|
This is with version ``1.8.2``","bug"
|
|
"6013","6086","Array contiguousness is not preserved during np.vstack np.hstack.
|
|
Hi, my current version of numpy is 1.9.2 .
|
|
It is not clear to me why array contiguousness is not preserved during vstack hstack operations.
|
|
I expect that `c` should be in the Fortran order.
|
|
It is interesting enough that when you have an array with more than two rows the behavior is quite expected .
|
|
I would highly appreciate any clarifications on this inconsistent behaviour.","bug"
|
|
"6013","6087","Passing in dictionary to np.log10 raises AttributeError.
|
|
.
|
|
I would expect appreciate a `TypeError` here instead, as the current error message is rather unintuitive.","bug"
|
|
"6013","6089","Error in tiling zero sized arrays.
|
|
Tiling of zero sized arrays fails.
|
|
.
|
|
This happens because, in the tile function, `n` is equal to zero.
|
|
I ll submit a pull request to fix that.","bug"
|
|
"6013","609","Production code is importing numpy testcases Trac 2153).
|
|
Original ticket http projects.scipy.org numpy ticket 2153 on 2012 06 06 by trac user ms4py, assigned to unknown.
|
|
Numpy is importing its testcases on a default import, this is a very bad style.
|
|
See as reference >>> class A ... pass ... >>> sys.modules[ unittest ] = A >>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File usr local Cellar python 2.7.3 lib python2.7 site packages numpy init .py , line 137, in <module> import add newdocs File usr local Cellar python 2.7.3 lib python2.7 site packages numpy add newdocs.py , line 9, in <module> from numpy.lib import add newdoc File usr local Cellar python 2.7.3 lib python2.7 site packages numpy lib init .py , line 4, in <module> from type check import File usr local Cellar python 2.7.3 lib python2.7 site packages numpy lib type check.py , line 8, in <module> import numpy.core.numeric as nx File usr local Cellar python 2.7.3 lib python2.7 site packages numpy core init .py , line 40, in <module> from numpy.testing import Tester File usr local Cellar python 2.7.3 lib python2.7 site packages numpy testing init .py , line 8, in <module> from unittest import TestCase ImportError cannot import name TestCase I know how to solve this Exception, that is not the point)","enhancement"
|
|
"6013","6095","Intel compilers not working with distutils.
|
|
I m having some issues using distutils to compile some C and Fortran code in another package.
|
|
I ve built NumPy from source as the Intel C compilers in Windows don t seem to be supported at all in 1.9.2.
|
|
I m using Python 2.7.10 32 bit) and Intel Parallel Studio XE 2015.
|
|
I had to make some changes to the code to get things working at all.
|
|
The first issue appeared to be in numpy distutils intelccompiler.py I had to change from numpy.distutils.msvc9compiler to from distutils.msvc9compiler as I was getting this error .
|
|
This got me a little further in the build process, but then I had a problem with the Fortran compiler .
|
|
I removed the arch IA 32 flag from IntelVisualFCompiler.get flags arch in numpy distutils fcompiler intel.py, which again got me a little further, but then it failed at the linking stage.
|
|
It appears that ifort gets called with arguments that look like this .
|
|
These are not valid arguments to ifort, so I get a lot of warnings like this .
|
|
and then a lot more spam about unknown options starting with O due to the OUT argument it expanded everything after O as separate optimisation options, like OU OT etc).
|
|
I think ifort expects commands that look more like the following .
|
|
Running this command works a lot better, but I still got a few unresolved external symbol errors which I think are probably unrelated to the issues above and I still haven t resolved.
|
|
However, I thought you should at least be made aware of the above fixes I needed to at least stand a chance of getting things working!","bug"
|
|
"6013","6099","Improve np.histogram performance for uniform bins.
|
|
For the use case where the number of bins is specified by an integer in ``np.histogram``, the use of ``searchsorted`` etc.
|
|
is very inefficient.
|
|
It s possible to get at least a factor of 5x speedup by being smarter about this case just with Python Numpy code, and I can get a factor of 30 speedup with Cython.
|
|
Would it be worth implementing a more efficient version of that special yet common) case?
|
|
If so, I d be happy to turn this issue into a pull request.","enhancement"
|
|
"6013","610","numpy.dot a, b, a) gives an incorrect result for a Trac 2154).
|
|
Original ticket http projects.scipy.org numpy ticket 2154 on 2012 06 08 by trac user ktdrv, assigned to unknown.
|
|
Unless this is expected but undocumented behavior, trying to dot product an array in place, results in a broken result >>> a = arange 1, 10, dtype = float64).reshape 3, 3) >>> a array [[ 1., 2., 3.
|
|
], [ 4., 5., 6.
|
|
], [ 7., 8., 9.]])
|
|
>>> dot a, identity 3), a) array [[ 0., 2., 3.
|
|
], [ 4., 5., 6.
|
|
], [ 7., 8., 9.]])","bug"
|
|
"6013","6104","Indexing with single boolean behaves inconsistently.
|
|
Related to 4353, 4980, but would like to ask the cases when the indexer is a single bool.
|
|
1.
|
|
In normal array, ``bool`` looks coerces to ``int``.
|
|
Is this intended, or should be regarded as a boolean indexing with different shape 4353)?
|
|
.
|
|
2.
|
|
0 dim array behaves quite strangely.
|
|
I understand this should raise ``IndexError`` if it coerces to ``int``.
|
|
.","bug"
|
|
"6013","6105","Add opt in collection of user data to numpy..
|
|
The idea is to set up a central data collection server.
|
|
Code would be added to several prominent user interface front ends to ask users whether they would like to opt in to data collection.
|
|
A standard software library would allow other software to check whether the user has opted in, and if so then to make it easy to aggregate data and submit it to the central server.
|
|
See comments and links in https github.com numpy numpy pull 6103.","bug"
|
|
"6013","6106","ma.empty zeros ones do not accept shape keyword.
|
|
Using 1.9.2, the following gives an unexpected error .
|
|
Removing the `shape` keyword behaves as expected.
|
|
The non ma versions of empty zeros ones do not show this behavior.","bug"
|
|
"6013","6107","Documentation guidelines dead link to Q A.
|
|
The [other points to consider section][1] in the documentation contains the following statement > Questions and Answers For general questions on how to write docstrings that are not answered in > this document, refer to http docs.scipy.org numpy Questions Answers .
|
|
If you try to open that URL you will quickly get the following error thrown at you > Not Found > > The requested URL numpy Questions Answers was not found on this server.
|
|
Please fix, as I am quite interested in that section.
|
|
[1] https github.com numpy numpy blob master doc HOWTO DOCUMENT.rst.txt other points to keep in mind","bug"
|
|
"6013","6108","mergesort with order keyword is unstable.
|
|
This is possibly just a documentation issue.
|
|
When using the `order` keyword to sort and specifying only a subset of the fields, the comparator will continue to tie break on the other fields see `numpy.core.
|
|
internal.
|
|
newnames`).
|
|
When using a stable sort e.g.
|
|
mergesort) I was expecting elements that tied on the specified fields to be left in their original order.
|
|
For example .
|
|
has reordered the elements.
|
|
Using numpy 1.9.2 on Ubuntu 14.04.","bug"
|
|
"6013","6109","astype dtype) converts NaN to np.iinfo dtype).min for int32 and int64, but 0 for int16.
|
|
This seems inconsistent .","bug"
|
|
"6013","611","error while pickling MaskedConstant Trac 2155).
|
|
Original ticket http projects.scipy.org numpy ticket 2155 on 2012 06 11 by trac user r0k3, assigned to pierregm.
|
|
One gets an error while trying to pickle an object that contains a MaskedConstant.","bug"
|
|
"6013","6110","Numpy fails to compile on MSVC 14 VS2015).
|
|
On the master branch of numpy, line 432 of arraytypes.c.src hosts the following line .
|
|
Which leads to the following compiler error .
|
|
casting ov to a char will allow compilation to continue .","bug"
|
|
"6013","6114","Inconsistency between isreal and iscmplex for datetime64 array.
|
|
Assuming following inputs.
|
|
.
|
|
``iscomplex`` return bool array against all inputs OK).
|
|
.
|
|
Otherwise, ``isreal`` returns bool scalar for ``datetime64`` array.
|
|
Even though the [doc] http docs.scipy.org doc numpy reference generated numpy.isreal.html) says it should return Boolean array of same shape as x .
|
|
.","bug"
|
|
"6013","6116","on Raspberry Pi pip install numpy upgrade fails.
|
|
In file included from numpy core include numpy npy common.h 5 0, from numpy core src private npy sort.h 6, from numpy core src npysort mergesort.c.src 31 numpy core include numpy numpyconfig.h 4 26 fatal error numpyconfig.h No such file or directory compilation terminated.
|
|
In file included from numpy core include numpy npy common.h 5 0, from numpy core src private npy sort.h 6, from numpy core src npysort mergesort.c.src 31 numpy core include numpy numpyconfig.h 4 26 fatal error numpyconfig.h No such file or directory compilation terminated.
|
|
error Command gcc pthread fno strict aliasing DNDEBUG g fwrapv O2 Wall Wstrict prototypes fPIC Ibuild src.linux armv7l 2.7 numpy core src private Inumpy core include Ibuild src.linux armv7l 2.7 numpy core include numpy Inumpy core src private Inumpy core src Inumpy core Inumpy core src npymath Inumpy core src multiarray Inumpy core src umath Inumpy core src npysort Inumpy core include I usr include python2.7 Ibuild src.linux armv7l 2.7 numpy core src private Ibuild src.linux armv7l 2.7 numpy core src private Ibuild src.linux armv7l 2.7 numpy core src private Ibuild src.linux armv7l 2.7 numpy core src private c build src.linux armv7l 2.7 numpy core src npysort mergesort.c o build temp.linux armv7l 2.7 build src.linux armv7l 2.7 numpy core src npysort mergesort.o failed with exit status 1 Rolling back uninstall of numpy Command usr bin python c import setuptools; file = home pi build numpy setup.py ;exec compile open file ).read ).replace \r\n , \n ), file , exec )) install single version externally managed record tmp pip SP4Uhk record install record.txt failed with error code 1 in home pi build numpy Storing complete log in root .pip pip.log pi raspberrypi","bug"
|
|
"6013","6119","Broken iadd behavior on views.
|
|
Using ` =` with a view of the same array sometimes is broken.
|
|
To reproduce run .
|
|
This produces a bunch of zeros correct result), followed by a bunch of finite numbers of order 1 wrong).
|
|
For me the result is wrong for `n > 90`.
|
|
Replacing `h = h.T` with `h = h h.T` or anything equivalent fixes the issue.
|
|
I have verified that the issue appears in linux pip installation of numpy v1.9.2 in python 2, as well as python 3.
|
|
In particular, it is reproducible on http try.jupyter.org.","bug"
|
|
"6013","612","array split does not preserve dtype of original array when subarray length is 0.
|
|
Trac 2156).
|
|
Original ticket http projects.scipy.org numpy ticket 2156 on 2012 06 12 by fengy research, assigned to unknown.
|
|
The original dtype is u4, but the output becomes f8.
|
|
In [4] numpy.array split zeros shape=4, dtype= u4 ), 10) Out[4] [array [0], dtype=uint32), array [0], dtype=uint32), array [0], dtype=uint32), array [0], dtype=uint32), array [], dtype=float64), array [], dtype=float64), array [], dtype=float64), array [], dtype=float64), array [], dtype=float64), array [], dtype=float64)] The following code seems to be causing the problem def replace zero by x arrays sub arys) for i in range len sub arys)) if len nx.shape sub arys[i])) == 0 sub arys[i] = nx.array []) elif nx.sometrue nx.equal nx.shape sub arys[i]),0)) sub arys[i] = nx.array []) return sub arys shouldn t those nx.array []) be replaced by nx.array [], dtype=sub arys[i].dtype)?","bug"
|
|
"6013","6120","BUG regression in np.histogram.
|
|
After gh 6100 was merged, 4 tests for ``scipy.stats.histogram`` which uses ``np.histogram`` under the hood) have started failing ====================================================================== FAIL test increased bins test stats.TestHistogram) Traceback most recent call last) File home rgommers Code scipy scipy stats tests test stats.py , line 984, in test increased bins decimal=2) File home rgommers Code numpy numpy testing utils.py , line 510, in assert almost equal raise AssertionError build err msg )) AssertionError Arrays are not almost equal to 2 decimals ACTUAL 0 DESIRED 10 ====================================================================== FAIL test reduced bins test stats.TestHistogram) Traceback most recent call last) File home rgommers Code scipy scipy stats tests test stats.py , line 949, in test reduced bins decimal=2) File home rgommers Code numpy numpy testing utils.py , line 510, in assert almost equal raise AssertionError build err msg )) AssertionError Arrays are not almost equal to 2 decimals ACTUAL 0.07500000000000001 DESIRED 1.8749999999999996 ====================================================================== FAIL test simple test stats.TestHistogram) Traceback most recent call last) File home rgommers Code scipy scipy stats tests test stats.py , line 876, in test simple decimal=2) File home rgommers Code numpy numpy testing utils.py , line 886, in assert array almost equal precision=decimal) File home rgommers Code numpy numpy testing utils.py , line 708, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 2 decimals mismatch 20.0 ) x array [ 1., 1., 1., 2., 2., 1., 1., 0., 1., 1.])
|
|
y array [ 1., 1., 1., 2., 1., 2., 1., 0., 1., 1.])
|
|
====================================================================== FAIL test weighting test stats.TestHistogram) Traceback most recent call last) File home rgommers Code scipy scipy stats tests test stats.py , line 923, in test weighting decimal=2) File home rgommers Code numpy numpy testing utils.py , line 886, in assert array almost equal precision=decimal) File home rgommers Code numpy numpy testing utils.py , line 708, in assert array compare raise AssertionError msg) AssertionError Arrays are not almost equal to 2 decimals mismatch 20.0 ) x array [ 4. , 0. , 4.5, 0.9, 0. , 0.3, 110.2, 0. , 0. , 42. ])
|
|
y array [ 4. , 0. , 4.5, 0.9, 0. , 0.3, 7. , 103.2, 0. , 42. ])
|
|
As this is a regression, it should be looked at for 1.10","bug"
|
|
"6013","6121","Issue with ctypeslib on a Raspberry Pi.
|
|
Hi, I am using Numpy 1.6.2 on a Raspberry Pi and I encountered the following unexpected behaviour when trying to convert some arrays to ctypes; the first element is obviously wrong for some sizes of arrays list np.ctypeslib.as ctypes np.array range 5),dtype=float))) [0.0, 1.0, 2.0, 3.0, 4.0] list np.ctypeslib.as ctypes np.array range 6),dtype=float))) [0.0, 1.0, 2.0, 3.0, 4.0, 5.0] list np.ctypeslib.as ctypes np.array range 7),dtype=float))) [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0] list np.ctypeslib.as ctypes np.array range 8),dtype=float))) [5.654701094650191e 264, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0] list np.ctypeslib.as ctypes np.array range 9),dtype=float))) [5.654701094189978e 264, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0] list np.ctypeslib.as ctypes np.array range 10),dtype=float))) [5.654701094633645e 264, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0] list np.ctypeslib.as ctypes np.array range 4,8),dtype=float))) [4.0, 5.0, 6.0, 7.0] list np.ctypeslib.as ctypes np.array range 4,12),dtype=float))) [5.654702860194292e 264, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0] I have the same issues with other types integer, etc.
|
|
).","bug"
|
|
"6013","6122","operator.eq works on string arrays, but not np.equal does not.
|
|
This is on numpy 1.9.2 I m having trouble building on master right now) .
|
|
This is slightly disconcerting, especially given that we were considering delegating all binary operations to the corresponding NumPy ufuncs in 5844.","bug"
|
|
"6013","6123","Add a parameter for user to control precision when `random.choice )` ?.
|
|
I use [apollo] https github.com Russell91 apollo issues 5) and get the issue `ValueError probabilities do not sum to 1` I used pickle seriealized the probs, the error can be reproduced by the following test code.
|
|
file `test.py` .
|
|
file `test.pkl` .
|
|
Maybe we can add a parameter, so the user can control the precision such as 1e 8, or 1e 4) when comare sum of probs to 1.
|
|
Or are there a way to overcome this issue?
|
|
file `numpy random mtrand mtrand.pyx`, line 1094 to line 1099. .","bug"
|
|
"6013","6124","np.rec.fromarrays doesn t infer the shape for multidimensional arrays.
|
|
The following .
|
|
Gives `ValueError array shape mismatch in array 1`.
|
|
Should it not produce a `recarray` with elements `rec[0] == a[0], b[0]`, etc.
|
|
?","bug"
|
|
"6013","6125","unpleasant interaction between float ndarrays and python fractions.
|
|
This is causing trouble downstream.
|
|
I don t know if it s a numpy flaw regression bug or just a problem with the downstream project.
|
|
.","bug"
|
|
"6013","6127","BUG rounding error in divmod.
|
|
The `ndarray` implementation of `divmod` seems to do some incorrect rounding .","bug"
|
|
"6013","6128","numpy.linalg.norm returns nan for an array of int16.
|
|
Example .
|
|
The example is from http stackoverflow.com questions 31685156 normalizing vector produces nan in numpy, although it is not clear in that question if the array has type int16.","bug"
|
|
"6013","613","numpy.linalg.svd segmentation fault on large matrix Trac 2157).
|
|
Original ticket http projects.scipy.org numpy ticket 2157 on 2012 06 12 by trac user renz, assigned to pv.
|
|
numpy.linalg.svd reaches a segmentation fault only on a large matrix.
|
|
For example import numpy u, w, v = numpy.linalg.svd numpy.random.randn 10004, 565) 1j numpy.random.randn 10004, 565), full matrices = False) is fine.
|
|
But u, w, v = numpy.linalg.svd numpy.random.randn 10004, 566) 1j numpy.random.randn 10004, 566), full matrices = False) causes a segmentation fault on Mac OS X Lion.
|
|
I did a gdb run Program received signal EXC BAD ACCESS, Could not access memory.
|
|
Reason KERN INVALID ADDRESS at address 0x000000011d7c5540 0x00007fff8e3e2e7b in zgePack A ConjTran Unaligned ) gdb) bt 0 0x00007fff8e3e2e7b in zgePack A ConjTran Unaligned ) 1 0x00007fff8e3e25f1 in zgemm repack Unaligned ) 2 0x00007fff8e3e1d8e in APL zgemm ) 3 0x00007fff8e04b6c1 in cblas zgemm ) 4 0x00007fff8decca65 in zgemm ) 5 0x00007fff92be5fa9 in zlarfb ) 6 0x00007fff92c4f164 in zunmqr ) 7 0x00007fff92c4cf9a in zunmbr ) 8 0x00007fff92b6bfb3 in zgesdd ) 9 0x00000001010e2bfd in lapack lite zgesdd ) 10 0x00000001000a3feb in PyEval EvalFrameEx ) 11 0x00000001000a72a7 in PyEval EvalCodeEx ) 12 0x00000001000a7458 in fast function ) 13 0x00000001000a409d in PyEval EvalFrameEx ) 14 0x00000001000a72a7 in PyEval EvalCodeEx ) 15 0x00000001000a7326 in PyEval EvalCode ) 16 0x00000001000c5c09 in PyRun StringFlags ) 17 0x00000001000c6bc3 in PyRun SimpleStringFlags ) 18 0x00000001000d9e32 in Py Main ) 19 0x0000000100000f14 in dyld stub strlen ) This is a brand new Mac Pro.
|
|
I have never run program this large on this machine.","bug"
|
|
"6013","6130","Feature request Support for arbitrary byte length integers.
|
|
I have binary files where information is stored in 1 to 8 byte s) long integers depening on the file).
|
|
Would be nice to be able to manually specify byte length of integer when reading from binary file and adding the integers to a structured) numpy array.
|
|
At the moment I can only read files where integers are either 1, 2, 4 or 8 bytes long.","enhancement"
|
|
"6013","6132","random.choice doesn t allow float128 probs.
|
|
I think this is due to the `same kind` casting in numpy 1.10, and I don t know whether this is considered a bug.
|
|
.","bug"
|
|
"6013","6133","BUG array priority > 0 in float subclass causes comparison with np.float64 to fail.
|
|
Consider a subclass of `float` .
|
|
Now we try to compare an instance of `Bar` with a numpy float64 .
|
|
Puzzled, we try the following .
|
|
Running `test` over a few choice values we find note that the ` array priority ` of `np.float64` is ` 1000000.0`) .
|
|
As you can see, if ` array priority ` is greater than `0`, the comparison fails.
|
|
Whether our custom ` array priority ` is greater or less than `np.float64` s ` array priority ` doesn t matter.
|
|
Some of my colleagues tried this on their machines and we found that this problem only happens with python `2.7.6` 64 bit) and numpy `1.9.2`.
|
|
Using either that python version or that numpy version alone does not reproduce the bug.
|
|
It only happens when both of those versions are used together.
|
|
Here is a table of python and numpy versions and whether or not this bug shows up .
|
|
Also [posted to Stackoverflow] http stackoverflow.com questions 31712866 why does setting a custom classs array priority above zero make comparison 31713063 31713063).","bug"
|
|
"6013","6134","Strange type resolution of numpy.sqrt.
|
|
I found that numpy.sqrt has a strange type mapping .
|
|
It causes sqrt to upcast float16 to float32.
|
|
It is also strange that f and d appear twice.
|
|
Is this definition behavior intended?
|
|
Note I checked it with NumPy 1.9.2 on Python 2.7.9 3.4.3.","bug"
|
|
"6013","6136","Floating point repr.
|
|
.
|
|
Python floats use the shortest repr that give the same value when eval d since https bugs.python.org issue1580 was accepted.
|
|
It would be nice if numpy did the same.
|
|
I realize it would require a nontrivial patch )","bug"
|
|
"6013","614","genfromtxt masks non missing data Trac 2158).
|
|
Original ticket http projects.scipy.org numpy ticket 2158 on 2012 06 13 by trac user hoesterlin, assigned to unknown.
|
|
Under some conditions genfromtxt seems to mask not only missing data, but also data that happens to have the value as the filling values or its default if no filling values is given).
|
|
While setting the filling values to an unlikely value circumvents this i still find it troubling.
|
|
Oddly, it only appears if I read more than one column.
|
|
In the attached example code the 1 which is the default filling value for integers) is masked unnecessarily.","bug"
|
|
"6013","6149","Confusing error message from np.diff when axis is invalid.
|
|
Perhaps an edge case, but... .
|
|
The same error occurs when erroneously passing `None` e.g.
|
|
forgetting a `return` at the end of a previous function) to `diff`, as it gets interpreted as a 0 dim array, so even the default `axis=0` triggers the error.","bug"
|
|
"6013","615","add.accumulate w out array same as input causes memcpy overlap Trac 2159).
|
|
Original ticket http projects.scipy.org numpy ticket 2159 on 2012 06 13 by trac user shauncutts, assigned to unknown.
|
|
source looks like N.add.accumulate count, axis = 1, out = count ) where count is 2 dim array.
|
|
valgrind reports ==6607== Source and destination overlap in memcpy 0x23cdfdc0, 0x23cdfdc0, 8) ==6607== at 0x4C23ADE memcpy mc replace strmem.c 838) ==6607== by 0xAA24AAA PyUFunc ReductionOp ufunc object.c 3263) ==6607== by 0xAA25CD9 PyUFunc GenericReduction ufunc object.c 3482) ==6607== by 0x41BAFC PyObject Call abstract.c 2522) ==6607== by 0x17684B64 pyx pw 9factfiber 4stat 6pmodel 1c 9partition 9find emm partition partition.c 8867 This was a call from cython, but I don t think that is important.","bug"
|
|
"6013","6155","Inconsistent dtype behavior of numpy.where ).
|
|
The numpy.where ) function was rewritten in numpy 1.9 to speed it up.
|
|
I traced it to this changeset.
|
|
https github.com numpy numpy commit 593e3c30c24f0c61a271dc883c614724d7a57e1e The weird thing is the 1.9 behavior changed the resulting dtype in some situations when using scalar values as the second or third argument.
|
|
To illustrate, I wrote a simple test script and ran it against both numpy 1.7 and 1.9.
|
|
Here are the results .
|
|
In 1.7 the resulting dtype is consistent but then in 1.9 the resulting dtype is influenced by the value of the int.
|
|
It appears it is somehow related to whether the value falls within the range of a short.
|
|
Ideally it would not upcast to float64 if not necessary, but an alternative solution would be to have it consistently upcast.
|
|
The same tests with numpy.ma.where ) consistently produce float64 results.
|
|
Regardless of solution, the behavior should be more consistent when passing in a scalar without a cast.","bug"
|
|
"6013","6158","IO Error using np.tofile ) or np.fromfile ) with TemporaryFile.
|
|
Hi there, Having an issue running a test suite for a library which uses the `np.fromfile )` with a `TemporaryFile`.
|
|
It is throwing `IOError first argument must be a string or open file`.
|
|
I was able to reproduce the issue using both `tofile` and `fromfile` this all works happily with a real file.
|
|
Machine .
|
|
Sample Code .
|
|
Exception .","bug"
|
|
"6013","616","memory corruption on argmax 1, out=...) Trac 2160).
|
|
Original ticket http projects.scipy.org numpy ticket 2160 on 2012 06 14 by trac user mattip, assigned to unknown.
|
|
Running this one liner crashes python with a glibc memory corruption numpy.arange 15).reshape 5,3)).argmax 1, out=numpy.ones 3,1), dtype=int)) on centos6 64 bit Python 2.6.6 r266 84292, Dec 7 2011, 20 48 22) [GCC 4.4.6 20110731 Red Hat 4.4.6 3)] on linux2 >>> numpy.version.version 2.0.0.dev 4fcb53e","bug"
|
|
"6013","6162","np.nanstd doesn t work with masked arrays.
|
|
This issue may either be a feature request for ``np.ma.nan `` functions or a bug report for ``np.nanstd``.
|
|
The latter fails with a ``ValueError`` for a masked array, even when all the mask values are ``False``, and there are no NaN values in the array.
|
|
All the other ``np.nan `` functions but ``np.nanvar`` work with this same array as expected.
|
|
.","enhancement"
|
|
"6013","6163","AttributeError numpy.ndarray object has no attribute tobytes .
|
|
I run the code below .
|
|
but the error below is happen.
|
|
Do you give the reason why ndarray don t have tobytes )??
|
|
.","bug"
|
|
"6013","617","np.log gives RuntimeWarning for arrays with nan values when array is float32 but not float64 Trac 2161).
|
|
Original ticket http projects.scipy.org numpy ticket 2161 on 2012 06 15 by trac user cmorton, assigned to unknown.
|
|
I can easily suppress the warning by changing np.seterr, but it seemed like there shouldn t be a different result between the two.
|
|
>>> import numpy as np >>> np.seterr all= raise ) >>> np.log np.array [1, np.e, np.e 2, 0, np.nan]).astype np.float64)) array [ 0., 1., 2., inf]) >>>np.log np.array [1, np.e, np.e 2, 0, np.nan]).astype np.float32)) Traceback most recent call last) File <pyshell 57> , line 1, in <module> np.log np.array [1, np.e, np.e 2, 0]).astype np.float32)) FloatingPointError divide by zero encountered in log Sorry if this ticket is missing something, it is the first one I have ever done.","bug"
|
|
"6013","6170","linspace retstep=True) does not return step size in some cases.
|
|
bins, step = linspace 1,4,1,retstep=True) gives ValueError need more than 1 value to unpack this happens when num<=1","bug"
|
|
"6013","6176","Cannot convert boolean arrays to ctypes with ctypeslib.
|
|
ctypeslib seems to be unable to convert Boolean numpy arrays to ctypes.
|
|
.
|
|
Leads to .
|
|
This is using Python 2.7.10 and numpy 1.9.2 on Scientific Linux release 6.6 Carbon).
|
|
My naive fix I don t know much about the numpy internals and if this will break anything elsewhere) was adding the bool type to the type list in ctypeslib.py .
|
|
Now this works .","bug"
|
|
"6013","6178","Singular matrix inverse does t throw an error .
|
|
Hi, Here is an example of inv that doesn t throw an error when dealing with singular matrix and return wrong results.
|
|
import numpy as np import scipy as si A=np.array [[1,2,3],[4,5,6],[7,8,9]]) b=np.array [15,15,15]) print np.dot np.linalg.inv A),b) array [ 32., 48., 24.]
|
|
wrong result np.linalg.solve A,b) the result of np.dot np.linalg.inv A),b) should be same as this.
|
|
array [ 39., 63., 24.])
|
|
correct result It is necessary to say that scipy throws an error though.
|
|
np.dot si.linalg.inv A),b) My numpy version is 1.9.2 and python 2.7.6.","bug"
|
|
"6013","6179","Does Numpy or Scipy support solving the generalized singular value decomposition?
|
|
.
|
|
Hi, does anyone know if Numpy, Scipy, or any other python library now supports solving the generalized singular value decomposition like the `gsvd` command in MATLAB)?
|
|
I could not find any command related to generalized singular value decomposition in Numpy and Scipy.
|
|
Thanks in advance!","bug"
|
|
"6013","618","A few potential errors in the indexing documentation Trac 2162).
|
|
Original ticket http projects.scipy.org numpy ticket 2162 on 2012 06 15 by lebigot, assigned to unknown.
|
|
The documentation on array indexing http docs.scipy.org doc numpy 1.6.0 reference arrays.indexing.html ) contains a few details that look strange to me, that I list here in their order of appearance in the text 1 In order to remain backward compatible with a common usage in Numeric, basic slicing is also initiated if the selection object is any sequence such as a list) containing ) .
|
|
Wouldn t it be more correct to write any non ndarray sequence such as a list) ?
|
|
2 About negative indexes if i < 0, it means n i i I guess the correct value is d i i d i is the size of the i th dimension, in this part of the text).
|
|
3 About the first slice number i Then, if i is not given it defaults to 0 for k > 0 and n for k < 0 .
|
|
I guess the correct value is n 1 = last element), for k<0, no?","bug"
|
|
"6013","6184","Normal division is inconsistent with augmented division =) and changes dtype.
|
|
Numpy version 1.9.2 Python version 3.4.0 Ubuntu 14.04 .
|
|
I would expect the normal division ) to work the same way as augmented division =), but it seems like numpy changes the dtype when using the normal division.
|
|
Is this by design, or a bug?","bug"
|
|
"6013","6186","numpy array to matrix conversion overhead.
|
|
There seems to be a large overhead when converting a numpy.array to a numpy.matrix.
|
|
I am trying to make a small function returning a matrix, but the overhead in creating the np.matrix is really significant.
|
|
See the follow code .
|
|
The output of the code above is .
|
|
Is there a way to create a numpy.matrix object with less overhead?
|
|
Using a global object and making a copy works, but is not my ideal solution.","bug"
|
|
"6013","6189","PyArray Type busted with ansii compilers and one file compile..
|
|
When compiling as one file, `NPY NO EXPORT` is `static`.
|
|
Static structures must be initialized with constants.
|
|
`PyObject HashNotImplemented` is not a constant.
|
|
.","bug"
|
|
"6013","619","BLAS matrix product dot) never used for ndim > 2 tensordot does not use BLAS) Trac 2163).
|
|
Original ticket http projects.scipy.org numpy ticket 2163 on 2012 06 15 by trac user thatistosay, assigned to unknown.
|
|
dotblas matrixproduct ) contains the comment This function doesn t handle dimensions greater than 2 and calls PyArray MatrixProduct2 ) for these cases.
|
|
This means BLAS is never used for calls to dot ) with arguments of ndim>2!!
|
|
In more detail...
|
|
If I want to contract a pair of tensor indices for ndim=3 such that A = np.rand d,D,D); B = np.rand d,D,D) AB[s,i,t,j] == sum A[s,i, ], B[t, ,j]) then currently, although it can be done in a single line res = sp.dot A,B) it can often be done much faster with explicit python!)
|
|
loops res = np.zeros d,d,D,D)) for s in xrange d) for t in xrange d) np.dot A[s], B[t], out=res[s,t]) res = np.rollaxis res, 2, 1) ..assuming dot ) is using optimized BLAS for ndim=2, and the dimensions are large enough so that calling BLAS is worth it.
|
|
In general, reproducing the behaviour of dot ) for ndim>2 is just a matter of calling GEMM in loops as above and then calling rollaxis ) once.
|
|
I therefore propose doing this within blasdot.c as far as possible to eliminate the use of python loops) so that ndim>2 dot ), and tensordot ), can benefit from BLAS.
|
|
Some comparisons of the two methods above attached script) dtype=complex128 AB[s,i,t,j] = sum A[s,i, ], B[t, ,j]) A.shape = 16, 512, 512); B.shape = 16, 512, 512) looping over 2D dot ) vs. 3D dot ) 24 about 4 times faster) A.shape = 20, 64, 64); B.shape = 20, 64, 64) looping over 2D dot ) vs. 3D dot ) 35 A.shape = 20, 48, 32); B.shape = 20, 32, 48) looping over 2D dot ) vs. 3D dot ) 45 A.shape = 32, 32, 16); B.shape = 32, 16, 32) looping over 2D dot ) vs. 3D dot ) 82 A.shape = 64, 10, 8); B.shape = 64, 8, 10) looping over 2D dot ) vs. 3D dot ) 158 slow python loops..) this was on a 4 core i7 system using ATLAS under heavy load)","enhancement"
|
|
"6013","6191","DOC numpy.ctypeslib.load library has no docstring..
|
|
The subject says it all.","bug"
|
|
"6013","6192","Pythnon 3 and numpy scalars have different comparison rules.
|
|
In particular Python 3 scalars will refuse to be compared against a list while numpy will compare elementwise .","bug"
|
|
"6013","6196","assert allclose no longer works with ndarray subclasses that use units.
|
|
This can be seen from the following test script .
|
|
`g` in this context is an instance of `YTArray`, a subclass of `ndarray` that uses a sympy powered unit system.
|
|
Under NumPy 1.9.2, this runs without error.
|
|
Under NumPy 1.10b1, I get the following traceback .
|
|
FWIW, this isn t isolated to `YTArray`, I get the same traceback if I use `from astropy.units import g`.
|
|
One small difference in yt this fails with `atol=0` the default) but astropy is ok with it due to a difference in the way we treat zero in our respective unit systems.
|
|
For yt it would be particularly annoying to update our test suite to adapt to this by making sure `atol` has units attached to it everywhere we use assert allclose, particularly since it s sometimes used indirectly.
|
|
Looking at the NumPy source, it seems the `compare` function used in `assert allclose` has changed since NumPy 1.9.2.
|
|
Would it be possible to revert to the old implementation, or generalize the one used here so it doesn t break unit consistency in this case, subtracting a scalar without units from a quantity passed in with units)?","bug"
|
|
"6013","6197","Confusing warning for median of empty array.. .
|
|
Oops.
|
|
[numpy 1.9.2]","bug"
|
|
"6013","6198","Inplace Operation Casting Problem.
|
|
x = np.array [1, 2, 3]) x = x 1.2 > array [ 1.2, 2.4, 3.6]) does not yield the same result x = np.array [1, 2, 3]) x = 1.2 > array [1, 2, 3]) For all operations , , , ) and probably all other binary operations.
|
|
Is this expected behaviour?
|
|
Tested with numpy 1.9.2 and 1.8.0","bug"
|
|
"6013","620","Incorrect overflow warning message for float128 Trac 2164).
|
|
Original ticket http projects.scipy.org numpy ticket 2164 on 2012 06 15 by trac user edcjones, assigned to unknown.
|
|
I use up to date Debian testing wheezy), amd64 architecture.
|
|
I have installed packages python3, version 3.23 rc1 2, and python3 numpy, version 1 1.6.2 1.
|
|
For float128, I get an overflow warning message, when I calculate 2 8192.
|
|
The maximum exponent should be 16384.
|
|
Program !
|
|
usr bin env python3.2 import numpy print float128 ) fi = numpy.finfo numpy.float128) print ibeta , fi.machar.ibeta) print maxexp , fi.machar.maxexp) print xmax , fi.machar.xmax) two = numpy.float128 2) big = numpy.float128 8191) x = numpy.power two, big) OK print 2 8191 , x) big = numpy.float128 8192) x = numpy.power two, big) Output float128 ibeta 2 maxexp 16384 xmax 1.18973149536e 4932 2 8191 5.4537406781e 2465 .
|
|
z97.py 16 RuntimeWarning overflow encountered in power x = numpy.power two, big)","bug"
|
|
"6013","6201","Clarify return value for rfft2.
|
|
The following example may leave users perplexed .
|
|
After all, ``rfft2`` is just meant to be a version of ``fft2`` that is more efficient for real input.
|
|
Now I m assuming that what is different is that the negative frequencies are missing along one axis for rfft2 since they are redundant, but at the same time it seems inconsistent to do it along one axis and not the other.
|
|
The asymmetric array is due to the fact the 2 d rfft is done as a 1d rfft followed by a normal fft.
|
|
Ideally, ``rfft2`` should be fixed to actually return a complete `` 10,10)`` fourier transform to avoid requiring the user to manually mirror the values to make it back into a 10,10) array.
|
|
``rfft2`` should be usable as a drop in replacement for ``fft2``.
|
|
However, I presume this could break a lot of code, so as a plan B, ``rfft2`` could be given an optional argument to return the complete array defaulting to ``False`` to preserve backward compatibility).
|
|
Finally, either way, the docstring of ``rfft2`` should definitely be clarified to explain the shape of the output.
|
|
If others agree, I can try and implement this.","bug"
|
|
"6013","6204","Scipy compile broken for dual arch OSX at numpy >= 1.9.0.
|
|
Our standard way to compile scipy is with the dual arch gfortran 4.2.3 in https github.com MacPython scipy wheels tree master archives and other places.
|
|
Scipy compiles without error against numpy==1.8.2, but raises the following error among many other warnings compiling against numpy==1.9.0 and later .","bug"
|
|
"6013","6205","recfunctions improvements.
|
|
I would like to suggest a few improvements to the functions in `np.lib.recfunctions`.
|
|
They are backwards incompatible, but given that there are some suggestions e.g.
|
|
5008) to move that functionality to the root namespace, perhaps at least the new functions in the root namespace can get the improvements.
|
|
drop the `usemask` kwarg from all functions, and just return a masked array iff.
|
|
the input array is masked at the very least, the default should be changed to `False`).
|
|
drop the `asrecarray` kwarg, and just return a recarray iff.
|
|
the input is a recarray.
|
|
drop the `rec ` functions, which are AFAICT just duplicates of the functions without the `rec ` prefix but with `asrecarray` set to True).","enhancement"
|
|
"6013","6207","Non standard Calendar Support.
|
|
The current `datetime64` API only supports the standard Gregorian calendar.
|
|
In the climate science community and presumably others), a number of other calendars systems are used, for example [Julian Calendar] https en.wikipedia.org wiki Julian calendar) [No Leap Day Calendar] http cfconventions.org Data cf conventions cf conventions 1.6 build cf conventions.html calendar) Gregorian calendar without leap years, i.e., all years are 365 days long.
|
|
[All Leap Day Calendar] http cfconventions.org Data cf conventions cf conventions 1.6 build cf conventions.html calendar) Gregorian calendar with every year being a leap year, i.e., all years are 366 days long.
|
|
I d like to propose adding calendar support to the `datetime64` API with the hope that changes here will trickle down into [pandas] https github.com pydata pandas) and [xray] https github.com xray xray).
|
|
I m not sure exactly how I d like to see the API and would like to hear from those familiar with `datetime64` on how it could be done.
|
|
Of course, the default calendar behavior should remain the Gregorian calendar.
|
|
xref https github.com xray xray issues 521, https github.com xray xray pull 523, https github.com pydata pandas issues 7307 cc shoyer rabernat ocefpaf","enhancement"
|
|
"6013","6209","bug in nansum with non float64 dtypes.
|
|
These non float ``nansum s`` are a little suspect xref https github.com pydata pandas pull 10815 .","bug"
|
|
"6013","621","Errors with for numpy.float16 Trac 2165).
|
|
Original ticket http projects.scipy.org numpy ticket 2165 on 2012 06 15 by trac user edcjones, assigned to unknown.
|
|
I use up to date Debian testing wheezy), amd64 architecture.
|
|
I have installed packages python3, version 3.23 rc1 2, and python3 numpy, version 1 1.6.2 1. makes mistakes for float16.
|
|
power works correctly.
|
|
The attached program outputs 2.0 5.9605e 08 4.0","bug"
|
|
"6013","6210","clang build error conflicting types for PyUFunc FromFuncAndData .
|
|
I am trying to build numpy master on a new Python 3.4.3 virtualenv with OSX 10.10.5 provided clang .
|
|
I get the following error when building with `python setup.py build` without any custom site.cfg) .
|
|
I have tried to google a bit and could not find any related issue.
|
|
Can other people build numpy under OSX 10.10.5?","bug"
|
|
"6013","6213","numpy 1.9.x on Py3 modifies `environ`.
|
|
See issue dvarrazzo py setproctitle 42 It seems that numpy 1.9 on Python 3 modifies the `environ` variable.
|
|
Reproduced on Debian Jessie and OS X 10.10.4.
|
|
Doesn t happen on numpy 1.8 and on Python 2.
|
|
Info to verify the behaviour in the issue above.
|
|
This breaks `setproctitle`, unless `setproctitle` is imported before `numpy`, in which case I wonder if `numpy` won t break... Is this a known behaviour of numpy 1.9 or did it happen accidentally?
|
|
Is there a workaround to avoid clobbering `environ`?
|
|
For a rationale about why clobbering `environ` could be undesirable, take a look at dvarrazzo py setproctitle 15.","bug"
|
|
"6013","6215","Wrong result for integer power on underflow.
|
|
Should be 0, but returns infinity .","bug"
|
|
"6013","6217","stable reversed indirect sort?.
|
|
Looks like there is no easy way to achieve a stable reversed indirect sort with the sorting functions in numpy argsort )[ 1] would give a reversed indirect sort.
|
|
argsort kind= mergesort )[ 1] does not give a stable reversed indirect sort.
|
|
[ 1] makes it unstable.
|
|
Would you consider including a reversed parameter to argsort?","bug"
|
|
"6013","622","npyio.py > recfromcsv seems to have a bug and ignores any dtypes provided Trac 2168).
|
|
Original ticket http projects.scipy.org numpy ticket 2168 on 2012 06 16 by trac user kobejohn, assigned to unknown.
|
|
I found this due to someone having a problem not being able to load large integers with recfromcsv.
|
|
After playing with it, I think it is simply ignoring all the dtype information provided.
|
|
[https github.com numpy numpy blob master numpy lib npyio.py L1859 Line 1859 in npyio.py, part of recfromcsv )] tries to either bring in the dtypes passed or set None.
|
|
However, it actually brings in a kwarg that I couldn t find anywhere else called update which looks suspiciously like a copied mistype from earlier in the line.
|
|
If I change update > dtype , then the types get picked up as expected.
|
|
I found this in the github repo although I may have been looking at the wrong branch) as well as 1.6.2","bug"
|
|
"6013","6221","Advanced indexing using arrays created by as strided.
|
|
Instead of using <code>np.indices< code>, I tried the following method which creates indices for advanced indexing using as strided, but it behaved quite strange at least in NumPy 1.9.2).
|
|
E.g.
|
|
<code>C=3< code> was fine but <code>C=4< code> gave error out of bound IndexError).
|
|
I was using NumPy 1.9.2 MKL) under Python 2.7.10 Anaconda 2.3.0 64 bit), and my system s NumPy 1.9.2 Ubuntu 12.04 with Python 2.7.3) also got the same problem.
|
|
However, NumPy 1.6.1 on another machine) seemed to be fine.
|
|
.","bug"
|
|
"6013","6225","Bug in assignment with integer indexing to array subset.
|
|
This fails with the expected `ValueError` >>> arr = np.zeros 10) >>> arr[5 7] = np.arange 1,10) As does this >>> idx = np.zeros 10, dtype=bool) >>> idx[5 7] = True >>> arr[idx] = np.arange 1,10) However, this silently completes the assignment >>> arr[[5,6]] = np.arange 1,10) >>> arr array [ 0., 0., 0., 0., 0., 1., 2., 0., 0., 0.])
|
|
As does this >>> arr[np.arange 10)] = [1,2] >>> arr array [ 1., 2., 1., 2., 1., 2., 1., 2., 1., 2.])
|
|
I am running numpy 1.9.2 with python 2.7.8.
|
|
According to the [documentation] http docs.scipy.org doc numpy user basics.indexing.html assigning values to indexed arrays) the last two examples should fail because the index array and the assignment array are not shape consistent.","bug"
|
|
"6013","6229","numpy.sign behaves oddly when dtype is numpy.object.
|
|
In [1] import numpy In [2] foo = numpy.array [ .1, 0, .1]) In [3] numpy.sign foo) Out[3] array [ 1., 0., 1.])
|
|
In [4] numpy.sign foo.astype numpy.object)) Out[4] array [1, 1, 1], dtype=object) The underlying data is still just a float In [5] type foo.astype numpy.object)[0]) Out[5] float This result when using an array of numpy.object does not seem desirable.
|
|
Either it should operate correctly on the underlying numbers or if called with invalid data strings, etc) it should throw an exception.
|
|
This way I could populate an array with heterogeneous data, say a mix of integers and floats, and still get a valid result.
|
|
I tested this on Mac OS 10.6.8 Python 3.4.2 numpy 1.9.2","bug"
|
|
"6013","623","Wrong result when calculating the mean masking nan values on a 64 bit system Trac 2169).
|
|
Original ticket http projects.scipy.org numpy ticket 2169 on 2012 06 19 by trac user knopfra, assigned to unknown.
|
|
The array provided in attachment as a pickle file can be used to detect the following issue 64 bit Linux system, python 2.7.3, numpy 1.6.1).
|
|
The array has 15606478 elements.
|
|
In [1] import numpy as np In [2] a = np.load data.pkl ) In [3] np.nanmin a)[[BR]] Out[3] 4.715836 In [4] np.nanmax a)[[BR]] Out[4] 4.7189121 In [5] idx = np.where np.isfinite a)) In [6] a[idx].mean )[[BR]] Out[6] 4.1792714738680736 In [7] from scipy.stats import nanmean In [8] nanmean a)[[BR]] Out[8] 4.1792714738680727 The mean value obtained is clearly wrong.
|
|
On a 32 bit system the result is the following In [1] import numpy as np In [2] a = np.load data.pkl ) In [3] idx = np.where np.isfinite a)) In [4] a[idx].mean )[[BR]] Out[4] 4.7184738182116019 and this time the mean value is correct.","bug"
|
|
"6013","6230","deprecate np.math.
|
|
This is just the standard math library.
|
|
.
|
|
Related https github.com numpy numpy pull 6103, https github.com numpy numpy issues 2448 issuecomment 21218184.","bug"
|
|
"6013","6231","add Welford s algorithm for stable low memory calculation of mean and variance.
|
|
[Welford s algorithm] https en.wikipedia.org wiki Algorithms for calculating variance Online algorithm) computes the mean and variance in a way that pays attention to stability while using less memory overhead than the current numpy implementation of variance or standard deviation.
|
|
See http stackoverflow.com questions 32135572, https github.com numpy numpy issues 1696, https github.com numpy numpy pull 3685.","enhancement"
|
|
"6013","6232","xe2x80x9dTwo fields with the same namexe2x80x9d should tell user which fieldname.
|
|
I am creating a structured dtype with 126 fields, constructed from a table defining a particular file format.
|
|
This results in a `ValueError two fields with the same name`, but it doesn t tell me what the fieldname is that occurs twice or more).
|
|
It would be quite helpful if it did.
|
|
.","bug"
|
|
"6013","6235","Negative dtype.itemsize.
|
|
While tracking down something else, I stumbled upon a way to make a ``dtype`` with an ``itemsize`` that is negative, which is surely a bug.
|
|
Essentially, with a structured ``ndarray`` one with fields), while size checking is done on each field individually and probably on their sum, it seems that overflow checking is not done.
|
|
Let X either be the bit size of the internal representation of ``itemsize``.
|
|
So far, I have found that even in 64 bit Python both 2.x and 3.x), it can be 32 bit or 64 bit depending on the build is there a build option controlling it when Numpy is built?).
|
|
So if you try to make a structured Numpy ``ndarray`` with one field that is too big by one byte, an error is thrown like it should .
|
|
Put it at the maximum size and there is no problem you will note that I am running the examples from a build where X is 64, but it was 32 bit on the machine I was using earlier today) as it should.
|
|
.
|
|
But make a second field of the same size and one gets a negative value of ``itemsize`` .
|
|
Which should not be allowed.
|
|
There was an overflow that was not detected.
|
|
I do not know what a negative ``itemsize`` will when the dtype is passed to other Numpy functions or other packages.
|
|
It is hard to imagine it happening when ``X == 64`` on accident, but when ``X == 32`` it is pretty easy, say one got a MAT file from someone with a structure with a total size greater than 2 GB but each field smaller than 2 GB and tries to import it as a structured ``ndarray`` with SciPy haven t tested this yet).","bug"
|
|
"6013","6237","High CPU usage.
|
|
I am running numpy 1.9.2 on a Wandboard quad ARM processor).
|
|
If I open python and type import numpy immediately my CPU usage pegs for three of the four processors.
|
|
I am not sure if this is a bug or not, but have not found anything googling the problem.
|
|
Thanks","bug"
|
|
"6013","6238","numpy.arange miscounts when step is large relative to value past stop.
|
|
.","bug"
|
|
"6013","6239","ctypeslib.ndpointer should support passing None as the argument.
|
|
Consider the following commonly used pattern.
|
|
We have a function with this signature void foo double array, int length) We call it twice.
|
|
First to find the required array length, then again to populate the array.
|
|
The call sequence looks like this int length; foo NULL, length); double array = malloc length sizeof array); foo array, length); When we try to wrap this using numpy and ctypes we wish to use `ctypeslib.ndpointer` to create the `argtypes` specification for the array.
|
|
That allows us to pass the numpy object s memory to the external function and avoid unnecessary copying.
|
|
However, this fails.
|
|
We write lib.foo None, length) which produces an error of this nature > ctypes.ArgumentError argument 1 argument must be an ndarray This could be solved if the type returned by `ctypeslib.ndpointer` had a `from param` that handled being passed `None`.
|
|
If instead of treating `None` as an error, `from param` returned `None` in that scenario, this calling pattern would work out.
|
|
As an illustration, I can subclass the type returned by `ctypeslib.ndpointer` myself to support this calling pattern.
|
|
DoubleArrayTypeBase = numpy.ctypeslib.ndpointer dtype=numpy.float64, ndim=1, flags= C CONTIGUOUS ) def from param cls, obj) if obj is None return obj return DoubleArrayTypeBase.from param obj) DoubleArrayType = type DoubleArrayType , DoubleArrayTypeBase,), { from param classmethod from param)} ) It would be great if this could be handled by numpy direcrtly.
|
|
Reference http stackoverflow.com q 32120178 505088","bug"
|
|
"6013","624","Wrong result when calculating the mean masking nan values on a 64 bit system Trac 2170).
|
|
Original ticket http projects.scipy.org numpy ticket 2170 on 2012 06 19 by trac user knopfra, assigned to unknown.
|
|
The array provided as a pickle file at the following link [http dl.dropbox.com u 30592272 data.pkl.gz] can be used to detect the following issue 64 bit Linux system, python 2.7.3, numpy 1.6.1).
|
|
The array has 15606478 elements.
|
|
In [1] import numpy as np In [2] a = np.load data.pkl ) In [3] np.nanmin a)[[BR]] Out[3] 4.715836 In [4] np.nanmax a)[[BR]] Out[4] 4.7189121 In [5] idx = np.where np.isfinite a)) In [6] a[idx].mean )[[BR]] Out[6] 4.1792714738680736 In [7] from scipy.stats import nanmean In [8] nanmean a)[[BR]] Out[8] 4.1792714738680727 The mean value obtained is clearly wrong.
|
|
On a 32 bit system the result is the following In [1] import numpy as np In [2] a = np.load data.pkl ) In [3] idx = np.where np.isfinite a)) In [4] a[idx].mean )[[BR]] Out[4] 4.7184738182116019 and this time the mean value is correct.","bug"
|
|
"6013","6240","dtype signature for arithmetic is not fixed.
|
|
NumPy operations generally follow the rule that the output dtype shape is a deterministic function of the inputs dtype shape.
|
|
However, binary arithmetic with 0d arrays appears to be an exception to these rules.
|
|
Consider .
|
|
Basically, whether the result is float32 depends on the size of the scalar argument.
|
|
I can see why this would be desirable for operations with Python scalars e.g., `array 1`), but when the 0d scalar argument is already typed this is highly surprising.","bug"
|
|
"6013","6241","moni.
|
|
Apologies for the unintended issue.","bug"
|
|
"6013","6246","numpy.fromfile OSError first argument must be an open file Python 3) .
|
|
Using Python 3 and Numpy 1.9.2 there is an issue with `numpy.fromfile` when called on an existing file object.
|
|
This code works fine in Python 2.7. .
|
|
I ve checked `fh.closed` is `False`, so the file is still open.
|
|
The error is thrown by [numpy.multiarraymodule.c 2119] https github.com numpy numpy blob c29733c228ef6168b2f811f51113e9660bf6c47d numpy core src multiarray multiarraymodule.c L2119).
|
|
Someone [mentioned these issues on Stack Overflow] http stackoverflow.com questions 30466656 using new io classes with numpy fromfile tofile in python 2 converting io class), but no clear solutions are present.
|
|
Any suggestions?","bug"
|
|
"6013","6247","Document numpy.genfromtxt with list of strings.
|
|
As discovered [with this question answer] http stackoverflow.com a 32217754 327026), `numpy.genfromtxt` can be used to read a list of strings.
|
|
However, the first argument `fname` is [currently documented] http docs.scipy.org doc numpy reference generated numpy.genfromtxt.html) to take file or str types, and the documentation does not mention of this feature.","bug"
|
|
"6013","6248","Implement round specialmethod for ndarrays?.
|
|
This would allow calling `round array)`.
|
|
While this would certainly be equivalent to `np.round array)`, a similar feature is already there with `abs array)` which calls ` abs `) and `np.abs array)`.","enhancement"
|
|
"6013","625","Corner case problem for Python int ) and numpy.float128 Trac 2171).
|
|
Original ticket http projects.scipy.org numpy ticket 2171 on 2012 06 19 by trac user edcjones, assigned to unknown.
|
|
I use up to date Debian testing wheezy) linux, amd64 architecture.
|
|
I have installed packages python3.2, version 3.2.3 rc2 1 and python3 numpy, version 1 1.6.2 1.
|
|
I got the same answers with packages python2.7, version 2.7.3 rc2 2.1, and python numpy, version 1 1.6.2 1.
|
|
Create a numpy.float128 scalar whose value is 2 64 1, and call it x .
|
|
int x) is incorrect.","bug"
|
|
"6013","6250","Reference leak when comparing structured scalars.
|
|
If you execute the following code, you ll it leaks a reference to the array .
|
|
The leak disappears if you remove the comparison `u == v`) or if you remove the character string member in the dtype the `d` member here).
|
|
I ve tried to diagnose this but I m ending up completely lost in the comparison logic which creates multiple temporary arrays).","bug"
|
|
"6013","6251","What is wrong if ``np.zeros 2) == 0`` returns ``array [False, True], dtype=bool)``?.
|
|
Hi, my first issue here in this great package so please bear with me ) As I encounter the above and this strange behaviour in a HPC environment.
|
|
.
|
|
prints .
|
|
.
|
|
.
|
|
NumPy compilation with .
|
|
Python 3.4.3 CPython GCC 4.3.4 [gcc 4 3 branch revision 152973 Linux 2.6.32.59 0.7 default x86 64 with SuSE 11 x86 64 .
|
|
.
|
|
By the way, under Python 2.7.9 on the same system with the same setup, this bug does not occur.","bug"
|
|
"6013","6253","np.linalg.pinv under ubuntu anaconda.
|
|
np.linalg.pinv covR) returns wrong results when covR > 1000x1000).
|
|
The above only happens under ubuntu under windows it worked ok).
|
|
I have not tested it on a second linux machine.
|
|
I temporary resolved the problems by switching np.linalg.pinv covR) to scipy.linalg.pinv2 covR).
|
|
p.s.
|
|
the problem starts when covR becomes 1001x1001), up to 1000x1000) everything is ok.","bug"
|
|
"6013","6255","np.sqrt.types is weird.
|
|
There are duplicated entries, and also `e >e` comes after `f >f`, which probably means there ll always be a conversion to float32.
|
|
.","bug"
|
|
"6013","6257","NumPy version written out incorrectly in egg info filename.
|
|
NumPy prefers using `distutils` to `setuptools` for the `install` step and thus creation of the `egg info` filename.
|
|
This means that the `egg info` filename gets the ` ` in the version replaced with ` ` and written out as `numpy 1.11.0.dev0 2329eae.egg info`, despite the fact that the version is `numpy 1.11.0.dev0 2329eae`.
|
|
This causes a problem for satisfying requirements for other packages, since `setuptools` sees this as a `LegacyVersion` after replacing the ` ` with a ` ` in the filename), and thus you can get errors like this .
|
|
One solution that might work is using `setuptools` for the `install` step, because `setuptools` creates the name with the ` ` in it and it thus gets parsed properly.
|
|
However, given the lengths taken to triage `distutils` vs `setuptools` in that step, I assume there is a reason why `distutils` is preferred.
|
|
And it looks like back in 2013 this was addressed https github.com scipy scipy commit a4e93fbd1c1e192b699e1ff545236dab7cffc579 diff 2eeaed663bd0d25b7e608891384b7298L185 Maybe `setuptools` would be okay to use now...?
|
|
Related `setuptools` ticket https bitbucket.org pypa setuptools issues 419 setuptools pkg resources fails to detect Any ideas for how to solve this problem?","bug"
|
|
"6013","626","npyio.py Make .squeeze ) an option for genfromtxt ) Trac 2172).
|
|
Original ticket http projects.scipy.org numpy ticket 2172 on 2012 06 19 by trac user lmeyn, assigned to unknown.
|
|
I m trying to read several data files with column headers using numpy.genfromtxt.
|
|
The issue is that some of the data files only have one row of data and squeeze operation used in genfromtxt returns an object with no rows shape = ) ).
|
|
I would suggest adding an argument, squeeze=True, and the replace the last three lines if unpack return output.squeeze ).T return output.squeeze ) with if squeeze output = output.squeeze ) if unpack return output.T return output This would keep the current behavior if the squeeze argument isn t specified, yet allow the squeeze operation to be avoided if desired.","bug"
|
|
"6013","6261","bug in searchsorted when multiple values are provided.
|
|
I m using Python 2.7.9 on Linux, numpy 1.9.2) The following code shows that min max indices are not maintained correctly in the C loop https github.com numpy numpy blob master numpy core src npysort binsearch.c.src L55 L61) .
|
|
Output .
|
|
I believe it s a 1.9 regression, with numpy 1.7.1 everything is OK.","bug"
|
|
"6013","6265","MANT Replace PyObject Compare with PyObject RichCompareBool.
|
|
See https github.com numpy numpy pull 6236 issuecomment 133768265.
|
|
Since our oldest supported Python version is 2.6, there is no longer a need for separate compilation paths for Python 2.x and 3.x, for Python object comparison, since both provide `PyObject RichCompareBool`.
|
|
It may also be a good idea to get rid of the `Pyobject Cmp` function, defined for Python 3.x in [`npy 3kcompat.h`] https github.com numpy numpy blob maintenance 1.10.x numpy core include numpy npy 3kcompat.h L324), which was recently discovered to be broken by pv, see https github.com numpy numpy issues 6229 issuecomment 133458966.
|
|
I ll try to put a PR together for this shortly, unless someone else wants the glory for themselves.","bug"
|
|
"6013","6266",".imag, .real, .angle don t call array wrap after their execution .
|
|
`.imag`, `.real`, `.angle` don t call ` array wrap ` after their execution Looks like a bug.
|
|
https github.com pydata pandas issues 10921 issuecomment 135846584","bug"
|
|
"6013","627","int ) segfaults for float128 Trac 2173).
|
|
Original ticket http projects.scipy.org numpy ticket 2173 on 2012 06 19 by trac user edcjones, assigned to unknown.
|
|
I use up to date Debian testing wheezy) linux, amd64 architecture.
|
|
I have installed packages python3.2, version 3.2.3 rc2 1 and python3 numpy, version 1 1.6.2 1.
|
|
I got the same answers with packages python2.7, version 2.7.3 rc2 2.1, and python numpy, version 1 1.6.2 1.
|
|
The second line causes a segmentation fault M = numpy.finfo numpy.float128).max i = M. int )","bug"
|
|
"6013","6270","Scipy 0.16.0 test failure with maintenance 1.10.x.
|
|
On Windows using numpy mkl, I get a scipy test failure with numpy maintenance 1.10.x because `np.histogram` returns different results for numpy 1.9.2 and 1.10.x .
|
|
Standalone code .
|
|
numpy 1.9.2 .
|
|
numpy 1.10.x .","bug"
|
|
"6013","6272","Automatic copy on overlap.
|
|
Tracking issue as a continuation to gh 6166 [ ] add copy on overlap to nditer, as optional flag [ ] use it for ufuncs [ ] check whether copy on overlap is respected in indexed assignment [ ] respect it for `dot )` [ ] check other places in numpy that may need checking Some initial code, https github.com numpy numpy compare master...pv copy overlap","bug"
|
|
"6013","6273","f2py broken on Mac OS X 10.10.5?.
|
|
<b> Edit somehow, the parameters undefined dynamic lookup bundle are not being input to the link command at the end on my workstation compile.
|
|
When that is added, the compile works.
|
|
somehow the numpy distutils fcompile gnu.py is an old version.
|
|
Sorry to have bothered this list < b> Not sure how this happened whether its an os update, or something to do with numpy update in the last 30 60 days.
|
|
I am using the conda environment from Continuum Whether I use gfortran 5.1, or 5.0, or 4.9, and I have tried numpy 1.9.1 and numpy 1.9.2, the classic example case FIB3.f) from the docs page no longer works.
|
|
e.g., the code from https sysbio.ioc.ee projects f2py2e usersguide fib3.f) right now, my typical environment is gfortran v Using built in specs.
|
|
COLLECT GCC=gfortran COLLECT LTO WRAPPER= usr local libexec gcc x86 64 apple darwin14.4.0 5.1.0 lto wrapper Target x86 64 apple darwin14.4.0 Configured with .. gcc 5.1.0 configure enable languages=c ,fortran Thread model posix gcc version 5.1.0 GCC) The errors I get using either the standard build or the the f2py c m fib3 fib3.f fails with this output.
|
|
Clearly the key is at the bottom ld symbol s) not found for architecture x86 64 is why this is failing, but looking at all the output I cannot understand where there is a problem with this arch issue.
|
|
running build running config cc unifing config cc, config, build clib, build ext, build commands compiler options running config fc unifing config fc, config, build clib, build ext, build commands fcompiler options running build src build src building extension fib3 sources f2py options [] adding build src.macosx 10.5 x86 64 2.7 fortranobject.c to sources.
|
|
adding build src.macosx 10.5 x86 64 2.7 to include dirs.
|
|
build src building npy pkg config files running build ext customize UnixCCompiler customize UnixCCompiler using build ext customize Gnu95FCompiler Found executable usr local bin gfortran customize Gnu95FCompiler customize Gnu95FCompiler using build ext building fib3 extension compiling C sources C compiler gcc fno strict aliasing I Volumes Master 1 Users wicker anaconda include arch x86 64 DNDEBUG g fwrapv O3 Wall Wstrict prototypes I usr local include I usr local include compile options Ibuild src.macosx 10.5 x86 64 2.7 I Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include I Volumes Master 1 Users wicker anaconda include python2.7 c gcc build src.macosx 10.5 x86 64 2.7 fortranobject.c In file included from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy ndarraytypes.h 1804 0, from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy ndarrayobject.h 17, from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy arrayobject.h 4, from build src.macosx 10.5 x86 64 2.7 fortranobject.h 13, from build src.macosx 10.5 x86 64 2.7 fortranobject.c 2 Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy npy 1 7 deprecated api.h 15 2 warning warning Using deprecated NumPy API, disable it by defining NPY NO DEPRECATED API NPY 1 7 API VERSION [ Wcpp] warning Using deprecated NumPy API, disable it by \ gcc build src.macosx 10.5 x86 64 2.7 fib3module.c In file included from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy ndarraytypes.h 1804 0, from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy ndarrayobject.h 17, from Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy arrayobject.h 4, from build src.macosx 10.5 x86 64 2.7 fortranobject.h 13, from build src.macosx 10.5 x86 64 2.7 fib3module.c 18 Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include numpy npy 1 7 deprecated api.h 15 2 warning warning Using deprecated NumPy API, disable it by defining NPY NO DEPRECATED API NPY 1 7 API VERSION [ Wcpp] warning Using deprecated NumPy API, disable it by \ build src.macosx 10.5 x86 64 2.7 fib3module.c 104 12 warning xe2x80x98f2py sizexe2x80x99 defined but not used [ Wunused function] static int f2py size PyArrayObject var, ...) compiling Fortran sources Fortran f77 compiler usr local bin gfortran Wall g ffixed form fno second underscore m64 fPIC O3 funroll loops Fortran f90 compiler usr local bin gfortran Wall g fno second underscore m64 fPIC O3 funroll loops Fortran fix compiler usr local bin gfortran Wall g ffixed form fno second underscore Wall g fno second underscore m64 fPIC O3 funroll loops compile options Ibuild src.macosx 10.5 x86 64 2.7 I Volumes Master 1 Users wicker anaconda lib python2.7 site packages numpy core include I Volumes Master 1 Users wicker anaconda include python2.7 c gfortran f77 fib3.f usr local bin gfortran Wall g L usr local lib L usr local lib build temp.macosx 10.5 x86 64 2.7 build src.macosx 10.5 x86 64 2.7 fib3module.o build temp.macosx 10.5 x86 64 2.7 build src.macosx 10.5 x86 64 2.7 fortranobject.o build temp.macosx 10.5 x86 64 2.7 fib3.o L usr local lib gcc x86 64 apple darwin14.4.0 5.1.0 L Volumes Master 1 Users wicker anaconda lib lgfortran o build lib.macosx 10.5 x86 64 2.7 fib3.so Undefined symbols for architecture x86 64 PyArg ParseTupleAndKeywords , referenced from f2py rout fib3 fib in fib3module.o PyCObject AsVoidPtr , referenced from initfib3 in fib3module.o F2PyCapsule AsVoidPtr in fortranobject.o PyCObject FromVoidPtr , referenced from fortran getattr in fortranobject.o F2PyCapsule FromVoidPtr in fortranobject.o PyCObject Type , referenced from initfib3 in fib3module.o F2PyCapsule Check in fortranobject.o PyComplex Type , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyDict DelItemString , referenced from fortran setattr in fortranobject.o PyDict GetItemString , referenced from fortran getattr in fortranobject.o PyDict New , referenced from PyFortranObject New in fortranobject.o PyFortranObject NewAsAttr in fortranobject.o fortran setattr in fortranobject.o PyDict SetItemString , referenced from initfib3 in fib3module.o fortran getattr in fortranobject.o F2PyDict SetItemString in fortranobject.o PyFortranObject New in fortranobject.o fortran setattr in fortranobject.o PyErr Clear , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o fortran repr in fortranobject.o F2PyDict SetItemString in fortranobject.o PyErr Format , referenced from initfib3 in fib3module.o fortran call in fortranobject.o PyErr NewException , referenced from initfib3 in fib3module.o PyErr Occurred , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o F2PyDict SetItemString in fortranobject.o PyErr Print , referenced from initfib3 in fib3module.o F2PyDict SetItemString in fortranobject.o PyErr SetString , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o PyExc AttributeError , referenced from initfib3 in fib3module.o fortran setattr in fortranobject.o PyExc ImportError , referenced from initfib3 in fib3module.o PyExc RuntimeError , referenced from initfib3 in fib3module.o fortran call in fortranobject.o PyExc TypeError , referenced from fortran call in fortranobject.o array from pyobj in fortranobject.o PyExc ValueError , referenced from array from pyobj in fortranobject.o PyImport ImportModule , referenced from initfib3 in fib3module.o PyMem Free , referenced from fortran dealloc in fortranobject.o PyModule GetDict , referenced from initfib3 in fib3module.o PyNumber Int , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyObject GetAttrString , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o fortran repr in fortranobject.o PySequence Check , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PySequence GetItem , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyString AsString , referenced from fortran repr in fortranobject.o PyString ConcatAndDel , referenced from fortran getattr in fortranobject.o PyString FromFormat , referenced from fortran repr in fortranobject.o PyString FromString , referenced from initfib3 in fib3module.o fortran repr in fortranobject.o fortran getattr in fortranobject.o PyType IsSubtype , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o array from pyobj in fortranobject.o PyType Type , referenced from initfib3 in fib3module.o Py BuildValue , referenced from f2py rout fib3 fib in fib3module.o Py FindMethod , referenced from fortran getattr in fortranobject.o Py InitModule4 64 , referenced from initfib3 in fib3module.o PyObject New , referenced from PyFortranObject New in fortranobject.o PyFortranObject NewAsAttr in fortranobject.o Py NoneStruct , referenced from f2py rout fib3 fib in fib3module.o fortran getattr in fortranobject.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o main , referenced from start in crt1.10.5.o ld symbol s) not found for architecture x86 64 collect2 error ld returned 1 exit status Undefined symbols for architecture x86 64 PyArg ParseTupleAndKeywords , referenced from f2py rout fib3 fib in fib3module.o PyCObject AsVoidPtr , referenced from initfib3 in fib3module.o F2PyCapsule AsVoidPtr in fortranobject.o PyCObject FromVoidPtr , referenced from fortran getattr in fortranobject.o F2PyCapsule FromVoidPtr in fortranobject.o PyCObject Type , referenced from initfib3 in fib3module.o F2PyCapsule Check in fortranobject.o PyComplex Type , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyDict DelItemString , referenced from fortran setattr in fortranobject.o PyDict GetItemString , referenced from fortran getattr in fortranobject.o PyDict New , referenced from PyFortranObject New in fortranobject.o PyFortranObject NewAsAttr in fortranobject.o fortran setattr in fortranobject.o PyDict SetItemString , referenced from initfib3 in fib3module.o fortran getattr in fortranobject.o F2PyDict SetItemString in fortranobject.o PyFortranObject New in fortranobject.o fortran setattr in fortranobject.o PyErr Clear , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o fortran repr in fortranobject.o F2PyDict SetItemString in fortranobject.o PyErr Format , referenced from initfib3 in fib3module.o fortran call in fortranobject.o PyErr NewException , referenced from initfib3 in fib3module.o PyErr Occurred , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o F2PyDict SetItemString in fortranobject.o PyErr Print , referenced from initfib3 in fib3module.o F2PyDict SetItemString in fortranobject.o PyErr SetString , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o PyExc AttributeError , referenced from initfib3 in fib3module.o fortran setattr in fortranobject.o PyExc ImportError , referenced from initfib3 in fib3module.o PyExc RuntimeError , referenced from initfib3 in fib3module.o fortran call in fortranobject.o PyExc TypeError , referenced from fortran call in fortranobject.o array from pyobj in fortranobject.o PyExc ValueError , referenced from array from pyobj in fortranobject.o PyImport ImportModule , referenced from initfib3 in fib3module.o PyMem Free , referenced from fortran dealloc in fortranobject.o PyModule GetDict , referenced from initfib3 in fib3module.o PyNumber Int , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyObject GetAttrString , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o initfib3 in fib3module.o fortran repr in fortranobject.o PySequence Check , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PySequence GetItem , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o PyString AsString , referenced from fortran repr in fortranobject.o PyString ConcatAndDel , referenced from fortran getattr in fortranobject.o PyString FromFormat , referenced from fortran repr in fortranobject.o PyString FromString , referenced from initfib3 in fib3module.o fortran repr in fortranobject.o fortran getattr in fortranobject.o PyType IsSubtype , referenced from int from pyobj.constprop.0 in fib3module.o f2py rout fib3 fib in fib3module.o array from pyobj in fortranobject.o PyType Type , referenced from initfib3 in fib3module.o Py BuildValue , referenced from f2py rout fib3 fib in fib3module.o Py FindMethod , referenced from fortran getattr in fortranobject.o Py InitModule4 64 , referenced from initfib3 in fib3module.o PyObject New , referenced from PyFortranObject New in fortranobject.o PyFortranObject NewAsAttr in fortranobject.o Py NoneStruct , referenced from f2py rout fib3 fib in fib3module.o fortran getattr in fortranobject.o array from pyobj in fortranobject.o fortran setattr in fortranobject.o main , referenced from start in crt1.10.5.o ld symbol s) not found for architecture x86 64 collect2 error ld returned 1 exit status error Command usr local bin gfortran Wall g L usr local lib L usr local lib build temp.macosx 10.5 x86 64 2.7 build src.macosx 10.5 x86 64 2.7 fib3module.o build temp.macosx 10.5 x86 64 2.7 build src.macosx 10.5 x86 64 2.7 fortranobject.o build temp.macosx 10.5 x86 64 2.7 fib3.o L usr local lib gcc x86 64 apple darwin14.4.0 5.1.0 L Volumes Master 1 Users wicker anaconda lib lgfortran o build lib.macosx 10.5 x86 64 2.7 fib3.so failed with exit status 1","bug"
|
|
"6013","6274","Invalid timedelta division on ARMv7l.
|
|
On x86 .
|
|
On a Raspberry Pi 2 with the numpy package from Anaconda) .
|
|
Note that it triggers with float division but not integer division.
|
|
.","bug"
|
|
"6013","6276","ravel and unravel are synonyms.
|
|
Ravel and unravel are synonyms , but the methods where they appear accomplish inverse operations.
|
|
See my [SO question] http stackoverflow.com questions 32336913 numpy ravel unravel naming rage) for details and example.
|
|
Can you consider introducing alternative methods that contain `unravel` in their name even functions containing ravel are IMHO not a happy choice)?
|
|
I propose to change e.g.
|
|
| before | after | | `unravel index )` | `index to coord )` `ravel multi index )` | `index to flat`","bug"
|
|
"6013","628","Byteswapped complex256 on Debian PPC randomly incorrect subtraction Trac 2174).
|
|
Original ticket http projects.scipy.org numpy ticket 2174 on 2012 06 20 by matthew brett, assigned to unknown.
|
|
PPC 32 bit on Debian Wheezy, numpy 1.6.2 and current numpy trunk In [2] np.
|
|
version Out[2] 1.7.0.dev 651ef74 import numpy as np arr = np.arange 10, dtype=np.complex256) bs arr = arr.byteswap ).newbyteorder S ) print arr print bs arr print arr bs arr print arr bs arr print arr bs arr The result of subtraction is wrong most 90 ) of the time; with the byteswapped array acting as if it is all zeros.
|
|
np devel)[mb312 joshlegacy tmp] python funny bs.py [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] About 10 of the time the arrays appear to behave correctly np devel)[mb312 joshlegacy tmp] python funny bs.py [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j] [ 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j] [ 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j] More rarely there is a mix of incorrect and correct.
|
|
Notice below that the first pass of `arr bs arr` is incorrect, and the second and third repeat of exactly the same statement on the same variables are correct.
|
|
np devel)[mb312 joshlegacy tmp] python funny bs.py [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 1.0 0.0j 2.0 0.0j 3.0 0.0j 4.0 0.0j 5.0 0.0j 6.0 0.0j 7.0 0.0j 8.0 0.0j 9.0 0.0j] [ 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j] [ 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j 0.0 0.0j] Float128 and other numpy dtypes appear to be correct.
|
|
I m happy to give access to this machine if someone needs to debug.","bug"
|
|
"6013","6282","ENH digitize to accept bins same shape as x.
|
|
As described in the [SO question] http stackoverflow.com questions 32188039 apply digitize to ndarray by row) I wanted to apply `digitize )` by row, where the bins also change by row.
|
|
In my application I have a 5000 by 9000 matrix and I want to digitize each row into the corresponding decile.
|
|
It takes approx.
|
|
4 5 seconds and I can live with that, but maybe the enhancement is simple enough and could provide some significant speedup?
|
|
I can give it a stab, if you point me in the right direction, or let it go if you reckon the improvement might be negligible with respect to the solution I adopted in the SO question.","enhancement"
|
|
"6013","6284","Segmentation fault when using Numpy arrays and astropy time objects .
|
|
The following example triggers a segfault in the last line when the array is 501 elements or longer, and an exception when it is shorter.
|
|
from astropy import time import numpy as np t max = 1258.2 phase = np.linspace 10, 10, num=501) date obs = time.Time np.random.random phase.shape), format= jd ) selected = phase == phase[320] selected = phase < phase[3] ph = np.array time.Time t max, format= jd ) date obs[selected]) print ph, type ph) phase[selected] = ph I am using Python 2.7 on Fedora 22, Numpy 1.9.1 and astropy 1.0.4 installed in a virtualenv.
|
|
This is what gdb has to say PyErr Restore type=<type at remote 0x309fd89e20>, value=<optimized out>, traceback=traceback entry=0x0) at usr src debug Python 2.7.10 Python errors.c 39 39 oldtype = tstate >curexc type;","bug"
|
|
"6013","6285","Regression in numpy.intersect1d comparing sets?.
|
|
Hi, Is this intentional or a regression between 1.7.1 and 1.9.1 I don t have intermediate versions).
|
|
I couldn t find anything in the release notes pointing directly to it.
|
|
Python 2.7.5 default, May 15 2013, 22 44 16) [MSC v.1500 64 bit AMD64)] .
|
|
Python 2.7.10 default, May 23 2015, 09 44 00) [MSC v.1500 64 bit AMD64)] .
|
|
Thanks and best regards, Martin","bug"
|
|
"6013","6286","log1p gives different result in Windows compared to Linux Mac for large argument.
|
|
On 64 bit Windows, Anaconda Python np.log1p 1.7976931348622732e 308) gives inf.
|
|
On Linux and Mac it gives 709.78.... More detail at http www.walkingrandomly.com ?p=5852","bug"
|
|
"6013","6287","npy copysign should be a macro.
|
|
GCC has a ` builtin copysign` that it inlines, so call like `copysign 1, x)` get specialized.
|
|
When using `npy copysign`, that doesn t happen.
|
|
It would be nice to conditionally .
|
|
in `npy math.h` so that the builtin gets used.","bug"
|
|
"6013","6289","Truncation behavior raises exception dependent on native Python int length.
|
|
When casting Python ints to NumPy arrays, the truncation of values to the dtype in the array will either succeed or raise OverflowError depending on the native int type of a particular Python compile.
|
|
For example .
|
|
Oddly, on a Windows 64 bit Anaconda distribution, the native int type is still 32 bits even though floats are natively 64 bit.
|
|
So on those systems, this will fail in the same way but succeed on a 64 bit OSX version, with 64 bit native ints.
|
|
.
|
|
Obviously, Python native ints are actually unlimited size numbers, but internally they are stored as fixed bit lengths until forced into the unlimited size format.
|
|
So in other words, when they happen to be stored in the fixed size optimization, truncation works, but otherwise it fails.
|
|
However, whether this will happen is non obvious to users, and inconsistent between platforms.
|
|
A workaround is to wrap every Python sequence used to initialize an np.array in the following function .
|
|
However, given that it is obviously possible to truncate even more than native sized ints to cast to NumPy arrays, doing so would be better default behavior.","bug"
|
|
"6013","629","pickle with masked array Trac 2175).
|
|
Original ticket http projects.scipy.org numpy ticket 2175 on 2012 06 21 by trac user eyurtsev, assigned to unknown.
|
|
An error message comes up when I try to pickle the result of numpy.mean on a masked array of nans.
|
|
I suspect that this is a bug.
|
|
For convenience a python file is attached with this code.
|
|
import pickle import numpy a = numpy.mean numpy.ma.masked invalid numpy.nan)) with open test.txt , wb ) as f pickle.dump a, f) Error Message usr lib python2.7 pickle.pyc in save self, obj) 304 reduce = getattr obj, reduce ex , None) 305 if reduce > 306 rv = reduce self.proto) 307 else 308 reduce = getattr obj, reduce , None) usr lib python2.7 dist packages numpy ma core.pyc in reduce self) 5471 return mareconstruct, 5472 self.
|
|
class , self.
|
|
baseclass, 0,), b ,), > 5473 self.
|
|
getstate )) 5474 5475 def deepcopy self, memo=None) usr lib python2.7 dist packages numpy ma core.pyc in getstate self) 5444 getmaskarray self).tostring cf), 5445 getmaskarray self).tolist ), > 5446 self.
|
|
fill value, 5447 ) 5448 return state AttributeError MaskedConstant object has no attribute fill value","bug"
|
|
"6013","6292","Unstable behaviour with augmented division.
|
|
I m loading a numpy array using scipy.misc.imread path) and then try to normalise it.
|
|
If I use img = img 255.0, I get the expected behaviour, but if I use img = 255.0 it does what seems to be an int division and I get a lot of zeros.","bug"
|
|
"6013","6294","odd boolean result for np.dtype.
|
|
numpy 1.9.2 python objects should report a bool of ``True`` except for empty ones .
|
|
https github.com libdynd dynd python issues 353","bug"
|
|
"6013","6295","BUG standard dtypes are not singletons.
|
|
I encountered the following while working on [scipy 5213] https github.com scipy scipy pull 5213) .
|
|
This has implications for things like using dtypes as keys in a dictionary, which is fairly common in the scipy unit tests.","bug"
|
|
"6013","6296","Support of multiple axes in max operator for c api.
|
|
In the python api there is support for multiple axes i.e.
|
|
a tuple of integers) in the max operator.
|
|
However, the c api only seems to support a single axis i.e an integer); see http docs.scipy.org doc numpy reference c api.array.html.
|
|
Is there an efficient implementation for dealing with multiple axes?
|
|
Of course, I could use a for loop to go over the axes, but I am wondering if there something more efficient.
|
|
Also, the argmax does not support multiple axes not even in the python api).
|
|
Is there going to be support in the future?
|
|
Does anybody know an efficient way to deal with multiple axes?","enhancement"
|
|
"6013","6297","Recent performance regression on np.dot linked to openBLAS with large matrices.
|
|
Not quite sure if it s an issue with openBLAS or numpy yet.
|
|
Regularly setup high performance environments from source on a monthly basis on the same hardware system and noticed a 2x performance regression that sometimes results in very long hanging calls to `np.dot` with large matrices i.e.
|
|
4096x4096 and up).
|
|
For reference, building both from source this morning results in the issue while using these commits from when last built works fine.
|
|
openBLAS 5408074941d5cc0f4aad180562cafcf4cf27a56d numpy d750cba93436f20d5457daec5ad5d5ffb65f22bd cython d6cf8959a58395daacb1e338644a698ef263ce98 in case its relevant) The system is ubuntu 14.04.3 on a 8 core haswell 5960x with 64 GB of ddr4.
|
|
When `np.dot` is called a task manager first shows load distributed across all cores as expected) and when hanging occurs it appears distributed across only 8 threads out of 16 with hyperthreading) to begin with and slowly those threads exit till it seems to stall for 10 minutes on only two threads remaining.
|
|
The probability of hanging appears to be linked to the size of the matrices, at 8192x8192 it has always hanged even waiting 100x the normal amount of time it takes) at 4096x4096 it happens 1 in 3 times and at 128x128 it never appears to hang.
|
|
Example code .","bug"
|
|
"6013","6299","numpy.fft slow when used in a subprocess.
|
|
I ve noticed that numpy.fft.fft and its variants) very slow about 10x) when used inside of a subprocess spawned by multiprocessing), as compared to outside of it Here is example code import numpy as np import multiprocessing as mproc import time import sys the producer function, which will run in the background and produce data def Producer dataQ) numFrames = 5 n = 0 while n < numFrames data = np.random.rand 3000, 200) dataQ.put data) send the datta to the consumer time.sleep 0.1) sleep for 0.5 second, so we dont overload CPU n = 1 the consumer function, which will run in the backgrounnd and consume data from the producer def Consumer dataQ) while True data = dataQ.get ) t1 = time.time ) fftdata = np.fft.rfft data, n=3000 5) tDiff = time.time ) t1 print Elapsed time is 0.3f tDiff) time.sleep 0.01) sys.stdout.flush ) the main program if name == main is necessary to prevent this code from being run only when this program is started by user if name == main data = np.random.rand 3000, 200) t1 = time.time ) fftdata = np.fft.rfft data, n=3000 5, axis=0) tDiff = time.time ) t1 print Elapsed time is 0.3f tDiff) generate a queue for transferring data between the producedr and the consumer dataQ = mproc.Queue 4) start up the processoso producerProcess = mproc.Process target=Producer, args=[dataQ], daemon=False) consumerProcess = mproc.Process target=Consumer, args=[dataQ], daemon=False) print starting up processes ) producerProcess.start ) consumerProcess.start ) time.sleep 10) let program run for 5 seconds producerProcess.terminate ) consumerProcess.terminate ) The output it produces on my machine is Elapsed time is 0.112 starting up processes Elapsed time is 0.716 Elapsed time is 0.749 Elapsed time is 0.831 Elapsed time is 0.908 Elapsed time is 0.955 Anybody know whats going on?
|
|
I ve tried other functions like convolve ), but they don t produce the same slowdown.
|
|
The CPU usage in both single and multiprrocess cases with heavy calls to fft ) is the same about 25 30 ), as reported by Windows Task Manager.
|
|
Anybond know what could be causing the discrepancy?","bug"
|
|
"6013","630","Add step parameter to linspace or endpoint parameter to arange) Trac 2176).
|
|
Original ticket http projects.scipy.org numpy ticket 2176 on 2012 06 28 by endolith, assigned to unknown.
|
|
There s no simple way to make a sequence based on step size that includes both endpoints.
|
|
For instance, to produce the sequence [1.0, 1.1, 1.2, 1.3, 1.4, 1.5], you d have to do something clunky like linspace 1.0, 1.5, 1.5 1.0) 0.1 1) or arange 1.0, 1.5 0.01, 0.1) it would be more convenient if you could just say linspace 1.0, 1.5, step=0.1) or maybe arange 1.0, 1.5, 0.1, endpoint=True) Of course, `step` could not be specified at the same time as `num`.","enhancement"
|
|
"6013","6300","mem overlap 538) VLA not compile with visual studio.
|
|
visual studio 2015 numpy 4091facf55e03d356a5d7ce219ae0c0494920df3 mem overlap 538) VLA not compile with visual studio numpy\core\src\private\mem overlap.c 538 line diophantine term t Ep[n]; npy int64 Epsilon[n], Gamma[n];","bug"
|
|
"6013","6301","FR increment counter on write check.
|
|
Presumably any block of code that modifies an ndarray is wrapped in a thread safe?)
|
|
check of the writable flag.
|
|
Would it be possible to hold a counter rather than a simple bool flag and then increment the counter whenever you test the flag?
|
|
Hopefully this would introduce only a tiny additional overhead, but would permit pseudo hashing to test whether an immutable array has changed since you last encountered it.","bug"
|
|
"6013","6303","Add optional out= argument to vectorized string functions; allow inplace string manipulations.
|
|
I was just poking over some old code I wrote, trying to understand why it existed.
|
|
It turned out to just be doing the equivalent of an `rsrtrip` on a string array.
|
|
However, I know I wanted to do it in place i.e.
|
|
just replacing all trailing spaces with nulls), rather than make a copy of the array, which is unfortunately what you get from `np.char.rstrip )`.
|
|
I don t think it would be hard to add an optional `out=` argument to all the string functions in `np.char` so that they can output to an existing array so long as that array has the correct dtype or can is castable in some way).
|
|
At the same time this would enable in place string operations, since looking at the ` vec string` function I don t think there s any reason the output array can t be the input array.
|
|
I d be happy to work on a patch if this sounds worthwhile.","enhancement"
|
|
"6013","6304","1.10.0b1 test error on Mac OS X Version 10.9.5.
|
|
I get one error using 1.10.0b1 on a Mac with OSX 10.9.5 .","bug"
|
|
"6013","6305","augmented assignment doesn t vectorize as expected.
|
|
The following code, written as a loop, a = np.array [0, 0, 0]) v = [1, 2, 1] for i in v a[i] = 1 results in a = [0 2 1].
|
|
I attempted to vectorize it, as follows a = np.array [0, 0, 0]) v = [1, 2, 1] a[v] = 1 which resulted in a = [0 1 1].
|
|
Why are these two versions not equivalent?","bug"
|
|
"6013","6308","BUG f2py character array variables now broken.
|
|
in current devel version when a character variable is defined using the asterisk form character 8 name and I then use common name 12) I get an error message from f2py like analyzevars character array character 8 name 12) is considered as character name 12,8) ; intent c) is forced.
|
|
if the first definition is changed to character 8) then the error goes away.
|
|
The old behaviour that is consistent with FORTRAN should be restored.
|
|
Any suggestion on bug fix?","bug"
|
|
"6013","631","np.histogramdd does not work with reverse sorted bin sequence Trac 2177).
|
|
Original ticket http projects.scipy.org numpy ticket 2177 on 2012 06 29 by trac user andyfaff, assigned to unknown.
|
|
The reference documentation for np.histogramdd does not specify that a supplied sequence for a bin specification needs to be sorted.
|
|
Whilst this may be kind of obvious it should be stated.
|
|
Moreover, and perhaps more importantly, it does not specify that the bin sequence is required to be sorted in ascending order.
|
|
Bin edges sorted in descending order do not work, the following results in a ValueError >>> a = 0.1 np.arange 21.)
|
|
10) >>> a = a[ 1] >>> b = np.random.randn 100) >>> c,d = np.histogramdd b, bins=[a]) Warning invalid value encountered in log10 Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib function base.py , line 314, in histogramdd decimal = int log10 dedges[i].min ))) 6 ValueError cannot convert float NaN to integer The obvious workaround for me is to check if the bin edges are reverse sorted.
|
|
If they are, reverse them and histogram, then reverse the output and bin edges.
|
|
However, it would be a lot easier if histogramdd worked with arbitrarily sorted bin edges on each dimension.
|
|
This kind of requirement for binning is quite common when dealing with images.","enhancement"
|
|
"6013","6311","cumsum improvements.
|
|
`cumsum` could support simultaneously summing over multiple axes xc3xa0 la `sum`).
|
|
I also wonder if people are really relying on it to sum over the flattened array in dimensions higher than one... if not, perhaps that could be a warning.","bug"
|
|
"6013","6313","Why numpy does not provide a 64 bits windows distribution ?.
|
|
Why numpy does not provide a 64 bits windows distribution ?
|
|
Yes I know, there are many sources to install a 64 bits numpy on windows, but why the official release does not come in a win 64 flavor ?
|
|
Thanks, Jose","bug"
|
|
"6013","6314","Missing byteswap when converting array with record dtype to opposite endian dtype..
|
|
Consider the following self contained snippet of code .
|
|
When I run this under Python 2.7 and NumPy 1.9.2 on my OS X 10.9 machine, I get the following output .
|
|
I d expect to get this output .
|
|
Doing the same with a scalar dtype gives me the expected behaviour, and replacing `value` with the tuple ` 1.1, 1.2)` also gives me the expected behaviour.
|
|
It looks as though there s a byteswap operation being missed at some level during the conversion.","bug"
|
|
"6013","6315","implicit ascii encoding for dtypes of kind S .
|
|
Greetings, Investigating PyTables PyTables 499 I came across the following numpy s behaviour that I find unexpected .
|
|
So it seems that the string ` ciao ` automagically becomes ` ciao .encode ascii )` on its way to `z[0]` which is surprising and kind of wrong anyway.
|
|
Said otherwise, I won t expect any implicit conversion between bytes and strings.","bug"
|
|
"6013","6316","numpy.convolve ) swap of input arguments are illogical when using same .
|
|
`convolve a, v, mode)` swaps the input arguments `a` and `v` if `v` has greater length than `a`.
|
|
When using the `mode = same ` the output is clipped to the size of `a`.
|
|
When the arguments have been swapped you can end up with the output being the length of `v`, while I would expect to always get the same behaviour when using same the output should be the length of the first argument.
|
|
Foe example, when using an averaging filter on an input vector if the input vector length `a`) decreases to below the filter length `v`) you will get the unexpected result of `y` having a greater length than the input `a`).
|
|
.","bug"
|
|
"6013","6317","Signed zero discrepancy vs. Python on float np.remainder.
|
|
.
|
|
It seems Python is right here, since the result is supposed to be of the same sign as the dividend, e.g.
|
|
.
|
|
And actually .","bug"
|
|
"6013","6318","np.reciprocal gives strange result for integer zero.
|
|
I was looking through numpy ufuncs, and noticed `np.reciprocal` treats integer zero strangely >>> np.reciprocal np.array 0)) 9223372036854775808 Probably should return 0 with a warning, like `np.int32 1) np.int32 0)`.
|
|
Thus for ints, reciprocal would return 1 if x==1, otherwise 0.","bug"
|
|
"6013","6319","1.9.3 release very soon.
|
|
I think the mailing list is down ATM.
|
|
It would be good to get out a release compatible with Python 3.5 as soon as possible, because Python 3.5 is the default download at Python.org.
|
|
I propose just merging https github.com numpy numpy commit 7d6aa8c721d5274ac57d0c87685d472cb1fd7948 for 1.9.3 and releasing in the next few days, and then add more bugfixes for 1.9.4 over the following week.
|
|
I m happy to do the release work for 1.9.3 if that s acceptable.","bug"
|
|
"6013","632","regression closes not owned handle Trac 2178).
|
|
Original ticket http projects.scipy.org numpy ticket 2178 on 2012 07 02 by yarikoptic, assigned to unknown.
|
|
While troubleshooting http bugs.debian.org cgi bin bugreport.cgi?bug=669466 which used to build test just fine I ran to conclusion that it is numpy s regression.
|
|
Following snippet !
|
|
usr bin python import numpy as np print Numpy version , np.
|
|
version from tempfile import mktemp outfile name = mktemp suffix= .np ) outfile = open outfile name, w ) np.savez outfile, data= LOVELY LOAD ) outfile.close ) print outfile name outfile = open outfile name, r , 10000) outfile.seek 0) print We seek before just fine print np.load outfile)[ data ] print Can we seek after?
|
|
outfile.seek 0) Only needed here to simulate closing reopening file print We could seek after as well produces on Debian squeeze > python tmp numpy closingbug.py Numpy version 1.4.1 home yoh .tmp tmpQltbIC.np We seek before just fine LOVELY LOAD Can we seek after?
|
|
We could seek after as well but on Debian sid > tmp numpy closingbug.py Numpy version 1.6.2 home yoh .tmp tmpdbfXKA.np We seek before just fine LOVELY LOAD Can we seek after?
|
|
Traceback most recent call last) File tmp numpy closingbug.py , line 20, in <module> outfile.seek 0) Only needed here to simulate closing reopening file ValueError I O operation on closed file and with current master > PYTHONPATH= HOME proj numpy tmp numpy closingbug.py Numpy version 1.8.0.dev e15d0bd home yoh .tmp tmpg1uQA4.np We seek before just fine LOVELY LOAD Can we seek after?
|
|
Traceback most recent call last) File tmp numpy closingbug.py , line 20, in <module> outfile.seek 0) Only needed here to simulate closing reopening file ValueError I O operation on closed file","bug"
|
|
"6013","6323","Automatic local variable representing start in slicing.
|
|
Hi, I think it would be great if we can have a local variable representing the start position in slicing.
|
|
For example, I m parsing a lot of json data and some of them are corrupted, and I find that I use the following frequently when I m debugging .
|
|
Which is a pain to type and edit.
|
|
If we have something like this .
|
|
and `np.
|
|
start` equals to `580374` inside the slicing notation.
|
|
I think this will provide a fluent work flow because I can easily do history search and edit only the first part.
|
|
I use these kind of thing a lot when I work with data, etc as well.
|
|
Of course, I can do the following, but it is less fluent .
|
|
I understand there can be problems with this and it may not be practical, but I wanted to hear what others think of this. )
|
|
Thank you!","bug"
|
|
"6013","6324","Export CBLAS functions.
|
|
I was wondering if it s possible for the relevant Numpy DLLs e.g.
|
|
lapack lite or umath linalg) to export the CBLAS functions, so that other software can call into them using e.g.
|
|
ctypes.
|
|
Under Linux they are exported by default, but not under Windows AFAIK you have to explicitly export symbols on Windows, by default global symbols in a shared library are private to the library).
|
|
Or perhaps I m missing another way of accessing those functions at runtime?","bug"
|
|
"6013","6325","mailing list is down again, and has been for a bit.... Feel silly filing a bug but I m not sure who to even ping I have a vague memory that usually when this happens then usually Ralf pings somebody?
|
|
Which is not very helpful right now ) Maybe... rkern knows to talk to?
|
|
scipy dev actually bounced email for me earlier today as an unknown address, so that s exciting too.
|
|
In the longer run I m starting to wonder if we should give up and move to python.org like matplotlib and nipy both did recently back in July I talked to Chip Parker Enthought briefly when I had some problem and he got in touch with me about it, and when I asked him about the instability he assured me that it was going to be sorted out any time now, but I am losing faith...)","bug"
|
|
"6013","6326","weighted percentile.
|
|
Support for weights in `percentile` would be nice to have.
|
|
A quick look suggests https github.com nudomarinero wquantiles; I d be happy to make a PR out of this implementation if there s interest.","enhancement"
|
|
"6013","6327","Bug in documentation for numpy.cov.
|
|
Concerning the parameters, http docs.scipy.org doc numpy reference generated numpy.cov.html numpy.cov says about rowvar rowvar int, optional If rowvar is non zero default), then each row represents a variable, with observations in the columns.
|
|
Otherwise, the relationship is transposed each column represents a variable, while the rows contain observations.
|
|
But looking at the source code, we have if rowvar N = X.shape[1] axis = 0 else N = X.shape[0] axis = 1 in other words, rowvar is boolean.
|
|
if rowvar is True default), then each row represents a variable and observations are in columns.
|
|
If rowvar is False, it means that each column represents a variable, and observations are in rows.
|
|
The current text is kind of confusion, as one might think rowvar can specify a range or specific rows or something.
|
|
The documentataion or rather the comment in source code) should be changed accordingly.
|
|
Note that interestingly, http docs.scipy.org doc numpy reference generated numpy.ma.cov.html has it right","bug"
|
|
"6013","6329","test gzip load error with Python 3.5.
|
|
I m trying out python 3.5, and wasn t sure numpy installed properly, so I ran the tests and got this error python c import numpy; numpy.test ) ... ERROR test io.test gzip load Traceback most recent call last) File Users me py env p3 lib python3.5 site packages nose case.py , line 198, in runTest self.test self.arg) File Users me py env p3 lib python3.5 site packages numpy lib tests test io.py , line 1675, in test gzip load assert array equal np.load f), a) File Users me py env p3 lib python3.5 site packages numpy lib npyio.py , line 381, in load fid.seek N, 1) back up File Users me py env p3 lib python3.5 site packages numpy lib npyio.py , line 52, in seek offset = self.offset offset AttributeError GzipFile object has no attribute offset Ran 5556 tests in 29.421s FAILED KNOWNFAIL=6, SKIP=18, errors=1) I wasn t sure if the tests are expected to run without failing or not on Python 3.5, thought I would report it just in case.
|
|
I have the following installed with this virtual environment .","bug"
|
|
"6013","633","Memmap children retain mmap reference in all cases Trac 2179).
|
|
Original ticket http projects.scipy.org numpy ticket 2179 on 2012 07 02 by trac user sveinugu, assigned to unknown.
|
|
Consider the following >>> a = np.memmap file , dtype= int32 ) >>> a 10).
|
|
mmap is a. mmap True fancy indexing doesn t return a view!)
|
|
>>> a[[1, 2, 3]].
|
|
mmap is a. mmap True >>> a.sum ).
|
|
mmap is a. mmap True Ideally one should be returning other ndarray objects, e.g.
|
|
numpy scalars in the case of sum ), but this requires a larger rewrite.
|
|
At least numpy version 1.5.1rc1 did this.
|
|
The only time the mmap should be kept as I am aware of) is when slicing.
|
|
One major problem with this the inability to garbage collect the mmap when a result of some operation is stored.
|
|
I propose a patch to temporarily fix the situation.
|
|
See also the following thread http mail.scipy.org pipermail numpy discussion 2012 July 063126.html","bug"
|
|
"6013","6331","Import Error No Such process on cygwin.
|
|
I have installed python numpy on cygwin in windows and I get the following error when trying to import numpy.
|
|
.","bug"
|
|
"6013","6333","Error in numpy.random.gumbel doc examples.
|
|
Copypaste from https github.com scipy scipy.org issues 106 In the second example code at http docs.scipy.org doc numpy reference generated numpy.random.gumbel.html The following line beta = np.std maxima) np.pi np.sqrt 6) should be beta = np.std maxima) np.sqrt 6) np.pi","bug"
|
|
"6013","6334","Generated Documentation Images 404.
|
|
Copypaste from https github.com scipy scipy.org issues 86 On this <a href= http docs.scipy.org doc numpy reference generated numpy.interp.html numpy.interp >page< a>, the links on the Source code, png, pdf) all result in 404 errors.
|
|
I haven t checked any other pages.","bug"
|
|
"6013","6335","Why np.isscalar hello ) is True?.","bug"
|
|
"6013","6337","The `f` attribute of the object returned by numpy.load is not documented.
|
|
When reading an `npz` file, `numpy.load` returns an object that has an attribute called `f`.
|
|
`f` has attributes with names and values corresponding to the names and values of the variables stored in the file.
|
|
Since `f` does not have a leading underscore, I assume `f` is intended to be part of the public API, but `f` is not documented in the docstring of `numpy.load`.
|
|
Here s an example see http stackoverflow.com questions 32682928 loading arrays from npz files in pythhon 32683037 32683037) .","bug"
|
|
"6013","6338","e7689a2600a28e75d187bbad06d730fb7e66bc5f fails on MacOS X 10.10.5 with Python 3.5.0.
|
|
Paris U.E.
|
|
), le 22 09 2015 Hi With Python 3.5.0 and numpy e7689a2600a28e75d187bbad06d730fb7e66bc5f I have the following failure .
|
|
Hubert Holin","bug"
|
|
"6013","6339","Support for Decimal in numpy.percentile.
|
|
It would be great if `numpy.percentile` supported Decimal.
|
|
Most numpy functions I ve tried are fine with `numpy.array`s containing `Decimal`s, but not numpy.percentile.
|
|
.
|
|
I m using Python 3.4.3 and numpy 1.9.3.
|
|
Thanks","bug"
|
|
"6013","634","remove import time exec calls from polynomial Trac 2181).
|
|
Original ticket http projects.scipy.org numpy ticket 2181 on 2012 07 06 by trac user dalke, assigned to charris.
|
|
numpy polynomial .py contains 5 files which do an exec of a 25K template, generated at import time.
|
|
The extra overhead of parsing and byte compiling takes roughly 25 of the entire import numpy overhead.
|
|
I changed the code so it s the responsibility of the person who edits the template to rerun the polytemplate.py file as a command line program.
|
|
This then goes through the correct files to find the special markers which indicate where the template should be inserted, get the file specific parameters, apply those to the template, and replacing the old code which was between the two markers with the newly evaluated template output.
|
|
In more details, currently there are 5 places under numpy polynomial which do something like exec polytemplate.substitute name= Hermite , nick= herm , domain= [ 1,1] ) I have changed those so there are start end markers, like REPLACE POLYTEMPLATE name= Hermite , nick= herm , domain= [ 1,1] ... END REPLACE I have edited polytemplate.py so that it implements a main which finds those block markers and replaces the content with the appropriate polytemplate substitution.
|
|
The performance of python c import time; t1=time.time ); import numpy; print time.time ) t1 goes from 0.079 seconds to 0.057 seconds best of 10), for an import speedup of 25 .","bug"
|
|
"6013","6343","zip gz files produced by python setup.py sdist may have missing files.
|
|
In particular, `cblasfiles.c` and `python xerbla.c` are missing when using `numpy vendor`.The problem probably originates here .
|
|
Which makes the list dependent on the environment.
|
|
Does anyone have a good idea about how to work around that?","bug"
|
|
"6013","6345","issus about numpy number comparison after accumulate.
|
|
In [1] import numpy In [2] a = numpy.array [0.9, 0.05, 0.03, 0.02]) In [3] a Out[3] array [ 0.9 , 0.05, 0.03, 0.02]) In [4] a = numpy.add.accumulate a) In [5] a Out[5] array [ 0.9 , 0.95, 0.98, 1. ])
|
|
In [6] a>0.95 Out[6] array [False, True, True, True], dtype=bool) In [7] a>numpy.float64 0.95) Out[7] array [False, True, True, True], dtype=bool) After accumulate the array, it seems that the precision of the number cause some issue while comparison.
|
|
Could you help me out this problem?
|
|
Thanks","bug"
|
|
"6013","6348","Broken f2py script in wheeled numpy 1.9.3.
|
|
After installing wheeled numpy 1.9.3, the f2py script starts with .
|
|
I don t know who travis is, but she s definitely not a user on my laptop... Numpy 1.9.2 worked fine as I recall.
|
|
Numpy compiled from master is of course fine as well.","bug"
|
|
"6013","635","Support UTF 8 in f2py Trac 2182).
|
|
Original ticket http projects.scipy.org numpy ticket 2182 on 2012 07 09 by trac user opoplawski, assigned to pearu.
|
|
f2py cannot handle UTF 8 input files.
|
|
See attached for example.","enhancement"
|
|
"6013","6351","numpy.linalg.cond error.
|
|
Hello there, For high dimension arrays, numpy.linalg.cond function does not return the right answer.
|
|
Eg >>> a=numpy.random.random 4,3,3)) >>> print numpy.linalg.cond a).shape 3,) It should be a 4 item array.
|
|
Qiming","bug"
|
|
"6013","6353","1.9 release notes do not mention API change to nansum.
|
|
While I agree that it s logical that the result is zero, apparently many users have code that expects otherwise.
|
|
This API change should be documented in the release notes.","bug"
|
|
"6013","6356","np.savetxt raises an Exception when passed a file handle that was opened with w .
|
|
I am trying to write an array to disk with `np.savetxt`, using a file handle opened with w .
|
|
I get the following Exception.
|
|
python 3.4.3, numpy 1.9.2, OS X).
|
|
It works when the file is opened with wb .
|
|
This is either a bug in numpy, or the documentation needs to say that the file has to be opened in binary mode.
|
|
>>> import numpy as np >>> a = np.arange 10.)
|
|
>>> with open test.dat , w ) as f ... np.savetxt f, a) TypeError Traceback most recent call last) <ipython input 57 cf77f423517e> in <module> ) 2 a = np.arange 10.)
|
|
3 with open test.dat , w ) as f > 4 np.savetxt f, a) Users anz Documents Andy programming dev3 lib python3.4 site packages numpy lib npyio.py in savetxt fname, X, fmt, delimiter, newline, header, footer, comments) 1085 else 1086 for row in X > 1087 fh.write asbytes format tuple row) newline)) 1088 if len footer) > 0 1089 footer = footer.replace \n , \n comments) TypeError must be str, not bytes","bug"
|
|
"6013","6357","argsort argsort )) behavior not stable?.
|
|
When sorting a list of exactly the same values, when the length is < 16, the order is preserved, when length > 16, the order is scattered.
|
|
>>> np.argsort np.argsort [2] 16)) array [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]) >>> np.argsort np.argsort [2] 17)) array [ 0, 14, 13, 12, 11, 10, 9, 15, 8, 6, 5, 4, 3, 2, 1, 7, 16]) This is avoided when it s sorted by mergesort >>> np.argsort np.argsort [2] 17,kind= mergesort ),kind= mergesort ) array [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16])","bug"
|
|
"6013","6358","Fix vdot or revert making ravel create less copies.
|
|
The changed `vdot` code relied on `ravel` always returning a contiguous array.
|
|
With the also changed `ravel` code to always return a view when possible, `ravel` can return arrays with a non contiguous memory layout .
|
|
Previously, this would return a contiguous copy with the exception of keeporder .
|
|
While the `vdot` code may be just be called buggy, the question is whether we should avoid potential problems for other C interfacing code and instead of changing `vdot` just revert `ravel` and document this behaviour.
|
|
Users who would prefer the new definition can be pointed to use `arr.reshape 1)` for most purposes keeporder is an exception).","bug"
|
|
"6013","6359","structured dtype descr method returns incorrect descriptor when void space is at the end of dtype built through a dict.
|
|
Using python 2.7.10, numpy 1.9.2 If I define a structured dtype in the following manner >>>my dtype = np.dtype { names [ A , B ], formats [ f4 , f4 ], offsets [0, 8], itemsize 16}) And I then try to create a new dtype from this dtypes descr >>>new dtype = np.dtype my dtype.descr) Then the two dtypes will not have the same itemsize >>>my dtype.itemsize 16 >>>new dtype.itemsize 12 Examining the descr of my dtype, we see that it is leaving off the 4 void bytes at the end >>>my dtype.descr [ A , <f4 ), , |V4 ), B , <f4 )] What should happen instead >>>my dtype.descr [ A , <f4 ), , |V4 ), B , <f4 ), , |V4 )] This has relevance for use of structured arrays with IPython.parallel, as this is how structured arrays are reconstructed when serialized and sent to engines in IPython.parallel.
|
|
A work around for the user of course is to define some field that marks the end of the structured data, but it seems that this should not be necessary.","bug"
|
|
"6013","636","Typo in f2py typedefs Trac 2183).
|
|
Original ticket http projects.scipy.org numpy ticket 2183 on 2012 07 09 by trac user pepijndevos, assigned to pearu.
|
|
in cfuncs.py it said insigned long long , and the definition for unsigned long long was moved into long long.
|
|
It also seems the Windows declaration was incorrect.","bug"
|
|
"6013","6364","linalg.cond documentation is incomplete or misleading for non square matrices.
|
|
The notes section says The condition number of x is defined as the norm of x times the norm of the inverse of x but the implementation is more general it returns nontrivial values for non square matrices.
|
|
The fix would not be as simple as replacing inverse with pseudoinverse in the notes, because the implementation distinguishes between structural rank vs. numerical rank.","bug"
|
|
"6013","6365","Error easy installing Numpy.
|
|
I m using easy install to install numpy and getting the install error shown below.
|
|
Can somebody please tell me what is causing this error and what I can do to fix it?
|
|
!
|
|
[image] https cloud.githubusercontent.com assets 14841518 10108734 aca07436 6390 11e5 8f69 9f8e019742bd.png)","bug"
|
|
"6013","6366","np.full and object arrays.
|
|
`np.full` says that the resulting array s dtype is `np.array fill value).dtype` without mention of object arrays, but .
|
|
and .","bug"
|
|
"6013","6367","Is `np.empty ..., dtype=object)` guaranteed to fill the array with None?.
|
|
Currently, `np.empty ..., dtype=object)` fills the array with `None`, which is certainly a very reasonable though undocumented, afaict) behavior.
|
|
Is this guaranteed?
|
|
Note that the other more obvious way of creating such an array, `np.fill ..., None)`, does not work, see 6366.)","bug"
|
|
"6013","637","npyio.py genfromtxt ) handles comments incorrectly with names=True Trac 2184).
|
|
Original ticket http projects.scipy.org numpy ticket 2184 on 2012 07 11 by trac user khaeru, assigned to unknown.
|
|
The documentation for `genfromtxt )` reads When the variables are named either by a flexible dtype or with names , there must not be any header in the file else a ValueError exception is raised).
|
|
and also If names is True, the field names are read from the first valid line after the first skip header lines.
|
|
The cause of this seems to be in [https github.com numpy numpy blob master numpy lib npyio.py L1347 numpy lib npyio.py at lines 1347 9] if names is True if comments in first line first line = asbytes ).join first line.split comments)[1 ]) The last line should read `first line = first line.split comments)[0]`.
|
|
With the current code, the input line Example comment line will be transformed to Example comment line resulting in columns named Example , comment and line this is what the warning in the documentation is about).
|
|
But also the input line ColumnA ColumnB ColumnC the column names precede this comment will be transformed to the column names precede this comment resulting in columns named the , column , names etc.
|
|
In this instance actual column names present in the file are inappropriately discarded.
|
|
By taking the `[0]` portion of the split instead of `[1 ]` Lines beginning with comments result in an empty string being passed to `split lines )` on L1350, producing no usable output and causing the `while not first values` loop to try the next line.
|
|
Partial line comments following actual heading names are discarded, instead of the names themselves.
|
|
As a result, files can have commented headers of any length and column names, simultaneously.","bug"
|
|
"6013","6370","Wrong axis order when indexing by int, slice, list).
|
|
Advanced indexing sometimes gives the result transposed w.r.t.
|
|
the expected result.
|
|
This occurs when indexing by int, slice, list), but not int, slice, slice) or slice, int, list).
|
|
Example .
|
|
Incorrect .
|
|
Correct .
|
|
The incorrect result is the transpose of the correct one .","bug"
|
|
"6013","6378","numpy eigh not deterministic.
|
|
To illustrate the problem I need to attach a .py python script file and a .npz numpy binary savez file and github won t let me attach either of these file types.
|
|
How can I do this?","bug"
|
|
"6013","6379","Inconsistent behaviour in array multiplication using =.
|
|
I saw similar reports in 6074 and 4707, but I didn t find any report matching this issue exacly.
|
|
When multiplying arrays `a` and `b`, `a =b` and `a=a b` are not equivalent if their types are different.
|
|
The same happens with addition.
|
|
In Numpy 1.8.2, on python 3.4, different behaviour is observed in the following two cases .
|
|
1 .
|
|
Output .
|
|
2 .
|
|
Output .
|
|
However, I m not sure whether this is a bug, or just a matter of semantics.","bug"
|
|
"6013","638","recarray scalars are not swapped on assignment Trac 2185).
|
|
Original ticket http projects.scipy.org numpy ticket 2185 on 2012 07 14 by trac user mbyt, assigned to unknown.
|
|
Hello, numpy seems to not swap the byte order when assigning scalars in record arrays !python >>> import numpy as np >>> >>> dt = np.dtype [ ... head , >u4 ), ... data , >u4 , 2), ... ]) >>> buf = np.recarray 1, dtype=dt) >>> buf[0][ head ] = 1 >>> buf[0][ data ][ ] = [1,1] >>> >>> h = buf[0][ head ] >>> d = buf[0][ data ][0] >>> buf[0][ head ] = h >>> buf[0][ data ][0] = d >>> print buf, \n , h.dtype.byteorder, d.dtype.byteorder [ 16777216L, array [1, 1], dtype=uint32))] = = The byte order of both, h and d is native.
|
|
However, when reassigning the scalar head in the recarray to h, the byte order is not swapped and h gets wrapped around to 16777216L should be 1).
|
|
Interestingly, the same is ok with an element of a vector d stays 1, thus is swapped).
|
|
This behavior is the same on windows 32bit, python 2.7.4, numpy 1.6.2. and on linux 32bit, python 2.6.6, numpy 1.4.1.
|
|
Keep up the good work and cheers,[[BR]] Moritz","bug"
|
|
"6013","6380","iOS arm) compile defines both BIG and LITTLE endian.
|
|
Compiling NumPy for iOS armv7 and arm64 architectures) .
|
|
does not correctly configure the endian defines.
|
|
Because this is an iOS application, I am cross compiling NumPy on OSX with the following setup NumPy Version v1.9.3 AND git branch 1.10.x Host OS OSX 10.10.5 Xcode 7.0 iOS Minimum Version 8.0 During compilation, errors are encountered which are fixed by forcing CPU endian detection instead of relying on endian.h.
|
|
Example in the 1.10.x branch .
|
|
Example from v1.9.3 .
|
|
I suspect this is related to the fact that arm supports both BIG and LITTLE endian modes.","bug"
|
|
"6013","6381","iOS compile fails without ATLAS BLAS LAPACK .
|
|
Compiling NumPy for iOS without ATLAS, BLAS, and LAPACK, the final application link fails with undefined symbols.
|
|
list of undefined symbols below).
|
|
I am cross compiling NumPy in this environment NumPy Version v1.9.3 AND git branch 1.10.x Host OS OSX 10.10.5 Xcode 7.0 iOS Minimum Version 8.0 NumPy Compile Environment .
|
|
It appears that .
|
|
correctly reads the environment and detects that ATLAS, LAPACK, and BLAS are disabled, however .
|
|
still returns a valid data structure.
|
|
Checks in the code only looks for a valid return value and does not check the content to verify the functionality is enabled.
|
|
I solved this in version v1.9.3 by making the following changes .
|
|
Compiling the branch .
|
|
commit a600a5c396f236ac4542ab33dc6bae816efadb82 required the following changes .
|
|
List of undefined symbols .","bug"
|
|
"6013","6387","Updating the docstring of loadtxt.
|
|
The docstring of the first argument of loadtxt currently states .
|
|
In fact the generator can yield strings .
|
|
Not making a patch yet because we may as well clarify encoding issues at the same time... or not?
|
|
Something like If the file or generator is yielding unicode strings, they will be decoded as...","enhancement"
|
|
"6013","6388","loadtxt with non matching number of fields in dtype.
|
|
The docstring of the `dtype` argument of `loadtxt` states .
|
|
In fact, it is currently dropping columns from the end if the dtype has fewer fields and raising a not so helpful ValueError if it has more .
|
|
I would suggest raising something like `ValueError Cannot load file with X fields into a dtype with Y fields )` in both cases.","bug"
|
|
"6013","639","failure to solve system of linear equations Trac 2186).
|
|
Original ticket http projects.scipy.org numpy ticket 2186 on 2012 07 16 by trac user al khwarizmi, assigned to pv.
|
|
Sadly, numpy.linalg.solve is giving me crap a = numpy.matrix [1 2 2 0; 1 0 0 2; 3 2 1 6; 0 1 0 1] ) b = numpy.matrix 0; 0; 0; 0 ) numpy.linalg.solve a, b) outputs a zero vector.
|
|
If it means anything, I m not just trying to break numpy; this is a real system of stoichiometric equations.
|
|
The computer algebra system Maxima outputs the correct answer.
|
|
In Maxima what I did was type in linsolve [r0 2 r1 = 2 p0, r0 = 2 p1, 3 r0 2 r1 = p0 6 p1, r1 = p1], [r0, r1, p0, p1]); It outputs [r0 = 2 r2, r1 = r2, p0 = 2 r2, p1 = r2] Thanks in advance for your hard work.","bug"
|
|
"6013","6390","BUG searchsorted on an object array of tuples returns incorrect result.
|
|
This is present in both NumPy 1.9.3 and master .
|
|
xref https github.com pydata pandas issues 11201","bug"
|
|
"6013","6393","Add indexing to ndindex.
|
|
It would be nice to make `ndindex` an abc.Sequence so that it supports indexing e.g., ` np.ndindex 2,3)[3]`), ` contains ` and ` reversed ` efficiently.","enhancement"
|
|
"6013","6395","masked array 2 operation differs from array 2 operation on cluster architecture.
|
|
Hi juliantaylor, it s me again.
|
|
I encounter a similar issue as before 6251) with the following code using now Python 3.4.3 GCC 4.7.2) with numpy 1.9.3 .
|
|
The output is .
|
|
numpy.show config ) .
|
|
The numpy tests pass, but this lets one scipy test fail, scipy scipy 5197 .
|
|
Do you think it is related.
|
|
This time I used the newer C compiler, though.","bug"
|
|
"6013","6398","generalizing meshgrid a little more.
|
|
A few versions ago, meshgrid added the option for more than two dimensions.
|
|
I would like to see intelligence in meshgrid such that if you only pass a single one dimensional coordinate array, that array will just be transparently returned back as it was passed in.
|
|
This way the usage of meshgrid can work for any arbitrary number of dimensions, including 1.
|
|
As it is now, if you have a code that uses meshgrid with a variable number of dimensions, you have to specifically catch the 1 dimensional case and skip meshgrid for that.
|
|
meshgrid should automatically do so for you.
|
|
Also, why does the documentation refer to the coordinate arrays as x1, x2, x3, ...., xn?
|
|
Should it start counting at 0, making them x0, x1, x2, ...., xn, like normal python conventions?","enhancement"
|
|
"6013","6399","Special casing particular scalars for np.power.
|
|
Pointed out on SO [here] http stackoverflow.com q 32903967 3923281).
|
|
For the scalar exponents 1, 0, 0.5, 1 and 2, it is significantly faster to write `arr exp` than `np.power arr, exp)` because the former catches these scalars and handles them using more appropriate functions.
|
|
For example .
|
|
Would it make sense for `np.power` to also handle these scalars in this way?","bug"
|
|
"6013","640","fixed dtype, shape) view of recarray broken Trac 2187).
|
|
Original ticket http projects.scipy.org numpy ticket 2187 on 2012 07 22 by rgommers, assigned to unknown.
|
|
Broken by https github.com numpy numpy pull 350 >>> import statsmodels.api as sm >>> dta = sm.datasets.macrodata.load ) >>> dta.data[[ infl , realgdp ]].view float,2)) Traceback most recent call last) File tmp.py , line 4, in <module> dta.data[[ infl , realgdp ]].view float,2)) File Users rgommers Code numpy numpy core records.py , line 496, in view return self.
|
|
array ).view dtype) ValueError new type not compatible with array.
|
|
Works with [578a4199] http github.com numpy numpy commit 578a4199), broken with [bc1005324] http github.com numpy numpy commit bc1005324).
|
|
Looks like a blocker for 1.7.0 to me.","bug"
|
|
"6013","6404","numpy.lib.Arrayterator doesn t appear in docs.
|
|
I just found out about this class from perusing the NumPy source code; it doesn t appear anywhere in the documentation, even though it s been around since 2008.","bug"
|
|
"6013","641","python3 regression for unique on dtype=object arrays with varying items types Trac 2188).
|
|
Original ticket http projects.scipy.org numpy ticket 2188 on 2012 07 23 by yarikoptic, assigned to unknown.
|
|
tested against current master present in 1.6.2 as well) If with python2.x series it works ok, without puking > python2.7 c import numpy as np; print repr repr np.unique np.array [1,2, None, str ])))) array [None, 1, 2, str], dtype=object) NB I will report a bug on repr here separately if not yet filed it fails with python3.x altogether > python3.2 c import numpy as np; print repr repr np.unique np.array [1,2,None, str ]))))) Traceback most recent call last) File <string> , line 1, in <module> File usr local lib python3.2 dist packages numpy lib arraysetops.py , line 194, in unique ar.sort ) TypeError unorderable types int ) > NoneType ) whenever IMHO it must operate correctly semantic of unique ) action should not imply ability to sort the elements","bug"
|
|
"6013","6410","setup.py clean breaks subsequent build ext.
|
|
After doing a `python setup.py clean`, I can t rebuild NumPy .
|
|
The problem is that `numpy config .py` and `numpy distutils config .py` persist.
|
|
It would be nice if `setup.py clean` could clean those up as well.","bug"
|
|
"6013","6412","Incorrect hashbang for Numpy wheels.
|
|
When installing Numpy via the wheels, the hashbang for the f2py script is incorrect, and is set to .
|
|
This causes the script to not work.
|
|
I think this should instead have .
|
|
cc matthew brett soylentdeen","bug"
|
|
"6013","6418","Allow pathlib.Path arguments.
|
|
`savez`, `savez compressed`, and `load` all take `str` as an argument, but give an error when passed a `pathlib.Path` argument as a filename.
|
|
Now that `pathlib.Path` is in the standard library, it would be nice if these functions could accept a `Path` instance as an argument, in addition to `str` and `file` objects.
|
|
Example .
|
|
.","enhancement"
|
|
"6013","6419","Test failures on py35.
|
|
Anyone know if the following test failures are known to be expected?
|
|
Feel free to close if they are, or if they are known artifacts of my lazy build procedure... Numpy built using openblas anaconda python3.5 on OSX yosemite.
|
|
Numpy built and installed via conda build with recipe given in following pull request https github.com conda conda recipes pull 436 .","bug"
|
|
"6013","642","segfault on searchsorted of wrong dtype in structured array Trac 2189).
|
|
Original ticket http projects.scipy.org numpy ticket 2189 on 2012 07 23 by trac user akhmerov, assigned to unknown.
|
|
numpy.searchsorted causes segfault when one searches in a structured array with a wrong dtype.
|
|
Bug is reproduced by python c import numpy as np; a = np.array [ a , 1)], dtype= S1, int ); np.searchsorted a, 1.2) gdb traceback 0 0x00007ffff5d63413 in ?? )
|
|
from usr lib python2.7 dist packages numpy core multiarray.so 1 0x00007ffff5d736a9 in ?? )
|
|
from usr lib python2.7 dist packages numpy core multiarray.so 2 0x000000000042a485 in PyEval EvalFrameEx ) 3 0x00000000004317f2 in PyEval EvalCodeEx ) 4 0x000000000042a998 in PyEval EvalFrameEx ) 5 0x00000000004317f2 in PyEval EvalCodeEx ) 6 0x000000000042a136 in PyRun StringFlags ) 7 0x0000000000432209 in PyRun SimpleStringFlags ) 8 0x000000000054c41a in Py Main ) 9 0x00007ffff68e576d in libc start main ) from lib x86 64 linux gnu libc.so.6 10 0x000000000041b931 in start )","bug"
|
|
"6013","6422","Atomic dtypes are not interned by unpickling.
|
|
When debugging an inconsistent behavior in `joblib.hash` joblib joblib 251) we noticed that atomic dtypes are always interned by default .
|
|
Note that this is not the case for compound dtypes .
|
|
This is all fine.
|
|
The problem however is that this behavior is not preserved when unpickling atomic dtypes .
|
|
Apparently it comes from the fact that the ` reduce ` method of the `dtype` class asks to build the new instance with the `copy=True` constructor argument to explicitly disable the dtype interning both for atomic and compound data types.
|
|
This makes it behave as follows .
|
|
This unpickling behavior is quite surprising.
|
|
Is this intended?
|
|
Why not make atomic dtypes always or never) interned in a consistent manner?","bug"
|
|
"6013","6423","Incorrect version in 1.10.0.post1 setup.py.
|
|
Hi, The version in setup.py has not been updated to reflect the source version 1.10.0.post1 MAJOR = 1 MINOR = 10 MICRO = 0 ISRELEASED = True VERSION = d. d. d MAJOR, MINOR, MICRO) Thanks Carlos","bug"
|
|
"6013","6425","ma.put issue.
|
|
The following snippet does not behave as expected import numpy.ma as ma z = ma.array [3.0, ma.masked]) data = ma.zeros 10) ma.put data, [1,2], z) print data[3] will throw an IndexError If after initializing z, I do z.mask = ma.getmaskarray z) then everything works fine.","bug"
|
|
"6013","6427","Build checks failing on configtest.c no return in nonvoid function .
|
|
I am currently trying to build numpy for openSUSE, and I am getting the following issue, which is considered a serious issue and thus results in a build failure E python numpy no return in nonvoid function configtest.c 10 This error is due to ` configtest.c` having a function that is non void, but doesn t return anything.
|
|
However, I am can t find ` configtest.c` it seems to be generated on the fly) so I can t figure out exactly where the problem is.","bug"
|
|
"6013","6428","msvc9 32 bit builds are unusable.
|
|
numpy 1.10.0 compiled with Visual Studio 2008 for 32 bit Python fails many tests only part of the output is shown) .
|
|
Reverting https github.com numpy numpy pull 6252 fixes this problem.
|
|
Builds using msvc9 64 bit, msvc10, msvc14, and icl do not show this problem.
|
|
I did not notice this before because I was testing the RCs with the Intel, not msvc, compilers.","bug"
|
|
"6013","643","ValueError in nanargmin Trac 2190).
|
|
Original ticket http projects.scipy.org numpy ticket 2190 on 2012 07 24 by trac user renaud.dussurget, assigned to unknown.
|
|
The function nanargmin, calling nanop returns ValueError cannot convert float NaN to integer when all values are masked nans).
|
|
This occurs when the nanop functions tries to fill the result array index res) with NaN values.
|
|
This, however, cannot work.
|
|
Indeed, the res array is initialized using the op argmin in our case) function.
|
|
This function returns an array of ints.
|
|
Thus the following line cannot work res[np.arange len mask all along axis)).compress mask all along axis)] = np.nan A workaround to this is to replace the return value NaNs) by 1 this will however affect the behaviour of the nansum function).","bug"
|
|
"6013","6431","NumpyVersion does not handle the `post` suffix correctly.. .","bug"
|
|
"6013","6435","setup.py seems to ignore site.cfg s runtime library path variable.
|
|
Hello, I m trying to build a version of numpy that links against a locally built version of openblas.
|
|
I added the `library dirs` and `include dirs` variable to the openblas section in my site.cfg.
|
|
Running `ldd` confirms that lapack lite.so links against libopenblas.so.
|
|
However, for some reason the `runtime library dirs` entry and `rpath` for that matter) in my site.cfg file seems to be ignored.
|
|
This means that I get an `ImportError` when importing numpy unless I define LD LIBRARY PATH before running the interpreter.
|
|
I fixed the issue by manually setting the rpath via `extra link args = Wl, rpath, path to openblas lib`.
|
|
Am I misinterpreting the meaning of the `runtime library dirs` variable or is this a bug?","bug"
|
|
"6013","644","numpy.test ) fails after running scipy.test ) Trac 2191).
|
|
Original ticket http projects.scipy.org numpy ticket 2191 on 2012 07 25 by trac user GeorgeSalt, assigned to unknown.
|
|
Python 2.7.1 r271 86832, Apr 11 2011, 18 13 53) [GCC 4.5.2] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy, scipy >>> numpy.test ) Running unit tests for numpy NumPy version 1.6.2 NumPy is installed in usr local lib python2.7 dist packages numpy Python version 2.7.1 r271 86832, Apr 11 2011, 18 13 53) [GCC 4.5.2] nose version 1.1.2 ....................................................................
|
|
Ran 3568 tests in 14.237s OK KNOWNFAIL=5, SKIP=4) <nose.result.TextTestResult run=3568 errors=0 failures=0> >>> scipy.test ) Running unit tests for scipy NumPy version 1.6.2 NumPy is installed in usr local lib python2.7 dist packages numpy SciPy version 0.10.1 SciPy is installed in usr local lib python2.7 dist packages scipy Python version 2.7.1 r271 86832, Apr 11 2011, 18 13 53) [GCC 4.5.2] nose version 1.1.2 ....................................................................
|
|
Ran 5103 tests in 43.796s OK KNOWNFAIL=13, SKIP=35) <nose.result.TextTestResult run=5103 errors=0 failures=0> >>> numpy.test ) Running unit tests for numpy NumPy version 1.6.2 NumPy is installed in usr local lib python2.7 dist packages numpy Python version 2.7.1 r271 86832, Apr 11 2011, 18 13 53) [GCC 4.5.2] nose version 1.1.2 .................................................................... ==================================================================== FAIL test doctests test polynomial.TestDocs) Traceback most recent call last) File usr local lib python2.7 dist packages numpy lib tests test polynomial.py , line 84, in test doctests return rundocs ) File usr local lib python2.7 dist packages numpy testing utils.py , line 988, in rundocs raise AssertionError Some doctests failed \n s \n .join msg)) AssertionError Some doctests failed File usr local lib python2.7 dist packages numpy lib tests test polynomial.py , line 32, in test polynomial Failed example p q Expected poly1d [ 0.33333333]), poly1d [ 1.33333333, 2.66666667])) Got poly1d [ 0.333]), poly1d [ 1.333, 2.667])) File usr local lib python2.7 dist packages numpy lib tests test polynomial.py , line 54, in test polynomial Failed example p.integ ) Expected poly1d [ 0.33333333, 1. , 3. , 0. ])
|
|
Got poly1d [ 0.333, 1. , 3. , 0. ])
|
|
File usr local lib python2.7 dist packages numpy lib tests test polynomial.py , line 56, in test polynomial Failed example p.integ 1) Expected poly1d [ 0.33333333, 1. , 3. , 0. ])
|
|
Got poly1d [ 0.333, 1. , 3. , 0. ])
|
|
File usr local lib python2.7 dist packages numpy lib tests test polynomial.py , line 58, in test polynomial Failed example p.integ 5) Expected poly1d [ 0.00039683, 0.00277778, 0.025 , 0. , 0. , 0. , 0. , 0. ])
|
|
Got poly1d [ 0. , 0.003, 0.025, 0. , 0. , 0. , 0. , 0. ])
|
|
Ran 3568 tests in 12.423s FAILED KNOWNFAIL=5, SKIP=4, failures=1) <nose.result.TextTestResult run=3568 errors=0 failures=1>","bug"
|
|
"6013","645","Inverse of a singular matrix Bug Trac 2192).
|
|
Original ticket http projects.scipy.org numpy ticket 2192 on 2012 07 25 by trac user Amit001, assigned to pv.
|
|
Hi, Numpy is not handling singular matrices properly.
|
|
>>> numpy.
|
|
version 1.6.1 for eg.
|
|
Numpy is giving the inverse of the below singular matrix.
|
|
a = array [[ 1, 2, 3], [ 5, 7, 9], [ 6, 9, 12]]) >>> det a) 5.3290705182007443e 015 >>> rank a) 2 >>> inv a) array [[ 5.62949953e 14, 5.62949953e 14, 5.62949953e 14], [ 1.12589991e 15, 1.12589991e 15, 1.12589991e 15], [ 5.62949953e 14, 5.62949953e 14, 5.62949953e 14]])","bug"
|
|
"6013","6452","np.datetime NaT ) == np.datetime NaT , us ) is no longer True on NumPy 1.10.
|
|
I m not quite sure exactly what happened, but whereas `np.datetime64 NaT ) == np.datetime64 NaT , us )` is `True` on NumPy 1.9, it s `False` on NumPy 1.10.
|
|
I tested this out on a few NumPy Python versions, and in some cases this comparison results in warnings.
|
|
In particular Python 2.7, NumPy 1.9 `True` No warning Python 3.4, NumPy 1.9 `True` `DeprecationWarning Implicitly casting between incompatible kinds.
|
|
In a future numpy release, this will raise an error.
|
|
Use casting= unsafe if this is intentional.` Python 2.7, NumPy 1.10 `False` `DeprecationWarning elementwise == comparison failed; this will raise an error in the future.` Python 3.4, NumPy 1.10 `False` No warning Based on git blame, it looks like [the latest deprecation warning] https github.com numpy numpy blob 9cc55dc7720a949cb3e6578805fe6f70906a700e numpy core src multiarray arrayobject.c L1429 L1430) introduced by njsmith in 5964. xref https github.com xray xray pull 618","bug"
|
|
"6013","6454","copysign fails to work inplace with integer arrays.
|
|
.
|
|
And then trying to change the sign in place which is the original motivation) .","bug"
|
|
"6013","6455","Weird behavior with float128 python long.
|
|
It appears operations between float128 and python longs have some strange corners .
|
|
The first one makes some sense, but the second one is strange.
|
|
Apparently, this can also trigger segfaults, cf.
|
|
https stackoverflow.com questions 33072998 how to multiply a super large number with a super small number in python ) Results are similar with Numpy 1.8.2","bug"
|
|
"6013","6458","Convolution of masked arrays doesn t work.
|
|
For example a = np.random.random 10) a[5] = 1e9 a = np.ma.masked equal a, 1e9) b = np.convolve a, np.ones 4) 4.)
|
|
gives incorrect results.","enhancement"
|
|
"6013","6459","error when making view of a recarray of an aligned record dtype.
|
|
A `TypeError view would access data parent array doesn t own` is raised when a view is created from a recarray and its dtype is an aligned record.
|
|
The following code demonstrate this error on py3.4 numpy1.10 .
|
|
The cause of the error is likely due to the loss of the `aligned` attribute at <https github.com numpy numpy blob b65cdbf41bdf11aa4732983555a32043f9abec13 numpy core records.py L502>.
|
|
The itemsize of the `aligned dtype` is 8 while the `nonaligned dtype` has itemsize 5.","bug"
|
|
"6013","646","Test failures if Python is compiled with pydebug Trac 2193).
|
|
Original ticket http projects.scipy.org numpy ticket 2193 on 2012 07 29 by trac user skrah, assigned to unknown.
|
|
Observed with Python 3.2.
|
|
Compile Python with .
|
|
configure with pydebug make Build NumPy from the source repository and run the tests Python 3.2.3 3.2 290f04722be3, Jul 29 2012, 13 41 41) [GCC 4.4.3] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy [176157 refs] >>> numpy.test ) Running unit tests for numpy NumPy version 1.8.0.dev Unknown NumPy is installed in home stefan usr lib python3.2 site packages numpy Python version 3.2.3 3.2 290f04722be3, Jul 29 2012, 13 41 41) [GCC 4.4.3] nose version 1.1.2 home stefan usr lib python3.2 site packages nose 1.1.2 py3.2.egg nose core.py 247 ResourceWarning unclosed file < io.TextIOWrapper name= home stefan usr lib python3.2 site packages nose 1.1.2 py3.2.egg nose usage.txt mode= r encoding= ANSI X3.4 1968 > os.path.dirname file ), usage.txt ), r ).read ) ........................S.................................................................................................python3.2 numpy core src multiarray common.c 161 PyArray DTypeFromObjectHelper Assertion ` PyObject ) temp)) >ob type)) >tp flags 1L<<27))) != 0) failed.
|
|
Aborted reverse i search)` .
|
|
configure prefix= home stefan usr Cwith pydebug make","bug"
|
|
"6013","6462","median of empty array now produces IndexError.
|
|
Recently upgraded from 1.9.3 to 1.10.0 in Python 2.7 but I get the same behavior in Python 3.5 on both Mac and Linux).
|
|
In 1.9.3, running `np.median np.array []))` produces a NaN and a warning of ` warnings.warn Mean of empty slice.
|
|
, RuntimeWarning)`.
|
|
Meanwhile, in 1.10.0, I m now getting IndexError index 1 is out of bounds for axis 0 with size 0 The traceback points to the line in python2.7 site packages numpy lib function base.pyc in median a, axis, out, overwrite input) > 3138 n = np.isnan part[..., 1]) Is this intended behavior now in 1.10.0?
|
|
I could switch all my code over to using `np.nanmedian`, which still produces the NaN result on empty arrays from 1.9.3, but this looks to be more of an accident bug based on the tracebacks.","bug"
|
|
"6013","6463","ENH float 4 IBM format support as part of the basic types such as f4 ).
|
|
I m aware that there are some ufunctions which convert float4ibm to ieee float and inverse.
|
|
The point of this enhancement request that there are many file formats mostly in Earth Science) dealing with float4ibm and doing something like .
|
|
looks much cleaner and possibly closer to the user) than .
|
|
My point is that ibm float is the same thing as ieee 4 byte and having one implemented and not another is unfair.","enhancement"
|
|
"6013","6464","ufunc cannot be cast with rule `same kind`.
|
|
Updgrading to 1.10, I noticed the following new behavior .
|
|
See also https github.com nipy dipy issues 730.
|
|
Is this an intended new behavior?
|
|
If so, what does it portend what should I stop doing?
|
|
)?","bug"
|
|
"6013","6467","performance regression for record array access in numpy 1.10.1.
|
|
It appears that access numpy record arrays by field name is significantly slower in numpy 1.10.1.
|
|
I have put below a simple example test that illustrates the issue.
|
|
I am aware that this particular example is much better accomplished by other means.
|
|
The point is that array access is slow, not that this a representative problem.)
|
|
The test script is .
|
|
Here are the running times for different versions of numpy .
|
|
esheldon has reproduced the relative timing differences on linux in addition to my tests which were with my mac.
|
|
I profiled the code for v1.10.1 and found this .
|
|
It appears that new code added at the python level for error checking is significantly degrading performance.","bug"
|
|
"6013","6468","numpy.interp uses left value even when x[0]==xp[0].
|
|
The following code should demonstrate that numpy.interp uses the padding value specified by the kwarg left even when the 0th element of the x values being interpolated onto is identical to the 0th element of the input x values.
|
|
This problem arose when I upgraded to numpy version 1.10.1 from version 1.9 .","bug"
|
|
"6013","6469","Better error message for float arg to np.histogram bins.
|
|
In numpy 1.9.x and previous this worked .
|
|
It looks like np.histogram is casting to int, because bins=2.1 and bins=2.9 both give 2 bins.
|
|
With numpy 1.10, I get the following error .
|
|
I can see the argument for not allowing floats that are not integers, but maybe a more explicit error message would help?
|
|
Found via https github.com nipy nipy issues 369","bug"
|
|
"6013","647","Gnu95FCompiler doesn t find libmingwex.a when using msvc gfortran Trac 2194).
|
|
Original ticket http projects.scipy.org numpy ticket 2194 on 2012 07 30 by trac user stevenwinfield, assigned to cournape.
|
|
This is actually a bug in 1.6.2, but I couldn t select that) A MingW installation is laid out as follows MingW | lib | | libmingw32.a | | libmingwex.a | | gcc | | mingw32 | | <version number> | | libgcc.a | mingw32 | lib | ...
|
|
In numpy distutils fcompiler gnu.py, in the get library dirs ) method of Gnu95FCompiler beginning line 299) the code first gets the libgcc directory then tries to navigate to the directory containing libmingwex.a by doing the following d = os.path.normpath self.get libgcc dir )) root = os.path.join d, os.pardir, os.pardir, os.pardir, os.pardir) mingwdir = os.path.normpath os.path.join root, target, lib )) target here is mingw32 , so mingwdir is MingW\mingw32\lib , which exists but doesn t contain libmingwex.a.
|
|
If target were to be removed from the last line then all would be fine.
|
|
This problem manifests itself when trying to build scipy, with the error message LINK fatal error LNK1181 cannot open input file mingw32.lib This is because a few MingW .a files are copied by libs with msvc and fortran ) in numpy distutils command build ext.py) to .lib files in a temporary build directory so that they can be found by msvc one of which should be libmingw32.a => mingw32.lib.
|
|
However, the .a file is never found and so the copying never takes place.","bug"
|
|
"6013","6471","dtype S0 , ))) raise unexpected exception.
|
|
Probably linked to 4955 Note that `dtype S`x` , )))` with x different from 0 works perfectly","bug"
|
|
"6013","6475","np.allclose returns memmap object when comparing a memmap array to a normal array.
|
|
This is a new behaviour in numpy 1.10 and according to the doc np.allclose is supposed to return a bool.
|
|
Is it an intended change?
|
|
Note that the returned object still has the right truthiness value so it should not be a problem except in conditions like `if np.allclose ...) is True `.
|
|
In our case a doctest fails but we could work around the failure if needed.
|
|
A snippet reproducing the behaviour .
|
|
Output with numpy 1.10.1 .
|
|
Output with numpy 1.9.3 .","bug"
|
|
"6013","6476","numpy.lib.read array header 1 0 is not returning anything.
|
|
Under numpy 1.9.3 the functions `numpy.lib.format.read array header 1 0` and `numpy.lib.format.read array header 2 0` are broken.
|
|
They are meant to return information from the header of the array file data as the docstring describes, but in fact they return None as they re missing a return statement.
|
|
The fix seems to just be to add a `return` in front of the call to ` read array header`.","bug"
|
|
"6013","6477","Augmented assignment and type conversion.
|
|
Hello.
|
|
Here s a bit from an iPython session In [1] import numpy as np In [2] x = np.arange 10) In [3] x Out[3] array [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]) In [4] x = x 2.0 In [5] x Out[5] array [ 0. , 0.5, 1. , 1.5, 2. , 2.5, 3. , 3.5, 4. , 4.5]) In [6] x = np.arange 10) In [7] x = 2.0 In [8] x Out[8] array [0, 0, 1, 1, 2, 2, 3, 3, 4, 4]) So, oddly, the in place assignment does integer division, while the copying assignment does float division.
|
|
I observe analogous behavior for the other arithmetic ops.
|
|
Is this expected behavior?","bug"
|
|
"6013","6479","np.equal can still return NotImplemented.
|
|
It looks like this didn t make it into the clean up in 5864 5964. .
|
|
There s a good chance njsmith is already aware of this, but I thought I would note it for future reference.
|
|
In any case, this will probably be easier to cleanup once we remove the deprecated behavior.
|
|
As reported by amueller on Gitter.","bug"
|
|
"6013","648","Memory leak in concatenate Patch is attached Trac 2195).
|
|
Original ticket http projects.scipy.org numpy ticket 2195 on 2012 07 31 by trac user cianci, assigned to unknown.
|
|
A temporary array is created with PyArray malloc, but not released with PyArray free) after use.
|
|
Parent [] http github.com numpy numpy commit ) 26fed25e7f040564e7be4f82d7cd8a8fcc8fa287 Attached patch merely inserts the necessary call to free, in both the success and failure cases.","bug"
|
|
"6013","649","The behavior of the npv function does not match the documentation Trac 2196).
|
|
Original ticket http projects.scipy.org numpy ticket 2196 on 2012 08 04 by trac user npgraham1, assigned to unknown.
|
|
The documentation for the npv indicates that it returns the true net present value, as given by the formula shown on the npv function s page.
|
|
This is the technically correct and universally accepted definition of net present value.
|
|
However, the function itself appears to mimic the Excel NPV function, which simply calculates the present value of a series of cash flows, and is known to be incorrect but persists for compatibility reasons).
|
|
Put differently, the documentation indicates that the summation is from 0 to n 1, but the function returns a summation of 1 to n. Since the current behavior is widely understood, it is probably acceptable to change the documentation to match, with a brief explanation of how to correctly compute NPV using the npv function.
|
|
Alternatively, the function could be easily) altered to match the documentation a brief note that this is different from several popular spreadsheet NPV functions might be useful).
|
|
Either way, one ought to change in order to match the other; as someone who studies finance, I d prefer the correct calculation, but either way is reasonable.","bug"
|
|
"6013","6491","Error in broadcasting stride tricks array.
|
|
We hit a new test failure when testing against numpy 1.10.1 https github.com nipy nibabel pull 358 The test failure boils down to this .
|
|
On Linux not OSX) I get the following error, for numpy 1.10.1 .
|
|
There is no error for the same code on numpy 1.9.3.
|
|
I think this is the same issue, arising in scipy https github.com scipy scipy pull 5374 issuecomment 148775998","bug"
|
|
"6013","6492","git tags for 1.10.0 and 1.10.1 missing.
|
|
Maybe I have misunderstood how git tags work I just learned about them), but I think we need git tags for the 1.10.0 and 1.10.1 releases and the rc s. At least, I don t see them in the output of `git tag`.
|
|
It looks like starting in 1.8.1 Julian Taylor has been tagging the non rc releases with annotated tags with his GPG signature, judging from the output of `git tag v v1.8.1`.","bug"
|
|
"6013","6493","Add a build step at the end of .travis.yml that uploads working wheels to a known location.
|
|
There are projects that want to run their tests against the latest numpy master, and it would be a lot easier for them if there was a wheel they could just download.
|
|
Some of them do `pip install git http ...path to numpy master`, but this has two disadvantages a) everyone has to rebuild from scratch all the time, b) sometimes hopefully rarely, but...) master is broken and this just doesn t work.
|
|
The solution is that we should have an `after success ` block in .travis.yml so that after each commit has passed its tests, we upload the corresponding wheel somewhere, probably one of these rackspace accounts that everyone seems to have access to.
|
|
Ideally we would only do this if all tests pass, but unfortunately this is impossible with travis currently https github.com travis ci travis ci issues 929 .
|
|
In the mean time, uploading each build variant iff its tests pass seems like a reasonable compromise.)
|
|
ogrisel volunteers to do this in a few weeks, so if it s, say, mid November and it hasn t happened, then someone should nag him ; ) https github.com numpy numpy issues 6462 issuecomment 148167879)","bug"
|
|
"6013","6494","Update release checklist to include more QA encouragement of QA for pre releases.
|
|
See discussion starting here https github.com numpy numpy issues 6462 issuecomment 148107774 It sounds like the actionable conclusions are we should make sure to test RCs against at least scipy, pandas, sklearn, skimage, astropy list subject to revision).
|
|
This could be done manually by the release manager juliantaylor might have hints?
|
|
), or some other way e.g.
|
|
by pinging the various package authors and waiting for confirmation that they ve tested their code before we make the final release this would also work to systematically handle dedicated QAers who aren t just running a single package s test, like C. Gohlke or people who want to test internal software stacks).
|
|
We should probably exhort people to test pre releases on twitter It might be good to start a public campaign to encourage everyone who uses `pip install` in their CI setup to add ` pre` to the options... this would dramatically increase the amount of testing that all pre releases get, not just numpy s. Doesn t work for those using conda in their CI, alas.)","bug"
|
|
"6013","6495","gfortran error unrecognized command line option xe2x80x98 ffpe summary=invalid,zeroxe2x80x99.
|
|
numpy.distutils adds a ` ffpe summary` command line option to gfortran, but it s apparently not supported by all versions of the compiler.
|
|
See https github.com scipy scipy issues 5380 Apparently this was added in b354c005 seems to affect also non windows platforms.","bug"
|
|
"6013","6496","Update pavement.py to sign release notes.. See https github.com scipy scipy pull 4918.","bug"
|
|
"6013","6497","numpy 1.10.1 reduce operation on recarrays.
|
|
Fails.","bug"
|
|
"6013","6498","Mention change in default casting rule in 1.10 release notes..","bug"
|
|
"6013","650","Division by zero when creating array from N dimensional new style buffer with zero length and no strides Trac 2197).
|
|
Original ticket http projects.scipy.org numpy ticket 2197 on 2012 08 08 by trac user batavus, assigned to unknown.
|
|
I believe there s a bug in function array from buffer 3118 in file numpy numpy core src multiarray ctors.c Consider the code which currently starts at line 1228 if view >strides != NULL) { for k = 0; k < nd; k) { strides[k] = view >strides[k]; } } else { d = view >len; for k = 0; k < nd; k) { d = view >shape[k]; strides[k] = d; } } if strides is NULL and shape is something containing at least on zero e.g.
|
|
[0] or [1,0,2]) such that view >len is zero, there will be a division 0 by 0 in d = view >shape[k]; I suggest replacing the line d = view >shape[k]; by if view >shape[k]) d = view >shape[k]; This should resolve the problem and be consistent with how strides is defined in numpy for such cases.
|
|
Also, it s probably a good idea to add a test for this issue.","bug"
|
|
"6013","6506","Generalise rot90 to rotate in any plane.
|
|
Hi.
|
|
A feature request I d like [rot90] http docs.scipy.org doc numpy reference generated numpy.rot90.html) generalised so you could rotate in any plane defined by a pair of axes.
|
|
Obviously you could keep 0, 1) as the default.","enhancement"
|
|
"6013","6507","Not all benchmark tests are listed in final comparison.
|
|
Six tests, 5 in final comparison.
|
|
.","bug"
|
|
"6013","6508","`python runtests.py bench` should take a commit argument..
|
|
It currently runs against the tip of master even if a COMMIT arguement is given.
|
|
Perhaps a bug?","bug"
|
|
"6013","651","np.loadtxt leaks memory Trac 2198).
|
|
Original ticket http projects.scipy.org numpy ticket 2198 on 2012 08 09 by tanriol, assigned to unknown.
|
|
The amount of memory leaked far exceeds the amount the loaded data takes and does not go away when the loaded array is deleted.
|
|
>>> import numpy as np >>> Python consuming 13M RAM >>> arr = np.zeros 10000, 5000), dtype= <i8 ) >>> Python consuming 394M RAM >>> np.savetxt array.txt , arr, d ) >>> Python consuming 395M RAM >>> del arr >>> Python consuming 14M RAM >>> arr = np.loadtxt array.txt , dtype= <i8 ) >>> Python consuming 2245M !)
|
|
>>> del arr >>> Python consuming 1863M !)
|
|
>>> import gc >>> gc.collect ) 6 >>> Python consuming 1863M !)
|
|
CPython 2.7.3, Numpy 1.6.2","bug"
|
|
"6013","6510","[Suggestion] Comparison functions for float point numbers.
|
|
I think this would be useful and easy to implement.
|
|
greater close a, b) = greater equal a, b) | isclose a, b) less close a, b) = less equal a, b) | isclose a, b) greater no close = greater a, b) isclose a, b) less no close = less a, b) isclose a, b) The results are element wise, just like the original functions.
|
|
I m not sure if it is useful enough to be a part of numpy.
|
|
If so, I will try to implement them and make a pull request.","enhancement"
|
|
"6013","6511","np.ctypeslib.as array leaks memory when used on a pointer.
|
|
I have a C function returning an array, which I convert to a numpy array using np.ctypeslib.as array pointer from C function, size of array,)).
|
|
This works as expected, but when I repeatedly call this function about a million times) I found a substantial increase of memory usage for my python process.
|
|
Running a sample script attached, output attached) through valgrind s memcheck output attached), it appears that the problem is in ctors.c, which calls the python C API function PyErr WarnEx.
|
|
The strange thing is that I never see a warning appear in my python output, so this could also be a bug in python.
|
|
For now I will try to work around this problem, but it would be great if this could be fixed.
|
|
Details of installation Ubuntu 15.04 x86 64 Python 3.4.3 installed from ubuntu repo), but problem arises with python 2.7.9 too also from ubuntu repo) Numpy 1.10.1 from pip), but problem also present in numpy 1.8.2 from ubuntu repo) Working example .
|
|
Output of script .
|
|
Relevant valgrind output .","bug"
|
|
"6013","6515","DOC missing space in arrays.classes.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 27 of `arrays.classes.rst`, see [here] https github.com numpy numpy blob master doc source reference arrays.classes.rst), there is a missing space, `arrayobject` > `array object`","bug"
|
|
"6013","6516","DOC typo in arrays.interface.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 313 of `arrays.interface.rst`, see [here] https github.com numpy numpy blob master doc source reference arrays.interface.rst), there is a typo, `asthetic` > `a e sthetic`.","bug"
|
|
"6013","6517","DOC another) typo in arrays.interface.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 158 of `arrays.interface.rst`, see [here] https github.com numpy numpy blob master doc source reference arrays.interface.rst), there is a typo, `appropiately` > `approp r iately`.","bug"
|
|
"6013","6518","DOC yet another) typo in arrays.interface.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
This may be debatable, but on line 74 of `arrays.interface.rst`, see [here] https github.com numpy numpy blob master doc source reference arrays.interface.rst), there is a potential) typo, `homogenous` > `homogen e ous`.
|
|
[This] http grammarist.com usage homogenous homogeneous ) suggests that the latter should be preferred over the former, even if it is in common use these days.","bug"
|
|
"6013","6519","DOC typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 237 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), fortran should be capitalized, `fortran` > ` F ortran`.","bug"
|
|
"6013","652","PySequence Check should return 0 for 0 d arrays Trac 2199).
|
|
Original ticket http projects.scipy.org numpy ticket 2199 on 2012 08 09 by trac user batavus, assigned to unknown.
|
|
0 d arrays are not iterable >>> list numpy.array 1)) Traceback most recent call last) File <stdin> , line 1, in <module> TypeError iteration over a 0 d array Consequently, they do not provide the sequence protocol PySequence Length and PySequence GetItem fail when called for a 0 d array.
|
|
However, PySequence Check returns 1.
|
|
It should return 0.","bug"
|
|
"6013","6520","DOC another) typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 271 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), there is a typo, `unitialized` > `uni ni tialized`.","bug"
|
|
"6013","6521","DOC yet another) typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 289 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), there is a typo, `deter e mined` > `determined`.","bug"
|
|
"6013","6522","DOC one more) typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 424 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), there is a typo, `r e aised` > `raised`.","bug"
|
|
"6013","6523","DOC and one more) typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 1033 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), there is a typo, `explict` > `explic i t`.","bug"
|
|
"6013","6524","DOC and yet one more) typo in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
On line 1662 of `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst), there is a typo, `compat a bility` > `compat i bility`.","bug"
|
|
"6013","6525","DOC and more and more) typos in c api.array.rst.
|
|
THIS ISSUE HAS BEEN CREATED AS PART OF THE BECOME AN OPEN SOURCE CONTRIBUTOR WORKSHOP.
|
|
IF YOU ARE NOT A WORKSHOP PARTICIPANT, PLEASE DO NOT SEND A PULL REQUEST TO FIX IT.
|
|
Several more typos in `c api.array.rst`, see [here] https github.com numpy numpy blob master doc source reference c api.array.rst).
|
|
On line 1832 `whther` > `wh e ther`.
|
|
On line 1916 `ouput` > `ou t put` On line 1917 `respons a bility` > `respons i bility` On line 2043 `specifi c e` > `specifie d ` On line 2113 ` e instein` > ` E instein` On line 2776 `compatib i le` > `compatible` On line 2908 `intialization` > `in i tialization` On line 3071 `ex c ecute` > `execute`","bug"
|
|
"6013","653","Setting the itemsize of a dtype in dict style constructor doesn t work Trac 2200).
|
|
Original ticket http projects.scipy.org numpy ticket 2200 on 2012 08 09 by trac user dpitch40, assigned to unknown.
|
|
From a Python 2.7 console session >>> import numpy as np >>> dt = np.dtype { names [ First ], formats [np.uint32], offsets [3], itemsize 8}) >>> dt.itemsize 7 Obviously the actual itemsize of the dtype doesn t match the one I specified.
|
|
It appears to set the itemsize to the smallest one that accommodates all of the elements of the type with their offsets.
|
|
I am using Numpy 1.6.2.","bug"
|
|
"6013","6530","1.10 regression partition errors out on empty input.
|
|
maintenance 1.9.x .
|
|
1.10.1 .
|
|
Noticed this because it s breaking some tests in patsy.","bug"
|
|
"6013","6532","numpy.inner return wrong inaccurate value sometimes... .
|
|
After updating bumpy from 1.9.2 py27 0 to 1.10.1 py27 0.
|
|
One of my class method returned a different and thus wrong result.
|
|
I localised the error to be from the numpy.inner fonction.
|
|
What is weird is that when I use the method within another class method declared in a different file, it works!
|
|
When I use the class after if name == main in the same file the class is written to test it it does not work even though the input data is exactly the same in each case.
|
|
When I declare the class within a Jupyter notebook, the same problem occurs, a wrong output is given by numpy.inner.","bug"
|
|
"6013","654","numpy.linalg and multiprocessing crash Trac 2201).
|
|
Original ticket http projects.scipy.org numpy ticket 2201 on 2012 08 14 by trac user agchang, assigned to pv.
|
|
There seems to be an issue with numpy.linalg being called in unison with python s multiprocessing module.
|
|
When invoking a numpy.linalg method from within a subprocess, e.g., through a worker pool and map ) function, the program hangs and ignores all interrupt signals.
|
|
The code provided below will demonstrate the case import numpy as np import multiprocessing as mp def foo x) print np.linalg.inv [[2,3],[2,2]]) this causes the crash print np.dot [[1,2],[3,4]],[[1,2],[3,4]]) this works fine def test ) print running... print np.
|
|
version print mp.
|
|
version vals = [1,2,3,4] pool = mp.Pool 1) This has an issue pool = mp.Pool mp.cpu count )) this has an issue pool.map foo, vals) if name == main test ) foo 1) this works fine By hang I mean the program becomes unresponsive and does not respond to interrupts Ctrl C).
|
|
I took a look using pdb and it seems that it hangs after the call waiter .acquire ) in threading.py the python system module) so I suspect some sort of deadlock?
|
|
multiprocessing.
|
|
version is 0.70a1 numpy.
|
|
version is 1.6.1 Some potentially related packages?
|
|
ii libblas3gf 1.2.20110419 2 Basic Linear Algebra Reference implementatio ii liblapack3gf 3.3.1 1 library of linear algebra routines 3 share ii python 2.7.3 0ubuntu2 interactive high level object oriented langu ii python numpy 1 1.6.1 6ubunt Numerical Python adds a fast array facility My kernel version is uname a) Linux agc 3.2.0 29 generic 46 Ubuntu SMP Fri Jul 27 17 03 23 UTC 2012 x86 64 x86 64 x86 64 GNU Linux Let me know for any additional information.
|
|
Thanks.","bug"
|
|
"6013","6542","Tests for exceptions needed in vstack, row stack, hstack, column stack, dstack, etc..
|
|
Example of exception for which there is no test.
|
|
.
|
|
The vstack, row stack, hstack, column stack, dstack, and hsplit functions, as well as their masked cousins, all seem to need such tests.","bug"
|
|
"6013","6549","Nansum function behavior.
|
|
Hello, I noticed the change regarding nan function and especially nansum function.
|
|
I think this choice is a big mistake.
|
|
I know that Matlab and R have made this choice but it is illogical and counterintuitive.
|
|
First argument is about logic.
|
|
An arithmetic operation between Nothing and Nothing cannot make a figure or an object.
|
|
Nothing Object can be an object or something else, but from nothing, it cannot ensue something else than nothing.
|
|
I hope you see what I mean.
|
|
Secondly, it s counterintuitive and not convenient.
|
|
Because, if you want to fill the result of nanfunction you can do that easily .
|
|
Whereas, if the result is already filled with zero on NaN full rows, you cannot replace the result of NaN full rows by NaN easily.
|
|
In the case above, you cannot because you lost information about NaN full rows.
|
|
I know it is tough to come back to a previous stage but I really think that it is wrong to absolutely fill with zeros the result of arithmetic operation containing NaN.
|
|
Thank for your work guys ; )","bug"
|
|
"6013","655","NumPy and matplotlib.pylab on Mac Trac 2202).
|
|
Original ticket http projects.scipy.org numpy ticket 2202 on 2012 08 14 by trac user Sattelmayer, assigned to unknown.
|
|
Python does not import pylab from matplotlib RuntimeError module compiled against API version 6 but this version of numpy is 4 Is there no version of numpy than 4?
|
|
Even NumPy 1.6. is version 4?
|
|
Can NumPy be updated to 6?
|
|
Details Python 2.7.3 v2.7.3 70274d53c1dd, Apr 9 2012, 20 52 43) [GCC 4.2.1 Apple Inc. build 5666) dot 3)] on darwin ... >>> import matplotlib >>> from matplotlib import pylab RuntimeError module compiled against API version 6 but this version of numpy is 4 Traceback most recent call last) File <stdin> , line 1, in <module> File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages matplotlib pylab.py , line 221, in <module> from matplotlib import mpl pulls in most modules File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages matplotlib mpl.py , line 1, in <module> from matplotlib import artist File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages matplotlib artist.py , line 6, in <module> from transforms import Bbox, IdentityTransform, TransformedBbox, TransformedPath File Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages matplotlib transforms.py , line 34, in <module> from matplotlib.
|
|
path import affine transform ImportError numpy.core.multiarray failed to import >>>","enhancement"
|
|
"6013","6550","array ...) casts mix lists of int and string to string automatically excepted dtype object).
|
|
The following code explains the issue >>> import numpy >>> print numpy.
|
|
version 1.10.1 >>> a = [[ a , 1], [2, 3]] >>> b = numpy.array a) >>> print repr a) [[ a , 1], [2, 3]] >>> print repr b) array [[ a , 1 ], [ 2 , 3 ]], dtype= |S1 ) I would have expected that not specifying a dtype would have given the result as if I had specify a dtype object) to be able to hold both ints and strings) >>> b = numpy.array a, dtype=numpy.dtype object)) >>> print repr b) array [[ a , 1], [2, 3]], dtype=object) as the doc http docs.scipy.org doc numpy reference generated numpy.array.html describes dtype data type, optional The desired data type for the array.
|
|
If not given, then the type will be determined as the minimum type required to hold the objects in the sequence.
|
|
This argument can only be used to xe2x80x98upcastxe2x80x99 the array.
|
|
For downcasting, use the .astype t) method.
|
|
But I may have missed some point in the docs","bug"
|
|
"6013","6551","matplotlib does not build with numpy master.
|
|
Something has changed in numpy.distutils that is causing matplotlib to not build.
|
|
I ve tried `git bisecting`, but the history seems a little convoluted and it isn t helping to find the issue.
|
|
.","bug"
|
|
"6013","6555","any all functions do not work with generators.
|
|
Numpy s any ) and all )) functions should work with generator expressions to be consistent with basic python.
|
|
I spent several hours trying to debug a case of this, which would work if I used a list, but not if I used a generator expression only to discover that it was because numpy s all ) function was being used instead of the usual python one that works identically with both generators and lists.
|
|
.","bug"
|
|
"6013","656","Runtime Error while twice usage Trac 2203).
|
|
Original ticket http projects.scipy.org numpy ticket 2203 on 2012 08 17 by trac user Lukasz16a, assigned to unknown.
|
|
The following simple code causes runtime error include <Python.h> include <numpy arrayobject.h> void f ) { import array ) ; } int main ) { for int i=0 ; i<2 ; i ) { Py Initialize ) ; f ) ; Py Finalize ) ; } } The problem is loading numpy.core...
|
|
I tested it on Visual Studio 2010 at Windows 7 Professional x64.","bug"
|
|
"6013","6563","f2py Intent out) broken in recent versions of numpy.
|
|
As of recently, the intent out) is no longer taken over into the signature file.
|
|
This concerns both the information from a Fortran90 variable declaration and the f2py hint with the syntax ..
|
|
The fib3.f example from the [f2py documentation] http docs.scipy.org doc numpy f2py getting started.html) is affected as well and can serve as a minimal example.
|
|
The reason seems to be that the function isintent out ) is not imported from auxfuncs.py into crackfortran.py .
|
|
Intent out) does hence pass not through the true intent list filter.","bug"
|
|
"6013","6564","Inconsistent indexing result depending on argument type.
|
|
The following results seem inconsistent to me numpy 1.9.2) <pre> >>> a = np.arange 8).reshape 2,2,2)) >>> a[ [[0], [0]] ] array [[0, 1]]) >>> a[ np.array [[0], [0]]) ] array [[[[0, 1], [2, 3]]], [[[0, 1], [2, 3]]]]) < pre> I ve briefly looked at the indexing manual http docs.scipy.org doc numpy reference arrays.indexing.html) and nothing seems to account for this.
|
|
Thanks, Lluis","bug"
|
|
"6013","6565","Document numpy.lib.stride tricks.as strided.
|
|
As far as I can tell `as strided` which can be rather useful, even though it is involved) is documented nowhere it doesn t even appear in the general index, probably because it isn t listed in `stride tricks.
|
|
all `.","bug"
|
|
"6013","6566","numpy 1.10 appears to break setting the fortran compiler when building scipy.
|
|
After moving to numpy 1.10.0 b2 in Fedora rawhide, we re seeing bad gfortran flags when building scipy Fortran f77 compiler usr bin gfortran Wall g ffixed form fno second underscore O2 g pipe Wall Werror=format security Wp, D FORTIFY SOURCE=2 fexceptions fstack protector strong param=ssp buffer size=4 grecord gcc switches specs= usr lib rpm redhat redhat hardened cc1 m32 march=i686 mtune=atom fasynchronous unwind tables fPIC O2 march=x86 64 DMS WIN64 mtune=generic msse2 Fortran f90 compiler usr bin gfortran Wall g fno second underscore O2 g pipe Wall Werror=format security Wp, D FORTIFY SOURCE=2 fexceptions fstack protector strong param=ssp buffer size=4 grecord gcc switches specs= usr lib rpm redhat redhat hardened cc1 m32 march=i686 mtune=atom fasynchronous unwind tables fPIC O2 march=x86 64 DMS WIN64 mtune=generic msse2 Fortran fix compiler usr bin gfortran Wall g ffixed form fno second underscore Wall g fno second underscore O2 g pipe Wall Werror=format security Wp, D FORTIFY SOURCE=2 fexceptions fstack protector strong param=ssp buffer size=4 grecord gcc switches specs= usr lib rpm redhat redhat hardened cc1 m32 march=i686 mtune=atom fasynchronous unwind tables fPIC O2 march=x86 64 DMS WIN64 mtune=generic msse2 It appears to be tacking on the march=x86 64 DMS WIN64 mtune=generic msse2 options from the mingw settings to the fortran compiler flags.","bug"
|
|
"6013","6569","Cannot run tests after `python setup.py build ext i`.
|
|
Here is what I mean .
|
|
The issue is pretty clear When `numpy init .py` is loaded, under [some conditions] https github.com numpy numpy blob master numpy init .py L159 L162) it will not set the ` version ` variable; these conditions occur when you do not use a `virtualenv` or `python setup.py install`, and just try and run `runtests.py`.
|
|
After 7b438fa90e53abe8b2f0356ec50daed6ab299794 [PR] https github.com numpy numpy pull 6461)), `nosetester.py` tries to access this ` version ` variable, which doesn t exist and so it crashes.
|
|
CC rgommers charris?","bug"
|
|
"6013","657","Errors obtained while running the NumPy test Trac 2205).
|
|
Original ticket http projects.scipy.org numpy ticket 2205 on 2012 08 21 by trac user alataza, assigned to unknown.
|
|
Hi!
|
|
I m a complete newbie to SciPy and NumPy.
|
|
I ve downloaded and installed these packages as I need them to run the Brian simulator.
|
|
I m using OSX 10.8 just installed from the scratch to be able to use the SciPy Superpack from Chris Fonnesbeck) with Fortan 2.7.2, XCode 4.4.1.
|
|
As per Fonnesbeck s reccomendation, I ve performed the test and several errors have arised.
|
|
I know that he s using development versions SciPy 0.12) and therefore may be bugs there, but I couldn t get Brian work just by installing the stable releases with my previous OSX version.
|
|
I m attaching the output generated by the test file in case it might help you in the development of NumPy.
|
|
Cheers!
|
|
Marc.","bug"
|
|
"6013","6571","test load refcount from numpy.lib.tests.test io.
|
|
The test test load refcount from numpy.lib.tests.test io seems to randomly fail even on the master branch.
|
|
.","bug"
|
|
"6013","6572","write 1.10.2 release notes..","bug"
|
|
"6013","6575","split produces empty arrays with wrong number of dimensions.
|
|
This line works on numpy 1.9.3 but errors on 1.10.1 current conda version) .
|
|
It s because the split call has an empty array with the wrong shape, as indicated in the FutureWarning that prints .
|
|
However, it looks like in 1.9.3 this wasn t the case .
|
|
This seems like a regression.
|
|
Is this intended behavior?
|
|
I m on OS X using conda and Python 2.7.10, and switching between versions with `conda install numpy` and `conda install numpy=1.9.3`.)","bug"
|
|
"6013","658","Floating point string conversions corrupted when using SSE2 Trac 2206).
|
|
Original ticket http projects.scipy.org numpy ticket 2206 on 2012 08 23 by trac user ekvamme, assigned to pv.
|
|
After running numpy.linalg functions, string conversions using str ) on some small floating point numbers are corrupted and return xe2x80x9cERRxe2x80x9d.
|
|
This bug occurs with all versions of Numpy win32 for Python2.7 with all releases of Python 2.7.
|
|
It only occurs with Numpy sse2 architecture enabled and does not occur with nosse or sse3 .
|
|
It was seen first on an older Pentum that only supported up to sse2 so it was the default Numpy installation architecture but the same failure is observed on any newer systems by using the arch sse2 argument when installing Numpy.
|
|
On an older system that has only up to sse2 support, the error is avoided by installing Numpy with the arch nosse argument.
|
|
A test script that demonstrates the error is as follows Demonstrate fail in numpy for Python 2.7 when installed with SSE2 support arch sse2) import numpy z = numpy.array [[15,0,672],[0,400,36],[0,0,1]]) beforestr = str 1e 12) numpy.linalg.inv z) This causes second str ) to return a different result afterstr = str 1e 12) if beforestr != afterstr print FAIL = , beforestr, afterstr else print PASS raw input Press any key to continue ) This has been shown to fail with the following versions Python 2.7.3 with Numpy 1.5.0, 1.5.1, 1.6.0, 1.6.1, 1.6.2, and 1.7.0b1 and Python 2.7 with Numpy 1.5.0 just to show all versions from the earliest releases of both Python 2.7 and Numpy for 2.7 fail).
|
|
This error does not occur with Numpy win32 for Python2.5.
|
|
The versions tried were Python 2.5.4 with Numpy 1.0.4, 1.1.0, 1.1.1, 1.2.0, 1.2.1, 1.3.0, 1.4.1, 1.5.0, 1.5.1, 1.6.0, 1.6.1, 1.6.2, and 1.7.0b1.
|
|
While the observation of the error has only been clearly observed on the string conversion of a float after running linalg, the concern is that other floating point computation problems may be generated by this but I have not discovered them yet.","bug"
|
|
"6013","6581","Memory leak when array contains circular references.
|
|
Memory is leaked when an array contains a circular reference .
|
|
This is because `PyArray` doesn t implement `tp traverse` for fairly reasonable reasons) but also leads to hard to track down memory leaks.
|
|
At first pass it seems reasonable to implement a `tp traverse` which only traverses if `dtype=object` but that does have a performance tradeoff.","enhancement"
|
|
"6013","6584","setting an array element with a sequence error could be improved..
|
|
This is a bit of a nitpick, but I think this would improve user friendlyness.
|
|
Currently ``np.asarray [[1, 2], [2, 3, 4]], dtype=np.float)`` yields setting an array element with a sequence.
|
|
I think it would be more helpful if it yielded something like Can t create array from lists.
|
|
Column length d of column d doesn t match column length d. I imagine the error bubbles up from somewhere deep in the code and makes perfect sense where it is raised.
|
|
It is just something that comes up a lot, and a more friendly and or informative message could be helpful.
|
|
So maybe we could catch it further up and include more information if possible.","enhancement"
|
|
"6013","6585","scipy.stats.ncx2 fails for nc=0.
|
|
For `nc = 0` it should reduce to the chi squared distribution, but `scipy.stats.ncx2.rvs 1, 0)` gives .
|
|
Similarly the pdf, etc.
|
|
also fail.","bug"
|
|
"6013","6589","DOC Is the where doc correct?.
|
|
Here http docs.scipy.org doc numpy reference generated numpy.where.html .
|
|
Do they actually need to have the same shape as condition?
|
|
Also from the docs .","bug"
|
|
"6013","659","SPAM ticket Trac 2207).
|
|
Original ticket http projects.scipy.org numpy ticket 2207 on 2012 08 24 by trac user wrick, assigned to unknown.
|
|
[contained spam redacted]","bug"
|
|
"6013","6590","Fortran Array problem in numpy 1.10.
|
|
Hi, I think I got a problem related to the numpy function np.asfortranarray.
|
|
I am using numpy 1.10 with a pacakge named SPAMS http spams devel.gforge.inria.fr ), it keep complaining that the input array is not a fortran array, but I used np.asfortranarray to convert it indeed.
|
|
Then I tried to downgrade to numpy 1.9.3, it worked quite well, so I think there is a bug introduced since 1.10.
|
|
To reproduce the problem .
|
|
To run the test, you need to cd spams python and them run the following code .
|
|
And if you are using numpy 1.10 then it will complain an error immediately, if you use numpy 1.9.3 then it will run the test.
|
|
Thanks in advance to look into this.
|
|
Best, Will","bug"
|
|
"6013","6591","inconsistencies in array construction from a list of ndarrays.
|
|
When looking at examples for https github.com numpy numpy issues 6584 issuecomment 152607473 I found the following weird or inconsistent behavior .
|
|
How much of this is intended, and if it is not intended then which open github issues are most relevant?","bug"
|
|
"6013","6597","Deprecate isfortran and a.fnc.
|
|
These are old functions that check if an array is Fortran contiguous and not C contiguous.
|
|
Their output depends on whether relaxed stride checking is in effect of not and they seem to have little use in current numpy.","bug"
|
|
"6013","6598","BUG `flatten` method does not validate its argument..
|
|
According to the docstring of the `flatten` method, the argument should be one of ` C `, ` F ` or ` A `.
|
|
I would expect a `ValueError` to be raised if the argument is not one of those values.
|
|
However, it appears to ignore invalid arguments.
|
|
For example, .
|
|
I noticed this while editing the NumPy for Matlab Users guide, where `y = x.flatten 1)` was suggested as the NumPy version of the Matlab code `y = x )`.
|
|
The guide has been fixed.)","bug"
|
|
"6013","6599","Install and packaging improvements todo list.
|
|
After the latest round of install pain and the recent thread stop supporting `` setup.py install ``; start requiring `` pip install .
|
|
`` instead , it s time to start fixing a few things.
|
|
Here s a list [ ] unconditionally import ``setuptools`` will fix gh 6551) [ ] error on ``python setup.py install``.
|
|
Provide a switch `` force`` for ``python setup.py install``.
|
|
Mailing list discussion [here] http article.gmane.org gmane.comp.python.numeric.general 61757) not going to fly for now, see [this email about issues] https mail.scipy.org pipermail numpy discussion 2015 November 074146.html) [ ] update all install docs [ ] make it even clearer in big bold letters in the docs that scientific distributions or things like Homebrew are preferred over building from source or even ``pip``.
|
|
[ ] error on ``python setup.py upload`` saying `` use `twine upload s` instead ``) [ ] error on ``python setup.py upload docs`` [ ] error on ``python setup.py easy install`` [ ] error on ``python setup.py test`` saying `` use `python runtests.py` instead ``) [ ] remove ``setupegg.py`` [ ] error on ``python setup.py clean`` saying `` use `git clean xdf` or Xdf ...) instead ``) [ ] change ``python setup.py help`` to first show numpy specific stuff before setuptools help info [ ] figure out a better way to test the various methods of installing numpy install from ``sdist`` install via ``bdist wheel`` ``easy install`` is still triggered via ``setup requires``).
|
|
Note, direct use of ``easy install`` by users is not supported meaning it may work but if it doesn t then don t expect support use ``pip`` instead).
|
|
the recommended ``pip install .`` in place builds ``pip install e``, ``build ext``, ``develop``) builds with `` config xxx build`` followed by ``pip install .`` with and without ``virtualenv`` can make a difference) what else?
|
|
[ ] when ``pip upgrade`` is released should be soon, maybe under a different name, see https github.com pypa pip pull 3194 and https github.com pypa pip issues 59), officially change our mind and recommend the use of ``install requires`` to packages depending on numpy.
|
|
[ ] figure out what to do about ``setup requires`` ``pip`` won t support that any time soon) [ ] fix up the metadata.
|
|
now ``sdist`` overwrites author with maintainer .
|
|
Consider using ``setup.cfg`` or a new static metadata format that may materialize soon.
|
|
[ ] do a check on version numbering correctness see [https github.com numpy numpy issues 6431 issuecomment 147112731) and make ``np.lib.NumpyVersion`` accessible in ``numpy.version`` For interested people needing a bit of context, the [Python Packaging User Guide] https packaging.python.org en latest ) is the authoritative doc about Python packaging in general.","bug"
|
|
"6013","660","chebval subtract half of first coefficient?
|
|
Trac 2208).
|
|
Original ticket http projects.scipy.org numpy ticket 2208 on 2012 08 28 by trac user smeared ink, assigned to charris.
|
|
I am very new to evaluating Chebyshev polynomials, but shouldn t chebval subtract off 0.5 first coefficient), as in Numerical Recipes, or here?
|
|
http mathworld.wolfram.com ChebyshevApproximationFormula.html which cites Numerical Recipes, actually).
|
|
I was finding unexpected differences in how numpy evaluated Chebyshev polynomials from some code I was trying to replicate, and I traced it back to this little thing.","bug"
|
|
"6013","6602","choice and dirichlet missing from numpy.random.
|
|
all .
|
|
Presumably numpy random info.py should include choice and dirichlet in the ` all ` variable?
|
|
Right now they aren t there.","bug"
|
|
"6013","6609","Building numpy on AIX with gcc and gfortran.
|
|
Hi, I have python 2.7.5 installed via binaries provided [here] http www.perzl.org aix index.php?n=main.python).
|
|
I also installed gcc and gfortran from the same site.
|
|
I got OpenBlas compiled and Cython compiled as well.
|
|
When I go to compile numpy, I get .
|
|
Why is it trying to use `xlc r` when we ve already specified `gcc` and `gfortran`?
|
|
Second question how can I link my compiled OpenBLAS for use with numpy?
|
|
Thanks.","bug"
|
|
"6013","661","Inconsistent dot product with BLAS Trac 2209).
|
|
Original ticket http projects.scipy.org numpy ticket 2209 on 2012 08 29 by trac user law, assigned to unknown.
|
|
Hello, I found some strang bug related to numpy dot product.
|
|
When using numpy sse2 or numpy sse3 I get different results for the same data and the same script.
|
|
The test script is as follows test script from numpy import a = [ 0.07738518658898276004, 0.96881291391416879399, 0.00769691931712315151, 0.08111387479193514494, 0.11885744586034743819, 0.05882337083773133551] a2 = [ \ [0, 0, 0, 0, 0, 0], \ [ 0.07738518658898276004, 0.96881291391416879399, 0.00769691931712315151, 0.08111387479193514494, 0.11885744586034743819, 0.05882337083773133551] \ ] b = [0.89887845358462115097, 0.08609082382649151099, 0.02124706755589483506, 0.22391191935591209239, 0.32810168298084929628, 0.16237978892078003268] set printoptions precision=20) ab = dot a, b) a2b = dot a2, b) ab i = inner array a), swapaxes array b), 1, 2)) a2b i = inner array a2), swapaxes array b), 1, 2)) r = 0 for i in range 6) r = r a[i] b[i] print dot .20f ab print dot2 .20f a2b[1] print inner .20f ab i print inner2 .20f a2b i[1] print sum prod .20f r And I get C \Documents and Settings\1>python c \temp\t0.py dot 0.08609082382649152487 dot2 0.08609082382649149712 inner 0.08609082382649152487 inner2 0.08609082382649149712 sum prod 0.08609082382649149712 C \Documents and Settings\1>python i c \temp\t0.py dot 0.08609082382649152487 dot2 0.08609082382649151099 inner 0.08609082382649152487 inner2 0.08609082382649151099 sum prod 0.08609082382649149712 >>> quit ) C \Documents and Settings\1>python c \temp\t0.py dot 0.08609082382649152487 dot2 0.08609082382649149712 inner 0.08609082382649152487 inner2 0.08609082382649149712 sum prod 0.08609082382649149712 C \Documents and Settings\1>python i c \temp\t0.py dot 0.08609082382649152487 dot2 0.08609082382649151099 inner 0.08609082382649152487 inner2 0.08609082382649151099 sum prod 0.08609082382649149712 >>> quit ) Surely, I don t change the script or anything else between runs.
|
|
It seems that the result depends on something, but I can t figure out what exactly.
|
|
In this particular example that difference caused by i switch, other time result changed when I add x or u, or just run script two time in a row.
|
|
I come to conclusion that this strange behavior related to dotblas, cause I tried to replace numpy sse3 with numpy 1.6.2 nosse numpy 1.7.0b1 nosse) the results are constant and match to simple sum of products C \Documents and Settings\1>python c \temp\t0.py dot 0.08609082382649149712 dot2 0.08609082382649149712 inner 0.08609082382649149712 inner2 0.08609082382649149712 sum prod 0.08609082382649149712 At the same time I see that random fluctuations with numpy 1.6.2 sse2, numpy 1.6.2 sse3, numpy 1.7.0b1 sse2, numpy 1.7.0b1 sse3 and on Windows XP, Windows 2008, Windows 7 tested on real and virtual machines AMD CPU)) I can perform other tests if it can be helpful.
|
|
Best regards, Alexander","bug"
|
|
"6013","6610","Overly restrictive error message in concatenate.
|
|
.
|
|
but in fact 1D arrays can be mixed with 2D arrays if the relevant dimensions match) .","bug"
|
|
"6013","6611","ma.dot no longer always returns a masked array in 1.10.
|
|
The function `numpy.ma.dot` in versions <1.10 always returned a masked array, but in 1.10 it does so only if one or both of the inputs are masked array.
|
|
I couldn t find this documented anywhere in the release notes, so perhaps it is a regression?
|
|
This broke some code in the wild that relied on `ma.dot` returning a masked array ajdawson eofs 34).
|
|
A simple code example, which returns a `numpy.ma.core.MaskedArray` on numpy <1.10 and a `numpy.ndarray` on numpy 1.10 .
|
|
It looks like the major change here was probably 5709.
|
|
Can someone confirm if this is intentional new behaviour or if this is a bug, charris abalkin?","bug"
|
|
"6013","6616","ndarray with buffer of integer 0 s and 64 bit dtype raises TypeError exception buffer is too small for requested array , but not float 0 s. Python v3.4.3 Numpy v1.10.1 Windows 7 Professional x64 np.ndarray shape= 2, 1), buffer=np.array [0, 0]), dtype=np.int64) => raises TypeError buffer is too small for requested array np.ndarray shape= 2, 1), buffer=np.array [0, 0]), dtype=np.float64) => raises TypeError buffer is too small for requested array But... np.ndarray shape= 2, 1), buffer=np.array [0.0, 0.0]), dtype=np.int64) => OK np.ndarray shape= 2, 1), buffer=np.array [0.0, 0.0]), dtype=np.float64) => OK","bug"
|
|
"6013","6617","Build numpy for Python3.5 get mathlib info raise RuntimeError Broken toolchain cannot link a simple C program ).
|
|
I installed python3.5 using the installer from python official website, then trying to install numpy, but get the following error I have Visual Studio 2015 installed).
|
|
.","bug"
|
|
"6013","6618","NPY FORTRANORDER in make fortran ) function, numpy.i.
|
|
Hi, in the process of converting numpy arrays to fortran contiguous ordering in a swig wrapper python< >armadillo), I noticed that the `make fortran` function in numpy does not seem to swap the memory layout of arrays.
|
|
In my use case I fixed the issue by substituting `NPY FORTRANORDER` with `NPY ARRAY F CONTIGUOUS` in the `PyArray FromArray` call.
|
|
I m not very familiar with the internal workings of numpy, but I thought I would share this in case it does represent an issue.
|
|
Best, Daniel","bug"
|
|
"6013","6619","Window 7 64 bit large int to float conversion lost precision on iterator passed as float .
|
|
Python 2.7.6 default, Nov 10 2013, 19 24 24) [MSC v.1500 64 bit AMD64)] Type copyright , credits or license for more information.
|
|
IPython 2.4.1 An enhanced Interactive Python.
|
|
?
|
|
> Introduction and overview of IPython s features.
|
|
quickref > Quick reference.
|
|
help > Python s own help system.
|
|
object?
|
|
> Details about object , use object??
|
|
for extra details.
|
|
guiref > A brief reference about the graphical user interface.
|
|
In [1] import numpy as np ... np.version.full version ... Out[1] 1.10.1 In [2] int32arr = np.array [39831850,39831850]) ... flt32arr = np.zeros 2), dtype = float32 ) ... flt32arr[ ] = int32arr ...
|
|
In [3] int32arr Out[3] array [39831850, 39831850]) In [4] flt32arr Out[4] array [ 39831848., 39831848.
|
|
], dtype=float32) 39831850 np.int goes to 39831848 as np.float32","bug"
|
|
"6013","662","Mixing regular IO with numpy.fromfile confuses file offset Trac 2210).
|
|
Original ticket http projects.scipy.org numpy ticket 2210 on 2012 09 04 by trac user allen ..., assigned to unknown.
|
|
I have a binary file which is written by a C program.
|
|
It is essentially a bunch of integer and single precision floating point values written one after another.
|
|
I m trying to read the file partly with the python `read )` function usually followed by `struct.unpack )`) and partly with `numpy.fromfile )`.
|
|
Generally, I m extracting the scalars with `read ) unpack ) ` and the arrays with `fromfile )`.
|
|
I ve discovered that `fromfile )` can become confused if the file itself is larger than a particular size.
|
|
On my Red Hat Enterprise Linux 6.3 and Ubuntu 12.04 64 bit systems, this size is 4096 bytes.
|
|
It appears to work OK on windows xp regardless of the file size.
|
|
I attached a simple program which writes a simple binary file and then reads it back.
|
|
It should produce the output offset0 125 [ 70.
|
|
65.
|
|
60.
|
|
55.
|
|
50.
|
|
45.
|
|
40.
|
|
35.
|
|
30.
|
|
25.
|
|
20.
|
|
15.
|
|
10.
|
|
5.
|
|
0.
|
|
5.
|
|
10.
|
|
15.
|
|
20.
|
|
25.
|
|
30.
|
|
35.
|
|
40.
|
|
45.
|
|
50.
|
|
55.
|
|
60.
|
|
65.
|
|
70.]
|
|
offset1 241 [ 80.
|
|
75.
|
|
70.
|
|
65.
|
|
60.
|
|
55.
|
|
50.
|
|
45.
|
|
40.
|
|
35.
|
|
30.
|
|
25.
|
|
20.
|
|
15.
|
|
10.
|
|
5.
|
|
0.
|
|
5.
|
|
10.
|
|
15.
|
|
20.
|
|
25.
|
|
30.
|
|
35.
|
|
40.
|
|
45.
|
|
50.
|
|
55.
|
|
60.]
|
|
offset2 357 [ 90.
|
|
85.
|
|
80.
|
|
75.
|
|
70.
|
|
65.
|
|
60.
|
|
55.
|
|
50.
|
|
45.
|
|
40.
|
|
35.
|
|
30.
|
|
25.
|
|
20.
|
|
15.
|
|
10.
|
|
5.
|
|
0.
|
|
5.
|
|
10.
|
|
15.
|
|
20.
|
|
25.
|
|
30.
|
|
35.
|
|
40.
|
|
45.
|
|
50.]
|
|
offset3 473 On linux I get offset0 125 [ 70.
|
|
65.
|
|
60.
|
|
55.
|
|
50.
|
|
45.
|
|
40.
|
|
35.
|
|
30.
|
|
25.
|
|
20.
|
|
15.
|
|
10.
|
|
5.
|
|
0.
|
|
5.
|
|
10.
|
|
15.
|
|
20.
|
|
25.
|
|
30.
|
|
35.
|
|
40.
|
|
45.
|
|
50.
|
|
55.
|
|
60.
|
|
65.
|
|
70.]
|
|
offset1 242 [ 1.78734834e 38 1.78698961e 38 1.78663088e 38 1.78627215e 38 1.78562643e 38 1.78490896e 38 1.78419150e 38 1.78347403e 38 1.78275657e 38 1.78203910e 38 1.78103465e 38 1.77959972e 38 1.77816479e 38 1.77644288e 38 1.77357302e 38 1.76898124e 38 0.00000000e 00 5.93486894e 39 5.98078669e 39 6.00948528e 39 6.02670444e 39 6.04105373e 39 6.05540303e 39 6.06544754e 39 6.07262218e 39 6.07979683e 39 6.08697148e 39 6.09414613e 39 6.10132078e 39] offset2 358 [ 1.78806581e 38 1.78770708e 38 1.78734834e 38 1.78698961e 38 1.78663088e 38 1.78627215e 38 1.78562643e 38 1.78490896e 38 1.78419150e 38 1.78347403e 38 1.78275657e 38 1.78203910e 38 1.78103465e 38 1.77959972e 38 1.77816479e 38 1.77644288e 38 1.77357302e 38 1.76898124e 38 0.00000000e 00 5.93486894e 39 5.98078669e 39 6.00948528e 39 6.02670444e 39 6.04105373e 39 6.05540303e 39 6.06544754e 39 6.07262218e 39 6.07979683e 39 6.08697148e 39] offset3 474 You see that the first array is read OK, but the file `offset1` following the `fromfile )` call is incorrect.
|
|
It should be 241, but is 242 instead.
|
|
I glanced at the C code which implements `fromfile )` but I didn t see anything obviously incorrect.
|
|
Except, it does make a copy of the underlying file handle to do the `fromfile )`.
|
|
I wondered if this was exposing a bug in GLIBC or the python file handling layer.
|
|
Thanks, Allen","bug"
|
|
"6013","6620","Interest in percentile function that matches behavior of Matlab IDL?.
|
|
Matlab uses linear interpolation based on the rank of the data see [Algorithm part of the Matlab documentation] http de.mathworks.com help stats prctile.html).
|
|
This algorithm is also used in some IDL libraries and can not be replicated using any of the interpolation schemes in the `numpy.percentile` function, see example at the bottom of this post.
|
|
It can easily be implemented as .
|
|
Example of differences between `numpy` and `matlab` .","bug"
|
|
"6013","6622","Compatibility problem with rpy2.
|
|
numpy appears to be overwriting the C level ``R Home`` from R when it gets imported, affecting rpy2 compatibility.
|
|
For example .
|
|
Importing numpy first and then importing rpy2 does not result in the error.
|
|
I ve reported this on rpy2, and the author seems to think this is likely numpy s fault https bitbucket.org rpy2 rpy2 issues 311 compatibility problem with numpy and pdf Version info .","bug"
|
|
"6013","6624","f2py does not correctly parse selected real kind precision 1.0d0)).
|
|
Consider the dummy declarations below which are valid Fortran.
|
|
Both bar1 and bar2 are of double precision, however the C code generated by f2py assumes that bar2 is typed NPY FLOAT which results to writes into foreign memory from the compiled Fortran module.
|
|
If you think it is a too seldomly employed form, please consider to make f2py refuse automatically guessing bar2 s type and require editing the .pyf file.
|
|
This is up to version 1.10.1 .","enhancement"
|
|
"6013","6625","numpy.percentile does not accept sequence of floats.
|
|
In numpy 1.8.2 numpy.percentile does not accept sequence of floats as it should, according to the documentation.
|
|
This is because it compares `q` with scalars, whereas `q` can be either a scalar or a sequence .","bug"
|
|
"6013","6629","Some functions in numeric.py call python all on ndarrays.. That is not efficient.
|
|
Unfortunately, the places where all is used can mix arrays and scalars, so the array method is not appropriate.
|
|
However, np.all could be imported from numpy or fromnumeric.","bug"
|
|
"6013","663","Alignment doesn t work for array of structs Trac 2211).
|
|
Original ticket http projects.scipy.org numpy ticket 2211 on 2012 09 05 by trac user Ihor.Melnyk, assigned to unknown.
|
|
array of subtype alignment t = np.dtype [ a , |i1 ), b , [ f0 , <i2 ), f1 , <f4 )], 2)], align=True) actual dtype [ a , |i1 ), b , [ f0 , <i2 ), f1 , <f4 )], 2,))] expected dtype [ a , |i1 ), , |V3 ), b , [ f0 , <i2 ), , |V2 ), f1 , <f4 )], 2)]","bug"
|
|
"6013","6631","Request currying `numpy.einsum` for repeated calls with the same subscript string?.
|
|
Hi, During my experience with the `numpy.einsum` function, I discovered that I frequently use it in the following pattern The zeroth argument to `einsum )`, the string describing the subscripts, is a fixed one, but the array arguments will vary during runtime.
|
|
That is to say, I apply the same summation rule to many different array arguments repeatedly.
|
|
I don t know if `einsum )` keeps its subscript parsing results in a cache, but if not, this could be a bit wasteful, because the same subscript description need to be unnecessarily parsed again and again.
|
|
So I wonder if it is possible to have a curriable version of `einsum` that, when called with the subscript description string, returns a callable that has already parsed the string.
|
|
This returned callable is then ready to be applied to the array arguments.
|
|
On the other hand, I m not sure about the performance impact of parsing.
|
|
Is this kind of optimization worth the effort?
|
|
Just my 2c.","bug"
|
|
"6013","6634","Problems encountered when trying to upgrade sagemath to numpy 1.10.1.
|
|
I have been looking on and off at moving sagemath beyond numpy 1.8.x and numpy 1.10.1 behaves better for us than the 1.9 series [ see http trac.sagemath.org ticket 17642].
|
|
However we have a strange conversion problem within numpy s `linspace` function .
|
|
It appears that `start` and `stop` are not converted to float by the multiplication by `1.`.
|
|
The initial suggestion in issue 3504 to use `float` to cast the value to the right types would work here.
|
|
Of course in that context `dt` becomes trivially `float`.","bug"
|
|
"6013","6635","No new ver.
|
|
of NumPy bin for Python32 to download .
|
|
There is no NumPy for Python 3.2.
|
|
Last ver.
|
|
was about 1.8.1 which I use now.","bug"
|
|
"6013","6636","memory leak in nested dtypes in numpy.recarray.
|
|
The following code snippet causes the peak memory to increase indefinitly.
|
|
I managed to trace it down to the dtype used in the recarray.
|
|
Basically for a nested dtype such as [ a , [ b , i )])] there is a memory leak.
|
|
.","bug"
|
|
"6013","6637","uint32 hash 2 32)) throws OverflowError in win amd64 builds, but nowhere else?.
|
|
Based on errors I ve seen from users, I believe the following code will reliably trigger an OverflowError Python int too large to convert to C long ) for people using win amd64 Numpy builds such as the unofficial ones at http www.lfd.uci.edu gohlke pythonlibs numpy) numpy.uint32 hash 2 32)) But, I believe this line will work without error on most other 32 bit or 64 bit pythons.
|
|
Anywhere the `hash )` value is more than 32 bits, `uint32 )` will still work to truncate xe2x80x93 except for these win amd64 builds.)
|
|
This discrepancy seems non ideal, and perhaps against the intended type casting functionality of `uint32 )`.
|
|
I know this may not be a core Numpy concern, but perhaps a win amd64 build expert who can resolve the issue is here, or maybe adding a test that this code does not throw an error) could prompt downstream builders to see and resolve the issue.","bug"
|
|
"6013","6638","Improve random variate generation of noncentral F distribution.
|
|
The noncentral F distribution has three parameters `dfnum`, degrees of freedom in numerator `dfden`, degrees of freedom in denominator `nonc`, noncentrality According to [http docs.scipy.org doc numpy reference generated numpy.random.noncentral f.html] http docs.scipy.org doc numpy reference generated numpy.random.noncentral f.html), the current implementation unnecessarily requires both `dfnum` and `dfden` be greater than 1, while they are allowed to be any positive value in theory.
|
|
This restriction stems from the fact that the implementation calls `random.noncentral chisquare`, which unnecessarily required that its `dof` argument must be greater than 1 and has been fixed in `numpy 1.10`; see 5766.
|
|
Since the implementation of `random.noncentral f` is straightforward, something like .
|
|
according to [https github.com numpy numpy blob master numpy random mtrand distributions.c L256 L260 ] https github.com numpy numpy blob master numpy random mtrand distributions.c L256 L260 ), this should be an easy fix.","bug"
|
|
"6013","6639","builtin isnan, builtin isinf, and bultin isinfinite broken on OS X with Intel compiler version 16.0.0.
|
|
Numpy uses builtin isnan, builtin isinf, and bultin isinfinite when available but these are not implemented correctly by version 16.0.0 of the Intel compiler installed on OS X 10.10.
|
|
This means that `numpy.isinf float inf ))` returns `False`, for instance.
|
|
Here is a minimal test case against the Intel compiler .
|
|
which when compiled with `icpc fp model strict test.cpp o test` prints 1 0 1 0 0 1 Here are some more details of my installation icpc v icpc version 16.0.0 gcc version 4.9.0 compatibility) uname a Darwin QuantumWises MacBook Pro.local 14.5.0 Darwin Kernel Version 14.5.0 Tue Sep 1 21 23 09 PDT 2015; root xnu 2782.50.1 1 RELEASE X86 64 x86 64","bug"
|
|
"6013","664","Add the 24 bit support in numpy.fromstring Trac 2212).
|
|
Original ticket http projects.scipy.org numpy ticket 2212 on 2012 09 05 by trac user bred, assigned to unknown.
|
|
In some real world applications, like audio or image processing should be possible that data in organized in words of 24 bit.
|
|
So would be useful to have an efficient 24bit string decoder.
|
|
Actually the situation is a bit depressing, for example if we read a 24bit audio file, we generate a string that is very difficult and slow to be reeded.
|
|
For example the code should look like ... import wave wave read = wave.open file name , r ) sample width = wave read.getsampwidth ) sample width == 3 !!!!
|
|
X = wave read.readframes wave read.getnframes ) ) Y = numpy.fromstring X, dtype= int24 ).reshape nframes , 2 ) If the sample width is 3 24bit) this very simple procedure don t work.
|
|
Tnx","enhancement"
|
|
"6013","6640","Python 3.5, building numpy master with MSVC 14.0 or visual studio 2015).
|
|
Hello, I tried building the numpy master or latest for python 3.5.
|
|
The reason behind it, because I wish to build it with open blas, on windows...
|
|
I know.
|
|
Even if I wish to just build it with said MSVC 14.0 I get the following compile error numpy\core\src\multiarray\mapping.c 1296) error C2036 void unknown size I apologize if I wasted your time over nothing and if the following snippet of code is pasted amateurly.
|
|
To just paste the snippet of the code from numpy\core\src\multiarray\mapping.c , where the problem occurs ... npy intp offset; ... view = PyArrayObject )PyArray NewFromDescr Py TYPE arr), fieldtype, PyArray NDIM arr), PyArray SHAPE arr), PyArray STRIDES arr), PyArray DATA arr) offset, < Here PyArray FLAGS arr), PyObject )arr); ... From the Data Type API page on docs.scipy.org I read that the npy intp type is a The enumeration value for a signed integer type which is the same size as a void ) pointer.
|
|
This is the type used by all arrays of indices.
|
|
I don t have any C C experience and I only found that the MSVC has problems over pointer type void.
|
|
Also because of that I tried using the casting char ) as this is supposedly the usual way to sort this out with MSVC.
|
|
And it didn t work, even though I somehow knew it wouldn t. Any thoughts on how to deal with this?","bug"
|
|
"6013","6641","Subsetting recarray by fields yields a structured array.
|
|
I would expect this to return a record array as well .","bug"
|
|
"6013","6646","ma.asarray does not default to C order as stated..
|
|
In fact, the `order` keyword is ignored.
|
|
.","bug"
|
|
"6013","6647","ma `a.filled ...)` does not return a copy as stated in the documentation.. .","bug"
|
|
"6013","665","Deadlock happening when using numpy scipy with python subinterpreters Trac 2213).
|
|
Original ticket http projects.scipy.org numpy ticket 2213 on 2012 09 12 by trac user raphael.de.feraudy, assigned to unknown.
|
|
Hello, I ve encountered a deadlock while trying to import a scipy subpackage using Python C API in my application.
|
|
I ve managed to build a minimal example showing this, which I join to this ticket.
|
|
According to what I ve found on the internet, I guess this is caused by the documented) unsupported use of both PyGilState functions in numpy code and Subinterpreters in my code As you will see in my example, the simple use of a new subinterpreter instead of the main default one is enough to produce the deadlock).
|
|
I use an ubuntu 12.04 Precise Pangolin 64bits distribution with standard python numpy 1.6.1) and python scipy 0.9.0) package , but the deadlock still happens with last revisions of git repositories.
|
|
I tried to fix this by myself, only using PyEval functions in a working copy, but didn t succeed.
|
|
My questions are Are you aware of this issue ?
|
|
Do you plan to address it in next milestone ?
|
|
Do you have any workaround to suggest ?","enhancement"
|
|
"6013","6654","f2py intent in,out) appears broken in NumPy 1.10.1.
|
|
Returning scalars using `intent in,out)` that was working up throught NumPy 1.9.x now returns `None` with NumPy 1.10.1.
|
|
Also, I had to switch to `intent in out)` no comma) to get the passed value modified but that still didn t get it to return anything.
|
|
Here is a stripped down example .
|
|
The problem was observed running NumPy 1.10.1 on 64 bit Windows 10 using 64 bit Python 2.7.10 and the NumPy package from http www.lfd.uci.edu gohlke pythonlibs.
|
|
My Linux systems don t have 1.10 yet.","bug"
|
|
"6013","6657","CPU loop in histogram on Mac OS 10.10.5 14F1021), python 2.7.
|
|
Dear Numpy gurus, Here is a link to download a zipped directory containing a notebook and some associated data.
|
|
Running the notebook using the file VVD990305K0002.dat works well, while the VVD990305K0001.dat yields a CPU loop.
|
|
Without a Timeout protection, my Mac Pro s RAM saturates, and finally the computer crashes in panic.
|
|
https dl.dropboxusercontent.com u 2021308 loop numpy.zip The stack is matching with 2MASS, checking calibration KeyboardInterrupt Traceback most recent call last) <ipython input 2 db69fe1dc1a6> in <module> ) 41 ax = plt.axes ) 42 > 43 hist dist match, bins= knuth , ax=ax, histtype= stepfilled , ec= k , fc= AAAAAA ) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages astroML plotting hist tools.pyc in hist x, bins, range, args, kwargs) 64 bins = bayesian blocks x) 65 elif bins in [ knuth , knuths ] > 66 dx, bins = knuth bin width x, True, disp=False) 67 elif bins in [ scott , scotts ] 68 dx, bins = scotts bin width x, True) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages astroML density estimation histtools.pyc in knuth bin width data, return bins, disp) 228 dx0, bins0 = freedman bin width data, True) 229 M0 = len bins0) 1 > 230 M = optimize.fmin knuthF, len bins0), disp=disp)[0] 231 bins = knuthF.bins M) 232 dx = bins[1] bins[0] opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages scipy optimize optimize.pyc in fmin func, x0, args, xtol, ftol, maxiter, maxfun, full output, disp, retall, callback) 375 return all retall} 376 > 377 res = minimize neldermead func, x0, args, callback=callback, opts) 378 if full output 379 retlist = res[ x ], res[ fun ], res[ nit ], res[ nfev ], res[ status ] opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages scipy optimize optimize.pyc in minimize neldermead func, x0, args, callback, xtol, ftol, maxiter, maxfev, disp, return all, unknown options) 461 xbar = numpy.add.reduce sim[ 1], 0) N 462 xr = 1 rho) xbar rho sim[ 1] > 463 fxr = func xr) 464 doshrink = 0 465 opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages scipy optimize optimize.pyc in function wrapper wrapper args) 283 def function wrapper wrapper args) 284 ncalls[0] = 1 > 285 return function wrapper args args)) 286 287 return ncalls, function wrapper opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages astroML density estimation histtools.pyc in call self, M) 153 154 def call self, M) > 155 return self.eval M) 156 157 def eval self, M) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages astroML density estimation histtools.pyc in eval self, M) 175 176 bins = self.bins M) > 177 nk, bins = np.histogram self.data, bins) 178 179 return self.n np.log M) opt local Library Frameworks Python.framework Versions 2.7 lib python2.7 site packages numpy lib function base.pyc in histogram a, bins, range, normed, weights, density) 263 for i in arange 0, len a), BLOCK) 264 sa = sort a[i i BLOCK]) > 265 n = np.r [sa.searchsorted bins[ 1], left ), 266 sa.searchsorted bins[ 1], right )] 267 else Cheers, JB","bug"
|
|
"6013","6658","numpy.ndarray.newbyteorder much slower in numpy1.10.
|
|
Recently christophreimer noticed that one of our dataset readers got noticeably slower in a new virtual environment.
|
|
This could be tracked down to calls of `numpy.ndarray.newbyteorder`.
|
|
I m running the anaconda builds for numpy 1.10.1 and numpy 1.9.3 for comparison numpy 1.9.3 .
|
|
numpy 1.10.1 .","bug"
|
|
"6013","666","impossible to define ufunc with builtin type input and user defined output Trac 2214).
|
|
Original ticket http projects.scipy.org numpy ticket 2214 on 2012 09 13 by fengy research, assigned to unknown.
|
|
My use case is 1. define a ufunc myufunc takes one in and one out.
|
|
2. define a user defined type mydtype 3. myufunc, when taking int64 as input, will give mydtype as output.
|
|
This is impossible to achieve for two reasons 1. with PyUFunc FromFuncAndData ... char types, ...) the user defined types can not be supplied to char types.
|
|
Furthermore, the types element in PyUFuncObject is a char .
|
|
2. with PyUFunc RegisterLoopForType, we can register a signature with int64 as input and mydtype and output in the userloops dictionary, with mydtype as the key in userloops.
|
|
However when we try to invoke the ufunc, an error is raised ufunc s not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule s , This is because the type resolver only looks for the dtypes in input operands, but the input operand is int64 and is never going to be found in userloops.
|
|
All in all, it becomes impossible to make use of myfunc.","bug"
|
|
"6013","6661","Bug ma.testutils.assert equal on nans throws AssertionError.
|
|
`np.ma.testutils.assert equal np.nan, np.nan)` throws an AssertionError.
|
|
In my opinion it should not consistent with the behavior of `np.testing.assert equal np.nan, np.nan)`.","bug"
|
|
"6013","6667","ma.make mask handles ma.nomask input incorrectly..
|
|
This was a nasty little bug to chase down.
|
|
I suspect `ma.mask or` is not alone in this.
|
|
.
|
|
Whereas .","bug"
|
|
"6013","6669","Problems with Intel MKL.
|
|
I ve been trying to get numpy working with Intel s MKL math library without much success.
|
|
The first problem was convincing numpy s build process to use MKL even though I have openblas installed on my system The problem is in `distutils system info.py` checks for the existence of openblas before checking whether MKL is configured.
|
|
Checking MKL first resolved that issue I can submit a patch if you re interested).
|
|
That leads to the second problem.
|
|
When linked with MKL, numpy fails several of it s tests.
|
|
Specifically .
|
|
I don t even know where to begin to isolate the problem.
|
|
Has anyone reported this before?
|
|
What other information would be helpful?
|
|
I m using numpy 1.10.1, python 3.5.0, gcc 4.8.3, MKL 11.3.","bug"
|
|
"6013","667","regression from 1.5.x) with return of operations on memmap ed arrays Trac 2215).
|
|
Original ticket http projects.scipy.org numpy ticket 2215 on 2012 09 14 by yarikoptic, assigned to unknown.
|
|
As http projects.scipy.org numpy ticket 2179 stated Ideally one should be returning other ndarray objects, e.g.
|
|
numpy scalars in the case of sum ), but this requires a larger rewrite.
|
|
At least numpy version 1.5.1rc1 did this.
|
|
The only time the mmap should be kept as I am aware of) is when slicing.
|
|
This issue remains not fixed thus opening a separate issue for the above.
|
|
Taking into account that np.asscalar ) called on scalars e.g.
|
|
which would be returned by np.sum on ndarrays) crashes, it becomes impossible to write code agnostic of ndarray memmap container types.
|
|
it is also not clear why array wide operations on mode= c memmap ed arrays which cannot really get reflected in the file storage, remain memmapped and do not result in normal ndarrays, but I guess that would be a separate issue...","bug"
|
|
"6013","6670","Typo at scipy manual.
|
|
At this page http docs.scipy.org doc numpy reference arrays.indexing.html It says Be sure to understand why this is occurs.
|
|
I think it should say Be sure to understand why this occurs.","bug"
|
|
"6013","6672","memory leak in np.clip ).
|
|
I found a usage pattern under which np.clip ) leaks memory.
|
|
You can reproduce it with the following code In [1] import numpy as np In [2] a = np.arange 10).reshape 5,2) In [3] a Out[3] array [[0, 1], [2, 3], [4, 5], [6, 7], [8, 9]]) In [4] b = np.array [[3,5],[0,6]]) In [5] np.clip a, b[ ,0], b[ ,1]) Out[5] array [[3, 1], [3, 3], [4, 5], [5, 6], [5, 6]]) In [6] for i in xrange 10000000) np.clip a, b[ ,0], b[ ,1], a) This problem was already present in numpy 1.9.0 installed from `numpy 1.9.0 win32 superpack python2.7.exe`), and I was now able to reproduce it in numpy 1.10.1 installed from Christoph Gohlke s [`numpy 1.10.1 mkl cp27 none win32.whl`] http www.lfd.uci.edu gohlke pythonlibs numpy) Windows binary), as well as in Linux numpy 1.10.1, local compile through pip).","bug"
|
|
"6013","6673","Unconsistent inplace division due to buffer size.
|
|
Numpy 1.8.2 with python 2.7.6 has unconsistent behaviour when doing inplace division of arrays exceeding the buffer size.
|
|
import numpy as np arr = np.random.rand 2, 5000) arr copy = arr.copy ) arr copy = arr copy arr copy[0] arr = arr[0] print np.sum arr != arr copy), arr.size np.sum np.isclose arr, arr copy)) Should print 0 for equal divisions.
|
|
This can be circumvented by increasing the buffersize, nevertheless it should be emphasized, that this can happen.","bug"
|
|
"6013","6675","optimized blas detection broken in master and 1.10. merge of gh 6183 broke cblas detection on linux distros .
|
|
cblas is required for numpy to use it in `dot`","bug"
|
|
"6013","6677","mmap slicing performance regression again).
|
|
mmap slicing has regressed in performance again, see the mmap slicing benchmark 1.8ms > 5ms only in master, maybe the overlap check?","bug"
|
|
"6013","6678","getting unexpected error from X.dtype = complex or Y = X.view complex)).
|
|
The following works with numpy 1.9.3 and python 3.4.3, but in numpy 1.10.1 and python 3.5.0 I get `ValueError new type not compatible with array.` on the `b.dtype = complex` line import numpy as np A = np.array [[1., 2.
|
|
], [3., 4.]])
|
|
A.dtype = complex this works a = np.array [[1., 2.]])
|
|
a.dtype = complex this works B = np.array [[1., 2.
|
|
], [3., 4.
|
|
]], order= F ) B.dtype = complex this works b = np.array [[1.
|
|
], [2.
|
|
]], order= F ) b.dtype = complex < In 1.10.1, gives `ValueError new type not compatible with array.` print A = , A) print a = , a) print B = , B) print b = , b) The output for the older version is as it should be I think) A = [[ 1.
|
|
2.j] [ 3.
|
|
4.j]] a = [[ 1.
|
|
2.j]] B = [[ 1.
|
|
3.j 2.
|
|
4.j]] b = [[ 1.
|
|
2.j]] Note that the 1.10.1 gave the same results for `A`, `a`, and `B`.
|
|
My current workaround is to flatten `b`, set the dtype, and reshape.
|
|
Ixe2x80x99ve seen this same behavior on Linux and on Windows.","bug"
|
|
"6013","668","Memory leak when an errors are raised ie.
|
|
comparison of array and empty list) Trac 2216).
|
|
Original ticket http projects.scipy.org numpy ticket 2216 on 2012 09 17 by trac user Hauke, assigned to unknown.
|
|
on 64bit Ubuntu 10.04, virtualenv 1.7.2, Python 2.7.3, Numpy 1.6.2, this import numpy as np r = np.random.rand 2) l = [] while True r == l causes growing memory usage reproducibly.
|
|
Could not be reproduced on Python 2.6.6, NumPy 1.4.1.
|
|
Also posted in http stackoverflow.com questions 12461413 why does comparison of a numpy array with a list consume so much memory Regards, Hauke Holtkamp","bug"
|
|
"6013","6680","view doesn t understand tuples.
|
|
When trying to convert an `n`x`2` array of floats into an `n`x`1` array of `2` tuples, .
|
|
NumPy bails out with .","bug"
|
|
"6013","6682","DOC Cython is not listed as build requirement.
|
|
Couldn t find it anywhere mentioned in the docs, but it seems that Cython is a build requirement of the development version.","bug"
|
|
"6013","6683","1.10.2 strides error.
|
|
I get this funny result with current master >>> np.empty 1,1), dtype= float64 ).strides 9223372036854775807, 9223372036854775807) In 1.10.1 on my system it correctly says ` 8,8)`.
|
|
Can anyone confirm this?
|
|
I initially suspected build issues on my side, but it seems to happen with a fresh build.
|
|
also, oops hit enter by accident before commenting.
|
|
This is an edit)","bug"
|
|
"6013","6685","Incorrect result for np.rint on Win32, numpy 1.10.1, MKL.
|
|
Sorry for the specificity of this message, but I m hoping it s the right place to file this.
|
|
With 32 bit numpy MKL on Windows .
|
|
This is true of conda numpy packages and Christoph s wheels though these seem to have disappeared now).","bug"
|
|
"6013","6687","numpy docs on data types do not seem to be comprehensive.
|
|
Is there a reason for some available data types like `np.float128` not being listed here?
|
|
https github.com numpy numpy blob master numpy doc basics.py","bug"
|
|
"6013","669","1d array can t be accessed with a tuple.
|
|
Trac 2217).
|
|
Original ticket http projects.scipy.org numpy ticket 2217 on 2012 09 18 by trac user laket, assigned to unknown.
|
|
Using a tuple to access elements of 1d array, I got IndexError, >>>c = numpy.array [1,2,3]) >>>c[ 0,1)] IndexError invalid index When I accessed 2d array with a tuple, I got expected result.
|
|
>>> c = numpy.array [[1,2],[3,4]]) >>> c[ , 0,1)] array [[1, 2], [3, 4]]) Python 2.7.3 Numpy 1.6.1 OS Ubuntu 12.04","bug"
|
|
"6013","6691","multiarray object has no attribute may share memory from recent commit.
|
|
I m getting an error on master Python 2.7.10) from >commit 8efc87ec599c0b3eac4e63bea6eda9023d8ed96d >Date Thu Nov 12 20 15 37 2015 0200 > > ENH reimplement may share memory in C to improve its performance .","bug"
|
|
"6013","6693","ifft RAM requirements.
|
|
Hi, I was trying to reduce the RAM requirementes of my program and noticed that the numpy.fft.ifft function needs a lot more RAM for the computation than numpy.fft.fft.
|
|
I don t understand why, as the formulas for fft and ifft are nearly the same... My numpy version is 1.8.2. .
|
|
I look at the RAM consumption by executing `top d.05 | grep ipython` in a second terminal window.","bug"
|
|
"6013","6694","lapack lite missing in sdist.
|
|
since 1e436a5a","bug"
|
|
"6013","670","numpy.any ignores imaginary part of arrays Trac 2218).
|
|
Original ticket http projects.scipy.org numpy ticket 2218 on 2012 09 25 by trac user mike.wimmer, assigned to unknown.
|
|
In numpy versions 1.6.1, 1.6.2 and the latest git version those I tried), numpy.any ignores the imaginary part of arrays completely >>> numpy.version.version 1.8.0.dev 6a847ef >>> numpy.any [1j]) False >>> numpy.any numpy.array [1j])) False >>> numpy.any numpy.array [0.1j])) False whereas >>> numpy.any numpy.array [1 1j])) True >>> numpy.any numpy.array [1 0j])) True Now, since >>> bool 1j) True I think this is a genuine bug.
|
|
Note that in numpy 1.5.1 all of the above returned True, i.e.
|
|
the imaginary part was taken into account then.","bug"
|
|
"6013","6701","Failure when installing Numpy dev on Windows with MSVC 2008 C .
|
|
I am trying to install the latest developer version of Numpy on AppVeyor Windows) with MSVC 2008 C and Python 2.7, and after compiling many files, I run into the following error .
|
|
The full log can be found here https ci.appveyor.com api buildjobs t0bimapv9p66yoxs log","bug"
|
|
"6013","6704","mkl lapack32 and mkl lapack64 are obsolete.
|
|
With newer versions of the intel compiler and its corresponding mkl library, the mkl lapack32 and mkl lapack64 flags are now obsolete.
|
|
See for example https mail.scipy.org pipermail scipy user 2012 July 032503.html The following patch will fix it, however it may break compatibility with older versions of intel compiler and mkl.
|
|
Since people who use intel compiler and mkl library are performance sensitive, and since intel compiler v16.0.0 improves performance noticeably compared to old versions, I claim that not being a big issue.
|
|
The following patch will fix it diff wu numpy distutils system info.py.orig numpy distutils system info.py numpy distutils system info.py.orig 2015 11 19 16 24 31.253412000 0700 numpy distutils system info.py 2015 11 19 16 27 46.243694265 0700 1039,7 1039,7 lapack libs = self.get libs lapack libs , [ mkl lapack ]) else lapack libs = self.get libs lapack libs , [ mkl lapack32 , mkl lapack64 ]) []) info = { libraries lapack libs} dict append info, mkl) Tested for intel v16.0.0 and MKL v11.3.0","bug"
|
|
"6013","6708","array constructor accepts instance of dtype instead of dtype for the dtype kwarg.
|
|
.
|
|
Same with `empty`, `zeros`, etc.
|
|
This is IMO a good place for a bug to hide.
|
|
Or perhaps there has always been this behavior?
|
|
Python 3.5, numpy 1.10.1.","bug"
|
|
"6013","6709","numpydoc convention when input can be an instance of a class.
|
|
For numpydoc What is the convention for parameter type if a parameter is an instance of specific class, e.g.
|
|
datetime.datetime?
|
|
Specifically, I have a parameter which can be either a float, datetime.date, or datetime.datetime.
|
|
How should I write this in the parameter section?
|
|
my param float or date or datetime my param float or datetime.date or datetime.datetime my param float or class \`datetime.date\` or class \`datetime.datetime\` my param float or class \` datetime.date\` or class \` datetime.datetime\` something else?
|
|
And of course, maybe a comma instead of the first or , but that s nitpicking which I like by the way) and not too important in this context unless I m mistaken).","bug"
|
|
"6013","671","Basic array like function like sin; cos ... are too slow with numpy MKL Trac 2219).
|
|
Original ticket http projects.scipy.org numpy ticket 2219 on 2012 09 28 by trac user bred, assigned to unknown.
|
|
If we perform a sin function on a very big array the computation is very slow respect to the real possibility offered by MKL.
|
|
Example x = np.random.rand 1e7) y = np.sin x) is too slow !!!
|
|
and it s not parallel!
|
|
I ve seen that mkl offers the functions for solving these operation in a performant manner.
|
|
Can you implement these functions in munpy in a better way?","enhancement"
|
|
"6013","6710","Breaking np.nan to num ) changes from 1.9.2 to 1.10.1.
|
|
This issue was originally filed with Pandas https github.com pydata pandas issues 11380) Previously when using Pandas 0.16.2 and Numpy 1.9.2 one could run various numpy funcs on a pandas Series and have the result returned back typed as a Series instead of an NDarray).
|
|
Example with numpy 1.9.2) .
|
|
renders the following output `dfn type <class pandas.core.series.Series >` However, when using Numpy 1.10.1, the result is now typed as an NDarray.
|
|
So far I ve only tested np.nan to num but perhaps there are other funcs that exhibit the same behaviour.
|
|
As advised by jreback, I am filing this as a numpy issue.","bug"
|
|
"6013","6711","docs for datetime units show incorrect time spans.
|
|
This can be seen in http docs.scipy.org doc numpy reference arrays.datetime.html datetime units The correct time spans should be .
|
|
rather than .","bug"
|
|
"6013","6712","arccosh is inaccurate when numpy is built with old glibc?.
|
|
An arccosh test is failing in one of the TravisCI scipy configurations that builds numpy.
|
|
I m speculating that this is because TravisCI runs a somewhat old Ubuntu with a glibc that has some trig function implementations that numpy has blacklisted.
|
|
If that is true, could this be fixed in a way that is better than bumping the scipy test tolerances or upgrading the TravisCI platform?
|
|
https github.com scipy scipy pull 4729","bug"
|
|
"6013","6718","f2py test fail in pip installed numpy 1.10.1 in virtualenv.
|
|
running `python c import numpy ; numpy.test ); ` i get the following error about test f2py .","bug"
|
|
"6013","6719","Building from git fails `Error compiling Cython file Pythonic division not allowed without gil`.
|
|
Building numpy from git fails when Cythonizing `numpy random mtrand mtrand.pyx`, due to an error `Pythonic division not allowed without gil, consider using cython.cdivision True)`.
|
|
This is Python 3.5.0, Cython 0.23.4, with `numpy` git revision `d94043f0fb93985fd4302eae9fcdca2d79239908` on branch `master`.
|
|
The same problem occurs on branch `maintenance 1.10.x` with commit `76db1ab4edef69e310426dccf95c15212b1ba4a6`, and on branch `maintenance 1.9.x` with commit `edb902cdc6573553afcf11047ecdfb447e444322`.
|
|
However, a regular `pip install numpy` which installs 1.10.1), works because releases include compiled cython files, but version control does not).
|
|
The failures occur as follows .
|
|
What s going wrong?","bug"
|
|
"6013","672","silently ignored size mismatch bug??)
|
|
Trac 2220).
|
|
Original ticket http projects.scipy.org numpy ticket 2220 on 2012 10 01 by nbecker, assigned to unknown.
|
|
n [19] u = np.arange 10) In [20] v = np.arange 10) In [21] u[v] = u In [22] u[v] = np.arange 11) silence...","bug"
|
|
"6013","6722","Testing Single element NumPy Array Against Empty Python List.
|
|
When I test a single element NumPy array against a regular Python list, `np.array [1]) != []`, the result evaluates to neither true nor false, while one of multiple elements evaluates to true.
|
|
E.g.
|
|
boo is printed with the following .
|
|
Using the interpreter .
|
|
This seems strange.
|
|
Was this intended?
|
|
Tested with Python 2.7.10 and 3.4.2, using Numpy 1.10.1.
|
|
Linux 4.2.6 200.fc22.x86 64 Fedora 22)","bug"
|
|
"6013","6723","Indexing multidimensional structured masked array fails to translate fill value correctly, leading to broadcasting errors ValueError when calling .filled ).
|
|
When using a structured masked array where one of the structure elements has a multidimensional dtype, indexing does not correctly set the `fill value` attribute for the new masked array.
|
|
This leads to a `fill value` with a shape that is incompatible with `.data` or `.mask`, which inevitably leads to broadcasting problems when `fill value` is used, such as when calling `filled )` .
|
|
Indeed, the fill value doesn t change upon indexing the masked array only its dtype does .","bug"
|
|
"6013","6724","ValueError when indexing numpy.ma.core.mvoid with multidimensional structured dtype.
|
|
When indexing a structured masked array with a multidimensional structured dtype, resulting in a `numpy.ma.core.mvoid`, an attempt at accessing a field results in a `ValueError`, as illustrated below on the current numpy repo .
|
|
The problem does not occur when the structured dtype has a scalar dimension .","bug"
|
|
"6013","6729","String representation of exotic masked structured arrays with ndim=0 fails with ValueError.
|
|
Even after the fixes in pull request 6094, there are still some cases where string representation fails, in particular where some members of a structured array are multi dimensional .
|
|
And without the `m = m.view ...)` line, `m.any )` would fail with `TypeError cannot perform reduce with flexible type`.
|
|
A bit more of an overview, illustrating what we would like to see .","bug"
|
|
"6013","673","numpy 1.6.2 fails to compile with python 3.3.0 xe2x80x98PyUnicodeObjectxe2x80x99 has no member named xe2x80x98strxe2x80x99 Trac 2221).
|
|
Original ticket http projects.scipy.org numpy ticket 2221 on 2012 10 03 by trac user tru, assigned to unknown.
|
|
numpy 1.6.2 python3 setup.py build fcompiler=gnu95 without site.cfg) fails to build git version as of 2012 10 03 commit e18e7441700db0ff2fd8f51901aa416c63e35cbc) compiles properly.
|
|
no issue with numpy 1.6.2 with python 3.2.3","bug"
|
|
"6013","6732","High memory peak when MaskedArray creates a mask.
|
|
Hi, We work with huge masked arrays, and noticed that there is a high memory peak when creating a masked array with `mask=False`.
|
|
In the example below you can see that when using `MaskedArray mask=np.ma.nomask ...)` the memory does not increase and the masked array use a view of the data array.
|
|
But when using `MaskedArray mask=False ...)` there is a memory peak of 15 times the size of the boolean mask which is created !
|
|
.
|
|
The issue comes from the line `mask = np.resize mask, data.shape)` in https github.com numpy numpy blob master numpy ma core.py L2770, where mask is just `array [False], dtype=bool)`.
|
|
And then `np.resize` call `concatenate a,) n copies)` which causes the memory peak https github.com numpy numpy blob master numpy core fromnumeric.py L1149).
|
|
Without knowing the reasons of this implementation, I wonder why the mask is not created simply with a `np.zeros dtype=bool, ...)` or `np.ones` depending the value of the mask parameter).","bug"
|
|
"6013","6739","performace issue when accessing large structured arrays.
|
|
numpy supports to create own `dtype` definitions.
|
|
I used this in order to store data together with some meta informations.
|
|
Irealized, that accessing the data then becomes VERY slow takes seconds!).
|
|
Since I couldnt find any warnings about large dtype definitions it is not clear to me if this is really the inteded behaviour .
|
|
on my laptop the 3rd line takes almost 4 seconds !
|
|
), although its just accessing a single scalar element.
|
|
Is this a bug or just a shortcoming, i could find the documentation of?
|
|
The jupyter notebook can be found here https gist.github.com skuschel 4d2085c3c377a0f833a0 numpy 1.10.1 on python3 or python2 with arch linux) Thanks a lot Stephan","bug"
|
|
"6013","674","a[ ] = broadcastedarray is wrong Trac 76).
|
|
Original ticket http projects.scipy.org numpy ticket 76 on 2006 04 21 by trac user tim.hochberg ..., assigned to trac user sasha.
|
|
>>> a = zeros [5,2]) >>> a[ ] = arange 5) >>> a array [[0, 1], [2, 3], [4, 0], [1, 2], [3, 4]])","enhancement"
|
|
"6013","6740","Faulty resize of empty record array.
|
|
It seems it is not possible to resize an empty record array while it is ok for regular array.
|
|
Not sure which solution is right but it might be better to have a consistent behavior.
|
|
.
|
|
on OSX El Capitan, python 3.5.0, numpy 1.10.1","bug"
|
|
"6013","6741","ctypeslib.as array aborts.
|
|
This script fails .
|
|
python 3.4.2 dbg numpy==1.10.1 Ubuntu 14.10 1.
|
|
Under release python this code runs without crashes, because of disabled assertions 2.
|
|
Under debug release python asserts failed `python3.4 ..
|
|
Objects abstract.c 2077 PyObject Call Assertion result != void )0) !PyErr Occurred )) || result == void )0) PyErr Occurred )) failed.` Can you please explain why?
|
|
P.S.
|
|
may be it can help If `sh = 1, 1)` change to `sh = 239` 1.
|
|
Release python will raise `TypeError shape must be a tuple` 2.
|
|
Debug python will also aborts","bug"
|
|
"6013","6743","recfunctions.merge arrays fails to merge some arrays with standard field names.
|
|
The following .
|
|
yields .
|
|
The problem is that merge arrays essentially flattens an array if it has only one field, which could lead to conflicts in some cases, e.g.
|
|
two single field arrays with identical field names second array has one field f0 while first array has more than one field first array has one field f1 while second array has more than one field numpy 1.8.2 on ubuntu 15.10)","bug"
|
|
"6013","6744","numpy irr should check whether the masked solution is empty.
|
|
In the current implementation, when no solution is returned by roots, the function return np.nan.
|
|
But if no real solution is returned, the function fails with an exception ValueError attempt to get argmin of an empty sequence The proposed fix is to add the following to https github.com numpy numpy blob master numpy lib financial.py L654 if res.size == 0 or res[mask].size == 0","bug"
|
|
"6013","6745","The arguments of randint ) and random integers ) do not broadcast..
|
|
This is either a request for an enhancement, or a bug report take your pick.)
|
|
Unlike most of the other distributions implemented in the `RandomState` class, `randint` and `random integers` do not broadcast their arguments.
|
|
For example, `hypergeometric` broadcasts .
|
|
But `randint` only accepts scalar integers .","bug"
|
|
"6013","6746","nan in float64 does not evaluate to nan.
|
|
Hello It is very likely that this is a nooby misunderstanding from my part.
|
|
But shouldn t `np.float64 np.nan) is np.nan` evaluate as `True` on Python3)?
|
|
These two do at least `np.isnan np.float64 np.nan))` and `np.float np.nan) is np.nan` Thanks, Niko","bug"
|
|
"6013","6749","polyval syntaxes don t match.
|
|
`numpy.polyval coeffs, x)` versus `numpy.polynomial.polynomial.polyval x, coeffs)` Is there any reason these don t match?
|
|
Given that these have both been around for a while, is there any sensible way to unify these?","bug"
|
|
"6013","675","NumPy SWIG typemaps don t compile with Visual Studio compiler Trac 77).
|
|
Original ticket http projects.scipy.org numpy ticket 77 on 2006 04 23 by alberts, assigned to unknown.
|
|
The SWIG typemaps included in `numpy numpy doc swig numpy.i` don t compile with the Visual Studio compiler.
|
|
The following diff fixes the problem Index numpy.i =================================================================== numpy.i revision 2392) numpy.i working copy) 290,8 290,8 define TYPEMAP IN1 type,typecode) typemap in) type IN ARRAY1, int DIM1) PyArrayObject array=NULL, int is new object) { int size[1] = { 1}; array = obj to array contiguous allow conversion input, typecode, is new object); int size[1] = { 1}; if !array || !require dimensions array,1) || !require size array,size,1)) SWIG fail; 1 = type ) array >data; 2 = array >dimensions[0]; 318,8 318,8 define TYPEMAP IN2 type,typecode) typemap in) type IN ARRAY2, int DIM1, int DIM2) PyArrayObject array=NULL, int is new object) { int size[2] = { 1, 1}; array = obj to array contiguous allow conversion input, typecode, is new object); int size[2] = { 1, 1}; if !array || !require dimensions array,2) || !require size array,size,1)) SWIG fail; 1 = type ) array >data; 2 = array >dimensions[0]; Diff made against revision 2392.","bug"
|
|
"6013","6750","np.mean with keepdims=True does not work on memmap.
|
|
I m using numpy 1.9.2 built locally with pip and linked to openblas) and there seems to be a strange bug with memmap.
|
|
I checked through recent issues pr and nothing seemed reported I don t feel like building the master to check this out as I tend to stay on the same version of everything to compare results between themselves).
|
|
Anyway .
|
|
So the memmap doesn t play well with the keepdims=True, but the numpy array version works as expected.","bug"
|
|
"6013","6751","error in documentation numpy.diff).
|
|
From the documentation on this page http docs.scipy.org doc numpy reference generated numpy.diff.html This example is given array [[2, 3, 4], [5, 1, 2]]) >>> np.diff x, axis=0) array [[ 1, 2, 0, 2]]) The correct answer is array [[ 3, 2, 2]]).","bug"
|
|
"6013","6752","Add `ndim` attribute to dtypes.
|
|
`ndarray`s have `shape` and `ndim` attributes.
|
|
`X.ndim` is short for `len X.shape)`.
|
|
`numpy.dtype`s have `.shape`, but no `.ndim`.
|
|
For consistency, numpy.dtype types should have an `ndim` attribute that behaves identically to the ndarray counterpart, such that `dt.ndim == len dt.shape)`.","bug"
|
|
"6013","6753","No warning for integer overflow in dot.
|
|
I just noticed that this .
|
|
spits out no warning about integer overflow.
|
|
Is it possible to generate one?","bug"
|
|
"6013","676","Allow objects in record arrays Trac 78).
|
|
Original ticket http projects.scipy.org numpy ticket 78 on 2006 04 25 by teoliphant, assigned to teoliphant.
|
|
It would be very nice to allow objects in record arrays.
|
|
To support this all of the void data type functions would probably need to be recursively defined like getitem and setitem are now).
|
|
Also every special check in the code for object arrays would be needed for void arrays with objects in them.
|
|
PyArray INCREF would need to be adapted to handle them and so would PyArray DECREF.
|
|
This is not absolutely essential, but it would be nice by 1.0 just in case fundamental changes are necessary to allow it.","enhancement"
|
|
"6013","6760","Masked arrays with nested structured multidimensional dtypes inconsistent between data and mask.
|
|
This may or may not be caused by the same problem as other `numpy.ma` related issues.
|
|
When a dtype contains a nested structured multidimensional sub dtype, the dtype of the mask is inconsistent with the dtype of the data .
|
|
For somewhat simpler dtypes, this problem does not occur .
|
|
This is bound to lead to problems somewhere .
|
|
It does appear to lead to a failure when calling `Z.filled`, for example .
|
|
), although it is also possible that the `dtype` I constructed wreaks havoc elsewhere as well.","bug"
|
|
"6013","6762","Numpy hangs forever when matching ufunc types for custom dtype.
|
|
I m using python 2.7 on numpy 1.10.1 with a custom dtype implementation.
|
|
Essentially the self >userloops is defined with types that do not match the specified types.
|
|
In such case numpy will hang indefinitely.
|
|
It seems like the function type tuple userloop type resolver in numpy numpy core src umath ufunc type resolution.c is incorrect.
|
|
It can get into an infinite loop around line 1745, which looks like .
|
|
I see that if matched = 0, then the loop does a continue without updating the funcdata.
|
|
Thus funcdata is never null and will run forever.","bug"
|
|
"6013","6764","Broken tests on Python 3.6 dev nightly) module inspect has no attribute getargspec .
|
|
While trying to run the continuous integration of numpy with `python nightly` in travis I found the following test failures caused by the use of a introspection API deprecated since Python 3.4 if I am not mistaken .","bug"
|
|
"6013","6765","MaskedArray attributes itemsize and nbytes should include mask.
|
|
Currently, the `MaskedArray` attributes `itemsize` and `nbytes` count only the size taken up by the data.
|
|
This affects the following attributes `itemsize` `nbytes` `strides`?
|
|
... others?
|
|
It would be more correct to add up the size by the data and the size by the mask.","bug"
|
|
"6013","6766","numpy.load OSError first argument must be an open file.
|
|
Using Python 3.5 and Numpy 1.10.1 I got an error at `numpy.load` when called on an opened file object.
|
|
.
|
|
The error is thrown at numpy lib format.py , line 649, in read array array = numpy.fromfile fp, dtype=dtype, count=count) OSError first argument must be an open file I checked with pdb The file was closed somewhere inside the `numpy.load` call.
|
|
Unfortunately, the error does not always occur.
|
|
I repeat loading the file many times in my script, the error occurs normally after several hundred loadings.
|
|
Does anyone have an explanation of why it is sometimes closed?","bug"
|
|
"6013","6767","Documentation np.fill diagonal val scalar implies scalar only?.
|
|
In the docs for `np.fill diagonal` http docs.scipy.org doc numpy reference generated numpy.fill diagonal.html), we have .
|
|
Which to my reading says that it is scalars only .
|
|
Turns out that the main assignment in `np.fill diagonal` https github.com numpy numpy blob v1.10.1 numpy lib index tricks.py L776) .
|
|
will take any object that can broadcast compatible.
|
|
So, something like val must be shape compatible with the diagonal of a would make sense.
|
|
If this is worth a fix, I can put a PR together for it.","bug"
|
|
"6013","6769","minor doc typos titles for r , c , s , ix miss the final underscore.
|
|
See e.g.
|
|
http docs.scipy.org doc numpy reference generated numpy.ix .html I mean the big title, not the one that shows the call signature.","bug"
|
|
"6013","677","Valgrind warnings for unicode changes Trac 79).
|
|
Original ticket http projects.scipy.org numpy ticket 79 on 2006 04 25 by stefanv, assigned to unknown.
|
|
Changes in r2405 Fix Unicode swapping problem) generates warnings under Valgrind.
|
|
It looks like an invalid memory write is made during from numpy import ulen = 1 ucs value = u \U0010FFFF ua = array [[[ucs value ulen] 2] 3] 4, dtype= U s ulen) ua2 = ua.newbyteorder ) The valgrind output is ==22099== Invalid write of size 1 ==22099== at 0x401D408 memcpy mac replace strmem.c 394) ==22099== by 0x4902A96 UNICODE copyswap arraytypes.inc 10216) ==22099== by 0x48FB166 PyArray Scalar arrayobject.c 1093) ==22099== by 0x48FB367 PyArray Return arrayobject.c 1130) ==22099== by 0x491338E array getfield arraymethods.c 291) ==22099== by 0x805943B PyObject Call in usr bin python2.4) ==22099== by 0x80B0422 PyEval CallObjectWithKeywords in usr bin python2.4) ==22099== Address 0x453B3FC is 0 bytes after a block of size 12 alloc d ==22099== at 0x401B422 malloc vg replace malloc.c 149) ==22099== by 0x48FB11D PyArray Scalar arrayobject.c 1023) ==22099== by 0x48FB367 PyArray Return arrayobject.c 1130) ==22099== by 0x491338E array getfield arraymethods.c 291) ==22099== by 0x805943B PyObject Call in usr bin python2.4) ==22099== by 0x80B0422 PyEval CallObjectWithKeywords in usr bin python2.4) ==22099== by 0x8059412 PyObject CallObject in usr bin python2.4)","bug"
|
|
"6013","6771","Numpy.rec.fromarrays losing dtype metadata between versions 1.9.2 and 1.10.1. in 1.9.2 if I run the following assume dtype has metadata associated with it) rtn = np.rec.fromarrays arrays, dtype=dtype, names=names) For some reason the dtype metadata is lost in the line above.
|
|
rtn.dtype = dtype the metadata is preserved even those fromarrays lost it, at least I can re set it).
|
|
In the latest version of numpy setting rtn.dtype to a dtype with metadata no longer works.
|
|
The metadata is lost.
|
|
Not sure if this is a bug or an intentional change.","bug"
|
|
"6013","6773","Benchmarks currently broken.
|
|
While trying to improve on 5514, I noticed that as of HEAD the benchmark suite is currently broken.
|
|
.
|
|
whether with asv 0.1.1 or HEAD) Perhaps more importantly, perhaps travis should be set to run the benchmark suite too just to be sure it runs)?","bug"
|
|
"6013","6777","Old style Sphinx C domain markup no longer working.
|
|
I am using Sphinx 1.3.1 to build the docs and getting a ton of errors like so .
|
|
It appears that these have been deprecated in favor of `c func`, etc.
|
|
for some time.
|
|
http sphinx doc.org ext oldcmarkup.html","bug"
|
|
"6013","678","Add more releases of the GIL Trac 80).
|
|
Original ticket http projects.scipy.org numpy ticket 80 on 2006 04 26 by teoliphant, assigned to unknown.
|
|
More releases of the Global Interpreter Lock) GIL need to be peppered throughout the code.
|
|
Whenever a loop is occurring that will make no calls to the Python C API, then the GIL can be released.
|
|
All significant loops should be considered.
|
|
Especially the fromfile and tofile code needs to be checked for release of the GIL.","enhancement"
|
|
"6013","6781","The travis ci script in maintenance 1.10.x needs fixing..
|
|
The recent travis move to google infrastruction breaks testing for maintenance 1.10.x.
|
|
The test script in master seems to work fine.","bug"
|
|
"6013","6784","FutureWarning elementwise comparison failed makes it hard to check whether a value exists in an array.
|
|
I have innumerable places in my code where I test whether a value exists in an array when I don t know either the type of the value or the dtype of the array.
|
|
In previous numpy versions I m at 1.10.1), mismatched types return False.
|
|
Now they generate a FutureWarning.
|
|
.
|
|
The same applies when comparing arrays elementwise .
|
|
I know the new behavior is deliberate, but I think practicality has been sacrificed for purity.
|
|
In the example above, it seems reasonable that .
|
|
returns False, because being a string type automatically means x cannot belong to an array of a different type.
|
|
It also makes sense that two arrays are not equal if they have different dtypes.
|
|
So how do I repair my code?
|
|
Do I add type checking for all value checks?
|
|
Do I convert the array to a list first?
|
|
Do I convert to an object dtype?
|
|
Do I use Try Except?
|
|
I don t like any of these options and am having trouble thinking of a clean fix.
|
|
For what it s worth, I view this change as a serious regression.","bug"
|
|
"6013","679","Numpy should be installable with setuptool s easy install Trac 81).
|
|
Original ticket http projects.scipy.org numpy ticket 81 on 2006 04 26 by dmcooke, assigned to dmcooke.
|
|
Currently, with the 0.9.6 release, easy install run as `easy install numpy` will find the right source file for 0.9.6 on sourceforge, but the setup.py fails.
|
|
When testing to get this working, this is useful easy install f http svn.scipy.org svn numpy trunk egg=numpy dev numpy==dev to grab the version out of subversion, or make an sdist .tar.gz, and put it on a local webserver.
|
|
I m going to try to get this working before the 0.9.8 release.","bug"
|
|
"6013","6790","dtype keyword for np.random.randint and np.random.rand.
|
|
As brought up at the mailinglist, this code ates up much memory spectrumArray = np.random.randint 0,255, 2 20,2 12)).astype np.uint8) > It has also crossed my mind that np.random.randint and np.random.rand could use an extra dtype keyword.
|
|
It didn t look easy to implement though.
|
|
See this thread on the mailinglist [[Numpy discussion] array of random numbers fails to construct] https mail.scipy.org pipermail numpy discussion 2015 December 074306.html)","enhancement"
|
|
"6013","6792","Doc typo in random.rand docstring replace propagate with populate .
|
|
I just noticed this in passing, after doing `help np.random.rand)` .
|
|
I suspect that propagate should be populate here.","bug"
|
|
"6013","6793","corrcoef[1, 0] != corrcoef[0, 1].
|
|
On Python 3.5 from Anaconda) with numpy 1.10.1, the following test fails import numpy as np import numpy.testing as npt np.random.seed 100) Only happens sometimes!
|
|
t = np.arange 1024) x = np.sin 10 t) np.random.rand t.shape[ 1]) y = np.sin 10 t) np.random.rand t.shape[ 1]) cc = np.corrcoef x, y) npt.assert equal cc[0, 1], cc[1, 0]) Is this by design?","bug"
|
|
"6013","6794","np.dot a, a.T) should be detected and dispatched to the BLAS syrk routine.
|
|
BLAS provides a specialized routine for the common `A A.T` operation.
|
|
We should detect this case same data pointer, reversed strides) and dispatch to `syrk` instead of `gemm`.
|
|
[see] http www.netlib.org lapack lug node145.html)) In addition to providing a nice speedup, this should presumably avoid issues like 6793, where `dot A, A.T)` returned a non exactly symmetric matrix and this caused confusion.","enhancement"
|
|
"6013","6796","Delinking Numpy from BLAS ATLAS etc.
|
|
.
|
|
Hello, I m trying to benchmark sparse vs. dense routines on equal footing.
|
|
Numpy inherently uses BLAS and other low level routines when doing matrix operations, significantly improving performance, which is great.
|
|
However, it s not an apples to apples comparison.
|
|
Is there a way to change the numpy config so that it only uses C C Fortran routines, so that it s comparable to scipy.sparse?
|
|
Thank you!","enhancement"
|
|
"6013","6798","Doc comment mistake in ifft.
|
|
https github.com numpy numpy blob master numpy fft fftpack.py L205 The indices for the pos to neg frequency overflows are wrong here.
|
|
They are described correctly in http docs.scipy.org doc numpy 1.10.0 reference routines.fft.html implementation details","bug"
|
|
"6013","680","Compiler warnings in arrayobject.h Trac 82).
|
|
Original ticket http projects.scipy.org numpy ticket 82 on 2006 04 26 by alberts, assigned to unknown.
|
|
GCC 4.0.2 says usr lib python2.4 site packages numpy core include numpy arrayobject.h 237 warning comma at end of enumerator list usr lib python2.4 site packages numpy core include numpy arrayobject.h 243 warning comma at end of enumerator list Patch Index arrayobject.h =================================================================== arrayobject.h revision 2414) arrayobject.h working copy) 233,13 233,13 PyArray INTNEG SCALAR=3, PyArray FLOAT SCALAR=4, PyArray COMPLEX SCALAR=5, PyArray OBJECT SCALAR=6, PyArray OBJECT SCALAR=6 } PyArray SCALARKIND; typedef enum { PyArray ANYORDER= 1, PyArray CORDER=0, PyArray FORTRANORDER=1, PyArray FORTRANORDER=1 } PyArray ORDER;","bug"
|
|
"6013","6800","rk beta in random lack of robust.
|
|
Even if the beta distribution of the parameters shouldn t < 0, function rk beta in distribution.c should check the arguments","bug"
|
|
"6013","6802","numpy.ndarray.
|
|
repr shorts incorrect inconsistent dtype for unions.
|
|
For a structured `ndarray` with a union style dtype such as constructed with ` base dtype, new dtype)`), the `dtype` is incorrectly shown by `ndarray.
|
|
repr `, as illustrated below .
|
|
Somehow both are correct as it has multiple dtypes if I understand things correctly I m new to unions), but it looks wrong to me to suggest that the dtype of `M[ A ]` is just `uint32`.","bug"
|
|
"6013","6803","isnan and isinf on field with union dtype mixing float and int bool fails with SystemError Objects dictobject.c 2172 bad argument to internal function.
|
|
Under some circumstances that I have not yet succeeded to reproduce across different machines configurations For a structured ndarray `M` with a field `A` with union dtype ` base dtype, new dtype)`, where `base dtype` derives from `numpy.floating` and `new dtype` derives from `numpy.integer` or `numpy.bool `, `isnan` and `isinf` on `M[ A ]` sometimes fails with `SystemError Objects dictobject.c 2172 bad argument to internal function`, as illustrated here .
|
|
It appears not to happen always, as it has happened that I retried the same code with some minutes in between and code that failed before succeeded on a second try, or code that failed first succeeded later.
|
|
.
|
|
Among other places, this leads to the same failure in ` str ` and ` repr `.
|
|
I have reproduced the failure on one machine 48 core, 20 TB RAM, mostly somewhat old libraries) using a self compiled Python 3.5.0 or self compiled 3.4.3 and numpy 1.9.2 or 1.11.0.dev0 fb0a35f.
|
|
I could not reproduce it on the same or other machines mostly normal computers) using binary installations.","bug"
|
|
"6013","6804","Segmentation Fault for large array as input to bincount ).
|
|
The following code produces a segmentation fault in my configuration .
|
|
platform linux 64 numpy version 1.93","bug"
|
|
"6013","6805","Reference leak in np.bincount with weights attribute.
|
|
This code should illustrate what s going on .","bug"
|
|
"6013","6806","setting entry in struct array with object field.
|
|
.
|
|
numpy 1.10.1 First reported by ldoddema in 6776.","bug"
|
|
"6013","6807","Windows testing errors for 1.10.2.
|
|
Christoph reports on 1.10.x On Python 2.x, 32 and 64 bit .
|
|
On Python 2.x, 32 bit .
|
|
The second looks like a promotion of `long` to `float`, not sure about the first.
|
|
Can someone running windows track these down?
|
|
I don t think they are serious, but it would be good to know where they come from.
|
|
Note that in the first test `assert ` is not used.","bug"
|
|
"6013","6809","f2py overwrites compilation flags if FFLAGS is set.
|
|
If the environment variable FFLAGS is set, f2py replaces its compilations flags by these, which leads to failures if they are incomplete.
|
|
Distutils behaviour is to append C XX)FLAGS instead, so it keeps the required flags.
|
|
My use case is setting ` march=` to get optimal performance in locally built packages, but f2py fails if the ` fPIC` flag is missing.","bug"
|
|
"6013","681","Proposed patch to return scalars from matrix methods with axis=None Trac 83).
|
|
Original ticket http projects.scipy.org numpy ticket 83 on 2006 04 27 by edschofield, assigned to unknown.
|
|
As of SVN revision 2431, matrix objects preserve orientation in methods that take axis arguments for axis=0 and axis=1.
|
|
For axis=None, they return a matrix of shape 1, 1).
|
|
For example >>> x = asmatrix random.uniform 0,1, 3,3))) >>> x.std ) matrix [[ 0.26890557]]) >>> x.argmax ) matrix [[4]]) I believe this behaviour is unfortunate, and that an operation aggregating a matrix over all dimensions should return a scalar.
|
|
Arguments for reducing dimensionality upon aggregation have been presented in several numpy discussion threads, most recently in [Broadcasting rules Ticket 76)].
|
|
This patch modifies this behaviour to return scalars as rank 0 arrays) instead.
|
|
It also removes some code duplication.
|
|
The behaviour with the patch is >>> x.std ) 0.29610630190701492 >>> x.std ).shape ) >>> x.argmax ) 3 A potential drawback to this patch is that matrix operations aggregating along all dimensions, which would now share the behaviour of numpy arrays, would be no longer be consistent with matrix operations that aggregate along only one dimension, which currently return matrices with height or width 1.
|
|
This could be an argument for introducing a new vector class to represent one dimensional data with orientation.
|
|
Matrix objects, by contrast, are inherently 2 d.","enhancement"
|
|
"6013","6810","ENV versus site.cfg for LAPACK.
|
|
WIth numpy 1.10.1 and reference netlib) lapack installed to .
|
|
Everything seems good .
|
|
However, trying to recreate that with a simple site.cfg file fails I ve also tried various permutations of library dirs and others to no avail) .
|
|
I think a hint to the issue is in the working output.txt file above when it says `Replacing lib names[0]== lapack with reflapack `.
|
|
My guess is that for some reason deep in numpy distutils system info.py) the LAPACK env variable causes the library name to be changed, but the [lapack] section does not.","bug"
|
|
"6013","6811","frombuffer segfault.
|
|
First noted in 6776 the following line segfaults python .
|
|
It is not completely clear how to handle this or whether this should be handled at all).
|
|
A suggestion I have no idea of whether this is enough or too much in practice) is to return a read only buffer when accessing the `data` attribute of object or struct including object) arrays as frombuffer respects this flag).","bug"
|
|
"6013","6812","random.randint failure with large int64s on windows.
|
|
clearly using platform ints internally.
|
|
But I think this should upcast in the case of overflow?
|
|
.","enhancement"
|
|
"6013","6818","expm1 buggy on some platforms.
|
|
As we found over in astropy astropy 4275, some implementations of `expm1` incorrectly return `NaN` instead of `INF` on large inputs it still correctly raises an overflow error, but returns an incorrect result).
|
|
The Numpy implementation of this function doesn t have this problem as far as I can tell.
|
|
However, per https sourceware.org bugzilla show bug.cgi?id=6814 some versions of glibc and, it seems, MSVCRT have this problem.
|
|
Numpy uses ` HAVE EXPM1` to determine if this function is provided by the system.
|
|
But it doesn t have a way to check if the implementation is buggy or not.
|
|
Numpy uses [`config.check func`] https github.com numpy numpy blob v1.10.1 numpy distutils command config.py) to check if a function is available and can be called.
|
|
But I couldn t see if there was any existing way to test the return value of a function, so that we can test it for bugs, etc.
|
|
That might be nice to have and I will provide an implementation if there s agreement on that.
|
|
Should Numpy be overriding buggy platform versions of some functions that are otherwise available?","bug"
|
|
"6013","682","Tests for matrix returns from matrix methods Trac 84).
|
|
Original ticket http projects.scipy.org numpy ticket 84 on 2006 04 27 by alberts, assigned to teoliphant.
|
|
Check that matrix s instance methods return a matrix instance where expected.
|
|
As of r2414, clip, std, var and imag fail this check.","enhancement"
|
|
"6013","683","Code generators should use line directives Trac 85).
|
|
Original ticket http projects.scipy.org numpy ticket 85 on 2006 04 28 by trac user sasha, assigned to unknown.
|
|
In order to make compilers and debuggers point to the original source rather than auto generated code, line preprocessor directives should be included in the generated files.","enhancement"
|
|
"6013","6833","Upload OSX wheels to pypi before release announcement.
|
|
I suggest that we build and upload OSX wheels for numpy releases, before announcing, to minimize the number of OSX users who accidentally pip install from source after the announcement.
|
|
I guess it s implied but not emphasized in the HOWTO RELEASE notes.
|
|
Is that policy acceptable?
|
|
Should I add something more to the release notes?","bug"
|
|
"6013","6835","Strange casting with integers.
|
|
So, sometimes `int`s get casted to `float`s.
|
|
It seems a bit unexpected.
|
|
There are some other weird cases too.
|
|
>>> import numpy >>> i = numpy.uint32 7) >>> i 7 >>> i 1 8 >>> i = numpy.uint64 7) >>> i As expected.
|
|
7 >>> i 1 Casts to float?!
|
|
8.0 >>> i = numpy.int64 7) >>> i 7 >>> i 1 This certainly doesn t cast to float.
|
|
8","bug"
|
|
"6013","6838","Use runtests.py for travis testing..
|
|
The proposal is to use `runtests.py` as a replacement for the current script `tools test installed numpy.py`.
|
|
I think that using `runtests.py` would take care of the virtual environment setup and generally clean up the testing scripts.","enhancement"
|
|
"6013","6839","NumPy for MATLAB users mathematical term abuse.
|
|
This article https docs.scipy.org doc numpy dev user numpy for matlab users.html uses rank for dimension , which is quite confusing for people who knows what the mathematical definition of rank of a matrix but not tensor rank) is.
|
|
I suggest we change rank n matrix to n dimensional matrix through the article.","bug"
|
|
"6013","684","x.fill []) dumps core when x is an object array Trac 86).
|
|
Original ticket http projects.scipy.org numpy ticket 86 on 2006 04 28 by trac user sasha, assigned to unknown.
|
|
>>> x = zeros 1, O ) >>> x.fill []) Program received signal SIGSEGV, Segmentation fault.","bug"
|
|
"6013","6844","numpy.ma.mean axis argument does not accept tuple like numpy.mean.
|
|
numpy 1.10.1 py27 0 Linux carson 3.13.0 71 generic 114 precise1 Ubuntu SMP Tue Dec 1 19 37 38 UTC 2015 x86 64 x86 64 x86 64 GNU Linux","bug"
|
|
"6013","685","Implemented test ufunclike as a normal test Trac 87).
|
|
Original ticket http projects.scipy.org numpy ticket 87 on 2006 04 28 by alberts, assigned to teoliphant.
|
|
I am in the process of identifying parts of the NumPy library where more testing is needed.
|
|
I am using trace.py which is included with Python to do coverage analysis of the code.
|
|
This tool doesn t like doctests, so I am converting the two files that contain doctests to normal test code.
|
|
I ll attach the patch against r2426 for test ufunclike.py to this ticket.
|
|
There is one test failure on my system.
|
|
Someone might want to investigate this.
|
|
Found 5 tests for main ...F. ====================================================================== FAIL check log2 main .test ufunclike) Traceback most recent call last) File C \home\albert\work2\numpy\numpy\lib\tests\test ufunclike.py , line 60, in check log2 assert array equal b, array [2.169925, 1.20163386, 2.70043972])) File C \Python24\Lib\site packages\numpy\testing\utils.py , line 204, in assert array equal assert cond,\ AssertionError Arrays are not equal mismatch 100.0 ) Array 1 [ 2.1699250014423126 1.2016338611696504 2.7004397181410922] Array 2 [ 2.1699250000000001 1.2016338600000001 2.7004397199999999] Ran 5 tests in 0.000s FAILED failures=1)","enhancement"
|
|
"6013","686","Implemented test polynomial as a normal test Trac 88).
|
|
Original ticket http projects.scipy.org numpy ticket 88 on 2006 04 29 by alberts, assigned to teoliphant.
|
|
I am in the process of identifying parts of the !NumPy library where more testing is needed.
|
|
I am using trace.py which is included with Python to do coverage analysis of the code.
|
|
This tool doesn t like doctests, so I am converting the two files that contain doctests to normal test code.
|
|
I ll attach the patch against r2426 for test polynomial.py to this ticket.","enhancement"
|
|
"6013","687","Make DESIRED and ACTUAL reprs line up Trac 89).
|
|
Original ticket http projects.scipy.org numpy ticket 89 on 2006 04 29 by alberts, assigned to teoliphant.
|
|
Having the DESIRED and ACTUAL reps line up on the screen when something is wrong with a test makes it easier to spot the differences.
|
|
Patch to be attached.","enhancement"
|
|
"6013","688","beta distribution in numpy.random works incorrectly Trac 90).
|
|
Original ticket http projects.scipy.org numpy ticket 90 on 2006 05 01 by tovrstra, assigned to unknown.
|
|
Hi, As was already noted on the mailing list, the beta distribution does not work correctly with all parameters http article.gmane.org gmane.comp.python.numeric.general 3310 The problem is located in trunk numpy random mtrand distributions.c beginning at line 182.
|
|
I have the impression that the part labeled Use Jonk s algorithm is not working correctly.
|
|
When I disable that part and always use the gamma formulation instead, it seems to work correctly.
|
|
A quick example of what is going wrong >>> import numpy.random >>> print numpy.histogram numpy.random.beta 1, 1, 10000))[0] [1868 1793 1508 1281 1085 838 679 491 311 146] This should in fact be a uniform distribution.
|
|
Does anyone know where Jonk s algorithm comes from?
|
|
Maybe it would be wiser to use to use Cheng s algorithm here?
|
|
I m not an expert.)
|
|
R. C. H. Cheng, Generating beta variates with nonintegral shape parameters, Communications of the ACM, 21 317 322 1978) A source code for Cheng s algorithm is available at http www.netlib.org random ranlib.c.tar.gz","bug"
|
|
"6013","689","cov modifies its argument Trac 91).
|
|
Original ticket http projects.scipy.org numpy ticket 91 on 2006 05 01 by huard, assigned to rkern.
|
|
Calling cov x) modifies x when the rowvar argument is not the logical one.
|
|
For instance, x = multivariate normal [0,0], [[1,.5],[.5, 2]],50) y = x.copy ) cov x, rowvar=1) cov y, rowvar=0) Now x has been modified by calling cov x, rowvar=1) and is no longer equal to y.
|
|
The culprit seems to be that copy is set to false.
|
|
[[BR]] David Huard <david.huard gmail.com>","bug"
|
|
"6013","690","Segfault or TypeError when asarray cannot allocate memory Trac 92).
|
|
Original ticket http projects.scipy.org numpy ticket 92 on 2006 05 03 by alberts, assigned to unknown.
|
|
I described the problem here http sourceforge.net mailarchive forum.php?thread id=10292417 forum id=4890","bug"
|
|
"6013","691","dtype with align=1 and invalid typestring segfaults Trac 93).
|
|
Original ticket http projects.scipy.org numpy ticket 93 on 2006 05 03 by alberts, assigned to unknown.
|
|
The following code segfaults with !NumPy r2462 due to the invalid typestring in formats.
|
|
import numpy as N descr = N.dtype { names [ a ], formats [ foo ]}, align=1)","bug"
|
|
"6013","692","Array descriptors not compared correctly causes asarray to copy data Trac 94).
|
|
Original ticket http projects.scipy.org numpy ticket 94 on 2006 05 03 by alberts, assigned to teoliphant.
|
|
There seems to be a problem with asarray when using more than one instance of an equivalent dtype.
|
|
Example code import numpy as N print N. version import time a = N.dtype { names [ i , j ], formats [N.float64, N.float64]}) b = N.dtype { names [ i , j ], formats [N.float64, N.float64]}) print a == b print a.descr == b.descr starttime = time.clock ) arr = N.zeros 5000,1000), dtype=a) for x in arr y = N.asarray x, dtype=a) print done in f seconds time.clock ) starttime,) starttime = time.clock ) for x in arr y = N.asarray x, dtype=b) print done in f seconds time.clock ) starttime,) On my system I get the following result 0.9.7.2462 False True done in 0.153871 seconds done in 8.726785 seconds So while the descrs are equal, Python doesn t seem to think that the dtypes are which is probably causing the problem.","bug"
|
|
"6013","693","digitize causes memory corruption Trac 95).
|
|
Original ticket http projects.scipy.org numpy ticket 95 on 2006 05 04 by stefanv, assigned to unknown.
|
|
Under Numpy version 2476, using Valgrind to run import numpy as N for i in range 100) N.digitize [1,2,3,4],[1,3]) N.digitize [0,1,2,3,4],[1,3]) points to some memory de allocation problems in compiled base.c ==14997== Invalid free ) delete delete[] ==14997== at 0x401CFCF free vg replace malloc.c 235) ==14997== by 0x490375A arraydescr dealloc arrayobject.c 8914) ==14997== by 0x4914450 array dealloc arrayobject.c 1477) ==14997== by 0x42B9191 arr digitize compiled base.c 202) ==14997== by 0x80B6266 PyEval EvalFrame in usr bin python2.4) ==14997== by 0x80B76BE PyEval EvalCodeEx in usr bin python2.4) ==14997== by 0x80B7904 PyEval EvalCode in usr bin python2.4) ==14997== by 0x80D946B PyRun FileExFlags in usr bin python2.4) ==14997== by 0x80D970B PyRun SimpleFileExFlags in usr bin python2.4) ==14997== by 0x8055B02 Py Main in usr bin python2.4) ==14997== Address 0x4949140 is not stack d, malloc d or recently) free d ==14997== ==14997== Invalid free ) delete delete[] ==14997== at 0x401CFCF free vg replace malloc.c 235) ==14997== by 0x490375A arraydescr dealloc arrayobject.c 8914) ==14997== by 0x4914450 array dealloc arrayobject.c 1477) ==14997== by 0x42B91A2 arr digitize compiled base.c 201) ==14997== by 0x80B6266 PyEval EvalFrame in usr bin python2.4) ==14997== by 0x80B76BE PyEval EvalCodeEx in usr bin python2.4) ==14997== by 0x80B7904 PyEval EvalCode in usr bin python2.4) ==14997== by 0x80D946B PyRun FileExFlags in usr bin python2.4) ==14997== by 0x80D970B PyRun SimpleFileExFlags in usr bin python2.4) ==14997== by 0x8055B02 Py Main in usr bin python2.4) ==14997== Address 0x4949140 is not stack d, malloc d or recently) free d ==14997== ==14997== Invalid free ) delete delete[] ==14997== at 0x401CFCF free vg replace malloc.c 235) ==14997== by 0x490375A arraydescr dealloc arrayobject.c 8914) ==14997== by 0x4925180 PyArray FromAny arrayobject.c 6856) ==14997== by 0x42B8E87 arr digitize compiled base.c 168) ==14997== by 0x80B6266 PyEval EvalFrame in usr bin python2.4) ==14997== by 0x80B76BE PyEval EvalCodeEx in usr bin python2.4) ==14997== by 0x80B7904 PyEval EvalCode in usr bin python2.4) ==14997== by 0x80D946B PyRun FileExFlags in usr bin python2.4) ==14997== by 0x80D970B PyRun SimpleFileExFlags in usr bin python2.4) ==14997== by 0x8055B02 Py Main in usr bin python2.4) ==14997== Address 0x4949140 is not stack d, malloc d or recently) free d ==14997== ==14997== Invalid free ) delete delete[] ==14997== at 0x401CFCF free vg replace malloc.c 235) ==14997== by 0x490375A arraydescr dealloc arrayobject.c 8914) ==14997== by 0x4925180 PyArray FromAny arrayobject.c 6856) ==14997== by 0x42B8EC4 arr digitize compiled base.c 169) ==14997== by 0x80B6266 PyEval EvalFrame in usr bin python2.4) ==14997== by 0x80B76BE PyEval EvalCodeEx in usr bin python2.4) ==14997== by 0x80B7904 PyEval EvalCode in usr bin python2.4) ==14997== by 0x80D946B PyRun FileExFlags in usr bin python2.4) ==14997== by 0x80D970B PyRun SimpleFileExFlags in usr bin python2.4) ==14997== by 0x8055B02 Py Main in usr bin python2.4) ==14997== Address 0x4949140 is not stack d, malloc d or recently) free d","bug"
|
|
"6013","694","scalarmathmodule changes break MSVC build Trac 96).
|
|
Original ticket http projects.scipy.org numpy ticket 96 on 2006 05 04 by alberts, assigned to teoliphant.
|
|
Build with MSVC is broken as of r2476.
|
|
Patch to be attached to this ticket.","bug"
|
|
"6013","695","boolscalars do not multiply correctly Trac 97).
|
|
Original ticket http projects.scipy.org numpy ticket 97 on 2006 05 04 by rkern, assigned to teoliphant.
|
|
>>> from numpy import >>> m = 5 > arange 10) >>> m array [True, True, True, True, True, False, False, False, False, False], dtype=bool) >>> m 2 array [2, 2, 2, 2, 2, 0, 0, 0, 0, 0], dtype=uint8) >>> tr = m[0] >>> tr 2 0 >>> type tr) <type boolscalar > This causes a problem with clip ) and float inputs >>> clip 5.0, 0.25, 1.0) 5.0 Importing the scalarmath module does not seem to help at this time.","bug"
|
|
"6013","696","Incorrect result when multiplying views of record arrays Trac 98).
|
|
Original ticket http projects.scipy.org numpy ticket 98 on 2006 05 07 by alberts, assigned to teoliphant.
|
|
The code import numpy as N rdt = N.dtype { names [ i , j ], formats [N.intc, N.float64]}) x = N.array [[ 1,5.
|
|
), 2,6.
|
|
), 1,0.
|
|
)]], dtype=rdt) y = N.array [[ 1,1.
|
|
), 2,2.
|
|
), 1,0.
|
|
)], [ 1,3.
|
|
), 2,4.
|
|
), 1,0.
|
|
)]], dtype=rdt) xv = x[ j ][ , 1] yv = y[ j ][ , 1] z = N.dot yv, N.transpose xv)) Expected result is `1 5 2 6=17`, `3 5 4 6=39`.
|
|
This is the result I obtain when running r2476 on Windows compiled with MSVC 7.1.
|
|
With r2476 on Linux on Fedora Core 4 with GCC 4.0.0 and on Stefan van der Walt s Ubuntu system with GCC 4.0.3, the result is `5, 15`.
|
|
This is wrong.
|
|
Making a copy of the array and then multiplying a = N.array xv) b = N.array yv) c = N.dot b, N.transpose a)) gives the expected result on both platforms.","bug"
|
|
"6013","697","Scalar types return wrong error when conversion from string fails Trac 99).
|
|
Original ticket http projects.scipy.org numpy ticket 99 on 2006 05 07 by alberts, assigned to teoliphant.
|
|
Works for `int`, breaks for `intp` when it shouldn t, and with the wrong error In [32] int 0xb72a7008 , 16) Out[32] 3073011720L In [33] N.intp 0xb72a7008 , 16) TypeError function takes exactly 1 argument 2 given) `0xb72a7008` was actually a string returned in by a !NumPy array s ` array data [0]` on Linux.
|
|
Breaks for `long` and `intp`, but `intp` raises the wrong error In [34] int 0x1 , 32) ValueError invalid literal for int ) 0x1 In [35] N.intp 0x1 , 32) TypeError function takes exactly 1 argument 2 given) Tested with r2476.
|
|
Please add a unit test when fixing this bug.
|
|
Thanks!","enhancement"
|
|
"6013","698","Incorrect integer conversion to string Trac 100).
|
|
Original ticket http projects.scipy.org numpy ticket 100 on 2006 05 07 by stefanv, assigned to teoliphant.
|
|
Casting an integer to a string does not produce the expected result.
|
|
For example, `N.cast[N.string0] N.intc 1))` returns ` ` instead of ` 1 `.
|
|
If cast is meant to be used differently, a docstring might throw some light on the subject.","enhancement"
|
|
"6013","699","field access in recarray broken when field name matches class method or attribute Trac 101).
|
|
Original ticket http projects.scipy.org numpy ticket 101 on 2006 05 09 by chanley, assigned to teoliphant.
|
|
It is not possible to access a named recarray field if that name matches the name of a recarray class method or attribute.
|
|
Example code is below >>> from numpy import rec >>> r = rec.fromrecords [[456, dbe ,1.2],[2, de ,1.3]],names= num,name,field ) >>> r.field num ) array [456, 2]) >>> r.field field ) <bound method recarray.field of recarray [ 456, dbe , 1.2), 2, de , 1.3)], dtype=[ num , <i4 ), name , |S3 ), field , <f8 )])> >>>","bug"
|
|
"6013","700","cannot create single field recarray using string format input Trac 102).
|
|
Original ticket http projects.scipy.org numpy ticket 102 on 2006 05 09 by chanley, assigned to unknown.
|
|
An error is raised when attempting to create a single field recarray using string format as input.
|
|
If there is more than one format specified in the string no error is created.
|
|
However, it is possible to create a single field recarray if the format is given as a member of a 1 element list.
|
|
Example code below In [1] from numpy import rec In [2] rr = rec.array None, formats= f4,i4,i8,f8 , shape=100) In [3] rr.dtype Out[3] dtype [ f1 , <f4 ), f2 , <i4 ), f3 , <i8 ), f4 , <f8 )]) In [4] rr = rec.array None, formats= f4,i4 , shape=100) In [5] rr.dtype Out[5] dtype [ f1 , <f4 ), f2 , <i4 )]) In [6] rr = rec.array None, formats= f4 , shape=100) exceptions.TypeError Traceback most recent call last) data sparty1 dev pyfits NP test <ipython console> data sparty1 dev site packages lib python numpy core records.py in array obj, formats, names, titles, shape, byteorder, aligned, offset, strides) 414 return recarray shape, formats, names=names, titles=titles, 415 buf=obj, offset=offset, strides=strides, > 416 byteorder=byteorder, aligned=aligned) 417 elif isinstance obj, str) 418 return fromstring obj, formats, names=names, titles=titles, data sparty1 dev site packages lib python numpy core records.py in new subtype, shape, formats, names, titles, buf, offset, strides, byteorder, aligned) 152 descr = formats 153 else > 154 parsed = format parser formats, names, titles, aligned) 155 descr = parsed.
|
|
descr 156 data sparty1 dev site packages lib python numpy core records.py in init self, formats, names, titles, aligned) 42 class format parser 43 def init self, formats, names, titles, aligned=False) > 44 self.
|
|
parseFormats formats, aligned) 45 self.
|
|
setfieldnames names, titles) 46 self.
|
|
createdescr ) data sparty1 dev site packages lib python numpy core records.py in parseFormats self, formats, aligned) 51 dtype = sb.dtype formats, aligned) 52 fields = dtype.fields > 53 keys = fields[ 1] 54 self.
|
|
f formats = [fields[key][0] for key in keys] 55 self.
|
|
offsets = [fields[key][1] for key in keys] TypeError unsubscriptable object In [8] rr = rec.array None, formats=[ f4 ], shape=100) In [9] rr.dtype Out[9] dtype [ f1 , <f4 )])","bug"
|
|
"6013","701","fromfile is broken on Windows Trac 103).
|
|
Original ticket http projects.scipy.org numpy ticket 103 on 2006 05 09 by alberts, assigned to unknown.
|
|
The following code reliably fails on Windows with r2489 import numpy as N x = N.randn 50000) print repr x.data[ 10]) x.tofile file temp.dat , wb )) y = N.fromfile temp.dat , dtype=N.float64) I get stuff like this \xffH\x1a\xad\xc5l\xeb?l\x84 50000 items requested but only 0 read E\x9e\xbc\xcb\xe6Z\xdf?\xa8\xba 50000 items requested but only 9 read =\xddZ\xfd \x88\xfa\xbf\x0c\xfa 50000 items requested but only 3 read When I write the file as follows `fromfile` still breaks f = file temp.dat , wb ) f.write x.data) f.close ) I ve checked that the file is on the disk.
|
|
Reading it like this f = file temp.dat , rb ) buf = f.read completefilesize) assert len buf) == completefilesize y = N.frombuffer buf, dtype=N.float64) always works.
|
|
`fread` is called only once in `PyArray FromFile`, which might not be returning all the data in the file in one call.
|
|
Alternatively, some weird combination of bits in the file might be breaking things, but this seems unlikely.","bug"
|
|
"6013","702","Array without string representation Trac 104).
|
|
Original ticket http projects.scipy.org numpy ticket 104 on 2006 05 10 by stefanv, assigned to unknown.
|
|
Under r2490, the following code creates an array that doesn t produce a valid string representation import numpy as N r = N.zeros shape= 3,3), dtype=[ x , f4 )]).view N.recarray) y = r.view f4 ) print Created y print y","bug"
|
|
"6013","703","boolean index array returns byteswapped values Trac 105).
|
|
Original ticket http projects.scipy.org numpy ticket 105 on 2006 05 10 by chanley, assigned to unknown.
|
|
When using a big endian array on a little endian OS, the use of a boolean array as an index array causes the resulting array to have byteswapped values.
|
|
Example code is below In [14] a = numpy.array [1,2,3,4,5,6,7,8,9], >f8 ) In [15] a Out[15] array [ 1., 2., 3., 4., 5., 6., 7., 8., 9.])
|
|
In [16] x = numpy.where a>2) a<6) ) In [17] x Out[17] array [2, 3, 4]),) In [18] a[x] Out[18] array [ 3., 4., 5.])
|
|
In [19] y = a>2) a<6) ) In [20] y Out[20] array [False, False, True, True, True, False, False, False, False], dtype=bool) In [21] a[y] Out[21] array [ 1.04346664e 320, 2.05531309e 320, 2.56123631e 320]) This bug was originally discovered by Erin Sheldon while testing pyfits.","bug"
|
|
"6013","704","Memory corruption with dot Trac 106).
|
|
Original ticket http projects.scipy.org numpy ticket 106 on 2006 05 10 by alberts, assigned to unknown.
|
|
Code import numpy as N x = N.randn 0, 1) y = N.randn 10, 1) z = N.dot x, N.transpose y)) Valgrind errors generated by Valgrind 3.1.1 on Fedora Core 4 with ATLAS 3.6.0 from Fedora Extras ==18465== 1 errors in context 10 of 11 ==18465== Use of uninitialised value of size 8 ==18465== at 0x1C09CD0C ATL daxpy xp1yp1aXbX in usr lib atlas libatlas.so.3.0) ==18465== ==18465== 1 errors in context 11 of 11 ==18465== Use of uninitialised value of size 8 ==18465== at 0x1BF45A7C dotblas matrixproduct dotblas.c 370) ==18465== by 0x1B95E1F2 PyCFunction Call in usr lib libpython2.4.so.1.0) ==18465== by 0x1B997091 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==18465== by 0x1B997EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==18465== by 0x1B998227 PyEval EvalCode in usr lib libpython2.4.so.1.0) ==18465== by 0x1B9B4559 within usr lib libpython2.4.so.1.0) ==18465== by 0x1B9B57D1 PyRun SimpleFileExFlags in usr lib libpython2.4.so.1.0) ==18465== by 0x1B9B6268 PyRun AnyFileExFlags in usr lib libpython2.4.so.1.0) ==18465== by 0x1B9BC16C Py Main in usr lib libpython2.4.so.1.0) ==18465== by 0x80485B9 main in usr bin python)","bug"
|
|
"6013","705","Directory name build srcipts pythonversion needs platform information Trac 107).
|
|
Original ticket http projects.scipy.org numpy ticket 107 on 2006 05 10 by pearu, assigned to pearu.
|
|
Otherwise when using the same filesystem via nfs, samba) for multiple platforms, the generated scripts may be broken for certain platforms.","bug"
|
|
"6013","706","core tests test defmatrix.py fails on revision 2495 Trac 108).
|
|
Original ticket http projects.scipy.org numpy ticket 108 on 2006 05 11 by trac user nich2o, assigned to unknown.
|
|
Results of python c import numpy; numpy.test 10) with revision 2495 ........................................................................................................................................................................................................................................................F.................................................................................................................... ====================================================================== FAIL check instance methods numpy.core.tests.test defmatrix.test matrix return) Traceback most recent call last) File usr lib python2.4 site packages numpy core tests test defmatrix.py , line 165, in check instance methods assert type b) is matrix AssertionError Ran 365 tests in 1.385s FAILED failures=1) matrix.all ) method returns a <type boolscalar > instead of <class numpy.core.defmatrix.matrix >","bug"
|
|
"6013","707","dot still corrupts memory with 0 length arrays Trac 109).
|
|
Original ticket http projects.scipy.org numpy ticket 109 on 2006 05 11 by alberts, assigned to teoliphant.
|
|
Code import numpy as N x = N.randn 0, 1) y = N.randn 0, 1) N.dot x, N.transpose y)) Valgrind 3.1.1 says ==31961== 1 errors in context 1 of 1 ==31961== Use of uninitialised value of size 8 ==31961== at 0x1BF45A7C dotblas matrixproduct dotblas.c 372) ==31961== by 0x1B95E1F2 PyCFunction Call in usr lib libpython2.4.so.1.0) ==31961== by 0x1B997091 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==31961== by 0x1B997EF7 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ==31961== by 0x1B998227 PyEval EvalCode in usr lib libpython2.4.so.1.0) ==31961== by 0x1B9B4559 within usr lib libpython2.4.so.1.0) ==31961== by 0x1B9B57D1 PyRun SimpleFileExFlags in usr lib libpython2.4.so.1.0) ==31961== by 0x1B9B6268 PyRun AnyFileExFlags in usr lib libpython2.4.so.1.0) ==31961== by 0x1B9BC16C Py Main in usr lib libpython2.4.so.1.0) ==31961== by 0x80485B9 main in usr bin python) 31961","bug"
|
|
"6013","708","Various Valgrind errors Trac 110).
|
|
Original ticket http projects.scipy.org numpy ticket 110 on 2006 05 11 by alberts, assigned to unknown.
|
|
When running the file to be attached, Valgrind shows up errors in 2 different contexts.
|
|
I tried reducing the test to a few lines, but removing much more code causes one of the bugs to disappear.
|
|
One of the errors seems to be the same one as in ticket 551.
|
|
The code basically calculates the top triangle of a symmetric matrix, including the diagonal, with some use of record arrays.
|
|
Valgrind errors used Valgrind 3.1.1) ==5803== 408 bytes in 2 blocks are possibly lost in loss record 21 of 38 ==5803== at 0x1B909222 malloc vg replace malloc.c 130) ==5803== by 0x1BEBCD1D PyArray NewFromDescr arrayobject.c 4444) ==5803== by 0x1BF1B322 construct loop ufuncobject.c 1017) ==5803== by 0x1BF1BB80 PyUFunc GenericFunction ufuncobject.c 1419) ==5803== by 0x1BF1CD1F ufunc generic call ufuncobject.c 2670) ==5803== by 0x1B9383B3 PyObject Call in usr lib libpython2.4.so.1.0) ==5803== by 0x1B93847B PyObject CallFunction in usr lib libpython2.4.so.1.0) ==5803== by 0x1BEAEC12 PyArray GenericBinaryFunction arrayobject.c 2771) ==5803== by 0x1B936299 within usr lib libpython2.4.so.1.0) ==5803== by 0x1B936CFF PyNumber Multiply in usr lib libpython2.4.so.1.0) ==5803== by 0x1BEDC620 PyArray Round multiarraymodule.c 268) ==5803== by 0x1BEDCA49 array round arraymethods.c 1490) ==5803== by 0x1B95E1CA PyCFunction Call in usr lib libpython2.4.so.1.0) ==5803== by 0x1B997091 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ... ==5803== 530672 6520 direct, 524152 indirect) bytes in 163 blocks are definitely lost in loss record 32 of 38 ==5803== at 0x1B909222 malloc vg replace malloc.c 130) ==5803== by 0x1BEB0C5F array alloc arrayobject.c 5582) ==5803== by 0x1BEBCADB PyArray NewFromDescr arrayobject.c 4393) ==5803== by 0x1BF1B322 construct loop ufuncobject.c 1017) ==5803== by 0x1BF1BB80 PyUFunc GenericFunction ufuncobject.c 1419) ==5803== by 0x1BF1CD1F ufunc generic call ufuncobject.c 2670) ==5803== by 0x1B9383B3 PyObject Call in usr lib libpython2.4.so.1.0) ==5803== by 0x1B93847B PyObject CallFunction in usr lib libpython2.4.so.1.0) ==5803== by 0x1BEAEC12 PyArray GenericBinaryFunction arrayobject.c 2771) ==5803== by 0x1B936299 within usr lib libpython2.4.so.1.0) ==5803== by 0x1B936CFF PyNumber Multiply in usr lib libpython2.4.so.1.0) ==5803== by 0x1BEDC620 PyArray Round multiarraymodule.c 268) ==5803== by 0x1BEDCA49 array round arraymethods.c 1490) ==5803== by 0x1B95E1CA PyCFunction Call in usr lib libpython2.4.so.1.0) ==5803== by 0x1B997091 PyEval EvalFrame in usr lib libpython2.4.so.1.0) ...","bug"
|
|
"6013","709","Byte order problem in arange Trac 111).
|
|
Original ticket http projects.scipy.org numpy ticket 111 on 2006 05 11 by trac user esheldon, assigned to unknown.
|
|
Little endian linux box >>> numpy.
|
|
version 0.9.7.2497 >>> x = numpy.arange 10,dtype= <f8 ) >>> x array [ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9.])
|
|
>>> x = numpy.arange 10,dtype= >f8 ) >>> x array [ 0.00000000e 000, 1.00000000e 000, 1.37186586e 303, 5.82360826e 011, 7.98920843e 292, 3.60319875e 021, 4.94303335e 282, 2.09830067e 031, 2.87854483e 272, 1.29367874e 041]) As expected the opposite occurs on my big endian mac, with <f8 arange failing.
|
|
Note, this works >>> d=numpy.array [1,2,3,4,5], dtype= >f8 ) >>> d array [ 1., 2., 3., 4., 5.])
|
|
So it may be specific to arange.","bug"
|
|
"6013","710","string representation of longfloat numpy arrays shows inf instead of real value Trac 112).
|
|
Original ticket http projects.scipy.org numpy ticket 112 on 2006 05 11 by trac user solkaa ..., assigned to unknown.
|
|
In [20] import numpy In [21] a = numpy.exp numpy.array [1000], dtype=numpy.longfloat)) In [22] print a[0] 1.97007111401704699387e 434 In [23] print repr a) array [ inf], dtype=float96) In [24] print str a) [ inf] In [25] print numpy.
|
|
version 0.9.6","bug"
|
|
"6013","711","Compiler warnings Trac 113).
|
|
Original ticket http projects.scipy.org numpy ticket 113 on 2006 05 11 by stephan tolksdorf, assigned to unknown.
|
|
Numpy currently builds with various compiler warnings not only in MSVC.
|
|
An easy way to turn on the output of compiler warnings is to comment out line 40 in distutils ccompiler.py.
|
|
)[[BR]][[BR]] Attached is a patch that takes care of many of the warnings generated for MSVC.
|
|
Almost all of them warn of unsafe typecasts.
|
|
One problem was that a strict compiler interprets all floating point literals without an f or l) suffix as double s, which leads to warnings in float functions.
|
|
[[BR]][[BR]] I m not 100 sure that I did the right thing with nc cosh and nc expm1, please double check.
|
|
[[BR]][[BR]] Maybe someone else could have a look at the remaining MSVC warnings.","bug"
|
|
"6013","712","Problems with building with MSVC and GCC under Cygwin Trac 114).
|
|
Original ticket http projects.scipy.org numpy ticket 114 on 2006 05 11 by stephan tolksdorf, assigned to stephan tolksdorf.
|
|
The setup script currently does not correctly recognize external libraries for example ATLAS) when building with MSVC because the GCC naming convention is assumed.
|
|
[[BR]] [[BR]] The build with the MinGW tools in Cygwin is broken because mno cygwin is not consistently added to the compiler flags.","bug"
|
|
"6013","713","Strange dot memory corruption on r2497 Trac 115).
|
|
Original ticket http projects.scipy.org numpy ticket 115 on 2006 05 12 by trac user nich2o, assigned to unknown.
|
|
On r2497 the following code from numpy import a = arange 11 11).astype f ) f = arange 11 11).astype f ) while 1 r = dot a,f) crashes after a while 35minutes on my 4400 1GB machine) glibc detected free ) invalid pointer 0x00002aaaacaf0760","bug"
|
|
"6013","714","itemsize of stringscalar types doesn t match that of its ndarray Trac 116).
|
|
Original ticket http projects.scipy.org numpy ticket 116 on 2006 05 12 by FrancescAlted, assigned to unknown.
|
|
With the latest versions of numpy, I m seeing the next problem with stringscalar itemsizes In [16] numpy.
|
|
version Out[16] 0.9.7.2499 In [17] s=numpy.array [ aa ], S3 ) In [18] ss=s[0] In [19] s.dtype.itemsize Out[19] 3 In [20] ss.dtype.itemsize Out[20] 2 i.e.
|
|
the itemsize of the stringscalar type is different from the ndarray itemsize from what it derives.
|
|
This used to work ok in previous releases of numpy In [18] numpy.
|
|
version Out[18] 0.9.7.2278 In [19] s=numpy.array [ aa ], S3 ) In [20] ss=s[0] In [21] s.dtype.itemsize Out[21] 3 In [22] ss.dtype.itemsize Out[22] 3 The new behaviour makes that numarray can t deduce correctly the itemsizes when importing string numpy arrays In [1] import numpy In [2] from numarray import strings In [3] strings.array numpy.array [ aa ], S1 )).itemsize ) Out[3] 1 ok In [4] strings.array numpy.array [ aa ], S3 )).itemsize ) Out[4] 2 should be 3!
|
|
Thanks","bug"
|
|
"6013","715","bdist rpm and config fc don t work together Trac 117).
|
|
Original ticket http projects.scipy.org numpy ticket 117 on 2006 05 13 by trac user swalton, assigned to pearu.
|
|
The bdist rpm command doesn t work with config fc, by which I mean that if one does python setup.py config fc fcompiler=absoft bdist rpm the GNU compiler is still used.
|
|
I ve looked at it somewhat, but I am just not enough of a distutils guru to even understand where the problem might be.
|
|
In old scipy , version 0.3.2, bdist rpm consisted in its entirety of !python from distutils.command.bdist rpm import bdist rpm as old bdist rpm class bdist rpm old bdist rpm) pass This worked, in that config fc and bdist rpm worked together.
|
|
In the new version, there is additional code to handle the fact that the setup files can be named something other than setup.py.
|
|
I can t see how this makes a difference, but apparently it does.","bug"
|
|
"6013","716","linspace should always return floats Trac 118).
|
|
Original ticket http projects.scipy.org numpy ticket 118 on 2006 05 16 by stefanv, assigned to unknown.
|
|
linspace returns arrays of different types, depending on the parameters.
|
|
E.g.
|
|
In [1] N.linspace 0,1,0).dtype Out[1] dtype <i4 ) In [2] N.linspace 0,1,1).dtype Out[2] dtype <f8 ) This is due to the special case, when only one element is returned without being divided by a float).","bug"
|
|
"6013","717","argmin and argmax fail for axis other than the last one Trac 119).
|
|
Original ticket http projects.scipy.org numpy ticket 119 on 2006 05 16 by trac user PauGargallo, assigned to unknown.
|
|
argmin and argmax methods don t work properly for arrays with more than 2 dimensions.
|
|
x.argmax axis=i) raise and error for any axis but the last one.
|
|
Pierre Barbier de Reuille sent this testing code to the numpy discussion mailing list from numpy import array, random, all a = random.normal 0, 1, 4,5,6,7,8 ) ) for i in xrange a.ndim ) amax = a.max i ) aargmax = a.argmax i ) raise an error axes = range a.ndim ) axes.remove i ) assert all amax == aargmax.choose a.transpose i, axes ) ) ) which currently fails in the first iteration, and the patch attached to this ticket.","bug"
|
|
"6013","718","array.transpose ) too lenient with its arguments Trac 120).
|
|
Original ticket http projects.scipy.org numpy ticket 120 on 2006 05 16 by baxissimo, assigned to unknown.
|
|
When not given enough axes, it just drops data from the matrix.
|
|
>>> a = numpy.rand 2,2) >>> a array [[ 0.96685836, 0.55643033], [ 0.86387107, 0.39331451]]) >>> a.transpose 1) array [ 0.96685836, 0.55643033]) >>> a array [[ 0.96685836, 0.55643033], [ 0.86387107, 0.39331451]]) >>> a.transpose 0) array [ 0.96685836, 0.86387107]) >>> a.transpose ) array [[ 0.96685836, 0.86387107], [ 0.55643033 , 0.39331451]])","bug"
|
|
"6013","719","array.transpose ) docstring is less than helpful Trac 121).
|
|
Original ticket http projects.scipy.org numpy ticket 121 on 2006 05 16 by baxissimo, assigned to unknown.
|
|
An explanation of the arguments and their effect would be nice.
|
|
>>> help a.transpose) Help on built in function transpose transpose ...) m.transpose <None>)","bug"
|
|
"6013","720","No doc for ones like Trac 122).
|
|
Original ticket http projects.scipy.org numpy ticket 122 on 2006 05 17 by trac user PauGargallo, assigned to stefanv.
|
|
ones like have no documentation.
|
|
>>> from numpy import >>> ones like.
|
|
doc y = ones like x) return 1 on the other hand, as ones like is implemented as a ufunc, help ones like) gives a standard ufunc class documentation.
|
|
>>> help ones like) Help on ufunc object ones like = class ufunc builtin .object) | Optimized functions make it possible to implement arithmetic with arrays efficiently | | Methods defined here | | call ...) | x. call ...) <==> x ...) | | repr ...) | x. repr ) <==> repr x) | | str ...) | x. str ) <==> str x) A very simple doc like Return an array of ones of the shape and typecode of a. should be enought and coherent with zeros like.","bug"
|
|
"6013","721","simple math operations with scalar are leaking and all functions, like mean, using them are doomed to leak) Trac 123).
|
|
Original ticket http projects.scipy.org numpy ticket 123 on 2006 05 17 by trac user nich2o, assigned to unknown.
|
|
Memory consumption is growing when doing from numpy import v = arange 10) while 1 r = mean v) Valgrind says ==14921== 24 bytes in 1 blocks are definitely lost in loss record 6 of 59 ==14921== at 0x4A1A885 malloc vg replace malloc.c 149) ==14921== by 0x5EB2D19 gentype alloc scalartypes.inc.src 269) ==14921== by 0x5ED4F31 PyArray ScalarFromObject scalartypes.inc.src 228) ==14921== by 0x74996AC long convert to ctype scalarmathmodule.c.src 485) ==14921== by 0x74ABE0A long divide scalarmathmodule.c.src 498) ==14921== by 0x4139B7 binary op1 abstract.c 371) ==14921== by 0x4158DD PyNumber Divide abstract.c 422) ==14921== by 0x5EEA74D PyArray Mean multiarraymodule.c 622) ==14921== by 0x5EEA895 array mean arraymethods.c 1214) ==14921== by 0x41452F PyObject Call abstract.c 1756) ==14921== by 0x475021 PyEval EvalFrame ceval.c 3835) ==14921== by 0x477BEE PyEval EvalCodeEx ceval.c 2736)","bug"
|
|
"6013","722","Valgrind errors when running test varstd Trac 124).
|
|
Original ticket http projects.scipy.org numpy ticket 124 on 2006 05 18 by alberts, assigned to unknown.
|
|
When running test varstd numpy.core.tests.test ma.test array methods) the following error is caught by Valgrind ==28623== Conditional jump or move depends on uninitialised value s) ==28623== at 0x46AD926 PyUFunc GenericFunction ufuncobject.c 1423) ==28623== by 0x46AEA8F ufunc generic call ufuncobject.c 2669) ==28623== by 0x2C53B3 PyObject Call in usr lib libpython2.4.so.1.0) ...","bug"
|
|
"6013","723","Memory leak when manipulating matrices Trac 125).
|
|
Original ticket http projects.scipy.org numpy ticket 125 on 2006 05 18 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.1 the following code leaks memory from numpy import matrix a = matrix [1.0], dtype= f8 ) assert type a.real) is matrix assert type a.imag) is matrix c,d = matrix [0.0]).nonzero ) assert type c) is matrix assert type d) is matrix Valgrind error ==29008== 64 40 direct, 24 indirect) bytes in 1 blocks are definitely lost in loss record 7 of 26 ==29008== at 0x400444E malloc vg replace malloc.c 149) ==29008== by 0x4543B2D array alloc arrayobject.c 5582) ==29008== by 0x454FA4D PyArray NewFromDescr arrayobject.c 4393) ==29008== by 0x455F413 PyArray Zeros multiarraymodule.c 4801) ==29008== by 0x455F4F6 array imag get arrayobject.c 5379) ==29008== by 0x2CF925 within usr lib libpython2.4.so.1.0) ==29008== by 0x2EDA4E PyObject GenericGetAttr in usr lib libpython2.4.so.1.0) ==29008== by 0x2ED1CD PyObject GetAttr in usr lib libpython2.4.so.1.0) ==29008== by 0x32140B PyEval EvalFrame in usr lib libpython2.4.so.1.0) ...","bug"
|
|
"6013","724","divmod <int32scalar>,100) returns nonsense and python segfaults Trac 126).
|
|
Original ticket http projects.scipy.org numpy ticket 126 on 2006 05 19 by pearu, assigned to teoliphant.
|
|
Example demonstrating the bug >>> from numpy import array >>> divmod array [0])[0],10) ) >>> divmod array [1])[0],10) 0,) >>> divmod array [2])[0],10) 0, 0) >>> divmod array [3])[0],10) 0, 0, <refcnt 3075274272 at 0xb74cf780>) >>> divmod array [4])[0],10) Segmentation fault","bug"
|
|
"6013","725","Improve speed of Numeric like indexing Trac 127).
|
|
Original ticket http projects.scipy.org numpy ticket 127 on 2006 05 19 by teoliphant, assigned to unknown.
|
|
Some optimizations have been added for Numeric like indexing.
|
|
The speed could be improved by performing Numeric like indexing first and then resorting to fancy indexing only on error.
|
|
Currently fancy indexing is checked for rather than just letting an error return show the need for it.
|
|
This could also open the way for fancy indexing to be disabled either in a super class or using a flag on the array.","enhancement"
|
|
"6013","726","reproducible crash of multiply.reduce Trac 128).
|
|
Original ticket http projects.scipy.org numpy ticket 128 on 2006 05 22 by trac user martin wiechert, assigned to unknown.
|
|
Below is the line that triggers the crash and a traceback.
|
|
Apparently the number of repetitions necessary can vary, e.g.
|
|
when I import from scipy instead of numpy, I need to repeat six times.
|
|
Python 2.4.3 1, May 12 2006, 05 35 54) [GCC 4.1.0 SUSE Linux)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> multiply.reduceat 15,15,15,15), 0,2)) array [225, 225]) >>> multiply.reduceat 15,15,15,15), 0,2)) glibc detected python free ) invalid pointer 0xb7a2eac0 ======= Backtrace ========= lib libc.so.6[0xb7c1a911] lib libc.so.6 libc free 0x84)[0xb7c1bf84] usr local lib libpython2.4.so.1.0 PyObject Free 0x51)[0xb7e56f31] usr local lib python2.4 site packages numpy core multiarray.so[0xb79e0d97] usr local lib python2.4 site packages numpy core multiarray.so[0xb79f9dca] usr local lib python2.4 site packages numpy core umath.so[0xb7983d9f] usr local lib libpython2.4.so.1.0 PyCFunction Call 0x11d)[0xb7e5364d] usr local lib libpython2.4.so.1.0 PyEval EvalFrame 0x4e8e)[0xb7e8f42e] usr local lib libpython2.4.so.1.0 PyEval EvalCodeEx 0x869)[0xb7e905c9] usr local lib libpython2.4.so.1.0 PyEval EvalCode 0x63)[0xb7e90643] usr local lib libpython2.4.so.1.0 PyRun InteractiveOneFlags 0x1fd)[0xb7eb512d] usr local lib libpython2.4.so.1.0 PyRun InteractiveLoopFlags 0x5b)[0xb7eb526b] usr local lib libpython2.4.so.1.0 PyRun AnyFileExFlags 0x47)[0xb7eb5a87] usr local lib libpython2.4.so.1.0 Py Main 0xbad)[0xb7ebbf3d] python main 0x32)[0x80485e2] lib libc.so.6 libc start main 0xdc)[0xb7bcc87c] python[0x8048521] ======= Memory map ======== 08048000 08049000 r xp 00000000 03 05 205745 usr local bin python 08049000 0804a000 rw p 00000000 03 05 205745 usr local bin python 0804a000 081ad000 rw p 0804a000 00 00 0 [heap] b7000000 b7021000 rw p b7000000 00 00 0 b7021000 b7100000 p b7021000 00 00 0 b71b4000 b7297000 rw p b71b4000 00 00 0 b7297000 b72b2000 r xp 00000000 03 05 212490 usr local lib python2.4 site packages numpy random mtrand.so b72b2000 b72b6000 rw p 0001a000 03 05 212490 usr local lib python2.4 site packages numpy random mtrand.so b72b6000 b72d0000 r xp 00000000 03 05 201845 usr lib libg2c.so.0.0.0 b72d0000 b72d1000 rw p 00019000 03 05 201845 usr lib libg2c.so.0.0.0 b72d1000 b72d4000 rw p b72d1000 00 00 0 b72e2000 b72eb000 r xp 00000000 03 05 212480 usr local lib python2.4 site packages numpy dft fftpack lite.so b72eb000 b72ec000 rw p 00008000 03 05 212480 usr local lib python2.4 site packages numpy dft fftpack lite.so b72ec000 b758c000 r xp 00000000 03 05 212489 usr local lib python2.4 site packages numpy linalg lapack lite.so b758c000 b758e000 rw p 0029f000 03 05 212489 usr local lib python2.4 site packages numpy linalg lapack lite.so b758e000 b75ef000 rw p b758e000 00 00 0 b75ef000 b75f2000 r xp 00000000 03 05 208618 usr local lib python2.4 lib dynload math.so b75f2000 b75f3000 rw p 00002000 03 05 208618 usr local lib python2.4 lib dynload math.so b75f3000 b75f5000 r xp 00000000 03 05 212481 usr local lib python2.4 site packages numpy lib compiled base.so b75f5000 b75f6000 rw p 00002000 03 05 212481 usr local lib python2.4 site packages numpy lib compiled base.so b75f6000 b7610000 r xp 00000000 03 05 212486 usr local lib python2.4 site packages numpy core scalarmath.so b7610000 b7611000 rw p 00019000 03 05 212486 usr local lib python2.4 site packages numpy core scalarmath.so b7611000 b7614000 r xp 00000000 03 05 208625 usr local lib python2.4 lib dynload mmap.so b7614000 b7615000 rw p 00003000 03 05 208625 usr local lib python2.4 lib dynload mmap.so b7615000 b7656000 rw p b7615000 00 00 0 b7656000 b765a000 r xp 00000000 03 05 208644 usr local lib python2.4 lib dynload strop.so b765a000 b765c000 rw p 00003000 03 05 208644 usr local lib python2.4 lib dynload strop.so b765c000 b765f000 r xp 00000000 03 05 208595 usr local lib python2.4 lib dynload cStringIO.so b765f000 b7660000 rw p 00003000 03 05 208595 usr local lib python2.4 lib dynload cStringIO.so b7660000 b7671000 r xp 00000000 03 05 208619 usr local lib python2.4 lib dynload cPickle.so b7671000 b7672000 rw p 00010000 03 05 208619 usr local lib python2.4 lib dynload cPickle.so b7672000 b7964000 r xp 00000000 03 05 212484 usr local lib python2.4 site packages numpy core dotblas.so b7964000 b7966000 rw p 002f1000 03 05 212484 usr local lib python2.4 site packages numpy core dotblas.so b7966000 b798e000 r xp 00000000 03 05 212487 usr local lib python2.4 site packages numpy core umath.so b798e000 b7991000 rw p 00027000 03 05 212487 usr local lib python2.4 site packages numpy core umath.so b7991000 b79d3000 rw p b7991000 00 00 0 b79d3000 b7a28000 r xp 00000000 03 05 212482 usr local lib python2.4 site packages numpy core multiarray.so b7a28000 b7a32000 rw p 00054000 03 05 212482 usr local lib python2.4 site packages numpy core multiarray.so b7a32000 b7a6d000 r xp 00000000 03 05 17777 lib libncurses.so.5.5 b7a6d000 b7a78000 rw p 0003a000 03 05 17777 lib libncurses.so.5.5 b7a78000 b7a79000 rw p b7a78000 00 00 0 b7a79000 b7aba000 r xp 00000000 03 05 17792 usr lib libncursesw.so.5.5 b7aba000 b7ac6000 rw p 00040000 03 05 17792 usr lib libncursesw.so.5.5 b7ac6000 b7af0000 r xp 00000000 03 05 18393 lib libreadline.so.5.1 b7af0000 b7af4000 rw p 0002a000 03 05 18393 lib libreadline.so.5.1 b7af4000 b7af5000 rw p b7af4000 00 00 0 b7af5000 b7af8000 r xp 00000000 03 05 208646 usr local lib python2.4 lib dynload readline.so b7af8000 b7af9000 rw p 000030Aborted","bug"
|
|
"6013","727","line directives do not use correctly escaped paths on windows Trac 129).
|
|
Original ticket http projects.scipy.org numpy ticket 129 on 2006 05 24 by stephan tolksdorf, assigned to unknown.
|
|
The MinGW32 build is currently broken because the line 1 directive which gets injected in umathmodule.c is not properly escaped.
|
|
The backslashes in the path need to be escaped.","bug"
|
|
"6013","728","Inconsistent behaviour between hstack and vstack Trac 130).
|
|
Original ticket http projects.scipy.org numpy ticket 130 on 2006 05 24 by stefanv, assigned to unknown.
|
|
Take two arrays, x of shape 3,3) and y of shape 3,) x = N.array [[1,2,3],[4,5,6],[7,8,9]]) y = N.array [0,0,0]) These can be stacked with vstack, but not with hstack.
|
|
hstack complains with ValueError arrays must have same number of dimensions which does not seem to deter vstack.
|
|
I find vstack s behaviour more useful, and would like it if hstack could behave in the same way.","bug"
|
|
"6013","729","numpy 0.9.8RPM doesn t install on FC5 Trac 131).
|
|
Original ticket http projects.scipy.org numpy ticket 131 on 2006 05 25 by trac user jdpipe, assigned to unknown.
|
|
Hi all The RPM package numpy 0.9.8 1.i586.rpm doesn t install on FC5.
|
|
It has a dependency on python base which doesn t exist on this platform.
|
|
We have python and python devel only.
|
|
Is there any chance you could release a specific version of the RPM that works on Redhat style machines?
|
|
Cheers JP","bug"
|
|
"6013","730","Suggestions to improvements for NumPy s web presence Trac 132).
|
|
Original ticket http projects.scipy.org numpy ticket 132 on 2006 05 31 by edschofield, assigned to unknown.
|
|
NumPy s web presence could be improved by 1.
|
|
Renaming SciPy Core > Numpy in http numeric.scipy.org new features.html 2.
|
|
Pointing www.numpy.org to numeric.scipy.org instead of the SF page 3.
|
|
Renaming numeric.scipy.org to numpy.scipy.org, and configuring numeric.scipy.org with a 301 HTTP redirect to numpy.scipy.org 4.
|
|
Replacing the Apache test page at http projects.scipy.org with a redirect to http scipy.org Developer Zone","enhancement"
|
|
"6013","731","Squeeze behaviour for 1d and 0d arrays Trac 133).
|
|
Original ticket http projects.scipy.org numpy ticket 133 on 2006 05 31 by edschofield, assigned to teoliphant.
|
|
I ve discovered a bug in my SciPy code for sparse matrix slicing that was caused by the following behaviour of squeeze ) >>> a = array [3]) array of shape 1,) >>> type a.squeeze )) <type int32scalar > That is, squeezing a 1 dim array returns an array scalar.
|
|
I think this is surprising, and suggest we change squeeze to return 0 dim arrays here instead.
|
|
Another, related, issue is whether this behaviour >>> b = array 3) 0 dim array >>> type a.squeeze )) <type int32scalar > should also be changed so squeeze leaves 0d arrays alone.
|
|
The current behaviour is less clearly inconsistent than the case above, but this change would better honour the intention stated in the docstring that squeeze merely eliminates any length 1 dimensions.","bug"
|
|
"6013","732","Suggested changes to array creation from sequences Trac 134).
|
|
Original ticket http projects.scipy.org numpy ticket 134 on 2006 05 31 by edschofield, assigned to timleslie.
|
|
NumPy currently does the following >>> s = set [1, 100, 10]) >>> a = numpy.array s) >>> a array set [1, 100, 10]), dtype=object) >>> a.shape ) Many functions in NumPy s functional interface, like numpy.sort ), inherit this behaviour >>> b = numpy.sort s) >>> b array set [1, 100, 10]), dtype=object) >>> b.shape ) I d like to propose two possible modifications to improve array construction from non list sequences 1.
|
|
We inspect whether the data has a len method.
|
|
If it does, and returns an integer, we construct an array out of the C equivalent of list data).
|
|
Others on this list have noted that NumPy also creates a rank 0 object array from generators >>> c = numpy.array i 2 for i in xrange 10)) >>> c array <generator object at 0xb6918b2c>, dtype=object) This proposal wouldn t affect this case, since generators do not in general have a len attribute.
|
|
2.
|
|
A stronger version of the above proposal if the data has an iter method, we construct an array out of the elements.
|
|
Creating an array from an infinite iterator would loop forever, just as list inf iterator) does.
|
|
Currently array iterator) returns a 0 d object array e.g.
|
|
as above), which is less useful and probably unintended.","enhancement"
|
|
"6013","733","Memory de allocation error in scalar types Trac 135).
|
|
Original ticket http projects.scipy.org numpy ticket 135 on 2006 05 31 by stefanv, assigned to unknown.
|
|
Running the test suite on r2554, Valgrind yields the following error ==31956== Invalid free ) delete delete[] ==31956== at 0x401CFCF free vg replace malloc.c 235) ==31956== by 0x48D759B gentype dealloc scalartypes.inc.src 281) ==31956== by 0x80B48C9 PyEval EvalFrame in usr bin python2.4) ==31956== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==31956== by 0x80B6F92 PyEval EvalFrame in usr bin python2.4) ==31956== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==31956== by 0x80FC03C within usr bin python2.4) ==31956== by 0x805946B PyObject Call in usr bin python2.4) ==31956== by 0x80B4AB9 PyEval EvalFrame in usr bin python2.4) ==31956== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==31956== Address 0x4FA5818 is 55,144 bytes inside a block of size 262,144 alloc d ==31956== at 0x401C422 malloc vg replace malloc.c 149) ==31956== by 0x807F3E2 PyObject Malloc in usr bin python2.4) ==31956== by 0x80813E0 PyString FromStringAndSize in usr bin python2.4) ==31956== by 0x80D3509 within usr bin python2.4) ==31956== by 0x80D35D8 within usr bin python2.4) ==31956== by 0x80D32B4 within usr bin python2.4) ==31956== by 0x80D35D8 within usr bin python2.4) ==31956== by 0x80D32A7 within usr bin python2.4) ==31956== by 0x80D3C12 PyMarshal ReadLastObjectFromFile in usr bin python2.4) ==31956== by 0x80D069A within usr bin python2.4)","bug"
|
|
"6013","734","bitwise {and|or}.reduce ) does not function correctly Trac 136).
|
|
Original ticket http projects.scipy.org numpy ticket 136 on 2006 06 02 by trac user bborcic, assigned to unknown.
|
|
The reduce method of bitwise {and|or} takes only the extremities into account, instead of reducing over the entire axis.
|
|
Illustration >>> from numpy import >>> bitwise or.reduce array [1,2,4,8,16])) 17 >>> bitwise or.reduce array [1,16])) 17 >>> bitwise and.reduce array [ 1, 2, 4, 8, 16])) 17","bug"
|
|
"6013","735","Patch use non deprecated interfaces internally Trac 137).
|
|
Original ticket http projects.scipy.org numpy ticket 137 on 2006 06 04 by edschofield, assigned to teoliphant.
|
|
This patch replaces the use of capitalized type names with the newer names internally within !NumPy and replaces uses of isinstance a, !ArrayType) with isinstance a, ndarray).
|
|
Apply with patch p0 < use nondeprecated names.patch in the base numpy directory.
|
|
Patch made against SVN r2580.","enhancement"
|
|
"6013","736","Patch un deprecate functional interfaces Trac 138).
|
|
Original ticket http projects.scipy.org numpy ticket 138 on 2006 06 04 by edschofield, assigned to teoliphant.
|
|
Travis has declared an intention not to remove Numeric like functional interfaces from !NumPy.
|
|
This patch moves the relevant functions from oldnumeric.py, whose contents are declared deprecated, to a new file, fromnumeric.py.
|
|
This patch also adds unit tests for whether oldnumeric.py supports the old capitalized type names.
|
|
Patch against SVN r2580, with patch from ticket 735 applied.
|
|
Apply with patch p0 < undeprecate functional interfaces.patch in the !NumPy source directory.","enhancement"
|
|
"6013","737","Patch un export deprecated names by default Trac 139).
|
|
Original ticket http projects.scipy.org numpy ticket 139 on 2006 06 04 by edschofield, assigned to teoliphant.
|
|
This patch removes the names in oldnumeric.py from the default numpy and numpy.core namespaces.
|
|
It also adds a top level function called Numeric ) that imports the oldnumeric.py names into the numpy namespace, for Numeric compatibility.
|
|
A script is also attached below that adds the line numpy.Numeric ) to all Python files in the specified directory.
|
|
The script could be merged easily into numpy lib convertcode.py, to ease the upgrade path from Numeric to !NumPy.
|
|
With this patch, and after running the script over the !SciPy source tree, !SciPy runs its 1518 unit tests without errors.","enhancement"
|
|
"6013","738","numpy.rec.fromarrays causes interpreter crash Trac 140).
|
|
Original ticket http projects.scipy.org numpy ticket 140 on 2006 06 06 by trac user tvaught, assigned to teoliphant.
|
|
numpy.rec.fromarrays causes the interpreter to crash when provided an object array as one of the arrays in the list.
|
|
It works for small record counts but fails at some point e.g, it works for 10 records C \wrk\svnrepos\travis\code\enfin>python Python 2.3.5 Enthought Edition 0.9.7 62, May 11 2005, 20 02 58) [MSC v.1200 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> count = 10 >>> a = numpy.zeros count, dtype= O ) >>> b = numpy.zeros count, dtype= f8 ) >>> c = numpy.zeros count, dtype= f8 ) >>> d = numpy.zeros count, dtype= f8 ) >>> e = numpy.zeros count, dtype= f8 ) >>> for i in range len a)) ... a[i] = 1,2,3,4,5,6,7,8,9) ... >>> numpy.rec.fromarrays [a,b,c,d,e], names= date,data1,data2,data3,data4 ) recarray [ 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0), 1, 2, 3, 4, 5, 6, 7, 8, 9), 0.0, 0.0, 0.0, 0.0)], dtype=[ date , |O4 ), data1 , <f8 ), data2 , <f8 ), data3 , <f8 ), data4 , <f8 )]) >>> Then it fails for 25 records Python 2.3.5 Enthought Edition 0.9.7 62, May 11 2005, 20 02 58) [MSC v.1200 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> count = 25 >>> a = numpy.zeros count, dtype= O ) >>> b = numpy.zeros count, dtype= f8 ) >>> c = numpy.zeros count, dtype= f8 ) >>> d = numpy.zeros count, dtype= f8 ) >>> e = numpy.zeros count, dtype= f8 ) >>> for i in range len a)) ... a[i] = 1,2,3,4,5,6,7,8,9) ... >>> numpy.rec.fromarrays [a,b,c,d,e], names= date,data1,data2,data3,data4 ) <<<CRASH>>> It only fails when the first column is an object array in this case, a tuple).","bug"
|
|
"6013","739","standard setuptools install idiom fails.
|
|
patch provided Trac 141).
|
|
Original ticket http projects.scipy.org numpy ticket 141 on 2006 06 08 by astraw, assigned to dmcooke.
|
|
A standard setuptools idiom fails with numpy numpy python c import setuptools; execfile setup.py ) install single version externally managed Running from numpy source directory.
|
|
usage c [global opts] cmd1 [cmd1 opts] [cmd2 [cmd2 opts] ...] or c help [cmd1 cmd2 ...] or c help commands or c cmd help error option single version externally managed not recognized The reason is that numpy.distutils.command.install derives from distutils.command.install, and neither recognize the single version externally managed option.
|
|
The following patch fixes the issue.
|
|
Index numpy distutils command install.py =================================================================== numpy distutils command install.py revision 2588) numpy distutils command install.py working copy) 1,4 1,10 from distutils.command.install import install as old install import sys if setuptools in sys.modules import setuptools.command.install as old install mod else import distutils.command.install as old install mod old install = old install mod.install from distutils.file util import write file class install old install)","bug"
|
|
"6013","740","5 functions still don t preserve matrix type Trac 142).
|
|
Original ticket http projects.scipy.org numpy ticket 142 on 2006 06 09 by trac user svetosch, assigned to unknown.
|
|
affected functions diff vstack hstack outer msort symptom given numpy matrices as inputs, these functions still return numpy arrays as opposed to the applicable rest of numpy s functions) Sven >>> n. version 0.9.8 >>> a matrix [[1, 0, 0], [0, 1, 0], [0, 0, 1]]) >>> b matrix [[0, 0, 0], [0, 0, 0]]) >>> n.diff a) array [[ 1, 0], [ 1, 1], [ 0, 1]]) >>> n.outer a,b) array [[0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0], [0, 0, 0, 0, 0, 0]]) >>> n.msort a) array [[0, 0, 0], [0, 0, 0], [1, 1, 1]]) >>> n.vstack [a,b]) array [[1, 0, 0], [0, 1, 0], [0, 0, 1], [0, 0, 0], [0, 0, 0]]) >>> n.hstack [a,b.T]) array [[1, 0, 0, 0, 0], [0, 1, 0, 0, 0], [0, 0, 1, 0, 0]])","bug"
|
|
"6013","741","Setting ufunc identity rases SystemError Trac 143).
|
|
Original ticket http projects.scipy.org numpy ticket 143 on 2006 06 14 by trac user sasha, assigned to unknown.
|
|
>>> add.identity = None Traceback most recent call last) File <stdin> , line 1, in ?
|
|
SystemError error return without exception set If identity is not writable, the exception should be TypeError attribute identity of numpy.ufunc objects is not writable Should we make attribute identity writable?","bug"
|
|
"6013","742","Floating point error mode regress test fails on OpenBSD Trac 144).
|
|
Original ticket http projects.scipy.org numpy ticket 144 on 2006 06 15 by trac user djm, assigned to unknown.
|
|
I receive the following error on OpenBSD with numpy 0.9.8 ====================================================================== FAIL test divideerr numpy.core.tests.test numeric.test seterr) Traceback most recent call last) File usr local lib python2.4 site packages numpy core tests test numeric.py , line 159, in test divideerr self.fail ) AssertionError I have a patch for this which I ll attach in a moment","bug"
|
|
"6013","743","Pointer errors in windows numpy.core Trac 145).
|
|
Original ticket http projects.scipy.org numpy ticket 145 on 2006 06 15 by trac user gregas, assigned to pearu.
|
|
I appologize for the vauguess of this since I m not at all expert with the numpy internals.
|
|
Still, I did not see this reported and it sure is annoying.
|
|
While running within the ActiveState Komodo IDE, I have the following problem I run from numpy.core import print foo If I place a breakpoint at line 2, I get a window popup saying the memory at 0x00000014 cannot be read .
|
|
This crashes the program being tested but not the IDE), but does not allow breakpoint debugging.
|
|
I also get a similiar memory error from from numpy.core import foo ) Where foo ) is undefined.
|
|
This behaviour exists in v 0.94 0.98, but not in v 0.92.
|
|
I would be happy to help with more information if somebody w more experience wants to ask specific questions.
|
|
I don t normally hang around here, but I can be reached at gregasmaximus gmail.com Thanks to all for their hard work on numpy and scipy.
|
|
Greg","bug"
|
|
"6013","744","Incompatible compiler options for fcompiler intelem using ifort v9.0 Trac 146).
|
|
Original ticket http projects.scipy.org numpy ticket 146 on 2006 06 16 by brickZA, assigned to unknown.
|
|
I m using intel f90 version 9.0 on and AMD64 Ubuntu dapper machine.
|
|
Trying to wrap some F90 code with f2py like this f2py fcompiler=intelem m geo c geomwrap.f90 I get ... ifort Command line warning extension M not supported ignored in option x ifort Command line error option arch SSE2 not supported ifort Command line warning extension M not supported ignored in option x ifort Command line error option arch SSE2 not supported error Command usr local bin ifort FR KPIC cm O3 unroll xM arch SSE2 I tmp tmpINquAT src.linux x86 64 2.4 I usr lib python2.4 site packages numpy core include I usr include python2.4 c c geomwrap.f90 o tmp tmpINquAT geomwrap.o module tmp tmpINquAT I tmp tmpINquAT failed with exit status 1 It seems that when using ifort in em64t mode one whould not specify arch SSE2.
|
|
Further testing reveiled that arch SSE is also illegal.
|
|
I attach a patch vs. a recent svn numpy to fix the problem.
|
|
I did this by overriding the get flags arch method in class IntelEM64TFCompiler.
|
|
I also removed checks for a bunch of architectures that could not possibly AFAIK) be em64t, and the xM option that causes a warning.","bug"
|
|
"6013","745","ValueError or segfault wrapping ALLOCATABLE f90 arrays using ifort on AMD64 EM64T Trac 147).
|
|
Original ticket http projects.scipy.org numpy ticket 147 on 2006 06 16 by brickZA, assigned to pearu.
|
|
I m using an AMD64 machine under Ubuntu Dapper with intel fortran v9.0, and numpy SVN 2 2629.
|
|
Wrapping the following code in file geomwrap.f90 MODULE mesh !
|
|
The x,y,z coodirnates of each node in the mesh INTEGER, DIMENSION , ), ALLOCATABLE node coordinates !
|
|
The 4 node indices per element that define all the mesh elements REAL 8), DIMENSION , ), ALLOCATABLE element nodes CONTAINS SUBROUTINE init mesh ) ALLOCATE node coordinates 3,5)) node coordinates ,1) = 0.5, 0.5, 0.5 ) node coordinates ,2) = 0.5, 0.5, 0.5 ) node coordinates ,3) = 0.5, 0.5, 0.5 ) node coordinates ,4) = 0.5, 0.5, 0.5 ) node coordinates ,5) = 0.5, 0.5, 0.5 ) ALLOCATE element nodes 4,2)) element nodes ,1) = 1, 2, 3, 4 ) element nodes ,2) = 1, 3, 4, 5 ) END SUBROUTINE init mesh END MODULE mesh using after applying the patch attached to Ticket 744) f2py fcompiler=intelem m geo c geomwrap.f90 The module seems to build fine, but the data variables can t be accesed.
|
|
If I allocate and initialize the data variables through the init mesh ) F90 subroutine, I get the following when I try to access them In [1] from geo import mesh In [2] mesh.init mesh Out[2] <fortran object at 0x2aaaaaadc148> In [3] mesh.init mesh ) In [4] mesh.element nodes exceptions.ValueError Traceback most recent call last) home brick akademie articles ACES2006 special <ipython console> ValueError negative dimensions are not allowed If I try to initialise them in python, I get a segfault In [1] from geo import mesh In [2] mesh.element nodes In [3] mesh.element nodes=[[1,2],[3,4]] Segmentation fault The exact same numpy version and sofware setup on an i386 machine works fine.
|
|
Also, an older pre numpy version of F2PY 2.46.243 2020) also works fine on the AMD64 machine.
|
|
This seems to be related to this report http cens.ioc.ee pipermail f2py users 2006 June 001264.html","bug"
|
|
"6013","746","PyArray Scalar should use getitem ) Trac 148).
|
|
Original ticket http projects.scipy.org numpy ticket 148 on 2006 06 19 by trac user matthieu.perrot, assigned to unknown.
|
|
In order to access single element of an array, the documentation says that the function getitem ) could be used.
|
|
I want to use it to dynamically build a Python wrapper out of an array of C datas.
|
|
Yet the PyArray Scalar functions doesn t call getitem ) in arrayobject.c lines 983 984) it shows Py INCREF PyObject )data)); return PyObject )data); while i think it should be return descr >f >getitem data, base); Changing that line works for me.
|
|
Thanks","bug"
|
|
"6013","747","error in compaq.py Trac 149).
|
|
Original ticket http projects.scipy.org numpy ticket 149 on 2006 06 20 by trac user bhoel, assigned to unknown.
|
|
Trying to compile numpy I get python Python 2.4.2 46, Sep 30 2005, 09 57 37) [MSC v.1200 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
python setup.py config fc fcompiler=compaqv build ...
|
|
Generating build\src\numpy\core\config.h No module named msvccompiler in numpy.distutils, trying from distutils.. 0 1 Traceback most recent call last) File setup.py , line 76, in ?
|
|
setup package ) File setup.py , line 69, in setup package setup config.todict ) ) File w \hoel\work\numpy 0.9.6\numpy\distutils\core.py , line 85, in setup return old setup new attr) File w \hoel\work\Python 2.4.2\dist\lib\distutils\core.py , line 149, in setup dist.run commands ) File w \hoel\work\Python 2.4.2\dist\lib\distutils\dist.py , line 946, in run commands self.run command cmd) File w \hoel\work\Python 2.4.2\dist\lib\distutils\dist.py , line 966, in run command cmd obj.run ) File w \hoel\work\Python 2.4.2\dist\lib\distutils\command\build.py , line 112, in run self.run command cmd name) File w \hoel\work\Python 2.4.2\dist\lib\distutils\cmd.py , line 333, in run command self.distribution.run command command) File w \hoel\work\Python 2.4.2\dist\lib\distutils\dist.py , line 966, in run command cmd obj.run ) File w \hoel\work\numpy 0.9.6\numpy\distutils\command\build src.py , line 84, in run self.build sources ) File w \hoel\work\numpy 0.9.6\numpy\distutils\command\build src.py , line 99, in build sources self.build extension sources ext) File w \hoel\work\numpy 0.9.6\numpy\distutils\command\build src.py , line 209, in build extension sources sources = self.generate sources sources, ext) File w \hoel\work\numpy 0.9.6\numpy\distutils\command\build src.py , line 267, in generate sources source = func extension, build dir) File numpy\core\setup.py , line 35, in generate config h library dirs = default lib dirs) File c \Python24\lib\distutils\command\config.py , line 278, in try run self.
|
|
check compiler ) File w \hoel\work\numpy 0.9.6\numpy\distutils\command\config.py , line 34, in check compiler dry run=self.dry run, force=1) File w \hoel\work\numpy 0.9.6\numpy\distutils\fcompiler\ init .py , line 641, in new fcompiler import module name) File w \hoel\work\numpy 0.9.6\numpy\distutils\fcompiler\compaq.py , line 47, in ?
|
|
class CompaqVisualFCompiler FCompiler) File w \hoel\work\numpy 0.9.6\numpy\distutils\fcompiler\compaq.py , line 66, in CompaqVisualFCompiler ar exe = MSVCCompiler ).lib AttributeError MSVCCompiler instance has no attribute lib This error can be fixed in compaq.py diff numpy distutils fcompiler compaq.py numpy distutils fcompiler compaq.py 66c66,68 < ar exe = MSVCCompiler ).lib > m = MSVCCompiler ) > m.initialize ) > ar exe = m.lib 86a89,97 > def get cc args self, pp opts, debug, before) > cc args = pp opts > if debug > cc args[ 0] = self.get flags debug ) > if before > cc args[ 0] = before > return cc args > >","bug"
|
|
"6013","748","Cant compile NumPy using VC 6.0 Trac 150).
|
|
Original ticket http projects.scipy.org numpy ticket 150 on 2006 06 20 by trac user bhoel, assigned to unknown.
|
|
Adding the fix from Ticket 747 to my numpy.distutils does not take me much further in the numpy build process.
|
|
I get python Python 2.4.2 46, Sep 30 2005, 09 57 37) [MSC v.1200 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
python setup.py config fc fcompiler=compaqv build ... C \PROGRA 1\MICROS 4\VC98\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG Ibuild\src.win32 2.4\numpy\core\src Inumpy\core\include Ibuild\src.win32 2.4\numpy\core Inumpy\core\src Inumpy\core\include Ic \Python24\include Ic \Python24\PC Tcnumpy\core\src\multiarraymodule.c Fobuild\temp.win32 2.4\Release\numpy\core\src\multiarraymodule.obj multiarraymodule.c build\src.win32 2.4\numpy\core\src\scalartypes.inc 1) warning C4129 c unrecognized character escape sequence build\src.win32 2.4\numpy\core\src\scalartypes.inc 1) warning C4129 s unrecognized character escape sequence build\src.win32 2.4\numpy\core\src\scalartypes.inc 1) warning C4129 s unrecognized character escape sequence build\src.win32 2.4\numpy\core\src\arraytypes.inc 1) warning C4129 c unrecognized character escape sequence build\src.win32 2.4\numpy\core\src\arraytypes.inc 1) warning C4129 s unrecognized character escape sequence numpy\core\src\arrayobject.c 1346) warning C4018 < signed unsigned mismatch numpy\core\src\arrayobject.c 4779) warning C4244 = conversion from int64 to int , possible loss of data numpy\core\src\arrayobject.c 4807) warning C4244 = conversion from int64 to int , possible loss of data numpy\core\src\multiarraymodule.c 1600) warning C4305 initializing truncation from const int to char numpycoresrcrraytypes.inc.src 571) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpycoresrcrraytypes.inc.src 571) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpycoresrcrraytypes.inc.src 571) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpycoresrcrraytypes.inc.src 631) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpycoresrcrraytypes.inc.src 631) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpycoresrcrraytypes.inc.src 631) error C2520 conversion from unsigned int64 to double not implemented, use signed int64 numpy\core\src\multiarraymodule.c 1700) warning C4761 integral size mismatch in argument; conversion supplied numpy\core\src\multiarraymodule.c 1700) warning C4761 integral size mismatch in argument; conversion supplied error Command C \PROGRA 1\MICROS 4\VC98\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG Ibuild\src.win32 2.4\numpy\core\src Inumpy\core\include Ibuild\src.win32 2.4\numpy\core Inumpy\core\src Inumpy\core\include Ic \Python24\include Ic \Python24\PC Tcnumpy\core\src\multiarraymodule.c Fobuild\temp.win32 2.4\Release\numpy\core\src\multiarraymodule.obj failed with exit status 2","bug"
|
|
"6013","749","binary repr breaks for input 0 Trac 151).
|
|
Original ticket http projects.scipy.org numpy ticket 151 on 2006 06 20 by stefanv, assigned to stefanv.
|
|
N.binary repr 0) raises an error in r2651.","bug"
|
|
"6013","750","unique should return sorted arrays Trac 152).
|
|
Original ticket http projects.scipy.org numpy ticket 152 on 2006 06 23 by huard, assigned to teoliphant.
|
|
unique x) sould return a sorted array instead of letting dict.keys ) dictate the order in which the array is returned.
|
|
Arguments in favor of this are Consistency with matlab.
|
|
Return identical results each time unique is called.","enhancement"
|
|
"6013","751","scipy 0.4.9 bdist rpm build fails Trac 153).
|
|
Original ticket http projects.scipy.org numpy ticket 153 on 2006 06 24 by trac user ckkart, assigned to unknown.
|
|
Building a scipy rpm of scipy 0.4.9 fails.
|
|
A normal build however runs without problems.
|
|
python setup.py bdist rpm yields the following error building extension scipy.fftpack.
|
|
fftpack sources Traceback most recent call last) File setup.py , line 50, in ?
|
|
setup package ) File setup.py , line 42, in setup package configuration=configuration ) File usr lib python2.4 site packages numpy distutils core.py , line 170, in setup return old setup new attr) File usr lib python2.4 distutils core.py , line 149, in setup dist.run commands ) File usr lib python2.4 distutils dist.py , line 946, in run commands self.run command cmd) File usr lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr lib python2.4 distutils command build.py , line 112, in run self.run command cmd name) File usr lib python2.4 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr lib python2.4 site packages numpy distutils command build src.py , line 87, in run self.build sources ) File usr lib python2.4 site packages numpy distutils command build src.py , line 106, in build sources self.build extension sources ext) File usr lib python2.4 site packages numpy distutils command build src.py , line 218, in build extension sources sources = self.f2py sources sources, ext) File usr lib python2.4 site packages numpy distutils command build src.py , line 412, in f2py sources raise ValueError r missing target file,)) ValueError fftpackmodule.c missing error Bad exit status from home ck testarea rpm tmp rpm tmp.24879 build) RPM build errors Bad exit status from home ck testarea rpm tmp rpm tmp.24879 build) error command rpmbuild failed with exit status 1","bug"
|
|
"6013","752","dtype fields contains invalid item Trac 154).
|
|
Original ticket http projects.scipy.org numpy ticket 154 on 2006 06 24 by alberts, assigned to teoliphant.
|
|
Consider the following example In [25] svm node dtype = \ .... N.dtype { names [ index , value ], .... formats [N.intc, N.float64]}, .... align=1) In [26] svm node dtype.fields.items ) Out[26] [ index , dtype <i4 ), 0)), 1, index , value )), value , dtype <f8 ), 8))] What is the field with key ` 1` doing in there?
|
|
In [27] N. version Out[27] 0.9.9.2674","bug"
|
|
"6013","753","RandomState.randn ) sometimes skips Trac 155).
|
|
Original ticket http projects.scipy.org numpy ticket 155 on 2006 06 24 by rkern, assigned to rkern.
|
|
Using this function with a seed of [11,21,699,1] from numpy import sum from numpy.random import RandomState def f seed) rs = RandomState seed) a = rs.rand 100) b = rs.randn 100) print a print b return sum a b) one can get several different responses 41.117761294827254, 40.397043045651962, 43.218652895636772.
|
|
The latter two come from a shifting of the gaussian results to the right by one element.
|
|
That is, respectively, the `b` arrays start like this respectively [ 1.3137503 1.09689975 0.86098048 [ 0.23073286 1.3137503 1.09689975 [ 3.05234271 1.3137503 1.09689975 The bug is that the `init by array )` function that seeds the !RandomState did not properly initialize the `has gauss` and `has binomial` flags.","bug"
|
|
"6013","754","numpy.dot A,B) is broken when A is not contiguous and B contains object O ) arrays Trac 156).
|
|
Original ticket http projects.scipy.org numpy ticket 156 on 2006 06 25 by fperez, assigned to unknown.
|
|
The attached script contains a full example reproducing the problem.","bug"
|
|
"6013","755","overflow in PyArray NEW allocation value leads to segfault Trac 157).
|
|
Original ticket http projects.scipy.org numpy ticket 157 on 2006 06 26 by bthyreau, assigned to unknown.
|
|
Hi, When creating arrays at C level, the allocation test sometimes fail probably because of an overflow.
|
|
Altough it should return a NULL pointer, the PyArray New function returns a dummy value.
|
|
To reproduce, try static PyObject enlargeYourArray PyObject self, PyObject null) { PyObject a = NULL; intp dims = intp ) malloc sizeof intp) 3); dims[0] = 16777216; dims[1] = 24; dims[2] = 1; a = PyArray SimpleNew 3, dims, PyArray DOUBLE); printf a = p\n , a); should be nil) return PyArray Return PyArrayObject ) a); } then in Python, you might want to call enlargeYourArray )[ 1, 1, 1] = 42 ) If it doesn t crash ie.
|
|
correctly returns NULL), try with some even larger values.
|
|
I m on IA32, Fedora4.
|
|
Note that in pure python, the arrays correctly raise a MemoryError exception regardless of how huge are the input dimensions.
|
|
Thanks","bug"
|
|
"6013","756","Memory allocation errors in arrayobject Trac 158).
|
|
Original ticket http projects.scipy.org numpy ticket 158 on 2006 06 26 by stefanv, assigned to teoliphant.
|
|
Caught in r2680 with numpy.test level=2) the test added for ticket 754).
|
|
Valgrind reports ==28011== 40 bytes in 1 blocks are possibly lost in loss record 14 of 134 ==28011== at 0x401C422 malloc vg replace malloc.c 149) ==28011== by 0x48E4DFB array alloc arrayobject.c 5955) ==28011== by 0x48EEBF4 PyArray NewFromDescr arrayobject.c 4703) ==28011== by 0x49160EB array subscript simple arrayobject.c 2414) ==28011== by 0x491D7DC array subscript arrayobject.c 2522) ==28011== by 0x491DEAC array subscript nice arrayobject.c 2707) ==28011== by 0x80B3E14 PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==28011== by 0x80FC03C within usr bin python2.4) ==28011== by 0x805946B PyObject Call in usr bin python2.4) ==28011== ==28011== 672 bytes in 1 blocks are possibly lost in loss record 97 of 134 ==28011== at 0x401C422 malloc vg replace malloc.c 149) ==28011== by 0x48EC896 PyArray IterNew arrayobject.c 7671) ==28011== by 0x48F6F6E flat copyinto arrayobject.c 7710) ==28011== by 0x48F7453 PyArray Flatten multiarraymodule.c 336) ==28011== by 0x49139C6 array ravel multiarraymodule.c 195) ==28011== by 0x80B62C6 PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==28011== by 0x80B6F92 PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B703A PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B703A PyEval EvalFrame in usr bin python2.4) ==28011== ==28011== at 0x401C422 malloc vg replace malloc.c 149) ==28011== by 0x48EC896 PyArray IterNew arrayobject.c 7671) ==28011== by 0x48F6F6E flat copyinto arrayobject.c 7710) ==28011== by 0x48F7453 PyArray Flatten multiarraymodule.c 336) ==28011== by 0x48F7506 array flatten arraymethods.c 1497) ==28011== by 0x80B62C6 PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==28011== by 0x80B6F92 PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B703A PyEval EvalFrame in usr bin python2.4) ==28011== by 0x80B703A PyEval EvalFrame in usr bin python2.4)","bug"
|
|
"6013","757","Views with int8 return invalid values Trac 159).
|
|
Original ticket http projects.scipy.org numpy ticket 159 on 2006 06 26 by alberts, assigned to teoliphant.
|
|
Consider the following example In [4] x = N.array [1,2,3,4,5,6],dtype=N.int32) In [5] y = x.view [ a , <i4 ), b , <i4 )]) In [6] y Out[6] array [ 1, 2), 3, 4), 5, 6)], dtype=[ a , <i4 ), b , <i4 )]) Everything works as expected.
|
|
Now with 1 byte ints In [1] x = N.array [1,2,3,4,5,6],dtype=N.int8) In [2] y = x.view [ a , <i1 ), b , <i1 )]) In [3] y Out[3] array [ \x01 , 2), \x03 , 4), \x05 , 6)], dtype=[ f1 , |S1 ), f2 , |i1 )]) Stefan noticed that changing the second field s name makes the problem go away In [8] x = N.array [1,2,3,4,5,6],dtype=N.int8) In [9] y = x.view [ a , <i1 ), bb , <i1 )]) In [10] y Out[10] array [ 1, 2), 3, 4), 5, 6)], dtype=[ a , |i1 ), bb , |i1 )]) For your reference In [11] N. version Out[11] 0.9.9.2674","bug"
|
|
"6013","758","iter ) support to core.records.record Trac 160).
|
|
Original ticket http projects.scipy.org numpy ticket 160 on 2006 06 26 by trac user parejkoj, assigned to unknown.
|
|
I recently posted on the Numpy discussion list, a bug feature request for iteration over individual records in a recarray http sourceforge.net mailarchive forum.php?thread id=14661520 forum id=4890 Travis Oliphant responded That is probably reasonable, but as yet is unsupported.
|
|
I am marking this an enhancement, even though I consider it a bug one should be able to iterate over an array, which is what a record behaves as.
|
|
Though probably not ideal, the following addition to core records.py does the job 137,6 137,12 raise AttributeError, record object has no \ attribute s attr def iter self) counter = 0 while counter < len self) yield self[counter] counter = 1 The recarray is almost identical to a standard array which supports named fields already) The biggest difference is that it can use attribute lookup to find the fields and it is constructed using","enhancement"
|
|
"6013","759","2D histogram Trac 161).
|
|
Original ticket http projects.scipy.org numpy ticket 161 on 2006 06 29 by huard, assigned to unknown.
|
|
This function returns a 2D histogram from arrays of samples.
|
|
There is no loop in the script so its pretty fast.
|
|
x =randn 1000) y = randn 1000) 2 H, xe, ye = histogram2d x,y, 10,10), normed=True)","enhancement"
|
|
"6013","760","problems with numpy.linalg.svd Trac 162).
|
|
Original ticket http projects.scipy.org numpy ticket 162 on 2006 06 30 by tovrstra, assigned to unknown.
|
|
Hi, This simple example should always work, since svd is supposed to be) a very stable algorithm.
|
|
import numpy a = numpy.random.uniform 1.0, 1.0, 160,260)) print numpy.linalg.svd a) For numpy 0.9.8 with LAPACK atlas 3.6.0), the output is On entry to DGESDD parameter number 12 had an illegal value and the script crashes.
|
|
For the latest development version of numpy revision 2707), also compiled with LAPACK atlas 3.6.0), the problem is the same When numpy is compiled without LAPACK export ATLAS=None export LAPACK=None python setup.py install version 0.9.8 hangs on the example program with 100 CPU usage.
|
|
I could not interrupt the test program with ctrl C. The latest development version without LAPACK) works perfect.
|
|
The problem vanishes for smaller array sizes.
|
|
To verify the resulting matrices, one can use this code import numpy a = numpy.random.uniform 1.0, 1.0, 160,260)) u,s,vh = numpy.linalg.svd a, full matrices=True) if a.shape[1] > a.shape[0] error1 = sum numpy.dot u s,vh[ a.shape[0]]) a).ravel ) 2) else error1 = sum numpy.dot u[ , a.shape[1]] s,vh) a).ravel ) 2) print error1 , error1 error2 = sum numpy.dot u.transpose ), u) numpy.identity a.shape[0], float)).ravel ) 2) print error2 , error2 error3 = sum numpy.dot vh.transpose ), vh) numpy.identity a.shape[1], float)).ravel ) 2) print error3 , error3 where the three errors should be very small <1e 20).
|
|
This verification works for the latest development release compiled without lapack.
|
|
There seems to be a small typo in the svd documentation.
|
|
The line a == dot u,dot S,vh)) should be replace by if a.shape[1] > a.shape[0] a == numpy.dot u s,vh[ a.shape[0]]) else a == numpy.dot u[ , a.shape[1]] s,vh)","bug"
|
|
"6013","761","numpy.testing.assert equal does not handle ndarrays in tuples Trac 163).
|
|
Original ticket http projects.scipy.org numpy ticket 163 on 2006 06 30 by brickZA, assigned to unknown.
|
|
If an ndarray is contained in a tuple rather than a list, assert equal won t work, since it does not realise it needs to recurse on the tuple elements.","bug"
|
|
"6013","762","Patch to build numpy 0.9.8 on Windows X64 AMD64) with MSVS2005 Trac 164).
|
|
Original ticket http projects.scipy.org numpy ticket 164 on 2006 07 01 by trac user GerardVermeulen, assigned to unknown.
|
|
I am assembling a Python system for Windows X64 AMD64) built with MS Visual Studio 2005 and I have patched Python ask, if you need the patches for Python) and numpy.
|
|
Comments on the patch for numpy 0.9.8 below 1. setup.py this system has no fabsf, rintf, and hypotf 2. arrayobject.c fixes a typo 3. umathmodule.c.src implements the missing fabsf, rintf, and hypotf numpy 0.9.8 numpy core setup.py.win64 2006 06 25 18 26 56.984375000 0200 numpy 0.9.8 numpy core setup.py 2006 06 25 18 46 24.375000000 0200 75,6 75,12 moredefs.append HAVE ISINF ) if config cmd.check func rint , kws args) moredefs.append HAVE RINT ) if config cmd.check func fabsf , kws args) moredefs.append HAVE FABSF ) if config cmd.check func rintf , kws args) moredefs.append HAVE RINTF ) if config cmd.check func hypotf , kws args) moredefs.append HAVE HYPOTF ) if sys.version[ 3] < 2.4 kws args[ headers ].append stdlib.h ) numpy 0.9.8 numpy core src arrayobject.c.win64 2006 06 25 18 23 48.765625000 0200 numpy 0.9.8 numpy core src arrayobject.c 2006 06 25 18 24 29.781250000 0200 451,7 451,7 elif SIZEOF INTP == SIZEOF INT descr = INT Descr; else descr = LONGLONG DESCR; descr = LONGLONG Descr; endif arr = NULL; numpy 0.9.8 numpy core src umathmodule.c.src.win64 2006 05 13 03 56 38.000000000 0200 numpy 0.9.8 numpy core src umathmodule.c.src 2006 06 25 18 43 48.578125000 0200 216,6 216,15 endif endif ifndef HAVE FABSF ifdef fabsf undef fabsf endif static float fabsf float x) { return float)fabs double) x)); } endif ifdef HAVE HYPOT if !defined NeXT) !defined MSC VER) 242,6 251,12 } endif ifndef HAVE HYPOTF static float hypotf float x, float y) { return float)hypot float) x), float) y)); } endif ifndef HAVE RINT static double 266,7 281,13 } endif ifndef HAVE RINTF static float rintf float x) { return float)rint double) x); } endif Define isnan, isinf, isfinite, signbit if needed Use fpclassify if possible","enhancement"
|
|
"6013","763","a.cumsum dtype=xxx) results bad when len a) > 10001 Trac 165).
|
|
Original ticket http projects.scipy.org numpy ticket 165 on 2006 07 03 by trac user cpl, assigned to unknown.
|
|
cumsum with any meaningful dtype, on any array larger than 10001 items, has incorrect output.
|
|
For the 10002 element case, the [1] and [ 1] entries are wrong; for a 10003 element array, items 1,2, 2, 1) are bad.
|
|
I briefly looked for PyArray BUFSIZE related botches, but am confused enough by the code flow to not have gotten very far.
|
|
Python 2.4.3 1, Apr 3 2006, 18 07 14) [GCC 4.0.1 Apple Computer, Inc. build 5247)] on darwin Type help , copyright , credits or license for more information.
|
|
>>> import numpy >>> numpy.
|
|
version 0.9.9.2707 >>> a = numpy.arange 10002) >>> a.cumsum dtype= f8 ) array [ 0.00000000e 00, 1.00010000e 04, 3.00000000e 00, ..., 4.99950000e 07, 5.00050000e 07, 0.00000000e 00]) >>> a.cumsum dtype= i2 ) array [ 0, 10001, 3, ..., 8968, 1032, 0], dtype=int16) >>> b = numpy.arange 30000) >>> a.cumsum dtype= f8 ) array [ 0.00000000e 00, 1.00010000e 04, 3.00000000e 00, ..., 4.99950000e 07, 5.00050000e 07, 0.00000000e 00])","bug"
|
|
"6013","764","assert array equals breaks with record array Trac 166).
|
|
Original ticket http projects.scipy.org numpy ticket 166 on 2006 07 04 by alberts, assigned to unknown.
|
|
Code from numpy.testing import assert array equal import numpy as N dt = N.dtype { names [ x , y ], formats [N.intc, N.intc]}) x = N.array [ 0,0)], dtype=dt) assert array equal x, x) Traceback Traceback most recent call last) File C \home\albert\work2\scipy\Lib\sandbox\svm\k.py , line 6, in ?
|
|
assert array equal x, x) File c \Python24\Lib\site packages\numpy\testing\utils.py , line 215, in assert array equal verbose=verbose, header= Arrays are not equal ) File c \Python24\Lib\site packages\numpy\testing\utils.py , line 198, in assert array compare reduced = comparision x, y).ravel ) AttributeError bool object has no attribute ravel In [23] N. version Out[23] 0.9.9.2731","bug"
|
|
"6013","765","Patch for histogram2d test Trac 167).
|
|
Original ticket http projects.scipy.org numpy ticket 167 on 2006 07 04 by alberts, assigned to unknown.
|
|
Patch from David Huard.","bug"
|
|
"6013","766","Docstring for transpose is wrong ) Trac 168).
|
|
Original ticket http projects.scipy.org numpy ticket 168 on 2006 07 05 by baxissimo, assigned to unknown.
|
|
See http svn.scipy.org svn numpy trunk numpy core src arraymethods.c search for doc transpose .
|
|
Replace example portion with Example \n >>> a = array [[1,2],[3,4]])\n >>> a\n array [[1, 2],\n [3, 4]])\n >>> a.transpose )\n array [[1, 3],\n [2, 4]])\n >>> a.transpose 1,0))\n array [[1, 3],\n [2, 4]])\n >>> a.transpose 1,0)\n array [[1, 3],\n [2, 4]])\n ;","bug"
|
|
"6013","767","meshgrid changes shapes of input arguments Trac 169).
|
|
Original ticket http projects.scipy.org numpy ticket 169 on 2006 07 05 by stefanv, assigned to stefanv.
|
|
Under r2727, this snippet import numpy as N xcoords=N.linspace 2,2,10) ycoords=N.linspace 1,1,10) print xcoords.shape print ycoords.shape X1,Y1 = N.meshgrid xcoords,ycoords) print xcoords.shape print ycoords.shape generates 10,) 10,) 1, 10) 10, 1) Looks like meshgrid is changing the shape of the input arguments.","bug"
|
|
"6013","768","svn snapshot version numbers don t sort properly with release versions Trac 170).
|
|
Original ticket http projects.scipy.org numpy ticket 170 on 2006 07 06 by astraw, assigned to unknown.
|
|
Currently, we the svn snapshot of numpy has version 0.9.9.2631 , while the next release will be 0.9.9 .
|
|
Ignoring for the moment that we may skip ahead to 1.0.)
|
|
Yet, all automatic verion dection schemes I know of would sort this in the opposite order e.g.
|
|
setuptools, Debian s dpkg system).
|
|
I propose that, after the next release, we don t bump the major.minor version number until the following release.
|
|
Thus, what I m proposing would end up with a timeline something like the following 0.9.9.2631 < relatively recent snapshot post 0.9.8, pre 0.9.9) 0.9.9 < release 0.9.9.3295 < future snapshot post 0.9.9) 1.0 < celebration time Another alternative, which I don t like as much, is to have the versions as they are now, but prefix the svn number with pre , so the snapshots would be 0.9.9pre2631 before the 0.9.9 release.
|
|
This has the advantage, however, of maintaining greatest consistency.
|
|
The reason I don t like this approach is that Debian s dpkg wouldn t get it right dpkg compare versions 0.9.9 lt 0.9.9pre2631 echo yes yes Whatever we do, we should do the same for scipy, also.
|
|
It has the same issue.","bug"
|
|
"6013","769","NumpyTest s test and testall methods return emtpy test suites when level is negative Trac 171).
|
|
Original ticket http projects.scipy.org numpy ticket 171 on 2006 07 06 by alberts, assigned to stefanv.
|
|
Code to reproduce In [1] import numpy In [2] from numpy.testing import NumpyTest In [7] NumpyTest numpy).testall level=9999) ...
|
|
Ran 452 tests in 0.641s OK Out[7] <unittest.TextTestRunner object at 0x01EA70D0> In [8] NumpyTest numpy).testall level= 9999) Out[8] <unittest.TestSuite tests=[]> According to the documentation, `level < 0` should scan for tests of `level=abs level)`.
|
|
The `abs level)` stuff should probably happen at the point where level is passed to ` get method names )`.
|
|
Note that `test )` and `testall )` are broken.
|
|
Tested with In [9] numpy.
|
|
version Out[9] 0.9.9.2750","bug"
|
|
"6013","770","Memory leak when pickling masked arrays Trac 172).
|
|
Original ticket http projects.scipy.org numpy ticket 172 on 2006 07 06 by alberts, assigned to unknown.
|
|
According to Valgrind 3.1.0, the following code from the test of pickling of masked arrays leaks memory from numpy.core.ma import arange, masked import pickle x = arange 12) x[4 10 2] = masked x = x.reshape 4,3) s = pickle.dumps x) y = pickle.loads s) Valgrind says ==1223== 8 bytes in 2 blocks are definitely lost in loss record 2 of 24 ==1223== at 0x40051F9 malloc vg replace malloc.c 149) ==1223== by 0x4736D52 PyArray IntpConverter multiarraymodule.c 3608) ==1223== by 0x5F62D23 within usr lib libpython2.4.so.1.0) ==1223== by 0x5F6424F within usr lib libpython2.4.so.1.0) ==1223== by 0x5F64660 PyArg ParseTuple in usr lib libpython2.4.so.1.0) ==1223== by 0x473CBDE array reconstruct multiarraymodule.c 5820) ==1223== by 0x5F15D2C PyCFunction Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with !NumPy 0.9.9.2763.","bug"
|
|
"6013","771","strange shape mismatch error when trying to assign values of an array Trac 173).
|
|
Original ticket http projects.scipy.org numpy ticket 173 on 2006 07 07 by trac user nich2o, assigned to unknown.
|
|
The following code was working with r2621 but doesn t work with r2763 import numpy a = numpy.empty 3,3,4)) a[ ] = numpy.arange 3 3 4) ValueError shape mismatch objects cannot be broadcast to a single shape I am using this kind of assignation in a super fast 2d convolution algorithm based on dotblas list comprehension and reshaping at every iteration is very slow... Is this behavior wanted needed ?","bug"
|
|
"6013","772","Supply docstrings for every function Trac 174).
|
|
Original ticket http projects.scipy.org numpy ticket 174 on 2006 07 07 by huard, assigned to unknown.
|
|
Every function in the numpy namespace should have a docstring.
|
|
To get a list of functions that miss a docstring, run the following code import numpy as np D = np.
|
|
dict ls = {} nodoc = [] for k in D.keys ) if hasattr D[k], call ) try ls[k] = len D[k].
|
|
doc ) except TypeError nodoc.append k) print nodoc","enhancement"
|
|
"6013","773","inappropriate mask set when raising 0 to a positive floating point power Trac 175).
|
|
Original ticket http projects.scipy.org numpy ticket 175 on 2006 07 10 by trac user fitz, assigned to unknown.
|
|
For MAs, raising 0.
|
|
1. sets the mask instead of returing zero.
|
|
See [http sourceforge.net mailarchive forum.php?thread id=14661503 forum id=4890 this] thread, which includes a patch.","bug"
|
|
"6013","774","Incorrect docstring for argsort, take Trac 176).
|
|
Original ticket http projects.scipy.org numpy ticket 176 on 2006 07 11 by stefanv, assigned to unknown.
|
|
The docstring for argsort states argsort a,axis= 1) return the indices into a of the sorted array along the given axis, so that take a,result,axis) is the sorted array.
|
|
however, `take a,a.argsort ))` doesn t work as expected.
|
|
To complicate matters, `take` doesn t have a docstring, which should be remedied.
|
|
See http numeric.scipy.org numpydoc numpy 9.html pgfId 36425 for more documentation on `take`.","bug"
|
|
"6013","775","matrixmultiply cornercase Trac 177).
|
|
Original ticket http projects.scipy.org numpy ticket 177 on 2006 07 12 by trac user martin wiechert, assigned to unknown.
|
|
The product of an n by 0 matrix A and a 0 by m matrix B, which should be an n by m zero matrix, is not computed correctly.
|
|
Indeed AB),,i, j,, = Sum,,k,, A,,i, k,, B,,k, j,, by definition.
|
|
In the cornercase the rhs sum is empty, thus zero.
|
|
>>> import numpy >>> numpy.version.version 0.9.9.2547 >>> a = numpy.zeros 2,0)) >>> numpy.matrixmultiply a, numpy.transpose a)) array [[ 0, 1211035256], [ 0, 0]])","bug"
|
|
"6013","776","Bad results of SVD on some non contiguous?)
|
|
arrays Trac 178).
|
|
Original ticket http projects.scipy.org numpy ticket 178 on 2006 07 13 by bthyreau, assigned to unknown.
|
|
Hi, The svd functions may return fully wrong values on the svn trunk Here 0.9.9.2800).
|
|
It worked with 0.9.8, still.
|
|
Short script to trigger the problem from numpy import from numpy.linalg import svd from numpy.random import rand K = rand 10003 2).reshape 10003, 2).astype float32) K3 = K[ ,0 1] r1 = svd K3.copy ), 0)[0] r2 = svd K3, 0)[0] r1 == r2).all ) The last line should return True, but returns False presently.
|
|
Note that strangely enough, it starts failing for me at sizes of about 10010.
|
|
Thanks","bug"
|
|
"6013","777","Issue with fancy indexing assignment and complex dtypes Trac 179).
|
|
Original ticket http projects.scipy.org numpy ticket 179 on 2006 07 14 by alberts, assigned to teoliphant.
|
|
The following assignment to a basic array indexed with a list works import numpy as N y = N.zeros 2,), N.intc) y[0], y[1] = 1, 1 y[[0,1]] = 2, 2 But with a more complex dtype, assignment fails import numpy as N dt = N.dtype { names [ i , j ], formats [N.intc, N.intc]}) x = N.zeros 2,), dt) x[0], x[1] = 1, 1), 1, 1) x[[0,1]] = 2, 2), 2, 2) With !NumPy 0.9.9.2813 I get the following Traceback most recent call last) File numpyidx.py , line 5, in ?
|
|
x[[0,1]] = 2, 2), 2, 2) TypeError an integer is required By the way, this still works though x[[0,1]][ i ] = 2, 2 x[[0,1]][ j ] = 2, 2","bug"
|
|
"6013","778","AttributeError numpy.dtype object has no attribute dtype Trac 180).
|
|
Original ticket http projects.scipy.org numpy ticket 180 on 2006 07 14 by nilswagner01, assigned to dmcooke.
|
|
File test pinv.py , line 4, in ?
|
|
b = linalg.pinv a) File usr lib64 python2.4 site packages numpy linalg linalg.py , line 426, in pinv if issubclass a.dtype.dtype, complexfloating) AttributeError numpy.dtype object has no attribute dtype","bug"
|
|
"6013","779","FAIL check csingle numpy.linalg.tests.test linalg.test det) Trac 181).
|
|
Original ticket http projects.scipy.org numpy ticket 181 on 2006 07 14 by nilswagner01, assigned to dmcooke.
|
|
This failure is restricted to 32 bit machines.
|
|
I cannot reproduce it on 64 bit machines.
|
|
====================================================================== FAIL check csingle numpy.linalg.tests.test linalg.test det) Traceback most recent call last) File usr lib python2.4 site packages numpy linalg tests test linalg.py , line 33, in check csingle self.do a, b) File usr lib python2.4 site packages numpy linalg tests test linalg.py , line 75, in do assert almost equal d, multiply.reduce ev)) File usr lib python2.4 site packages numpy linalg tests test linalg.py , line 17, in assert almost equal old assert almost equal a, b, decimal=decimal, kw) File usr lib python2.4 site packages numpy testing utils.py , line 154, in assert almost equal assert round abs desired actual),decimal) == 0, msg AssertionError Items are not equal ACTUAL 8.8817841970012523e 16 4j) DESIRED 4.1718351440067636e 08 4j) Ran 420 tests in 1.319s FAILED failures=1) <unittest.TextTestRunner object at 0xb1ef56cc> >>> numpy.
|
|
version 0.9.9.2821","bug"
|
|
"6013","780","numpy build fails with ...libfrtbegin.a...undefined reference to MAIN Trac 182).
|
|
Original ticket http projects.scipy.org numpy ticket 182 on 2006 07 14 by trac user timcera, assigned to unknown.
|
|
Background I want to use the built in lapack.
|
|
site.cfg [atlas] library dirs = atlas libs = [blas] [lapack] My Python is not in a standard location.
|
|
Command line python setup.py build build ext L sjr beodata software relink package manager64 vapp lib python2.4 config lpython2.4 Result ... creating build temp.linux x86 64 2.4 numpy linalg compile options DNO ATLAS INFO=2 Inumpy core include Ibuild src.linux x86 64 2.4 numpy core Inumpy core src Inumpy core include I beodata software relink package manager64 app Python 2.4.3 include python2.4 c gcc numpy linalg lapack litemodule.c usr bin g77 L beodata software relink package manager64 vapp lib L beodata software relink package manager vapp lib build temp.linux x86 64 2.4 numpy linalg lapack litemodule.o L usr lib L sjr beodata software relink package manager64 vapp lib python2.4 config lpython2.4 llapack lg2c o build lib.linux x86 64 2.4 numpy linalg lapack lite.so usr bin ld skipping incompatible usr lib liblapack.so when searching for llapack usr bin ld skipping incompatible usr lib liblapack.a when searching for llapack usr bin ld skipping incompatible usr lib libm.so when searching for lm usr bin ld skipping incompatible usr lib libm.a when searching for lm usr bin ld skipping incompatible usr lib libm.so when searching for lm usr bin ld skipping incompatible usr lib libm.a when searching for lm usr bin ld skipping incompatible usr lib libc.so when searching for lc usr bin ld skipping incompatible usr lib libc.a when searching for lc usr lib gcc x86 64 redhat linux 3.4.3 libfrtbegin.a frtbegin.o) .text 0x1e) In function `main undefined reference to `MAIN collect2 ld returned 1 exit status usr bin ld skipping incompatible usr lib liblapack.so when searching for llapack usr bin ld skipping incompatible usr lib liblapack.a when searching for llapack usr bin ld skipping incompatible usr lib libm.so when searching for lm usr bin ld skipping incompatible usr lib libm.a when searching for lm usr bin ld skipping incompatible usr lib libm.so when searching for lm usr bin ld skipping incompatible usr lib libm.a when searching for lm usr bin ld skipping incompatible usr lib libc.so when searching for lc usr bin ld skipping incompatible usr lib libc.a when searching for lc usr lib gcc x86 64 redhat linux 3.4.3 libfrtbegin.a frtbegin.o) .text 0x1e) In function `main undefined reference to `MAIN collect2 ld returned 1 exit status error Command usr bin g77 L beodata software relink package manager64 vapp lib L beodata software relink package manager vapp lib build temp.linux x86 64 2.4 numpy linalg lapack litemodule.o L usr lib L sjr beodata software relink package manager64 vapp lib python2.4 config lpython2.4 llapack lg2c o build lib.linux x86 64 2.4 numpy linalg lapack lite.so failed with exit status 1","bug"
|
|
"6013","781","longdouble on OS X doesn t work Trac 183).
|
|
Original ticket http projects.scipy.org numpy ticket 183 on 2006 07 18 by dmcooke, assigned to teoliphant.
|
|
>>> longdouble 3.0) 0.0 This is for an iBook G4 running OS X 10.4 Tiger).
|
|
I ve added a test case in r2830 for this.
|
|
`clongdouble` doesn t work either.
|
|
Note that on this machine, longdouble == float128.","bug"
|
|
"6013","782","Valgrind errors in MA tests Trac 184).
|
|
Original ticket http projects.scipy.org numpy ticket 184 on 2006 07 18 by alberts, assigned to teoliphant.
|
|
`check testArrayMethods` in `test ma.py` causes the following Valgrind errors ==2419== Invalid read of size 4 ==2419== at 0x46663BA PyArray ArgMax multiarraymodule.c 3116) ==2419== by 0x46666E7 array argmax arraymethods.c 163) ==2419== by 0x5F15CFC PyCFunction Call in usr lib libpython2.4.so.1.0) ... ==2419== Address 0x791F47C is 28 bytes inside a block of size 40 free d ==2419== at 0x4004E41 free vg replace malloc.c 235) ==2419== by 0x464A72E array dealloc arrayobject.c 1885) ==2419== by 0x466648B PyArray ArgMax multiarraymodule.c 3115) ==2419== by 0x46666E7 array argmax arraymethods.c 163) ==2419== by 0x5F15CFC PyCFunction Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with !NumPy 0.9.9.2833.
|
|
I use the following script to run Valgrind !
|
|
bin sh valgrind \ tool=memcheck \ leak check=yes \ error limit=no \ suppressions= home albert work tmp valgrind python.supp \ num callers=20 \ freelist vol=536870912 \ v \ python c import numpy as N; N.testing.NumpyTest N).testall level=9999,verbosity=9999)","bug"
|
|
"6013","783","Invalid array creation from list of numpy.poly1d objects Trac 185).
|
|
Original ticket http projects.scipy.org numpy ticket 185 on 2006 07 18 by trac user fwasilew, assigned to teoliphant.
|
|
This is a creation phase bug.
|
|
Array constructor seems to be interpreting `poly1d` as array because of `poly1d.
|
|
array ` method.
|
|
>>> import numpy >>> p = numpy.poly1d [3]) >>> numpy.array [[p, p 2], [p 3, p 4]], dtype=object) array [], shape= 2, 2, 0), dtype=object) >>> p = numpy.poly1d [2,1]) >>> numpy.array [[p, p 2], [p 3, p 4]], dtype=object) array [[[1], [2]], [[3], [4]]], dtype=object) >>> >>> p = numpy.poly1d [3,2,1]) >>> numpy.array [[p, p 2], [p 3, p 4]], dtype=object) array [[[[poly1d [3, 2, 1]), poly1d [6, 4, 2])], [poly1d [3, 2, 1]), poly1d [6, 4, 2])]], [[poly1d [3, 2, 1]), poly1d [6, 4, 2])], [poly1d [3, 2, 1]), poly1d [6, 4, 2])]]], [[[poly1d [9, 6, 3]), poly1d [12, 8, 4])], [poly1d [9, 6, 3]), poly1d [12, 8, 4])]], [[poly1d [9, 6, 3]), poly1d [12, 8, 4])], [poly1d [9, 6, 3]), poly1d [12, 8, 4])]]]], dtype=object) >>> p. array ) array [3, 2, 1]) >>> len p) but len returns degree of polynomial 2","bug"
|
|
"6013","784","Creating matrix of dtype=object fails, problem of ndarray.
|
|
new Trac 186).
|
|
Original ticket http projects.scipy.org numpy ticket 186 on 2006 07 18 by trac user fwasilew, assigned to teoliphant.
|
|
I assume that using objects as matrix cells values is allowed otherwise creating matrix from object array should be prohibited and more descriptive exceptions raised).
|
|
>>> numpy.matrix [[None, 1], [1, None]], dtype=object) ... TypeError cannot construct an object array from buffer data This exception is raised in matrix ` new ` constructor, when trying to create ndarray object with values from other <array of dtype=object>.
|
|
arr = numpy.array [[None, 1], [1, None]], dtype=object) numpy.ndarray.
|
|
new numpy.matrix, arr.shape, arr.dtype, buffer=arr) The following two work fine >>> numpy.matrix numpy.array [[None, 1], [1, None]], dtype=object)) matrix [[None, 1], [1, None]], dtype=object) >>> numpy.ndarray.
|
|
new numpy.ndarray, 2,2), object) array [[None, None], [None, None]], dtype=object)","bug"
|
|
"6013","785","Subclassing of numpy.matrix fails when passing ndarray to matrix.
|
|
new constructor Trac 187).
|
|
Original ticket http projects.scipy.org numpy ticket 187 on 2006 07 18 by trac user fwasilew, assigned to teoliphant.
|
|
Let s take the simplest case class submat numpy.matrix) pass This works fine >>> submat [[1,2], [3,4]]) matrix [[1, 2], [3, 4]]) >>> type submat [[1,2], [3,4]])) <class main .submat > >>> isinstance submat [[1,2], [3,4]]), submat) True But the following fails >>> submat numpy.array [[1,2], [3,4]])) matrix [[1, 2], [3, 4]]) >>> type submat numpy.array [[1,2], [3,4]]))) <class numpy.core.defmatrix.matrix > >>> isinstance submat numpy.array [[1,2], [3,4]])), submat) False The problem is tracked to the `matrix.
|
|
new ` constructor >>> class submat numpy.matrix) def new self) return numpy.matrix.
|
|
new submat, [[1,2], [3,4]] ) >>> type submat )) <class main .submat > >>> class submat numpy.matrix) def new self) return numpy.matrix.
|
|
new submat, numpy.array [[1,2], [3,4]])) >>> type submat )) <class numpy.core.defmatrix.matrix >","bug"
|
|
"6013","786","dtype should have nice looking str representation Trac 188).
|
|
Original ticket http projects.scipy.org numpy ticket 188 on 2006 07 19 by sebhaase, assigned to teoliphant.
|
|
>>> numpy.array [1,2]) [1 2] >>> .dtype <i4 It would be nice if it said int32 instead ?
|
|
I don t now how important the < is.
|
|
Thanks, Sebastian Haase.","enhancement"
|
|
"6013","787","Histograms 1d, 2d, nd) Trac 189).
|
|
Original ticket http projects.scipy.org numpy ticket 189 on 2006 07 19 by huard, assigned to teoliphant.
|
|
This patch provides two new functions histogram1d and histogramdd.
|
|
The goal of histogram1d is to replace histogram eventually.
|
|
It computes a one dimensional histogram column wise or row wise depending on the axis keyword.
|
|
It is slighly different from histogram in that it return the bin edges, instead of only the leftmost ones.
|
|
Also, values outside the bin edges are not counted.
|
|
histogramdd computes a d dimensional histogram from an NxD array or a sequence of D arrays.
|
|
Tests for both functions are included.
|
|
The patch also removes the dependence on histogram from histogram2d and corrects a bug.","enhancement"
|
|
"6013","788","Memory error on unicode array comparison Trac 190).
|
|
Original ticket http projects.scipy.org numpy ticket 190 on 2006 07 19 by stefanv, assigned to unknown.
|
|
Using r2831 of numpy, this snippet that excercises comparison to a unicode array causes a memory error import numpy as N a = N.array hello ,N.unicode ) b = N.array world ) print a == b Valgrind reports the following ==2214== Invalid read of size 4 ==2214== at 0x80C8A08 PyCodec LookupError in usr bin python2.4) ==2214== by 0x8099031 within usr bin python2.4) ==2214== by 0x809EFBD PyUnicodeUCS4 DecodeASCII in usr bin python2.4) ==2214== by 0x809F225 PyUnicodeUCS4 Decode in usr bin python2.4) ==2214== by 0x809F453 PyUnicodeUCS4 FromEncodedObject in usr bin python2.4) ==2214== by 0x807D34F PyObject Unicode in usr bin python2.4) ==2214== by 0x48FE4CD UNICODE setitem arraytypes.inc.src 271) ==2214== by 0x48FF76C STRING to UNICODE arraytypes.inc.src 738) ==2214== by 0x4911FF4 PyArray CastTo arrayobject.c 7238) ==2214== by 0x49133B6 PyArray FromArray arrayobject.c 7512) ==2214== Address 0x4 is not stack d, malloc d or recently) free d ==2214== ==2214== Process terminating with default action of signal 11 SIGSEGV) ==2214== Access not within mapped region at address 0x4 ==2214== at 0x80C8A08 PyCodec LookupError in usr bin python2.4) ==2214== by 0x8099031 within usr bin python2.4) ==2214== by 0x809EFBD PyUnicodeUCS4 DecodeASCII in usr bin python2.4) ==2214== by 0x809F225 PyUnicodeUCS4 Decode in usr bin python2.4) ==2214== by 0x809F453 PyUnicodeUCS4 FromEncodedObject in usr bin python2.4) ==2214== by 0x807D34F PyObject Unicode in usr bin python2.4) ==2214== by 0x48FE4CD UNICODE setitem arraytypes.inc.src 271) ==2214== by 0x48FF76C STRING to UNICODE arraytypes.inc.src 738) ==2214== by 0x4911FF4 PyArray CastTo arrayobject.c 7238) ==2214== by 0x49133B6 PyArray FromArray arrayobject.c 7512)","bug"
|
|
"6013","789","No overflow error for character arrays Trac 191).
|
|
Original ticket http projects.scipy.org numpy ticket 191 on 2006 07 19 by stefanv, assigned to unknown.
|
|
I m tentatively filing this, although I am not sure exactly what the expected behaviour should be.
|
|
I expect the following snippet to raise an error x = N.array [ 1.1 ]) N.seterr over= raise ) x[0] = 1.111","bug"
|
|
"6013","790","Comparison fails for record arrays with complex type descriptors Trac 192).
|
|
Original ticket http projects.scipy.org numpy ticket 192 on 2006 07 19 by stefanv, assigned to unknown.
|
|
Under r2831, dt = N.dtype [ x ,int), y ,float), z , S3 )]) a = N.rec.array [[1,1.1, 1.1 ]],dtype=dt) a == a raises an AttributeError with the message ` bool object has no attribute logical and `.","bug"
|
|
"6013","791","Imports of subpackages in numpy init .py should be explicit Trac 193).
|
|
Original ticket http projects.scipy.org numpy ticket 193 on 2006 07 20 by rkern, assigned to teoliphant.
|
|
The pkgload import machinery introduces fragility in the import process and interferes with tools like py2exe and py2app.
|
|
The attached patch does the imports explicitly.
|
|
All tests pass.","bug"
|
|
"6013","792","Flat assignment broken Trac 194).
|
|
Original ticket http projects.scipy.org numpy ticket 194 on 2006 07 20 by stefanv, assigned to unknown.
|
|
Under r2852, r2855, r2854, executing x = N.empty 3,1)) x[ ] = N.arange 3) results in array [[ 2.
|
|
], [ 2.
|
|
], [ 2.]])
|
|
Under older versions r2358), the result is array [[0], [1], [2]])","bug"
|
|
"6013","793","Non critical Valgrind issues Trac 195).
|
|
Original ticket http projects.scipy.org numpy ticket 195 on 2006 07 20 by alberts, assigned to unknown.
|
|
This ticket is for keeping track of Valgrind warnings that might not indicate real problems in NumPy.","bug"
|
|
"6013","794","Fatal error constructing object array Trac 196).
|
|
Original ticket http projects.scipy.org numpy ticket 196 on 2006 07 21 by stefanv, assigned to unknown.
|
|
Trying to create an object array by doing import numpy as N dt = N.dtype [ x ,int), y ,N.object )]) x = N.array [1, object ],dt) is met with severe resistance in r2863.","bug"
|
|
"6013","795","segmentation fault Trac 197).
|
|
Original ticket http projects.scipy.org numpy ticket 197 on 2006 07 21 by nilswagner01, assigned to teoliphant.
|
|
numpy.test 1,10) results in a segfault.
|
|
Here is a backtrace test boolean numpy.core.tests.test numeric.test index) ... ok test divideerr numpy.core.tests.test numeric.test seterr) Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 16384 LWP 17793)] 0x00002aaaaadf4cf0 in pthread lock ) from lib64 libpthread.so.0 gdb) bt 0 0x00002aaaaadf4cf0 in pthread lock ) from lib64 libpthread.so.0 1 0x00002aaaaadf4725 in sem post GLIBC 2.2.5 ) from lib64 libpthread.so.0 2 0x00002aaaaac76059 in PyThread release lock lock=<value optimized out>) at thread pthread.h 338 3 0x00002aaaabff40da in error handler method=2, errobj=0x2aaaadf98cb0, errtype=0x2aaaabfff08f divide by zero , retstatus=1) at ufuncobject.c 519 4 0x00002aaaabff42ea in PyUFunc handlefperr errmask=2, errobj=0x2aaaadf98cb0, retstatus=1) at ufuncobject.c 547 5 0x00002aaaabffb46d in PyUFunc GenericFunction self=0x559710, args=<value optimized out>, mps=0x7fffffc0d2b0) at ufuncobject.c 1450 6 0x00002aaaabffb540 in ufunc generic call self=0x559710, args=0x2aaaadf98710) at ufuncobject.c 2694 7 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 8 0x00002aaaaabfa883 in PyObject CallFunction callable=0x559710, format=<value optimized out>) at abstract.c 1793 9 0x00002aaaaabf8508 in binary op1 v=0x888970, w=0x888b40, op slot=24) at abstract.c 371 10 0x00002aaaaabf8771 in binary op v=0x888970, w=0x888b40, op slot=<value optimized out>, op name=0x2aaaaac8b2c8 ) at abstract.c 422 11 0x00002aaaaac521f1 in PyEval EvalFrame f=0x7a9480) at ceval.c 1060 12 0x00002aaaaac53b97 in PyEval EvalFrame f=0x5d5cd0) at ceval.c 3629 13 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbeece0, globals=<value optimized out>, locals=<value optimized out>, args=0x2aaaae2bdd10, argcount=2, kws=0x82c7d0, kwcount=0, defs=0x2aaaabbf7b28, defcount=1, closure=0x0) at ceval.c 2730 14 0x00002aaaaac0e9af in function call func=0x2aaaabc05758, arg=0x2aaaae2bdcf8, kw=<value optimized out>) at funcobject.c 548 15 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 16 0x00002aaaaac532e2 in PyEval EvalFrame f=0x5e81c0) at ceval.c 3824 17 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbeed50, globals=<value optimized out>, locals=<value optimized out>, args=0x2aaaae1338d8, argcount=2, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at ceval.c 2730 18 0x00002aaaaac0e9af in function call func=0x2aaaabc057d0, arg=0x2aaaae1338c0, kw=<value optimized out>) at funcobject.c 548 19 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 20 0x00002aaaaac02131 in instancemethod call func=<value optimized out>, arg=0x2aaaae1338c0, kw=0x0) Type <return> to continue, or q <return> to quit at classobject.c 2431 21 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 22 0x00002aaaaac5380d in PyEval EvalFrame f=0x6f7290) at ceval.c 3755 23 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbde7a0, globals=<value optimized out>, locals=<value optimized out>, args=0x2aaaae2bcd10, argcount=2, kws=0x0, kwcount=0, defs=0x2aaaabc09628, defcount=1, closure=0x0) at ceval.c 2730 24 0x00002aaaaac0e9af in function call func=0x2aaaabc07d70, arg=0x2aaaae2bccf8, kw=<value optimized out>) at funcobject.c 548 25 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 26 0x00002aaaaac02131 in instancemethod call func=<value optimized out>, arg=0x2aaaae2bccf8, kw=0x0) at classobject.c 2431 27 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 28 0x00002aaaaac33b0a in slot tp call self=<value optimized out>, args=0x2aaaadf0ca90, kwds=0x0) at typeobject.c 4526 29 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 30 0x00002aaaaac5380d in PyEval EvalFrame f=0x5d5850) at ceval.c 3755 31 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbf27a0, globals=<value optimized out>, locals=<value optimized out>, args=0x2aaaae2bd530, argcount=2, kws=0x71e290, kwcount=0, defs=0x0, defcount=0, closure=0x0) at ceval.c 2730 32 0x00002aaaaac0e9af in function call func=0x2aaaabc06050, arg=0x2aaaae2bd518, kw=<value optimized out>) at funcobject.c 548 33 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 34 0x00002aaaaac532e2 in PyEval EvalFrame f=0x5d2c70) at ceval.c 3824 35 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbf2810, globals=<value optimized out>, locals=<value optimized out>, args=0x2aaaad88b6e0, argcount=2, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at ceval.c 2730 36 0x00002aaaaac0e9af in function call func=0x2aaaabc060c8, arg=0x2aaaad88b6c8, kw=<value optimized out>) at funcobject.c 548 Type <return> to continue, or q <return> to quit 37 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 38 0x00002aaaaac02131 in instancemethod call func=<value optimized out>, arg=0x2aaaad88b6c8, kw=0x0) at classobject.c 2431 39 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 40 0x00002aaaaac33b0a in slot tp call self=<value optimized out>, args=0x2aaaae136f10, kwds=0x0) at typeobject.c 4526 41 0x00002aaaaabfa760 in PyObject Call func=<value optimized out>, arg=<value optimized out>, kw=<value optimized out>) at abstract.c 1751 42 0x00002aaaaac5380d in PyEval EvalFrame f=0x633460) at ceval.c 3755 43 0x00002aaaaac53b97 in PyEval EvalFrame f=0x67a970) at ceval.c 3629 44 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaabbe4b20, globals=<value optimized out>, locals=<value optimized out>, args=0x62baf8, argcount=3, kws=0x62bb10, kwcount=0, defs=0x2aaaabc0f2a8, defcount=2, closure=0x0) at ceval.c 2730 45 0x00002aaaaac53aba in PyEval EvalFrame f=0x62b960) at ceval.c 3640 46 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaaab24b90, globals=<value optimized out>, locals=<value optimized out>, args=0x541af0, argcount=2, kws=0x541b00, kwcount=0, defs=0x2aaaacc9a890, defcount=2, closure=0x0) at ceval.c 2730 47 0x00002aaaaac53aba in PyEval EvalFrame f=0x541960) at ceval.c 3640 48 0x00002aaaaac55404 in PyEval EvalCodeEx co=0x2aaaacca33b0, globals=<value optimized out>, locals=<value optimized out>, args=0x0, argcount=0, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at ceval.c 2730 49 0x00002aaaaac556d2 in PyEval EvalCode co=<value optimized out>, globals=<value optimized out>, locals=<value optimized out>) at ceval.c 484 50 0x00002aaaaac70719 in run node n=<value optimized out>, filename=<value optimized out>, globals=0x503b50, locals=0x503b50, flags=<value optimized out>) at pythonrun.c 1265 51 0x00002aaaaac71bc7 in PyRun InteractiveOneFlags fp=<value optimized out>, filename=0x2aaaaac95e73 <stdin> , flags=0x7fffffc0faf0) at pythonrun.c 762 52 0x00002aaaaac71cbe in PyRun InteractiveLoopFlags fp=0x2aaaab809e00, filename=0x2aaaaac95e73 <stdin> , flags=0x7fffffc0faf0) at pythonrun.c 695 53 0x00002aaaaac7221c in PyRun AnyFileExFlags fp=0x2aaaab809e00, filename=0x2aaaaac95e73 <stdin> , closeit=0, flags=0x7fffffc0faf0) at pythonrun.c 658 Type <return> to continue, or q <return> to quit 54 0x00002aaaaac77b25 in Py Main argc=<value optimized out>, argv=0x7fffffc127bc) at main.c 484 55 0x00002aaaab603ced in libc start main ) from lib64 libc.so.6 56 0x00000000004006ea in start ) at start.S 113 57 0x00007fffffc0fb88 in ?? )
|
|
58 0x00002aaaaabc19c0 in rtld errno ) from lib64 ld linux x86 64.so.2","bug"
|
|
"6013","796","Extension build fails with C source when using MSVC Trac 198).
|
|
Original ticket http projects.scipy.org numpy ticket 198 on 2006 07 21 by alberts, assigned to unknown.
|
|
numpy.distutils does a few special things when building extensions that include C code.
|
|
This includes the removal of command line arguments by calling the following function at line 200 of numpy distutils ccompiler.py !python try self.compiler so.remove Wstrict prototypes ) except ValueError pass This fails when building with MSVC AttributeError MSVCCompiler instance has no attribute compiler so The block mentioned above catches !ValueError, but not !AttributeError.
|
|
I m not sure if we want to catch just !AttributeError or !ValueError and !AttributeError.
|
|
Tested with !NumPy 1.1.2880.","bug"
|
|
"6013","797","append to an array Trac 199).
|
|
Original ticket http projects.scipy.org numpy ticket 199 on 2006 07 21 by trac user nil, assigned to unknown.
|
|
An append function would be similar to concatenate , except that it would save memory use and data copy.
|
|
`a.append b)` would be like `a = concatenate a,b))`, but faster... expected result >> a = array [1, 2, 3]) >> b = array [4, 5, 6]) >> a.append b) >> print a [1 2 3 4 5 6] >> a.append b, b, b)) >> print a [1 2 3 4 5 6 4 5 6 4 5 6 4 5 6]","enhancement"
|
|
"6013","798","split an array Trac 200).
|
|
Original ticket http projects.scipy.org numpy ticket 200 on 2006 07 21 by trac user nil, assigned to unknown.
|
|
An split function would be similar to slice extraction , except that it would save memory use and data copy by just redistributing the memory to different arrays).
|
|
` b,c) = a.split n)` would be like `b = a[ n]; c = a[n ]; del a;` but faster... expected result >> a = array [1, 2, 3, 4, 5, 6]) >> b, c) = a.split 3) >> print b [1 2 3] >> print c [4 5 6] >> a = array [1, 2, 3, 4, 5, 6, 7, 8, 9]) >> print a.split 3, 4)) array [1, 2, 3]), array [4]), array [5, 6, 7, 8, 9]))","enhancement"
|
|
"6013","799","Clean up and unify r ,c ,hstack,vstack etc Trac 201).
|
|
Original ticket http projects.scipy.org numpy ticket 201 on 2006 07 23 by baxissimo, assigned to unknown.
|
|
== SUMMARY == make r behave like vstack plus range literals make column stack only transpose its 1d inputs.
|
|
rename r ,c to v ,h or something else) to make their connection with vstack and hstack clearer.
|
|
Maybe vs and hs would be better?
|
|
make a new vertsion of c that acts like column stack and make r an alias for v just as row stack is an alias for vstack) so that theres a nice parallel v <=>vstack, h <=>hstack, c <=>column stack, r <=>row stack == DISCUSSION == Here s essentially what these different methods do vstack tup) concatenate map atleast 2d,tup), axis=0 ) hstack tup) concatenate map atleast 1d,tup),axis=1 ) column stack tup) arrays = map transpose,map atleast 2d,tup) ) concatenate arrays,1) note that column stack transposes everything not just 1 d inputs.
|
|
The current docstring only claims that it works with 1d inputs, so making it do something reasonable with 2d inputs will hopefully not break much code.)
|
|
The above 3 are pretty much exactly the code used by numpy.
|
|
That s all there is to those 3 functions.
|
|
For r and c I m summarizing, but effectively they seem to be doing something like r [args] concatenate map atleast 1d,args),axis=0 ) c [args] concatenate map atleast 1d,args),axis= 1 ) c behaves almost exactly like hstack with the addition of range literals being allowed.
|
|
r is mostly like vstack, but a little different since it effectively uses atleast 1d, instead of atleast 2d.
|
|
So you have >>> numpy.vstack 1,2,3,4)) array [[1], [2], [3], [4]]) but >>> numpy.r [1,2,3,4] array [1, 2, 3, 4]) However for cases like that with just 0 d or 1 d inputs, c behaves identically to r , so if you wanted to get a 1 d output you could have just used c .
|
|
The current situation is confusing because r , and c act similar to the existing hstack vstack methods but different enough to make the distinction hard to remember.
|
|
Ideally these changes will make it so that all the user needs to do to change from one of the stack methods to the equivalent that supports slice literals, or back again, is a simple syntactical transformation hstack args) ) <==> h [ args ] vstack args) ) <==> v [ args ] column stack args) ) <==> c [ args ] row stack args) ) <==> r [ args ] == Code == The following is code that implements the above suggestions in a file called altstack.py.
|
|
Mostly the code is copied from numpy lib shape base.py and numpy lib index tricks.py.
|
|
I d be happy to provide a patch against numpy svn with these changes if there s interest.
|
|
!python import sys import types import numpy import numpy.core.numeric as nx from numpy.core.numeric import asarray, ScalarType import numpy.lib.function base import numpy.core.defmatrix as matrix makemat = matrix.matrix ascol helper function to convert only 1d inputs to columns.
|
|
def ascol v) Turns 1 d inputs into a column vectors.
|
|
For all other inputs acts like atleast 2d.
|
|
Q Should this also transpose any 1xN) s to be columns?
|
|
The current thinking is if you have a 1 d then you haven t really decided whether it s a row or col, and this method is asserting that it should be a column.
|
|
But if it s already a 2 d row, then it s probably a row for a reason, and you should transpose it explicitly if you want a column.)
|
|
arr = numpy.array v,copy=False,subok=True) if arr.ndim<2 return numpy.transpose numpy.atleast 2d arr)) else return arr Modified column stack that leaves >2d data alone def column stack tup) Stack 1D arrays as columns into a 2D array Description Take a sequence of 1D and 2D arrays and stack them as columns to make a single 2D array.
|
|
All arrays in the sequence must have the same first dimension.
|
|
2D arrays are stacked as is, just like with hstack, 1D arrays are turned into 2D columns first.
|
|
Arguments tup sequence of 1D and 2D arrays.
|
|
All arrays must have the same first dimension.
|
|
Examples >>> import numpy >>> a = array 1,2,3)) >>> b = array 2,3,4)) >>> numpy.column stack a,b)) array [[1, 2], [2, 3], [3, 4]]) arrays = map ascol,tup) return nx.concatenate arrays,1) Slightly modified version of numpy.lib.index tricks.concatenator Only difference is the addition of mapper parameter and self.mapper attribute.
|
|
This allows for running the data through a given function e.g.
|
|
atleast 2d) before calling concatenate ).
|
|
Makes it so concatenator can generate results like vstack and column stack class concatenator object) Translates slice objects to concatenation along an axis.
|
|
def retval self, res) if self.matrix oldndim = res.ndim res = makemat res) if oldndim == 1 and self.col res = res.T self.axis = self.
|
|
axis self.mapping = self.
|
|
mapping self.matrix = self.
|
|
matrix self.col = 0 return res def init self, axis=0, mapping=None, matrix=False) self.
|
|
axis = axis self.
|
|
mapping = mapping self.
|
|
matrix = matrix self.axis = axis self.mapping = mapping self.matrix = matrix self.col = 0 def getitem self,key) if isinstance key, str) frame = sys.
|
|
getframe ).f back mymat = matrix.bmat key,frame.f globals,frame.f locals) return mymat if type key) is not tuple key = key,) objs = [] scalars = [] final dtypedescr = None for k in range len key)) scalar = False if type key[k]) is slice step = key[k].step start = key[k].start stop = key[k].stop if start is None start = 0 if step is None step = 1 if type step) is type 1j) size = int abs step)) newobj = function base.linspace start, stop, num=size) newobj = numpy.linspace start, stop, num=size) else newobj = nx.arange start, stop, step) elif type key[k]) is str if key[k] in rc ) self.matrix = True self.col = key[k] == c ) continue try self.axis = int key[k]) continue except ValueError, TypeError) raise ValueError, unknown special directive elif type key[k]) in ScalarType newobj = asarray [key[k]]) scalars.append k) scalar = True else newobj = key[k] objs.append newobj) if isinstance newobj, nx.ndarray) and not scalar if final dtypedescr is None final dtypedescr = newobj.dtype elif newobj.dtype > final dtypedescr final dtypedescr = newobj.dtype if final dtypedescr is not None for k in scalars objs[k] = objs[k].astype final dtypedescr) if self.mapping is not None objs = map self.mapping, objs) res = nx.concatenate tuple objs),axis=self.axis) return self.
|
|
retval res) def getslice self,i,j) res = nx.arange i,j) return self.
|
|
retval res) def len self) return 0 separate classes are used here instead of just making r = concatentor 0), etc.
|
|
because otherwise we couldn t get the doc string to come out right in help r ) The concatenator objects are now v like vstack r like row stack but row stack==vstack, so r ==v ) h like hstack c like column stack the modified version above) Note that r was used both for making a row and for stacking rows previously, but the old c could also be used for this purpose.
|
|
Now if you want to make a row, h is the way to do it.
|
|
class v class concatenator) Translates slice objects to vertical row wise) concatenation.
|
|
Similar to vstack with the additional ability to use slices.
|
|
For example >>> v [[1,2,3], 3 0 1, [4,5,6]] array [[1, 2, 3], [3, 2, 1], [4, 5, 6]]) def init self) concatenator.
|
|
init self, 0, numpy.atleast 2d) v = v class ) r = v class h class concatenator) Translates slice objects to horizontal concatenation.
|
|
Similar to hstack with the additional ability to use slices.
|
|
For example >>> h [ 1, 10 7 1, [2,3,4] ] array [ 1, 10, 9, 8, 2, 3, 4]) >>> h [[[1],[2],[3]], [[4],[5],[6]]] array [[1, 4], [2, 5], [3, 6]]) def init self) concatenator.
|
|
init self, 1) h = h class ) class c class concatenator) Translates slice objects to column wise concatenation.
|
|
Similar to column stack with the additional ability to use slices.
|
|
For example >>> c [[1,2,3], [4,5,6]] array [[1, 4], [2, 5], [3, 6]]) >>> c [ [1,2], [4,5], [[7,8],[9,10]] ] array [[ 1, 4, 7, 8], [ 2, 5, 9, 10]]) def init self) concatenator.
|
|
init self, 1, ascol) c = c class )","enhancement"
|
|
"6013","800","record arrays bug Trac 202).
|
|
Original ticket http projects.scipy.org numpy ticket 202 on 2006 07 23 by sebhaase, assigned to unknown.
|
|
Sorry that I don t have a better summary.
|
|
While trying to reproduce a possible recarray bug I reported on the mailing list I ran into a more basic bug >>> a= N.array [1,2,3],dtype=N.int32) >>> a [1 2 3] >>> r = N.rec.array a, shape=1, formats=[ 3i4 ], names=[ d ]) >>> r [ array [1, 2, 3]),),)] >>> a [ array [1, 2, 3]),),)] It looks like a get somehow overwritten Is there maybe only a view ) call missing inside rec.array ?","bug"
|
|
"6013","801","prevent matlib from modifying numpy s namespace Trac 203).
|
|
Original ticket http projects.scipy.org numpy ticket 203 on 2006 07 24 by astraw, assigned to unknown.
|
|
This patch prevents matlib from modifying numpy s namespace.","bug"
|
|
"6013","802","misleading error message in array shape=10, dtype=int32) Trac 204).
|
|
Original ticket http projects.scipy.org numpy ticket 204 on 2006 07 24 by sebhaase, assigned to sebhaase.
|
|
in numarray a giving a scalar instead of a tuple for a 1d shape was accepted in numpy I get a uninterpretable error message >>> N.array shape=2,dtype=N.int32) Traceback most recent call last) File <input> , line 1, in ?
|
|
TypeError function takes at least 1 argument 0 given)","enhancement"
|
|
"6013","803","unique doesn t work with empty arrays Trac 205).
|
|
Original ticket http projects.scipy.org numpy ticket 205 on 2006 07 24 by alberts, assigned to teoliphant.
|
|
Unique on an empty array should return an empty array.
|
|
In [1] import numpy as N In [2] N.unique N.array [])) Out[2] array [10682776]) In [3] N. version Out[3] 1.1.2881","bug"
|
|
"6013","804","Strange errors with dtypes when align=True Trac 206).
|
|
Original ticket http projects.scipy.org numpy ticket 206 on 2006 07 24 by alberts, assigned to unknown.
|
|
As far as I can figure, all of the following dtypes should work, but some don t. !python import numpy as N works dt = N.dtype [ x , <i4 ), y , <f8 )], align=False) raises TypeError data type not understood dt = N.dtype [ x , <i4 ), y , <f8 )], align=True) works dt = N.dtype [ a , <i4 ), b , <f8 )], align=False) raises ValueError mismatch in size of old and new data descriptor dt = N.dtype [ a , <i4 ), b , <f8 )], align=True) works dt = N.dtype [ x , <i4 ), , |V4 ), y , <f8 )], align=False) raises TypeError data type not understood dt = N.dtype [ x , <i4 ), , |V4 ), y , <f8 )], align=True) works dt = N.dtype { names [ x , y ], formats [N.intc, N.float64]}, align=False) works dt = N.dtype { names [ x , y ], formats [N.intc, N.float64]}, align=True) Tested with !NumPy 1.1.2881.","bug"
|
|
"6013","805","Visual Studio build broken Trac 207).
|
|
Original ticket http projects.scipy.org numpy ticket 207 on 2006 07 26 by alberts, assigned to unknown.
|
|
It seems the recent longdouble fixes have broken the build with Visual Studio.
|
|
For r2895 on the ver1.0 branch I get C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\cl.exe c nologo Od MD W3 GX Z7 DNDEBUG Ibuild\src.win32 2.4\numpy\core\src Inumpy\core\include Ibuild\src.win32 2.4\numpy\core Inumpy\core\src Inumpy\core\include Ic \Python24\include Ic \Python24\PC Tcnumpy\core\src\multiarraymodule.c Fobuild\temp.win32 2.4\Release\numpy\core\src\multiarraymodule.obj multiarraymodule.c numpy\core\src\scalartypes.inc.src 912) error C2036 void unknown size numpy\core\src\multiarraymodule.c 5126) warning C4018 <= signed unsigned mismatch I ll see if I can come up with a fix, but maybe somebody else has some ideas.","bug"
|
|
"6013","806","union1d fails on union of 2 empty sets Trac 208).
|
|
Original ticket http projects.scipy.org numpy ticket 208 on 2006 07 26 by trac user lane, assigned to unknown.
|
|
In [28] numpy.union1d numpy.array []), numpy.array [])) exceptions.IndexError Traceback most recent call last) home lane ZBSC py <ipython console> usr lib python2.4 site packages numpy lib arraysetops.py in union1d ar1, ar2) 131 def union1d ar1, ar2 ) 132 Union of 1D arrays with unique elements.
|
|
> 133 return unique1d numpy.concatenate ar1, ar2) ) ) 134 135 usr lib python2.4 site packages numpy lib arraysetops.py in unique1d ar1, retindx) 73 ar = numpy.array ar1 ).flatten ) 74 ar.sort ) > 75 return ar.compress ediff1d ar, 1 ) != 0) 76 77 usr lib python2.4 site packages numpy lib arraysetops.py in ediff1d ar1, to end, to begin) 45 elif to end 46 ed = numpy.empty ar1.shape, dtype = ar1.dtype ) > 47 ed[ 1] = to end 48 ed[ 1] = dar1 49 elif to begin IndexError index out of bounds","bug"
|
|
"6013","807","OSX ERROR check singleton numpy.lib.tests.test getlimits.test longdouble) Trac 209).
|
|
Original ticket http projects.scipy.org numpy ticket 209 on 2006 07 26 by trac user Andrew J, assigned to unknown.
|
|
I get the following error OSX Universal build of 2.4.3 using gcc 4.0) under `numpy.
|
|
version == 0.9.9.2909`.
|
|
Perhaps related to the fixed longdouble error elsewhere?
|
|
In [4] numpy.test 1000) Found 13 tests for numpy.core.umath Found 8 tests for numpy.lib.arraysetops Found 1 tests for numpy.lib.polynomial Found 8 tests for numpy.core.records Found 26 tests for numpy.core.numeric Found 5 tests for numpy.distutils.misc util Found 4 tests for numpy.lib.getlimits Found 30 tests for numpy.core.numerictypes Found 4 tests for numpy.core.scalarmath Found 10 tests for numpy.lib.twodim base Found 1 tests for numpy.lib.ufunclike Found 47 tests for numpy.lib.shape base Found 4 tests for numpy.lib.index tricks Found 32 tests for numpy.linalg.linalg Found 42 tests for numpy.lib.type check Found 3 tests for numpy.dft.helper Found 148 tests for numpy.core.multiarray Found 36 tests for numpy.core.ma Found 10 tests for numpy.core.defmatrix Found 39 tests for numpy.lib.function base Found 0 tests for main ..............................................................E........................................................................................................................................................................................................................................................................................................................................................................................................................ ====================================================================== ERROR check singleton numpy.lib.tests.test getlimits.test longdouble) Traceback most recent call last) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy 0.9.9.2909 py2.4 macosx 10.3 fat.egg numpy lib tests test getlimits.py , line 33, in check singleton ftype = finfo longdouble) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy 0.9.9.2909 py2.4 macosx 10.3 fat.egg numpy lib getlimits.py , line 49, in new obj = object.
|
|
new cls).
|
|
init dtype) File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy 0.9.9.2909 py2.4 macosx 10.3 fat.egg numpy lib getlimits.py , line 75, in init numpy longfloat precision floating \ File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy 0.9.9.2909 py2.4 macosx 10.3 fat.egg numpy lib machar.py , line 127, in init raise RuntimeError, could not determine machine tolerance \ RuntimeError could not determine machine tolerance for negep , locals ) > { it 33, float to str <function <lambda> at 0x317a270>, one array [1.0], dtype=float128), temp1 array [nan], dtype=float128), title numpy longfloat precision floating point number , self <numpy.lib.machar.MachAr object at 0x317cab0>, two array [2.0], dtype=float128), float to float <function <lambda> at 0x317acb0>, irnd 1, float conv <function <lambda> at 0x3171f70>, tempa array [inf], dtype=float128), beta array [ 2.1474836e 09], dtype=float128), ibeta 2147483648, a array [ 0.0], dtype=float128), b array [0.0], dtype=float128), negep 1, temp array [1.0], dtype=float128), i 35, betain array [ 4.6566129e 10], dtype=float128), itemp 2147483648, zero array [0.0], dtype=float128), betah array [ 1.0737418e 09], dtype=float128), int conv <function <lambda> at 0x317a0b0>} Ran 471 tests in 2.739s FAILED errors=1)","bug"
|
|
"6013","808","dtypes with invalid names should raise an exception Trac 210).
|
|
Original ticket http projects.scipy.org numpy ticket 210 on 2006 07 28 by alberts, assigned to unknown.
|
|
Trying to construct the following dtype should probably raise a !TypeError !python dt = N.dtype { names [123], formats [N.intc]}) If not, one can create an array but not access its fields !python In [2] import numpy as N In [3] dt = N.dtype { names [123], formats [N.intc]}) In [4] x = N.empty 5,), dt) In [5] x[123] IndexError index out of bounds Tested with !NumPy 1.0b2.dev2915.","bug"
|
|
"6013","809","descr subscripting segfault on Windows Trac 211).
|
|
Original ticket http projects.scipy.org numpy ticket 211 on 2006 07 28 by alberts, assigned to unknown.
|
|
The following code causes a segmentation fault on Windows, but not on Linux.
|
|
Tested with !NumPy 1.0b2.dev2917.
|
|
!python import numpy as N dt = N.dtype { names [123], formats [N.intc]}) dt[123] The descr subscript function in arrayobject.c is being called.
|
|
What happens 1.
|
|
`if self >names)` evaluates to true 1.
|
|
`if PyString Check op) || PyUnicode Check op))` evaluates to false 1.
|
|
`int value = value = PyArray PyIntAsInt op)` in else 1.
|
|
`if !PyErr Occurred ))` evaluates to true 1.
|
|
`size = PyTuple GET SIZE self >names)` sets size to 1 1.
|
|
`if value < 0 || value >= size)` evaluates to true and `PyErr Format` is executed 1.
|
|
At this point, an error should probably be returned.
|
|
1.
|
|
`name = PyTuple GET ITEM self >names, value)` puts an invalid object in name because value is an invalid key 1.
|
|
`descr subscript self, name)` is called 1.
|
|
`if self >names)` evaluates to true 1.
|
|
`PyString Check op)` segfaults op is name, which is invalid)","bug"
|
|
"6013","810","unique semantic problem unique1d, union1d etc.
|
|
fail with inf Trac 212).
|
|
Original ticket http projects.scipy.org numpy ticket 212 on 2006 07 29 by mforbes, assigned to unknown.
|
|
== Synopsis == >>> a=unique1d array [inf,inf])) >>> a array [ inf, inf]) arraysetops.py implies the semantics that Elements a and b are unique iff a b == 0.
|
|
To me, a more intuitive semantic would be Elements a and b are unique iff a!=b.
|
|
== Details == The problem is that union1d ultimately uses ediff1d which computes the differences and check for these equalling zero.
|
|
For inf, subtraction is not well defined and returns a nan >>> inf inf nan Thus, there is a fundamental difference with floating point numbers between being the same and having a difference of zero.
|
|
== Solution == A solution would require some restructuring of the arraysetops module.
|
|
I am hesitant to start this until I understand the design decisions made with the present code as fixing this bug may affect performance.
|
|
The codes author is Robert Cimrman, but I do not know how to contact him.)
|
|
Perhaps the easiest solution would be to check for the case of inf and inf after.
|
|
Since the arrays are sorted, this would take O log N)) time, then a simple slice would be needed.
|
|
At least the functions using ediff1d should be properly documented to include a comment Two elements a and b are considered unique iff a b == 0.
|
|
Note, with this semantic change, a similar issue would arise with nan since nan != nan is True, but nan is not usually a useful entry in a sorted array so this should be acceptable.)
|
|
== Rational == Arrays with inf and inf are perfectly reasonable for representing infinite ranges.
|
|
These can function correctly if the asymptotics of the problem are well defined and it can be preferable to have true inf and 0=1 inf rather than finfo float).max etc.
|
|
Michael.","bug"
|
|
"6013","811","SharedLibrary builder for numpy.distutils Trac 213).
|
|
Original ticket http projects.scipy.org numpy ticket 213 on 2006 07 29 by alberts, assigned to unknown.
|
|
A !SharedLibrary builder will be very useful for people who want to experiment with ctypes.
|
|
The [http www.ctcms.nist.gov oof oof2 index.html OOF2] folks already have such a beast.
|
|
Provided there aren t any licensing issues, maybe this could be integrated in numpy.distutils.","enhancement"
|
|
"6013","812","Build broken with MSVC Trac 214).
|
|
Original ticket http projects.scipy.org numpy ticket 214 on 2006 07 29 by alberts, assigned to unknown.
|
|
python setup.py config compiler=msvc build compiler=msvc bdist wininst Traceback Traceback most recent call last) File setup.py , line 89, in ?
|
|
setup package ) File setup.py , line 82, in setup package configuration=configuration ) File C \home\albert\work2\numpy\numpy\distutils\core.py , line 174, in setup return old setup new attr) ...
|
|
File C \home\albert\work2\numpy\numpy\distutils\fcompiler\gnu.py , line 98, in get library dirs d = self.get libgcc dir ) File C \home\albert\work2\numpy\numpy\distutils\fcompiler\gnu.py , line 88, in get libgcc dir status, output = exec command self.compiler f77 TypeError can only concatenate list not str ) to list I don t have a site.cfg or anything like that.
|
|
Tested with r2922.","bug"
|
|
"6013","813","Bug in linalg.eigvals Trac 215).
|
|
Original ticket http projects.scipy.org numpy ticket 215 on 2006 07 31 by nilswagner01, assigned to dmcooke.
|
|
import numpy as N import scipy as S import scipy.linalg A = N.array [0.,1.
|
|
],[ 1., 1.5])) for module in N,S print module.linalg.eigvals A)","bug"
|
|
"6013","814","crash with fromiter and ndindex Trac 216).
|
|
Original ticket http projects.scipy.org numpy ticket 216 on 2006 07 31 by bthyreau, assigned to unknown.
|
|
Hi, This code seems to trigger a crash, tested on 1.0b2.dev2927 import numpy numpy.fromiter numpy.ndindex 5,5)), numpy.float ) glibc detected python double free or corruption fasttop) 0x093e01a8 Thanks","bug"
|
|
"6013","815","Scalar types comparison fails Trac 217).
|
|
Original ticket http projects.scipy.org numpy ticket 217 on 2006 07 31 by alberts, assigned to unknown.
|
|
Tested with !NumPy 1.0b2.dev2927.
|
|
!python In [90] dt = N.dtype N.intp) In [93] dt.type Out[93] <type int32scalar > In [94] N.sctypes[ int ][2] Out[94] <type int32scalar > Seems to be the same.
|
|
Let s compare with `N.sctypes[ int ]` which contains !python In [95] N.sctypes[ int ] Out[95] [<type int8scalar >, <type int16scalar >, <type int32scalar >, <type int64scalar >] Works fine !python In [91] [dt==x for x in N.sctypes[ int ]] Out[91] [False, False, True, False] Doesn t work with dt.type though !python In [92] [dt.type==x for x in N.sctypes[ int ]] Out[92] [False, False, False, False]","bug"
|
|
"6013","816","Crash with string view Trac 218).
|
|
Original ticket http projects.scipy.org numpy ticket 218 on 2006 07 31 by alberts, assigned to unknown.
|
|
!python In [2] N. version Out[2] 1.0b2.dev2927 In [4] import numpy as N In [5] N.array [8224]).view S ) crashes","bug"
|
|
"6013","817","NameError PyArray ContiguousFromObject on FC3; affects mtrand Trac 219).
|
|
Original ticket http projects.scipy.org numpy ticket 219 on 2006 08 01 by trac user TomLoredo, assigned to unknown.
|
|
C code that calls PyArray ContiguousFromObject both my own and within numpy) compiles fine but gives a NameError at runtime on Fedora Core 3 Python 2.4.3).
|
|
Calling random.set state displays the bug.
|
|
If I edit mtrand.pyx to change its 3 PyArray ContiguousFromObject calls to PyArray ContiguousFromAny, run generate mtrand c.py, and reinstall, the error goes away.
|
|
A sample IPython session is copied below.
|
|
I deleted my previous numpy package from site packages before installing this one.
|
|
I do not have this problem on OS X 10.3.9, MacPy 2.4.1).
|
|
I am unclear of the status of PyArray ContiguousFromObject.
|
|
It is no longer mentioned in the Guide To Numpy.
|
|
Is it deprecated?
|
|
Tom Loredo In [1] import numpy In [2] numpy.
|
|
version Out[2] 1.0b2.dev2941 In [3] from numpy import random In [4] s=random.get state ) In [5] s Out[5] MT19937 , array [ 1367602364, 599621924, 1947610538, 686169956, 1451278673, 1648572372, 1175619180, 205273143, 13145610, 1292301198, [snip] 716751809, 622130452, 123887232, 2052494269]), 624) In [6] random.set state s) exceptions.NameError Traceback most recent call last) home loredo research exoplanets rv <ipython console> home loredo research exoplanets rv mtrand.pyx in mtrand.RandomState.set state ) NameError PyArray ContiguousFromObject","bug"
|
|
"6013","818","Build fails on Alpha Linux Trac 220).
|
|
Original ticket http projects.scipy.org numpy ticket 220 on 2006 08 02 by alberts, assigned to unknown.
|
|
Build log to be attached.
|
|
The first time I run the build, it fails when detecting the math libraries, but ` usr lib libm.
|
|
[a,so]` are present on the system.
|
|
Running the build again doesn t cause the config step to fail.
|
|
Instead the build continues and fails with numpy core src multiarraymodule.c In function `initmultiarray numpy core src multiarraymodule.c 6276 `NPY ALLOW THREADS undeclared first use in this function) numpy core src multiarraymodule.c 6276 Each undeclared identifier is reported only once numpy core src multiarraymodule.c 6276 for each function it appears in.)","bug"
|
|
"6013","819","test divideerr fails on NetBSD Trac 221).
|
|
Original ticket http projects.scipy.org numpy ticket 221 on 2006 08 02 by alberts, assigned to unknown.
|
|
====================================================================== FAIL test divideerr numpy.core.tests.test numeric.test seterr) Traceback most recent call last) File numpy core tests test numeric.py , line 195, in test divideerr self.fail ) AssertionError Tested with GCC 3.3.3 NetBSD nb3 20040520) on the !SourceForge Compile Farm.
|
|
This same test fails on Linux when the test suite is run under Valgrind.","bug"
|
|
"6013","820","rstrip on chararray causes Attribute Error Trac 222).
|
|
Original ticket http projects.scipy.org numpy ticket 222 on 2006 08 02 by chanley, assigned to unknown.
|
|
>>> from numpy.core import >>> r = rec.fromrecords [[456, dbe ,1.2],[2, de ,1.3]],names= col1,col2,col3 ) >>> r recarray [ 456, dbe , 1.2), 2, de , 1.3)], dtype=[ col1 , <i4 ), col2 , |S7 ), col3 , <f8 )]) >>> r.field 1) chararray [dbe , de ], dtype= |S7 ) >>> a = r.field 1) >>> a chararray [dbe , de ], dtype= |S7 ) >>> b = a.rstrip ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File data sparty1 dev site packages lib python numpy core defchararray.py , line 263, in rstrip return self.
|
|
generalmethod rstrip , broadcast self, chars)) File data sparty1 dev site packages lib python numpy core defchararray.py , line 138, in generalmethod if chk.dtype is object and chk.item ) is None AttributeError NoneType object has no attribute dtype >>> In numarray, chararrays always truncate trailing whitespace.
|
|
Without the ability to strip these trailing blanks pyfits will have problems working with numpy.
|
|
Chris","bug"
|
|
"6013","821","reshape should be able to guess the last dimension Trac 223).
|
|
Original ticket http projects.scipy.org numpy ticket 223 on 2006 08 03 by bthyreau, assigned to unknown.
|
|
Hi, I have an array Y, so that shape Y) == 125, 46, 63, 53) And I often need to reshape it in Y.reshape 125, 46 63 53) But it s boring to type and i d like to be able to just type Y.reshape 125,)) and it guess the last dimension from the size..
|
|
Thanks !","enhancement"
|
|
"6013","822","Fancy indexing assignment to array with fields fails Trac 224).
|
|
Original ticket http projects.scipy.org numpy ticket 224 on 2006 08 03 by alberts, assigned to unknown.
|
|
The code !python import numpy as N dt = N.dtype { names [ index , value ], formats [N.intc, N.float64]}) x = N.zeros 3,), dt) x[ value ][[0,1]] = [10., 20.]
|
|
print x y = N.zeros 3,), dt) y[[0,1]][ value ] = [10., 20.]
|
|
print y With !NumPy 1.0b2.dev2945 this prints [ 0, 10.0) 0, 20.0) 0, 0.0)] [ 0, 0.0) 0, 0.0) 0, 0.0)] In other words, if you do fancy indexing and then assign to a specific field, the array isn t changed.
|
|
Without fancy indexing both ways work fine !python In [55] x Out[55] array [ 0, 10.0), 0, 20.0), 0, 0.0)], dtype=[ index , <i4 ), value , <f8 )]) In [56] x[ value ][0] = 100 In [57] x Out[57] array [ 0, 100.0), 0, 20.0), 0, 0.0)], dtype=[ index , <i4 ), value , <f8 )]) In [58] x[0][ value ] = 200 In [59] x Out[59] array [ 0, 200.0), 0, 20.0), 0, 0.0)], dtype=[ index , <i4 ), value , <f8 )])","enhancement"
|
|
"6013","823","bincount does not accept input of type > N.uint16 Trac 225).
|
|
Original ticket http projects.scipy.org numpy ticket 225 on 2006 08 03 by stefanv, assigned to teoliphant.
|
|
Under r2944 In [22] N.bincount N.array [1],N.uint16)) Out[22] array [0, 1]) In [23] N.bincount N.array [1],N.uint32)) exceptions.TypeError Traceback most recent call last) home stefan work scipy hough <ipython console> TypeError array cannot be safely cast to required type","bug"
|
|
"6013","824","doc is single line too long Trac 226).
|
|
Original ticket http projects.scipy.org numpy ticket 226 on 2006 08 04 by sebhaase, assigned to unknown.
|
|
>>> a.byteswap.
|
|
doc m.byteswap False) Swap the bytes in the array.
|
|
Return the byteswapped array.
|
|
If the first argument is TRUE, byteswap in place and return a reference to self.
|
|
should be more like >>> a.byteswap.
|
|
doc m.byteswap False) Swap the bytes in the array.
|
|
Return the byteswapped array.
|
|
If the first argument is TRUE, byteswap in place and return a reference to self.
|
|
there are likely more cases.
|
|
Sebastian Haase","bug"
|
|
"6013","825","fromfunction ) should get a dtype argument Trac 227).
|
|
Original ticket http projects.scipy.org numpy ticket 227 on 2006 08 04 by sebhaase, assigned to unknown.
|
|
Hi, in numarray na.fromfuntion had an optinal type argument.
|
|
This should prevent a temporay array in a construct like this a= fromfunction lambda y,x x sin y), 256,256), float32) versus ) a= fromfunction lambda y,x x sin y), 256,256)).astype float32) Also it s shorter.
|
|
This could be implemented by just passing that dtype argument to indices.
|
|
This is how we did this in numarray).
|
|
But I guess that if dtype=int32 would be specified, would the sin of that not automatically convert to float64 !?
|
|
Sebastian Haase","bug"
|
|
"6013","826","Backward compatibility plan Trac 228).
|
|
Original ticket http projects.scipy.org numpy ticket 228 on 2006 08 04 by teoliphant, assigned to teoliphant.
|
|
The plan for compatibility with Numeric Numarray is 1) Create functions in oldnumeric and numarray that create and return NumPy arrays but do it using the syntax of the old packages i.e.
|
|
typecode=, savespace=).
|
|
Also the name spaces when doing from numpy.oldnumeric import should only include names that were in Numeric or Numarray already 2) Create 2 scripts for Numeric conversion and 2 for Numarray conversion a) The first conversion script just changes the imports to the compatibility layer and alters attribute and method access to be compatible with NumPy arrays.
|
|
This script should work most of the time.
|
|
b) The second conversion script tries to actually make the changes to the code so it will work without the compatibility layer.
|
|
This script will be harder to get to work perfectly and will probably require some manual help from the user.","bug"
|
|
"6013","827","numpy.random.poisson 0) should return 0 Trac 229).
|
|
Original ticket http projects.scipy.org numpy ticket 229 on 2006 08 04 by sebhaase, assigned to rkern.
|
|
>>> N.random.poisson 0) Traceback most recent call last) File <input> , line 1, in ?
|
|
File mtrand.pyx , line 741, in mtrand.RandomState.poisson ValueError lam <= 0 >>> N.random.poisson na.arange 10)) Traceback most recent call last) File <input> , line 1, in ?
|
|
TypeError only length 1 arrays can be converted to Python scalars.
|
|
To my knowledge the poisson statistics is defined to return always 0, if a mean of 0 is given as argument.
|
|
This makes sense if you think about quantum shot noise for zero photons should should always measure excactly 0)","bug"
|
|
"6013","828","a 2 not executed as a a if a.dtype = int32 Trac 230).
|
|
Original ticket http projects.scipy.org numpy ticket 230 on 2006 08 04 by sebhaase, assigned to unknown.
|
|
>>> a=N.random.poisson N.arange 1e6) 1) >>> U.timeIt a 2 ) 0.59 >>> U.timeIt a a ) 0.01 >>> a.dtype int32 float64, float32 work OK giving equal times for both cases.
|
|
I tested this on Linux 32 bit, Debian sarge) Sebastian Haase","bug"
|
|
"6013","829","Various types missing from sctypes lists Trac 231).
|
|
Original ticket http projects.scipy.org numpy ticket 231 on 2006 08 05 by alberts, assigned to teoliphant.
|
|
I did some testing with regards to ticket 815 and the changes in r2951.
|
|
!python fix in r2951 added intp to the list In [41] [N.intp==x for x in N.sctypes[ int ]] Out[41] [False, False, False, False, True] `integer` i.e.
|
|
`<type integerscalar >`) is still missing from the list !python In [46] [N.integer==x for x in N.sctypes[ int ]] Out[46] [False, False, False, False, False] Tested with !NumPy 1.0b2.dev2969.","bug"
|
|
"6013","830","NumPy breaks pydoc Trac 232).
|
|
Original ticket http projects.scipy.org numpy ticket 232 on 2006 08 05 by alberts, assigned to unknown.
|
|
Start the pydoc HTTP server c \Python24\lib\pydoc.py p 4392 Access the !NumPy documentation http localhost 4392 numpy.html problem in numpy RuntimeError digitize method already has a docstring Tested with !NumPy 1.0b2.dev2969.","bug"
|
|
"6013","831","NumPy breaks pydoc Trac 233).
|
|
Original ticket http projects.scipy.org numpy ticket 233 on 2006 08 05 by alberts, assigned to unknown.
|
|
Start the pydoc HTTP server c \Python24\lib\pydoc.py p 4392 Access the !NumPy documentation http localhost 4392 numpy.html problem in numpy RuntimeError digitize method already has a docstring Tested with !NumPy 1.0b2.dev2969.","bug"
|
|
"6013","832","IntelVisualFCompiler is broken Trac 234).
|
|
Original ticket http projects.scipy.org numpy ticket 234 on 2006 08 06 by alberts, assigned to unknown.
|
|
Patch to be attached.
|
|
Setting ar exe from MSVCCompiler ).lib seems to be unnecessary.","bug"
|
|
"6013","833","r , c , hstack, vstack, column stack should be made more consistent Trac 235).
|
|
Original ticket http projects.scipy.org numpy ticket 235 on 2006 08 07 by baxissimo, assigned to unknown.
|
|
Here s essentially what these different methods do vstack tup) concatenate map atleast 2d,tup), axis=0 ) hstack tup) concatenate map atleast 1d,tup),axis=1 ) column stack tup) arrays = map transpose,map atleast 2d,tup) ) concatenate arrays,1) note that column stack transposes everything not just 1 d inputs, so it doesn t do quite what I would expect, i.e.
|
|
only transposing 1 d inputs) The above 3 are pretty much exactly the code used by numpy.
|
|
That s all there is to those 3 functions.
|
|
For r and c I m summarizing, but effectively they do something like r [args] concatenate map atleast 1d,args),axis=0 ) c [args] concatenate map atleast 1d,args),axis=1 ) c behaves almost exactly like hstack with the addition of range literals being allowed.
|
|
r is most like vstack, but a little different since it effectively uses atleast 1d, instead of atleast 2d.
|
|
So you have >>> numpy.vstack 1,2,3,4)) array [[1], [2], [3], [4]]) but >>> numpy.r [1,2,3,4] array [1, 2, 3, 4]) However for cases like that with just 0 d or 1 d inputs, c behaves identically to r , so if you wanted to get a 1 d output you could have just used c .
|
|
I think the right thing to do would be to make r behave more like vstack.
|
|
That would make things more consistent, and make for less for the user to remember.
|
|
After making that change, to make things even more consistent, it might make sense to rename r and c to v and h instead.
|
|
Then it s easy to remember v is like vstack , h is like hstack.
|
|
Furthermore, I propose that column stack should only transpose its 1d inputs.
|
|
Stack colums defnitely doesn t imply to me that something that already has columns will be transposed.
|
|
Currently it is documented to only work on 1d inputs, so hopefully that s a change that wouldn t affect too many people.
|
|
The function in numpy lib shape base.py could be replaced with this def column stack tup) def transpose 1d array) if array.ndim<2 return nx.transpose atleast 2d array)) else return array arrays = map transpose 1d,map atleast 1d,tup)) return nx.concatenate arrays,1) If r , and c get renamed to v , h , then c could be re introduced with behavior similar to column stack.
|
|
SUMMARY make r behave like vstack plus range literals make column stack only transpose its 1d inputs.
|
|
rename r ,c to v ,h or something else) to make their connection with vstack and hstack clearer.
|
|
Maybe vs and hs would be better?
|
|
make a new vertsion of c that acts like column stack so that there s a nice parallel v <=>vstack, h <=>hstack, c <=>column stack","enhancement"
|
|
"6013","834","reduceat cornercase Trac 236).
|
|
Original ticket http projects.scipy.org numpy ticket 236 on 2006 08 07 by trac user martin wiechert, assigned to unknown.
|
|
<operator>.reduceat does not handle repeated indices correctly.
|
|
When an index is repeated the neutral element of the operation should be returned.
|
|
In the example below [0, 10], not [1, 10], is expected.
|
|
In [1] import numpy In [2] numpy.version.version Out[2] 1.0b1 In [3] a = numpy.arange 5) In [4] numpy.add.reduceat a, 1,1)) Out[4] array [ 1, 10])","enhancement"
|
|
"6013","835","reduceat should handle outlier indices gracefully Trac 237).
|
|
Original ticket http projects.scipy.org numpy ticket 237 on 2006 08 07 by trac user martin wiechert, assigned to unknown.
|
|
to the very least <operator>.reduceat a, I) should accept len a) in I usage example def weighted histo binborders, data) sd = sort data) I = sd.searchsorted binborders) Now elements of I may range between 0 and len sd).
|
|
return add.reduceat sd, I)","enhancement"
|
|
"6013","836","memmap segfault on byteswap maybe because read only) Trac 238).
|
|
Original ticket http projects.scipy.org numpy ticket 238 on 2006 08 07 by sebhaase, assigned to teoliphant.
|
|
With a being a memmap ed array, I get a segmentation fault when I call either `m.byteswap )` or `m.byteswap Treu)` It might be that this is related to `m` being memmap ed as read only","bug"
|
|
"6013","837","creation of object arrays not as advertised Trac 239).
|
|
Original ticket http projects.scipy.org numpy ticket 239 on 2006 08 10 by stefanv, assigned to teoliphant.
|
|
From the mailing list post at http thread.gmane.org gmane.comp.python.numeric.general 6992 focus=6992 one expects the following behaviour array [[1,2],3,4],dtype=object) returns a 1 d array of shape 3,) array [[1,2],[3,4]],dtype=object) returns a 2 d array of shape 2,2) array [ 1,2), 3,4)],dtype=object) returns a 1 d array of shape 2,) array [],dtype=object) returns a 0 d array of shape ) array [[],[],[]],dtype=object) returns a 1 d array of shape 3,) array [[3,4],[5,6],None],dtype=object) returns a 1 d array of shape 3,) but some of these return incorrect results, for example array [[1,2],3,4],dtype=object) yields a 3,2) array.","bug"
|
|
"6013","838","Correct the byte swaping with unicode strings Trac 240).
|
|
Original ticket http projects.scipy.org numpy ticket 240 on 2006 08 10 by trac user Harbort, assigned to unknown.
|
|
The function UNICODE copyswap and UNICODE copynswap contained a very small bug that prevented the swap to be performed correctly.
|
|
I attach a patch with the very small) correction.","bug"
|
|
"6013","839","maximum and nan identity for maximum Trac 241).
|
|
Original ticket http projects.scipy.org numpy ticket 241 on 2006 08 11 by trac user martin wiechert, assigned to unknown.
|
|
How should maximum handle nans?
|
|
I think there are only two reasonable behaviours.
|
|
Either nans should be ignored or the result should be nan.
|
|
I personally favour a keyword switch over a separate function nanmax.)
|
|
In the current implementation the outcome depends on the order of elements >>> import numpy >>> numpy.version.version 1.0b1 >>> from numpy import >>> a = array 1, nan)) >>> maximum.reduce a) 1.0 >>> maximum.reduce a [ 1]) nan This is certainly confusing if not a bug.
|
|
As a related request I d like to see an identity element for maximum, namely inf.
|
|
All the above of course also holds for minimum.","enhancement"
|
|
"6013","840","Long command lines in build don t work on Windows Trac 242).
|
|
Original ticket http projects.scipy.org numpy ticket 242 on 2006 08 12 by alberts, assigned to dmcooke.
|
|
As I mentioned on the !SciPy mailing list http projects.scipy.org pipermail scipy user 2006 August 008910.html Windows doesn t like long command lines, as documented here http blogs.msdn.com oldnewthing archive 2003 12 10 56028.aspx The command line that results when building LAPACK from source is more than 70000 characters long, which causes compilation to fail on Windows.
|
|
In the original distutils, the spawn nt method will have to be fixed.
|
|
However, it seems that numpy.distutils already overrides the spawn methods in original distutils, so this problem can be fixed in numpy distutils exec command.py.
|
|
As a reference, SCons already has a mechanism for dealing with this.
|
|
The relevant code is contained in the !TempFileMunge class, here http scons.tigris.org source browse scons trunk src engine SCons Platform init .py?rev=1582 view=markup","bug"
|
|
"6013","841","Memory leak in around Trac 243).
|
|
Original ticket http projects.scipy.org numpy ticket 243 on 2006 08 12 by alberts, assigned to teoliphant.
|
|
According to Valgrind 2.4.0, the following code leaks !python import numpy x = numpy.zeros 1,)) y = [0] from numpy.core import around decimal = 6 around abs x y),decimal) <= 10.0 decimal) Valgrind error ==15583== 56 40 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 8 of 28 ==15583== at 0x40051F9 malloc vg replace malloc.c 149) ==15583== by 0x4434AF2 array alloc arrayobject.c 6445) ==15583== by 0x443C59A PyArray NewFromDescr arrayobject.c 5063) ==15583== by 0x443E4B3 PyArray Empty multiarraymodule.c 5036) ==15583== by 0x4460CD8 PyArray Round multiarraymodule.c 293) ==15583== by 0x4460E2C array round arraymethods.c 1652) ==15583== by 0x5F15CFC PyCFunction Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with !NumPy 1.0b2.dev3000.","bug"
|
|
"6013","842","Build fails with Intel Visual Fortran compiler Trac 244).
|
|
Original ticket http projects.scipy.org numpy ticket 244 on 2006 08 13 by alberts, assigned to pearu.
|
|
I am building NumPy inside the Build Environment for Fortran IA 32 applications with the Intel Fortran Compiler version 9.1.
|
|
I am using the following command python setup.py config compiler=msvc fcompiler=intelv build clib compiler=msvc fcompiler=intelv build ext compiler=msvc fcompiler=intelv bdist wininst Because I am building inside this environment, the Intel compiler should be able to find all its libraries.
|
|
After applying the patch in ticket 840, the !NumPy build proceeds but the get version magic tries to link a dummy Fortran file which fails with various errors libifcore.lib libifcoremain.obj) error LNK2019 unresolved external symbol MAIN referenced in function main c \docume 1\albert\locals 1\temp\tmpwlplvf dummy.o fatal error LNK1120 1 unresolved externals Later the compiler is invoked again and this fails with LINK fatal error LNK1104 cannot open file ifconsol.lib","bug"
|
|
"6013","843","ndpointer test and fixes Trac 245).
|
|
Original ticket http projects.scipy.org numpy ticket 245 on 2006 08 14 by alberts, assigned to teoliphant.
|
|
Patch to be attached.
|
|
Fixes TypeError when raising TypeError for shape mismatch.
|
|
Raise a TypeError if an invalid object is passed in the flags argument.
|
|
Currently setting ndpointer s flags from an array s flags doesn t work.
|
|
Instead, one has to pass arr.flags.num.
|
|
Maybe this should be fixed.","bug"
|
|
"6013","844","chararray comparisons fail Trac 246).
|
|
Original ticket http projects.scipy.org numpy ticket 246 on 2006 08 14 by chanley, assigned to unknown.
|
|
Assuming that the numpy chararray is to behave like numarray s chararray, comparisons should ignore trailing white space.
|
|
Please see the following example.
|
|
For numpy In [1] from numpy import char In [2] a = char.array G185M , G185M , G185M )) In [3] a Out[3] chararray [G185M, G185M, G185M], dtype= |S8 ) In [4] print a == G185M [True False False] For numarray In [5] from numarray import strings as char In [6] a = char.array G185M , G185M , G185M )) In [7] a Out[7] CharArray [ G185M , G185M , G185M ]) In [8] print a == G185M [1 1 1]","bug"
|
|
"6013","845","GCC 4.1.1 warnings for ufuncobject.c Trac 247).
|
|
Original ticket http projects.scipy.org numpy ticket 247 on 2006 08 14 by alberts, assigned to teoliphant.
|
|
GCC 4.1.1 emits the following warnings when compiling ufuncobject.c in r3015 numpy core src ufuncobject.c In function construct reduce numpy core src ufuncobject.c 1873 warning value computed is not used numpy core src ufuncobject.c 1764 warning outsize may be used uninitialized in this function","bug"
|
|
"6013","846","linalg.det doesn t check return value of LAPACK function Trac 248).
|
|
Original ticket http projects.scipy.org numpy ticket 248 on 2006 08 14 by alberts, assigned to unknown.
|
|
The value assigned to the local variable `results` in the det function in linalg\linalg.py isn t checked for an error condition.
|
|
If no error can occur here, the return value from the LAPACK function should simply be ignored.
|
|
Found in NumPy 1.0b3.dev3013.
|
|
Found with pychecker.","bug"
|
|
"6013","847","strnlen in arrayobject.c is a GNU extension Trac 249).
|
|
Original ticket http projects.scipy.org numpy ticket 249 on 2006 08 14 by alberts, assigned to teoliphant.
|
|
A call to the strnlen function was introduced in arrayobject.c in r3016.
|
|
According to the strnlen manual page, this function is a GNU extension, so it not generally available especially not on Windows).","bug"
|
|
"6013","848","finfo equivalent for integer types Trac 250).
|
|
Original ticket http projects.scipy.org numpy ticket 250 on 2006 08 16 by alberts, assigned to unknown.
|
|
An equivalent to finfo for integer types would be very useful.
|
|
The main info I m interested in is maximum and minimum.
|
|
This is useful when converting between floating point and integer representation when you want to normalize your data before converting it.","enhancement"
|
|
"6013","849","pickling numpy types Trac 251).
|
|
Original ticket http projects.scipy.org numpy ticket 251 on 2006 08 16 by trac user IljyaKalai, assigned to unknown.
|
|
Trying to pickle numpy types raises an error.
|
|
>>> import numpy >>> import pickle >>> pickle.dumps numpy.float32) PicklingError Can t pickle <type float32 arrtype > it s not found as builtin .float32 arrtype I believe this is because the types module attribute is set to builtin instead of numpy.
|
|
>>> import numpy >>> numpy.float32.
|
|
module builtin For anyone else needing this functionality, I am using the following work around >>> import numpy >>> import pickle >>> import builtin >>> builtin .float32 arrtype = numpy.float32 >>> pickle.dumps numpy.float32) c builtin \nfloat32 arrtype\np0\n.
|
|
The name binding must also be done when unpickling.","bug"
|
|
"6013","850","numpy oldnumeric alter code1.py doesn t work.
|
|
[Fix included] Trac 252).
|
|
Original ticket http projects.scipy.org numpy ticket 252 on 2006 08 17 by zachrahan, assigned to unknown.
|
|
Line 67 of numpy oldnumeric alter code1.py is in error fstr = func2 re[name].sub \\1B\\2 ,fstr) It should be fstr = func re[name].sub \\1B\\2 ,fstr) because there is no func2 re symbol.","bug"
|
|
"6013","851","numpy oldnumeric alter code1.py should convert old style types.
|
|
Trac 253).
|
|
Original ticket http projects.scipy.org numpy ticket 253 on 2006 08 17 by zachrahan, assigned to unknown.
|
|
Recently types like numpy.Float32 and numpy.Float were deprecated in favor of numpy.float32 and numpy.float .
|
|
These changes should be reflected in the alter code .py routines.
|
|
Specifically, old style capitalized type names should be substituted for their current counterparts.","enhancement"
|
|
"6013","852","ma numpy arithmetic not commutative Trac 254).
|
|
Original ticket http projects.scipy.org numpy ticket 254 on 2006 08 18 by trac user guyer, assigned to unknown.
|
|
>>> numpy.
|
|
version 1.0b3.dev3032 >>> import numpy >>> from numpy.core import ma >>> a = ma.zeros 4,1)) >>> a[2,0] = ma.masked >>> b = numpy.zeros 4,2)) >>> a b Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy core ma.py , line 944, in mul return multiply self, other) File System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy core ma.py , line 379, in call return masked array result, m) File System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy core ma.py , line 1601, in masked array return array a, mask=mask, copy=0, fill value=fill value) File System Library Frameworks Python.framework Versions 2.3 lib python2.3 site packages numpy core ma.py , line 588, in init raise MAError, Mask and data not compatible.
|
|
numpy.core.ma.MAError Mask and data not compatible.
|
|
>>> b a array data = [[ 0.00000000e 00 0.00000000e 00] [ 0.00000000e 00 0.00000000e 00] [ 1.00000000e 20 1.00000000e 20] [ 0.00000000e 00 0.00000000e 00]], mask = [[False False] [False False] [True True] [False False]], fill value=1e 20)","bug"
|
|
"6013","853","Better declaration of the PyArray SETITEM macro Trac 255).
|
|
Original ticket http projects.scipy.org numpy ticket 255 on 2006 08 18 by FrancescAlted, assigned to unknown.
|
|
In include numpy arrayobject.h I think that the next patch should be applied numpy core include numpy arrayobject.h.orig 2006 08 18 12 43 24.000000000 0200 numpy core include numpy arrayobject.h 2006 08 18 12 44 45.000000000 0200 1258,8 1258,8 PyArrayObject ) obj)) >descr >f >getitem char )itemptr, \ PyArrayObject )obj); define PyArray SETITEM obj,itemptr,v) \ obj) >descr >f >setitem PyObject )v, char ) itemptr), \ PyArrayObject ) obj)); PyArrayObject ) obj)) >descr >f >setitem PyObject )v, char ) itemptr), \ PyArrayObject ) obj)); Cheers!","bug"
|
|
"6013","854","Error when importing a complex buffer from numarray Trac 256).
|
|
Original ticket http projects.scipy.org numpy ticket 256 on 2006 08 18 by FrancescAlted, assigned to unknown.
|
|
Hi, When trying to create a nested recarray from a numarray buffer with the next script import numpy print numpy version > , numpy.
|
|
version import numarray print numarray version > , numarray.
|
|
version import numarray.records Create a rather complex plain) recarray formats=[ 2,)Int32 , 1Complex64 , 1Float64 , 1a2 , 2,)Complex64 , 2,)Float64 , 1a2 , 1UInt8 , 1a4 , 1a2 , 1Complex64 , 2, 2)Float64 , 1UInt8 , 2,)UInt32 ] rna=numarray.records.array None, shape=1, formats=formats) Get a nested numpy recarray from the above buffer dtype=numpy.dtype [ x , <i4 , 2,)), Info , [ value , <c16 ), y2 , <f8 ), Info2 , [ name , |S2 ), value , <c16 , 2,)), y3 , <f8 , 2,))]), name , |S2 ), z2 , <u1 )]), color , |S4 ), info , [ Name , |S2 ), Value , <c16 )]), y , <f8 , 2, 2)), z , <u1 ), z3 , <u4 , 2,))]) r = numpy.array rna,dtype=dtype) print numpy result > , r I get a strange error with NumPy 1.0b2) numpy version > 1.0b2 numarray version > 1.5.1 descr > [ title1 , x ), <i4 , 2,)), title2 , Info value ), <c8 ), title3 , Info y2 ), <f8 )] descr2 > Traceback most recent call last) File numpy bug2.py , line 14, in ?
|
|
print descr2 > , numpy.rec.format parser formats2,names,titles).
|
|
descr File usr lib python2.4 site packages numpy core records.py , line 44, in init self.
|
|
parseFormats formats, aligned) File usr lib python2.4 site packages numpy core records.py , line 53, in parseFormats dtype = sb.dtype formats, aligned) TypeError data type not understood However, with earlier versions 0.9.8) this used to work well numpy version > 0.9.8 numarray version > 1.5.1 numpy result > [ array [ 1209358944, 1209358944]), 9.1961225082056555e 270 9.1336194581362765e 270j), 9.1938696256247779e 270, \x18o , array [ 1.03758533e 226 1.42324107e 227j, 2.20974089e 218 1.13138734e 221j]), array [ 2.37269122e 227, 1.62122697e 224])), \x13\x08 , 160), n\x13\x08\xac , n\x13 , 1.4055534309680069e 202 9.8467770594688458e 111j)), array [[ 9.84677706e 111, 9.84677706e 111], [ 4.50991060e 215, 4.51670943e 215]]), 8, array [135491364, 135491352], dtype=uint32))] Thanks","bug"
|
|
"6013","855","numpy.rec.format parser should understand numarray notation Trac 257).
|
|
Original ticket http projects.scipy.org numpy ticket 257 on 2006 08 18 by FrancescAlted, assigned to teoliphant.
|
|
It seems that numpy.rec.format parser doesn t understand the numarray notation for data types as can be seen in the next script import numpy print numpy version > , numpy.
|
|
version import numarray print numarray version > , numarray.
|
|
version import numarray.records formats=[ 2,)int32 , 1complex64 , 1float64 ] numpy notation names=[ x , Info value , Info y2 ] titles = [ title1 , title2 , title3 ] print descr > , numpy.rec.format parser formats,names,titles).
|
|
descr So far, so good, but... shouldn t numarray notation should be supported as well?
|
|
formats2=[ 2,)Int32 , 1Complex64 , 1Float64 ] numarray notation print descr2 > , numpy.rec.format parser formats2,names,titles).
|
|
descr and the output in my machine is numpy version > 1.0b2 numarray version > 1.5.1 descr > [ title1 , x ), <i4 , 2,)), title2 , Info value ), <c8 ), title3 , Info y2 ), <f8 )] descr2 > Traceback most recent call last) File numpy bug2.py , line 14, in ?
|
|
print descr2 > , numpy.rec.format parser formats2,names,titles).
|
|
descr File usr lib python2.4 site packages numpy core records.py , line 44, in init self.
|
|
parseFormats formats, aligned) File usr lib python2.4 site packages numpy core records.py , line 53, in parseFormats dtype = sb.dtype formats, aligned) TypeError data type not understood Cheers","enhancement"
|
|
"6013","856","fix numerictypes to make sure all types are registered Trac 258).
|
|
Original ticket http projects.scipy.org numpy ticket 258 on 2006 08 18 by teoliphant, assigned to unknown.
|
|
The numerictypes module does not define several variables that it used to like bool8) Fix.","bug"
|
|
"6013","857","Ctrl C interrupt handling Trac 259).
|
|
Original ticket http projects.scipy.org numpy ticket 259 on 2006 08 19 by teoliphant, assigned to unknown.
|
|
Add macros adapted from SAGE for Interrupt handling during computing.","enhancement"
|
|
"6013","858","Add mechanism for registering objects to be deallocated and memory to be freed at Python exit Trac 260).
|
|
Original ticket http projects.scipy.org numpy ticket 260 on 2006 08 19 by teoliphant, assigned to unknown.
|
|
A couple of API s to set up freelists of both Python Objects to be DECREF d and memory pointers to be freed on Python exit would be useful.
|
|
We could then use the atexit module to call the freeing function.","enhancement"
|
|
"6013","859","Add API for replacing a UFunc signature Trac 261).
|
|
Original ticket http projects.scipy.org numpy ticket 261 on 2006 08 19 by teoliphant, assigned to unknown.
|
|
Need an API for replacing the UFUNC 1 d loops for a particular signature.
|
|
The same thing can already be done easily with the data type pointers.","enhancement"
|
|
"6013","860","F2py might be copying unnecessarily Trac 262).
|
|
Original ticket http projects.scipy.org numpy ticket 262 on 2006 08 19 by teoliphant, assigned to unknown.
|
|
f2py seems to be copying memory even when it doesn t need to.
|
|
The comparison with pyrex ctypes weave showed over head likely due to copying even when given Fortran arrays.","bug"
|
|
"6013","861","Inconsistency in type shapes 1 and 1,) Trac 263).
|
|
Original ticket http projects.scipy.org numpy ticket 263 on 2006 08 19 by FrancescAlted, assigned to unknown.
|
|
I found what I think is an inconsistent behaviour in the shape definition for datatypes in numpy.
|
|
Look at this >>> dt1=numpy.dtype numpy.int32) >>> dt2=numpy.dtype numpy.int32, 1,))) >>> dt1.shape 1,) >>> dt2.shape 1,) I think that dt1 should have a different shape than dt2.
|
|
In particular, dt1 should have a shape == 1 instead of 1,).
|
|
The difference can be significant in cases like >>> numpy.arange 1,dtype=dt1) array [0]) >>> numpy.arange 1,dtype=dt2) array [[0]]) and, perhaps more importantly >>> type numpy.arange 1,dtype=dt1)[0]) <type int32scalar > >>> type numpy.arange 1,dtype=dt2)[0]) <type numpy.ndarray > in one case you get an int32scalar and in the other case an ndarray object, although looking at the dtype object you cannot notice the difference and you should, IMO.
|
|
Using 1.0b2 version, but 0.9.8 has the same behavior.","bug"
|
|
"6013","862","linspace endpoint not always honored Trac 264).
|
|
Original ticket http projects.scipy.org numpy ticket 264 on 2006 08 21 by baxissimo, assigned to unknown.
|
|
It is not always the case that linspace start,stop,num)[ 1] == stop The following little example should output [], but it doesn t. In [390] filter lambda x x[1]!=0.0, [ i, 1.0 numpy.linspace 0,1,i)[ 1]) for i in range 2,200) ]) Out[390] [ 50, 1.1102230246251565e 016), 99, 1.1102230246251565e 016), 104, 1.1102230246251565e 016), 108, 1.1102230246251565e 016), 162, 1.1102230246251565e 016), 188, 1.1102230246251565e 016), 197, 1.1102230246251565e 016), 198, 1.1102230246251565e 016)] I know it s not a good idea to count on floating point equality in general, but it doesn t seem too much to expect that the first and last values returned by linspace are exactly the values asked for if they both have exact floating point representations.
|
|
The simplest fix is to just add in a y[ 1] = stop in the case of endpoint=True.
|
|
So the code would become def linspace start, stop, num=50, endpoint=True, retstep=False) Return evenly spaced numbers.
|
|
Return num evenly spaced samples from start to stop .
|
|
If endpoint is True, the last sample is stop .
|
|
If retstep is True then return the step value used.
|
|
num = int num) if num <= 0 return array [], float) if endpoint if num == 1 return array [float start)]) step = stop start) float num 1)) y = nx.arange 0, num) step start y[ 1] = float stop) else step = stop start) float num) y = nx.arange 0, num) step start if retstep return y, step else return y Note, the above change would have the effect of making linspace start,stop,1) return stop, rather than start.
|
|
If this is considered undesirable, it s a simple matter to special case the num==1 case for backwards compatibility.
|
|
I.e.
|
|
if num>1 y[ 1] = float stop)","bug"
|
|
"6013","863","Big endian complex array s imag returns invalid values Trac 265).
|
|
Original ticket http projects.scipy.org numpy ticket 265 on 2006 08 22 by alberts, assigned to unknown.
|
|
!python In [11] N.arange 4, dtype= <c8 ).imag.max ) Out[11] 0.0 In [10] N.arange 4, dtype= >c8 ).imag.max ) Out[10] 4.6006029882248069e 041 Tested with 1.0b4.dev3050.
|
|
Reported by Sebastian Haase.","bug"
|
|
"6013","864","Problems creating chararrays from numarray chararray objects Trac 266).
|
|
Original ticket http projects.scipy.org numpy ticket 266 on 2006 08 23 by FrancescAlted, assigned to unknown.
|
|
Hi, I don t know if this is a bug or not, but this represents a change over NumPy 0.9.8 and I thought that I should report it The next script import numpy print numpy version > , numpy.
|
|
version import numarray print numarray version > , numarray.
|
|
version import numarray.strings sna=numarray.strings.array None, itemsize=5, shape= 4,2)) print itemsize reported by numpy , numpy.asarray sna).itemsize sna=numarray.strings.array a 40, itemsize=5, shape= 4,2)) print itemsize reported by numpy , numpy.asarray sna).itemsize gives the next output in 1.0b2 also in 1.0b4.dev3053) numpy version > 1.0b2 numarray version > 1.5.1 itemsize reported by numpy 1 itemsize reported by numpy 5 where the first itemsize reported by numpy is wrong.
|
|
However, this used to work well in numpy 0.9.8 numpy version > 0.9.8 numarray version > 1.5.1 itemsize reported by numpy 5 itemsize reported by numpy 5 FWIW, here is a workaround that allows creating a numpy chararray from a numarray chararray buffer sna=numarray.strings.array None, itemsize=5, shape= 4,2)) dtype = numpy.dtype |S s sna.itemsize )) snp=numpy.ndarray buffer=sna.
|
|
data, dtype=dtype, shape=sna.shape) print itemsize reported by numpy , snp.itemsize sna=numarray.strings.array a 40, itemsize=5, shape= 4,2)) dtype = numpy.dtype |S s sna.itemsize )) snp=numpy.ndarray buffer=sna.
|
|
data, dtype=dtype, shape=sna.shape) print itemsize reported by numpy , snp.itemsize which gives numpy version > 1.0b2 numarray version > 1.5.1 itemsize reported by numpy 5 itemsize reported by numpy 5 that is what I think is the correct answer.
|
|
Regards,","bug"
|
|
"6013","865","Build fails on Windows with MSVC Trac 267).
|
|
Original ticket http projects.scipy.org numpy ticket 267 on 2006 08 24 by alberts, assigned to teoliphant.
|
|
The recent signal handling changes have caused some compile errors with MSVC C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG Ibuild\src.win32 2.4\numpy\core\src Inumpy\core\include Ibuild\src.win32 2.4\numpy\core Inumpy\core\src Inumpy\core\include IC \Python24\include IC \Python24\PC Tcnumpy\core\src\multiarraymodule.c Fobuild\temp.win32 2.4\Release\numpy\core\src\multiarraymodule.obj multiarraymodule.c c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 95) error C2061 syntax error identifier NPY SIGINT BUF c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 95) error C2059 syntax error ; c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 100) error C2065 SIG IGN undeclared identifier c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 100) warning C4047 function PyOS sighandler t differs in levels of indirection from int c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 101) warning C4013 longjmp undefined; assuming extern returning int c \home\albert\work2\numpy\numpy\core\include\numpy\npy interrupt.h 101) error C2065 NPY SIGINT BUF undeclared identifier numpy\core\src\arraytypes.inc.src 1045) error C2036 void unknown size numpy\core\src\arraytypes.inc.src 1045) error C2036 void unknown size numpy\core\src\arraytypes.inc.src 1045) error C2036 void unknown size numpy\core\src\multiarraymodule.c 5459) warning C4018 <= signed unsigned mismatch numpy\core\src\multiarraymodule.c 6441) error C2065 SIGINT undeclared identifier numpy\core\src\multiarraymodule.c 6441) warning C4013 setjmp undefined; assuming extern returning int error Command C \Program Files\Microsoft Visual Studio .NET 2003\Vc7\bin\cl.exe c nologo Ox MD W3 GX DNDEBUG Ibuild\src.win32 2.4\numpy\core\src Inumpy\core\include Ibuild\src.win32 2.4\numpy\core Inumpy\core\src Inumpy\core\include IC \Python24\include IC \Python24\PC Tcnumpy\core\src\multiarraymodule.c Fobuild\temp.win32 2.4\Release\numpy\core\src\multiarraymodule.obj failed with exit status 2 The setjmp undefined error can probably be sorted out by including the ANSI standard header setjmp.h.
|
|
The void unknown size error has cropped up before.
|
|
I think this was solved by casting the offending pointer to char .","bug"
|
|
"6013","866","Version of recent g95 release is not being detected correctly Trac 268).
|
|
Original ticket http projects.scipy.org numpy ticket 268 on 2006 08 24 by trac user ellisonbg ..., assigned to unknown.
|
|
G95 has release a new version that has a slightly different format of its version string > g95 version G95 GCC 4.0.3 g95 0.90!)
|
|
Aug 22 2006) The old version string looked like G95 GCC 4.0.3 g95!)
|
|
May 22 2006) Fixing will require changing the following line at the top of numpy distutils fcompiler g95.py version pattern = r G95 \ GCC ?P<gccversion>[\d.]
|
|
)|.
|
|
?)
|
|
\ g95!\) ?P<version>.
|
|
)\).
|
|
I am not sure if the version should be 0.90 or Aug 22 2006","bug"
|
|
"6013","867","scipy.take does not function correctly Trac 269).
|
|
Original ticket http projects.scipy.org numpy ticket 269 on 2006 08 29 by trac user kantor, assigned to unknown.
|
|
I came across what I believe is a bug with the scipy.take function.
|
|
I was having trouble with optimize.leastsq.
|
|
On one computer powerpc) it was able to compute the covariance matrix of a least squares fit, however on a x86) this failed with an error about inverse expecting a square matrix.
|
|
I traced this to the fact that on the powerpc scipy.take scipy.eye 2),[1,0]) correctly returns the matrix [[0,1],[1,0]], however on the x86 machine the same code only returns [0,1].
|
|
Oddly directly using Numeric.take works correctly on x86 gives the full matrix).
|
|
This makes the leastsq function when used with full output) in scipy optimize minpack.py fail because the computed permutation matrix is just a vector on x86.
|
|
These errors were on system running ubuntu in all cases.
|
|
The scipy and numpy packages were built from the source on Andrew Staw s deb repository.","bug"
|
|
"6013","868","array creation from a list containing objects fails.
|
|
Trac 270).
|
|
Original ticket http projects.scipy.org numpy ticket 270 on 2006 08 31 by charris, assigned to unknown.
|
|
In [2] a = array [1, A , None]) exceptions.TypeError Traceback most recent call last) home charris <ipython console> TypeError expected a readable buffer object In [3] a = array [1, A , None], dtype=object) In [4] a.shape Out[4] 3,) Statement [2] should be the same as [3].","bug"
|
|
"6013","869","Bug in ma.repeat Trac 271).
|
|
Original ticket http projects.scipy.org numpy ticket 271 on 2006 09 04 by trac user PGM, assigned to unknown.
|
|
Folks, The `repeat` function in MA doesn t work with the default `axis=None`.
|
|
indeed, shape af)[None] doesn t work).
|
|
Cf code below version 1.0b4) ........................................................... >>> MA.array [1],mask=False).repeat 10,0) array [1 1 1 1 1 1 1 1 1 1]) >>> MA.array [1],mask=False).repeat 10) usr lib64 python2.4 site packages numpy core ma.py in repeat a, repeats, axis) 1544 af = filled a) 1545 if isinstance repeats, types.IntType) > 1546 repeats = tuple [repeats] shape af)[axis])) 1547 1548 m = getmask a) TypeError tuple indices must be integers > usr lib64 python2.4 site packages numpy core ma.py 1546)repeat ) 1545 if isinstance repeats, types.IntType) > 1546 repeats = tuple [repeats] shape af)[axis])) 1547","bug"
|
|
"6013","870","lexsort memory leak Trac 272).
|
|
Original ticket http projects.scipy.org numpy ticket 272 on 2006 09 05 by alberts, assigned to teoliphant.
|
|
According to Valgrind 3.1.0, the code from check lexsort in test regression.py leaks memory.
|
|
Short snippet to reproduce !python import numpy as N N.lexsort N.array [1,2])) Valgrind says ==24663== 2,856 1,424 direct, 1,432 indirect) bytes in 2 blocks are definitely lost in loss record 21 of 27 ==24663== at 0x40051F9 malloc vg replace malloc.c 149) ==24663== by 0x5F170D0 PyMem Malloc in usr lib libpython2.4.so.1.0) ==24663== by 0x445831D PyArray LexSort multiarraymodule.c 2379) ==24663== by 0x4458F14 array lexsort multiarraymodule.c 6367) ==24663== by 0x5F15CFC PyCFunction Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with !NumPy 1.0rc1.dev3120.","bug"
|
|
"6013","871","crash in broadcast code segfault) Trac 273).
|
|
Original ticket http projects.scipy.org numpy ticket 273 on 2006 09 06 by bthyreau, assigned to unknown.
|
|
Hi.. Just found a crash condition.
|
|
cat crash.py from numpy import print version a = zeros 3, 1), int) a[[1,2]] = 1 python crash.py 1.0rc1.dev3124 Erreur de segmentation According to valgrind ==17992== Process terminating with default action of signal 11 SIGSEGV) ==17992== Access not within mapped region at address 0x4 ==17992== at 0x454FAC9 PyArray BroadcastToShape arrayobject.c 8694) ==17992== by 0x4583563 array ass sub arrayobject.c 2417) ==17992== by 0x80B20DE PyEval EvalFrame ceval.c 1474) The example wouldn t crash if the a array were 3,) shaped Thanks!","bug"
|
|
"6013","872","Speed up N D Boolean indexing Trac 274).
|
|
Original ticket http projects.scipy.org numpy ticket 274 on 2006 09 06 by teoliphant, assigned to unknown.
|
|
Currently N D Boolean indexing works by creating intermediate integer index arrays for N>1.
|
|
This should be avoided when possible so that Boolean indexing is faster for N>1.","enhancement"
|
|
"6013","873","numpy.oldnumeric.mlab.eye fails with NameError global name nn is not defined Trac 275).
|
|
Original ticket http projects.scipy.org numpy ticket 275 on 2006 09 08 by trac user jks ..., assigned to unknown.
|
|
The eye function in numpy.oldnumeric.mlab which is used by matplotlib.numerix for compatibility reasons, and thus imported in ipython pylab) fails with an error >>> import numpy.oldnumeric.mlab as mlab >>> mlab.eye 3) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File Library Frameworks Python.framework Versions 2.4 lib python2.4 site packages numpy oldnumeric mlab.py , line 24, in eye m = nn.equal nn.subtract.outer nn.arange N), nn.arange M)), k) NameError global name nn is not defined This happens in numpy 1.0b5 btw, this version is missing from the Trac version menu).","bug"
|
|
"6013","874","problem building numpy after svn checkout with subversion 1.4 Trac 276).
|
|
Original ticket http projects.scipy.org numpy ticket 276 on 2006 09 12 by ddale, assigned to unknown.
|
|
This morning I upgraded to subversion 1.4.
|
|
The svn website claims that version 1.4 is compatible with servers running all previous versions of svn, but ever since the update I havent been able to build numpy python setup.py build Running from numpy source directory.
|
|
non existing path in numpy distutils site.cfg No module named svn version Traceback most recent call last) File setup.py , line 89, in ?
|
|
setup package ) File setup.py , line 82, in setup package configuration=configuration ) File home darren src numpy numpy distutils core.py , line 144, in setup config = configuration ) File setup.py , line 48, in configuration config.add subpackage numpy ) File home darren src numpy numpy distutils misc util.py , line 753, in add subpackage caller level = 2) File home darren src numpy numpy distutils misc util.py , line 736, in get subpackage caller level = caller level 1) File home darren src numpy numpy distutils misc util.py , line 683, in get configuration from setup py config = setup module.configuration args) File .
|
|
numpy setup.py , line 8, in configuration config.add subpackage f2py ) File home darren src numpy numpy distutils misc util.py , line 753, in add subpackage caller level = 2) File home darren src numpy numpy distutils misc util.py , line 736, in get subpackage caller level = caller level 1) File home darren src numpy numpy distutils misc util.py , line 683, in get configuration from setup py config = setup module.configuration args) File numpy f2py setup.py , line 39, in configuration config.make svn version py ) File home darren src numpy numpy distutils misc util.py , line 1298, in make svn version py self.add data files , generate svn version py ))) File home darren src numpy numpy distutils misc util.py , line 1281, in generate svn version py assert revision is not None, hmm, why I am not inside SVN tree???
|
|
AssertionError hmm, why I am not inside SVN tree???
|
|
I tried deleting my old checkout and grabbing a new one, but that didn t help.
|
|
I think numpy s distutils may be incompatible with svn s new working copy format, which is mentioned at http subversion.tigris.org svn 1.4 releasenotes.html.","bug"
|
|
"6013","875","memory error in assignment to flat object array Trac 277).
|
|
Original ticket http projects.scipy.org numpy ticket 277 on 2006 09 14 by stefanv, assigned to unknown.
|
|
Neilen Marais notified me of this problem in r3154.
|
|
The following snippet tries to assign an invalid value to a flattened object array import numpy as N x = N.zeros 2,2),dtype=object) x.flat = [] which results in the following errors under Valgrind ==22358== Invalid read of size 4 ==22358== at 0x450F750 array flat set arrayobject.c 161) ==22358== by 0x807DDE1 PyObject GenericSetAttr in usr bin python2.4) ==22358== by 0x807D4C3 PyObject SetAttr in usr bin python2.4) ==22358== by 0x80B37FA PyEval EvalFrame in usr bin python2.4) ==22358== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==22358== by 0x80B7964 PyEval EvalCode in usr bin python2.4) ==22358== by 0x80D94CB PyRun FileExFlags in usr bin python2.4) ==22358== by 0x80D976B PyRun SimpleFileExFlags in usr bin python2.4) ==22358== by 0x8055B32 Py Main in usr bin python2.4) ==22358== by 0x4082EA1 libc start main in lib tls i686 cmov libc 2.3.6.so) ==22358== Address 0x5B152BC is 0 bytes after a block of size 4 alloc d ==22358== at 0x401C422 malloc vg replace malloc.c 149) ==22358== by 0x44F42D1 PyArray NewFromDescr arrayobject.c 5299) ==22358== by 0x450E216 PyArray FromAny arrayobject.c 7227) ==22358== by 0x450EE9B array flat set arrayobject.c 6468) ==22358== by 0x807DDE1 PyObject GenericSetAttr in usr bin python2.4) ==22358== by 0x807D4C3 PyObject SetAttr in usr bin python2.4) ==22358== by 0x80B37FA PyEval EvalFrame in usr bin python2.4) ==22358== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==22358== by 0x80B7964 PyEval EvalCode in usr bin python2.4) ==22358== by 0x80D94CB PyRun FileExFlags in usr bin python2.4) ==22358== ==22358== Invalid read of size 1 ==22358== at 0x401E24D memmove mac replace strmem.c 486) ==22358== by 0x450F112 array flat set arrayobject.c 6482) ==22358== by 0x807DDE1 PyObject GenericSetAttr in usr bin python2.4) ==22358== by 0x807D4C3 PyObject SetAttr in usr bin python2.4) ==22358== by 0x80B37FA PyEval EvalFrame in usr bin python2.4) ==22358== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==22358== by 0x80B7964 PyEval EvalCode in usr bin python2.4) ==22358== by 0x80D94CB PyRun FileExFlags in usr bin python2.4) ==22358== by 0x80D976B PyRun SimpleFileExFlags in usr bin python2.4) ==22358== by 0x8055B32 Py Main in usr bin python2.4) ==22358== Address 0x5B152BC is 0 bytes after a block of size 4 alloc d ==22358== at 0x401C422 malloc vg replace malloc.c 149) ==22358== by 0x44F42D1 PyArray NewFromDescr arrayobject.c 5299) ==22358== by 0x450E216 PyArray FromAny arrayobject.c 7227) ==22358== by 0x450EE9B array flat set arrayobject.c 6468) ==22358== by 0x807DDE1 PyObject GenericSetAttr in usr bin python2.4) ==22358== by 0x807D4C3 PyObject SetAttr in usr bin python2.4) ==22358== by 0x80B37FA PyEval EvalFrame in usr bin python2.4) ==22358== by 0x80B771E PyEval EvalCodeEx in usr bin python2.4) ==22358== by 0x80B7964 PyEval EvalCode in usr bin python2.4) ==22358== by 0x80D94CB PyRun FileExFlags in usr bin python2.4) ==22358==","bug"
|
|
"6013","876","ValueError trying to set with zero sized array Trac 278).
|
|
Original ticket http projects.scipy.org numpy ticket 278 on 2006 09 15 by nilswagner01, assigned to teoliphant.
|
|
numpy.test 1,10) results in ... ====================================================================== ERROR check manyways numpy.lib.tests.test arraysetops.test aso) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test arraysetops.py , line 128, in check manyways a = numpy.fix nItem 10 numpy.random.random nItem ) ) File usr lib64 python2.4 site packages numpy lib ufunclike.py , line 23, in fix y[x<0] = y[x<0] 1 ValueError trying to set with zero sized array ====================================================================== ERROR check setmember1d numpy.lib.tests.test arraysetops.test aso) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test arraysetops.py , line 80, in check setmember1d c = setmember1d a, b ) File usr lib64 python2.4 site packages numpy lib arraysetops.py , line 122, in setmember1d perm[ii 1] = perm[ii] ValueError trying to set with zero sized array ====================================================================== ERROR check asym numpy.lib.tests.test twodim base.test histogram2d) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test twodim base.py , line 155, in check asym H, xed, yed = histogram2d x,y, 6, 5), range = [[0,6],[0,5]], normed=True) File usr lib64 python2.4 site packages numpy lib twodim base.py , line 218, in histogram2d xbin[on edge x] = 1 ValueError trying to set with zero sized array ====================================================================== ERROR check norm numpy.lib.tests.test twodim base.test histogram2d) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test twodim base.py , line 166, in check norm H, xed, yed = histogram2d x,y,[[1,2,3,5], [1,2,3,5]], normed=True) File usr lib64 python2.4 site packages numpy lib twodim base.py , line 218, in histogram2d xbin[on edge x] = 1 ValueError trying to set with zero sized array ====================================================================== ERROR check simple numpy.lib.tests.test twodim base.test histogram2d) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test twodim base.py , line 141, in check simple H = histogram2d x, y, xedges, yedges))[0] File usr lib64 python2.4 site packages numpy lib twodim base.py , line 218, in histogram2d xbin[on edge x] = 1 ValueError trying to set with zero sized array ====================================================================== ERROR check simple numpy.lib.tests.test function base.test histogramnd) Traceback most recent call last) File usr lib64 python2.4 site packages numpy lib tests test function base.py , line 360, in check simple H, edges = histogramnd x, 2,3,3), range = [[ 1,1], [0,3], [0,3]]) File usr lib64 python2.4 site packages numpy lib function base.py , line 185, in histogramnd Ncount[i][on edge] = 1 ValueError trying to set with zero sized array Ran 492 tests in 0.609s FAILED errors=6)","bug"
|
|
"6013","877","a key should be added to typeDict Trac 279).
|
|
Original ticket http projects.scipy.org numpy ticket 279 on 2006 09 15 by FrancescAlted, assigned to unknown.
|
|
I think that typeDict should receive a new a entry for string types.
|
|
For example, the next works In [12] numpy.
|
|
version Out[12] 1.0b5.dev3084 In [13] numpy.dtype i2 ) Out[13] dtype <i2 ) In [14] numpy.typeDict[ i ] Out[14] <type numpy.int32 > but this don t In [15] numpy.dtype a2 ) Out[15] dtype |S2 ) In [16] numpy.typeDict[ a ] <type exceptions.KeyError > Traceback most recent call last) home faltet carabos repositori projectes GESTA 2006 <ipython console> in <module> ) <type exceptions.KeyError > a Thanks","bug"
|
|
"6013","878","GCC 4.1.1 warning in ufuncobject.c Trac 280).
|
|
Original ticket http projects.scipy.org numpy ticket 280 on 2006 09 17 by alberts, assigned to teoliphant.
|
|
The following warning should probably be looked at numpy core src ufuncobject.c In function PyUFunc RegisterLoopForType numpy core src ufuncobject.c 3215 warning cmp may be used uninitialized in this function","bug"
|
|
"6013","879","Remove the dft subpackage by 1.0 final Trac 281).
|
|
Original ticket http projects.scipy.org numpy ticket 281 on 2006 09 17 by alberts, assigned to teoliphant.
|
|
The dft package should be removed before releasing 1.0 final.","bug"
|
|
"6013","880","convert from dict ignores align switch when offsets are passed Trac 282).
|
|
Original ticket http projects.scipy.org numpy ticket 282 on 2006 09 17 by trac user martin wiechert, assigned to unknown.
|
|
Even if one explicitly specifies offsets it is often desirable to have the record as whole aligned.
|
|
Example python d = { formats [ <f8 , <i4 ], offsets [8,16], names [ a , b ]} C descr = convert from dict d, 1); As the align switch is ignored descr >elsize will be 20, which will typically prompt unaligned access messages on IA64 if an array with descr as dtype is used.
|
|
align is ignored, because maxalign is only updated in the else clause of the offsets test multiarraymodule.c line 4555 numpy1.0rc1.dev3171)).
|
|
Thus if offsets are given maxalign stays zero and totalsize is not adapted line 4615).","bug"
|
|
"6013","881","Inconsisteny in enumerated type mapping Trac 283).
|
|
Original ticket http projects.scipy.org numpy ticket 283 on 2006 09 18 by FrancescAlted, assigned to unknown.
|
|
Hi, I think that I ve detected an inconsistency in NumPy enumerated type mapping >>> numpy.array 1, dtype=numpy.int32).dtype.num 7 >>> numpy.array 1, dtype=numpy.int ).dtype.num 7 IMO, numpy.int32 should return 5 NPY INT) instead of 7 NPY LONG), even in 32 bit platforms like mine).
|
|
In fact, dtype i is returning the correct enum code >>> numpy.array 1, dtype= i ).dtype.num 5 Similarly, I think that the next map is wrong >>> numpy.typeDict[7] <type numpy.int32 > >>> numpy.typeDict[ l ] <type numpy.int32 > and I d say that the correct values should be >>> numpy.typeDict[7] <type numpy.int > >>> numpy.typeDict[ l ] <type numpy.int > This would allow better portability of NPY INT and NPY LONG between 32 and 64 bit platforms.
|
|
Thanks!","bug"
|
|
"6013","882","Glitch in sort function and method) description Trac 284).
|
|
Original ticket http projects.scipy.org numpy ticket 284 on 2006 09 19 by FrancescAlted, assigned to charris.
|
|
Hi, In the description for the sort ) function, the worst case for quicksort is O `n 2`) and not O n) as stated.
|
|
However, the description for argsort ) is correct.
|
|
Using NumPy 1.0b5 here.
|
|
Sorry if this has been corrected already in SVN.","bug"
|
|
"6013","883","f2py fails on F90 modules Trac 285).
|
|
Original ticket http projects.scipy.org numpy ticket 285 on 2006 09 19 by stefanv, assigned to pearu.
|
|
Using f2py test.f90 to wrap MODULE chirp transform SUBROUTINE chrft x, y, wr, wi, isize, lwork, itype, ifault) RETURN END SUBROUTINE chrft END MODULE chirp transform fails with Traceback most recent call last) File home stefan bin f2py , line 6, in ?
|
|
f2py.main ) File home stefan lib python2.4 site packages numpy f2py f2py2e.py , line 548, in main run main sys.argv[1 ]) File home stefan lib python2.4 site packages numpy f2py f2py2e.py , line 351, in run main raise TypeError, All blocks must be module blocks but got s `postlist[i][ block ]`) TypeError All blocks must be module blocks but got module This is f2py r3181.","bug"
|
|
"6013","884","numpy.test ) misses some tests on Python 2.5 Trac 286).
|
|
Original ticket http projects.scipy.org numpy ticket 286 on 2006 09 20 by alberts, assigned to unknown.
|
|
Checked with !NumPy 1.0rc1.dev3194, Python 2.4.3 and Python 2.5.
|
|
C \>C \Python24\python c import numpy; numpy.test ) Found 5 tests for numpy.distutils.misc util Found 3 tests for numpy.lib.getlimits Found 31 tests for numpy.core.numerictypes Found 32 tests for numpy.linalg Found 13 tests for numpy.core.umath Found 4 tests for numpy.core.scalarmath Found 9 tests for numpy.lib.arraysetops Found 42 tests for numpy.lib.type check Found 165 tests for numpy.core.multiarray Found 3 tests for numpy.fft.helper Found 36 tests for numpy.core.ma Found 12 tests for numpy.lib.twodim base Found 10 tests for numpy.core.defmatrix Found 1 tests for numpy.lib.ufunclike Found 4 tests for numpy.ctypeslib Found 40 tests for numpy.lib.function base Found 1 tests for numpy.lib.polynomial Found 8 tests for numpy.core.records Found 26 tests for numpy.core.numeric Found 4 tests for numpy.lib.index tricks Found 46 tests for numpy.lib.shape base Found 0 tests for main <snip> Ran 495 tests in 0.938s OK Python 2.5 is missing 165 tests and fails to import the Unicode tests.
|
|
C \>C \Python25\python c import numpy; numpy.test ) Found 5 tests for numpy.distutils.misc util Found 3 tests for numpy.lib.getlimits Found 31 tests for numpy.core.numerictypes Found 32 tests for numpy.linalg Found 13 tests for numpy.core.umath Found 4 tests for numpy.core.scalarmath Found 9 tests for numpy.lib.arraysetops Found 42 tests for numpy.lib.type check Warning FAILURE importing tests for <module numpy.core.multiarray from ...ages\\numpy\\core\\multiarray.pyd > C \Python25\Lib\site packages\numpy\core\tests\test multiarray.py 355 ImportError No module named test unicode in <module>) Found 3 tests for numpy.fft.helper Found 36 tests for numpy.core.ma Found 1 tests for numpy.lib.ufunclike Found 12 tests for numpy.lib.twodim base Found 10 tests for numpy.core.defmatrix Found 4 tests for numpy.ctypeslib Found 40 tests for numpy.lib.function base Found 1 tests for numpy.lib.polynomial Found 8 tests for numpy.core.records Found 26 tests for numpy.core.numeric Found 4 tests for numpy.lib.index tricks Found 46 tests for numpy.lib.shape base Found 0 tests for main <snip> Ran 330 tests in 0.859s OK The following code runs 627 tests on 2.4 and 2.5 !python import numpy as N N.testing.NumpyTest N).testall level=9999, verbosity=9999)","bug"
|
|
"6013","885","ndpointer test failures Trac 287).
|
|
Original ticket http projects.scipy.org numpy ticket 287 on 2006 09 20 by alberts, assigned to unknown.
|
|
The ndpointer tests fail because they weren t updated when ndpointer was moved to the ctypeslib package.
|
|
Example failure ====================================================================== ERROR check dtype numpy.lib.tests.test ndpointer.test ndpointer) Traceback most recent call last) File C \Python25\Lib\site packages\numpy\lib\tests\test ndpointer.py , line 9, in check dtype p = N.ndpointer dtype=dt) AttributeError module object has no attribute ndpointer The following test methods fail check dtype check flags check ndim check shape","bug"
|
|
"6013","886","Wrong message error for 0 dim slicing Trac 288).
|
|
Original ticket http projects.scipy.org numpy ticket 288 on 2006 09 22 by FrancescAlted, assigned to unknown.
|
|
Hi, I think the next message is wrong In [64] c=numpy.array 1) In [65] c[ ] <type exceptions.ValueError > Traceback most recent call last) home faltet carabos repositori ceei <ipython console> in <module> ) <type exceptions.ValueError > cannot slice a scalar Instead, the message should be something like <type exceptions.IndexError > cannot slice a 0 d array Thanks","bug"
|
|
"6013","887","Array constructor for dtype S doesn t accept numpy arrays as input Trac 289).
|
|
Original ticket http projects.scipy.org numpy ticket 289 on 2006 09 22 by FrancescAlted, assigned to unknown.
|
|
Hi, I don t know if this is expected In [104] numpy.array [ abc ], dtype= S ) Out[104] array [abc], dtype= |S3 ) so far so good, but In [105] numpy.array numpy.array [ abc ]), dtype= S ) <type exceptions.ValueError > Traceback most recent call last) home faltet carabos repositori ceei <ipython console> in <module> ) <type exceptions.ValueError > Empty data type I think the array constructor would allow a numpy array as input as well.
|
|
Thanks","bug"
|
|
"6013","888","Some inconsistency in i typecode Trac 290).
|
|
Original ticket http projects.scipy.org numpy ticket 290 on 2006 09 22 by FrancescAlted, assigned to unknown.
|
|
Hi, I find the next to be rather strange In [118] numpy.dtype numpy.int32) == numpy.int32 Out[118] True In [119] numpy.dtype i ) == numpy.int32 Out[119] True In [120] numpy.array [1], dtype=numpy.int32).dtype.type == numpy.int32 Out[120] True In [121] numpy.array [1], dtype= i ).dtype.type == numpy.int32 Out[121] False i.e.
|
|
using a i typecode doesn t generate a true numpy.int32, despite the fact that the dtype constructor seems to create a valid one.
|
|
The version of numpy here is 1.0b5.dev3084.
|
|
Cheers,","bug"
|
|
"6013","889","repmat doesn t preserve matrix subclass Trac 291).
|
|
Original ticket http projects.scipy.org numpy ticket 291 on 2006 09 23 by baxissimo, assigned to unknown.
|
|
Subject says it all.
|
|
>>> repmat matrix [[1,2,3]]), 2,2) array [[...]]) I think the asarray in the first line just needs to be changed to asanyarray .
|
|
Version is 1.0rc1.","bug"
|
|
"6013","890","Make repmat work for arbitrary dimensions Trac 292).
|
|
Original ticket http projects.scipy.org numpy ticket 292 on 2006 09 23 by baxissimo, assigned to unknown.
|
|
repmat ) should work with inputs of any shape and take a tuple for the output number of repeats.
|
|
If this is considered too big a change to replace repmat s 3 argument signature with an incompatible 2 arg one, then a new function reparray should be added.
|
|
An example implementation is below def reparray A, tup) Repeat an array the number of times given in the integer tuple, tup.
|
|
Similar to repmat, but works for arrays of any dimension.
|
|
reparray A, m,n)) is equivalent to repmat A,m,n) If tup has length d, the result will have dimension of max d, A.ndim).
|
|
If tup is scalar it is treated as a 1 tuple.
|
|
If A.ndim < d, A is promoted to be d dimensional by prepending new axes.
|
|
So a shape 3,) array is promoted to 1,3) for 2 D replication, or shape 1,1,3) for 3 D replication.
|
|
If this is not the desired behavior, promote A to d dimensions manually before calling this function.
|
|
If d < A.ndim, tup is promoted to A.ndim by appending 1 s to it.
|
|
Thus for an A.shape of 2,3,4,5), a tup of 2,2) is treated as 2,2,1,1) Examples >>> a = array [0,1,2]) >>> reparray a,2) array [0, 1, 2, 0, 1, 2]) >>> reparray a, 1,2)) array [[0, 1, 2, 0, 1, 2]]) >>> reparray a, 2,2)) array [[0, 1, 2, 0, 1, 2], [0, 1, 2, 0, 1, 2]]) >>> reparray a, 2,1,2)) array [[[0, 1, 2, 0, 1, 2]], [[0, 1, 2, 0, 1, 2]]]) See Also repmat, repeat if numpy.isscalar tup) tup = tup,) d = len tup) c = numpy.array A,copy=False,subok=True,ndmin=d) shape = list c.shape) n = c.size for i, nrep in enumerate tup) if nrep!=1 c = c.reshape 1,n).repeat nrep,0) dim in = shape[i] dim out = dim in nrep shape[i] = dim out n = dim in return c.reshape shape) A more complete version with two implementations and tests and timing code is attached.","enhancement"
|
|
"6013","891","Add a roll function or shift or rotate) Trac 293).
|
|
Original ticket http projects.scipy.org numpy ticket 293 on 2006 09 25 by baxissimo, assigned to unknown.
|
|
Roll shift rotate is a function commonly available in other array languages.
|
|
Numpy should have it too.
|
|
Especially now that there s a rollaxis .
|
|
Here s an implementation that i ve spent a little time tweaking and timing various variations on.
|
|
def roll y,shift,axis) Roll the elements in the array by shift positions along the given axis.
|
|
from numpy import asanyarray,concatenate,arange y = asanyarray y) n = y.shape[axis] shift = n does the right thing for negative shifts, too return y.take concatenate arange shift,n),arange shift))), axis) This would be a good candidate for shape base.py, probably.
|
|
Having this as a function also opens the door for a possibly faster version implemented directly in C at some point in the future.","enhancement"
|
|
"6013","892","Add canned report for tickets submitted by logged in user Trac 294).
|
|
Original ticket http projects.scipy.org numpy ticket 294 on 2006 09 25 by baxissimo, assigned to unknown.
|
|
Right now there s a view my tickets entry in the canned reports.
|
|
I expected this to show me all the ones I ve submitted, but apparenly it is my tickets from the maintainer s point of view, not the user s. It would be nice if there were also a canned report for viewing all the tickets submitted by the logged in user.
|
|
And probably view my tickets should be replaced by something less ambiguous, like view tickets I m assigned or view tickets I submitted .","enhancement"
|
|
"6013","893","Double sort in numpy.median Trac 295).
|
|
Original ticket http projects.scipy.org numpy ticket 295 on 2006 09 26 by FrancescAlted, assigned to unknown.
|
|
numpy.median is computing a sort twice for arrays with an even number of elements.
|
|
The next patch is a cure for this .
|
|
numpy lib function base.py 2006 09 26 10 05 27.000000000 0200 usr local lib python2.5 site packages numpy lib function base.py 2006 09 26 18 13 14.000000000 0200 1046,11 1046,10 median m) returns a median of m along the first dimension of m. sorted = msort m) index = sorted.shape[0] 2 if sorted.shape[0] 2 == 1 return sorted[int sorted.shape[0] 2)] return sorted[index] else sorted = msort m) index = sorted.shape[0] 2 return sorted[index 1] sorted[index]) 2.0 Also, and although median shouldn t have to be well defined in other values different from numbers, it can be forced to work decently for string, object and boolean arrays.
|
|
Behind is a patch for this.
|
|
.
|
|
numpy lib function base.py 2006 09 26 10 05 27.000000000 0200 usr local lib python2.5 site packages numpy lib function base.py 2006 09 26 18 13 27.000000000 0200 1046,11 1046,10 median m) returns a median of m along the first dimension of m. sorted = msort m) if sorted.shape[0] 2 == 1 return sorted[int sorted.shape[0] 2)] index = sorted.shape[0] 2 if sorted.shape[0] 2 == 1 or sorted.dtype.kind in [ S , O , b ] return sorted[index] else sorted = msort m) index = sorted.shape[0] 2 return sorted[index 1] sorted[index]) 2.0 Feel free to apply whatever patch you consider more appropriate.","bug"
|
|
"6013","894","All scalar constructors should return default values Trac 296).
|
|
Original ticket http projects.scipy.org numpy ticket 296 on 2006 09 27 by FrancescAlted, assigned to unknown.
|
|
Now, some scalar constructors without parameters do return defaults In [26] numpy.int32 ) Out[26] 0 but others don t In [28] numpy.int16 ) <type exceptions.TypeError > Traceback most recent call last) home faltet python.nobackup numpy <ipython console> in <module> ) <type exceptions.TypeError > function takes exactly 1 argument 0 given) They all should return default values.","bug"
|
|
"6013","895","Nested sequences with different lengths issues Trac 297).
|
|
Original ticket http projects.scipy.org numpy ticket 297 on 2006 09 27 by FrancescAlted, assigned to unknown.
|
|
!NumPy is not detecting string nested sequence with different lengths >>> numpy.array [ aaa , [ bbb , ccc ]]) array [aaa, [ b], dtype= |S3 ) numarray does >>> numarray.strings.array [ aaa , [ bbb , ccc ]]) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File usr lib python2.4 site packages numarray strings.py , line 1117, in array padc=padc, kind=kind) File usr lib python2.4 site packages numarray strings.py , line 977, in fromlist shape, itemsize = slistShape slist, itemsize=itemsize, shape=shape) File usr lib python2.4 site packages numarray strings.py , line 912, in slistShape shape items = slistShape0 slist) File usr lib python2.4 site packages numarray strings.py , line 887, in slistShape0 raise ValueError Nested sequences with different lengths. )
|
|
ValueError Nested sequences with different lengths.
|
|
and I think this last message is quite adequate for the issue.
|
|
Furthermore, I d say that a message like numarray is better for numerical nested sequences with different lengths.
|
|
NumPy >>> numpy.array [1, [2, 3]]) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError an integer is required numarray >>> numarray.array [1, [2, 3]]) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File usr lib python2.4 site packages numarray numarraycore.py , line 417, in array return fromlist sequence,type,shape) File usr lib python2.4 site packages numarray numarraycore.py , line 267, in fromlist arr.fromlist seq) ValueError Nested sequences with different lengths.","bug"
|
|
"6013","896","lexsort segfault sorting strings Trac 298).
|
|
Original ticket http projects.scipy.org numpy ticket 298 on 2006 09 28 by trac user ericjones, assigned to unknown.
|
|
I ve been using the new record arrays and lexsort from numpy quite a lot lately.
|
|
Very cool stuff.
|
|
Using the nightly egg for numpy from here I believe it is up to date...) http code.enthought.com enstaller eggs numpy nightly py2.4 win32.egg I get segfaults when using lexsort on character arrays.
|
|
A lot of my columns in record arrays are string based, so sorting the arrays based on these columns would be really handy.
|
|
Here is an example that crashes for me.
|
|
C \wrk\mt\trunk\src\lib\mt\statement\tests>python Python 2.4.3 Enthought Edition 1.0.0 553, Aug 2 2006, 12 09 59) [MSC v.1310 32 bit Intel)] on win32 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import lexsort >>> lst = [1,2,3] >>> lexsort lst,)) array [0, 1, 2]) >>> lst = [ abc , cde , fgh ] >>> lexsort lst,)) <seg fault> thanks, eric","bug"
|
|
"6013","897","Downcast unsafe when using array protocol?
|
|
Trac 299).
|
|
Original ticket http projects.scipy.org numpy ticket 299 on 2006 09 28 by FrancescAlted, assigned to unknown.
|
|
In numpy it seems that downcasting is possible >>> i1 = numpy.array [0,1], dtype= Int8 ) >>> i1[...] = numpy.array [8,9], dtype= Int32 ) >>> i1 array [8, 9], dtype=int8) even if an overflow does occur >>> i1[...] = numpy.array [800,900], dtype= Int32 ) >>> i1 array [ 32, 124], dtype=int8) However, you cannot do the same with other objects compliant with the array protocol >>> i1 = numpy.array [0,1], dtype= Int8 ) >>> i1[...] = numarray.array [8,9], dtype= Int8 ) >>> i1 array [8, 9], dtype=int8) Good.
|
|
But... >>> i1[...] = numarray.array [8,9], dtype= Int16 ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError array cannot be safely cast to required type >>> i1[...] = numarray.array [8,9], dtype= Int32 ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError array cannot be safely cast to required type The same happens with Numeric objects and for all kind of downcastings.
|
|
It would be nice if numpy could support this feature.","bug"
|
|
"6013","898","ValueError setting an array element with a sequence.
|
|
Trac 300).
|
|
Original ticket http projects.scipy.org numpy ticket 300 on 2006 09 30 by nilswagner01, assigned to unknown.
|
|
numpy.test 1) results in ====================================================================== ERROR check invalid subscript assignment numpy.core.tests.test multiarray.test zero rank) Traceback most recent call last) File usr lib python2.4 site packages numpy core tests test multiarray.py , line 169, in check invalid subscript assignment self.failUnlessRaises TypeError, assign, a, ), ) File usr lib python2.4 unittest.py , line 319, in failUnlessRaises callableObj args, kwargs) File usr lib python2.4 site packages numpy core tests test multiarray.py , line 166, in assign x[i] = v ValueError setting an array element with a sequence.
|
|
Ran 497 tests in 1.038s FAILED errors=1) I am using the latest svn version.
|
|
1.0.dev3235","bug"
|
|
"6013","899","Make power and divide return floats from int inputs like true divide) Trac 301).
|
|
Original ticket http projects.scipy.org numpy ticket 301 on 2006 09 30 by stefanv, assigned to unknown.
|
|
The power function returns an incorrect) value for negative powers of integers In [1] N.power 3, 2) Out[1] 0 Unlike integer division, which is well defined, I am not sure what the behaviour of power should be.
|
|
Tim Hochberg s comment on the topic can be seen on the mailing list thread at http thread.gmane.org gmane.comp.python.numeric.general 8418 focus=8418 In short, it offers the possibility of making numpy.power behave like Python s math.pow, which always returns floats, whereas x y will be equivalent to the Python integer behaviour.","enhancement"
|
|
"6013","900","Fancy indexing and python 2.5 issue Trac 302).
|
|
Original ticket http projects.scipy.org numpy ticket 302 on 2006 10 02 by FrancescAlted, assigned to unknown.
|
|
Hi, I think I ve detected a bug with fancy indexing and python 2.5.
|
|
The next exposes the problem With python 2.4 >>> numpy.array [1,2])[numpy.array [0])] array [1]) >>> numpy.array [1,2])[numpy.array [0])] array [1]) >>> type numpy.array [1,2])[numpy.array [0])]) <type numpy.ndarray > >>> numpy.array [1,2])[numpy.int32 0)] 1 >>> type numpy.array [1,2])[numpy.int32 0)]) <type numpy.int32 > But, with python 2.5 >>> numpy.array [1,2])[numpy.array [0])] 1 >>> type numpy.array [1,2])[numpy.array [0])]) <type numpy.int32 > >>> numpy.array [1,2])[numpy.int32 0)] 1 >>> type numpy.array [1,2])[numpy.int32 0)]) <type numpy.int32 > IMO, indexing with and array should always give an array as output, not a scalar, so I d say that output for python2.5 is wrong.
|
|
I m using numpy 1.0.dev3239.","bug"
|
|
"6013","901","numpy.distutils fails to find installed libraries on Darwin MacOSX Trac 303).
|
|
Original ticket http projects.scipy.org numpy ticket 303 on 2006 10 02 by trac user RJVB, assigned to unknown.
|
|
numpy.distutils fails to find installed libraries on Darwin MacOSX because Darwin has its own extension for shared libraries, dylib.
|
|
This can be remedied by modifying system info.py such that the library extensions method reads def library extensions self) static exts = [ .a ] if sys.platform == win32 static exts.append .lib ) .lib is used by MSVC if self.search static first exts = static exts [so ext] else exts = [so ext] static exts if sys.platform == cygwin exts.append .dll.a ) RJVB Darwin s shared lib extension is .dylib if sys.platform == darwin exts.append .dylib ) return exts","enhancement"
|
|
"6013","902","numpy.distutils fails to find installed libraries on Darwin MacOSX Trac 304).
|
|
Original ticket http projects.scipy.org numpy ticket 304 on 2006 10 02 by trac user RJVB, assigned to unknown.
|
|
numpy.distutils fails to find installed libraries on Darwin MacOSX because Darwin has its own extension for shared libraries, dylib.
|
|
This can be remedied by modifying system info.py such that the library extensions method reads def library extensions self) static exts = [ .a ] if sys.platform == win32 static exts.append .lib ) .lib is used by MSVC if self.search static first exts = static exts [so ext] else exts = [so ext] static exts if sys.platform == cygwin exts.append .dll.a ) RJVB Darwin s shared lib extension is .dylib if sys.platform == darwin exts.append .dylib ) return exts","enhancement"
|
|
"6013","903","Warning messages compiling on 64 bit system, python 2.5 Trac 305).
|
|
Original ticket http projects.scipy.org numpy ticket 305 on 2006 10 02 by trac user gpk, assigned to dmcooke.
|
|
Compiler warnings with gcc 3.3.5 on Suse opteron x86 64 2.6.11 kernel.
|
|
Numpy 1.0rc1.
|
|
The following warnings were returned following python setup.py build, using python 2.5 .
|
|
gcc configtest.c configtest.c In function `main configtest.c 50 warning int format, different type arg arg 3) configtest.c 57 warning int format, different type arg arg 3) configtest.c 72 warning int format, different type arg arg 3) gcc configtest.c gcc pthread configtest.o o configtest configtest.o .text 0xd) In function `main users gpk DOWNLOADS numpy 1.0rc1 configtest.c 5 undefined reference to `exp gcc numpy core src multiarraymodule.c In file included from numpy core src multiarraymodule.c 66 numpy core src arrayobject.c In function `PyArray Item INCREF numpy core src arrayobject.c 167 warning passing arg 2 of `PyDict Next from incompatible pointer type numpy core src arrayobject.c In function `PyArray Item XDECREF numpy core src arrayobject.c 194 warning passing arg 2 of `PyDict Next from incompatible pointer type In file included from numpy core src arrayobject.c 508, from numpy core src multiarraymodule.c 66 numpy core src scalartypes.inc.src At top level numpy core src scalartypes.inc.src 1651 warning initialization from incompatible pointer type numpy core src scalartypes.inc.src 1652 warning initialization from incompatible pointer type numpy core src scalartypes.inc.src 1653 warning initialization from incompatible pointer type numpy core src scalartypes.inc.src 1654 warning initialization from incompatible pointer type numpy core src scalartypes.inc.src In function `object arrtype getsegcount numpy core src scalartypes.inc.src 2124 warning passing arg 2 of pointer to function from incompatible pointer type numpy core src scalartypes.inc.src In function `initialize numeric types numpy core src scalartypes.inc.src 2345 warning assignment from incompatible pointer type gcc numpy core src multiarraymodule.c In file included from numpy core src arrayobject.c 509, from numpy core src multiarraymodule.c 66 numpy core src arraytypes.inc.src In function `STRING setitem numpy core src arraytypes.inc.src 328 warning passing arg 3 of `PyString AsStringAndSize from incompatible pointer type numpy core src arraytypes.inc.src In function `VOID copyswapn numpy core src arraytypes.inc.src 1236 warning passing arg 2 of `PyDict Next from incompatible pointer type numpy core src arraytypes.inc.src In function `VOID copyswap numpy core src arraytypes.inc.src 1264 warning passing arg 2 of `PyDict Next from incompatible pointer type numpy core src arraytypes.inc.src In function `VOID nonzero numpy core src arraytypes.inc.src 1487 warning passing arg 2 of `PyDict Next from incompatible pointer type In file included from numpy core src multiarraymodule.c 66 numpy core src arrayobject.c In function ` void compare numpy core src arrayobject.c 4526 warning passing arg 2 of `PyDict Next from incompatible pointer type gcc build src.linux x86 64 2.5 numpy core src umathmodule.c In file included from numpy core src umathmodule.c.src 2023 numpy core src ufuncobject.c In function `PyUFunc GetPyValues numpy core src ufuncobject.c 795 warning long int format, different type arg arg 4) numpy core src ufuncobject.c 795 warning long int format, different type arg arg 5) numpy core src ufuncobject.c In function `ufunc frompyfunc numpy core src ufuncobject.c 2987 warning passing arg 3 of `PyString AsStringAndSize from incompatible pointer type gcc numpy numarray capi.c numpy numarray capi.c In function `getBufferSize numpy numarray capi.c 131 warning passing arg 2 of pointer to function from incompatible pointer type gcc numpy random mtrand mtrand.c numpy random mtrand mtrand.c In function ` pyx f 6mtrand cont0 array numpy random mtrand mtrand.c 238 warning label ` pyx L4 defined but not used numpy random mtrand mtrand.c 236 warning label ` pyx L3 defined but not used numpy random mtrand mtrand.c In function ` pyx f 6mtrand cont1 array numpy random mtrand mtrand.c 440 warning label ` pyx L12 defined but not used numpy random mtrand mtrand.c 438 warning label ` pyx L11 defined but not used numpy random mtrand mtrand.c 395 warning label ` pyx L9 defined but not used numpy random mtrand mtrand.c 393 warning label ` pyx L8 defined but not used numpy random mtrand mtrand.c 351 warning label ` pyx L6 defined but not used numpy random mtrand mtrand.c 349 warning label ` pyx L5 defined but not used ...and similar warnings in other pyx f 6mtrand functions numpy random mtrand mtrand.c In function ` pyx f 6mtrand discd array numpy random mtrand mtrand.c 1460 warning ` pyx v oa data might be used uninitialized in this function numpy random mtrand mtrand.c In function ` pyx f 6mtrand discnp array numpy random mtrand mtrand.c 1013 warning ` pyx v op data might be used uninitialized in this function ... and other similar warnings in mtrand.c.","bug"
|
|
"6013","904","Location of arrayobject.h Trac 306).
|
|
Original ticket http projects.scipy.org numpy ticket 306 on 2006 10 02 by trac user gpk, assigned to unknown.
|
|
When numpy 1.0rc1 is installed via python setup.py install prefix= HOME local HOME= users gpk) when python 2.5 is installed in HOME local, then arrayobject.h ends up in a rather deep location users gpk local lib python2.5 site packages numpy core include numpy arrayobject.h .
|
|
That d be OK, except that when one tries to install another package via python setup.py gpk img python prefix= HOME local the search path for gcc does not include anything near to the actual location of arrayobjects.h.
|
|
I get this gcc pthread fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes fPIC I users gpk local include I users gpk local include I users gpk include I users gpk local include python2.5 c gpk avg py.cc o build temp.linux x86 64 2.5 gpk avg py.o So, in the C extension, I would have to include <site packages numpy core include numpy arrayobject.h> which seems a bit cumbersome.
|
|
This might be a distutils problem, in that ... site packages is not searched.
|
|
However, it also seems to be a poor choice of a location for arrayobject.h","bug"
|
|
"6013","905","Problems in creating recarrays coming from other arrays Trac 307).
|
|
Original ticket http projects.scipy.org numpy ticket 307 on 2006 10 02 by FrancescAlted, assigned to unknown.
|
|
It seems that numpy cannot create recarrays from other numpy arrays >>> a=numpy.array [ 1,3), 2,3)], dtype= i4,i4 ) >>> a array [ 1, 3), 2, 3)], dtype= [ f0 , <i4 ), f1 , <i4 )] ) >>> b=numpy.array [numpy.array [1,3]),numpy.array [2,3])], dtype= i4,i4 ) >>> b TypeError expected a readable buffer object In the same way, it should be nice to be able to put Numeric or numarray objects and not only numpy.","bug"
|
|
"6013","906","Array protocol issues in creating recarrays Trac 308).
|
|
Original ticket http projects.scipy.org numpy ticket 308 on 2006 10 02 by FrancescAlted, assigned to unknown.
|
|
Hi, numpy is having problems creating recarrays from numarray columns >>> numpy.rec.array [numpy.array [1,3]),numpy.array [2,3])], dtype= i4,i4 ) recarray [ 1, 2), 3, 3)], dtype= [ f0 , <i4 ), f1 , <i4 )] ) Fine, but... >>> numpy.rec.array [numarray.array [1,3]),numarray.array [2,3])], dtype= i4,i4 ) recarray [[ 1209141984, 1209141984), 105, 4)], bad data!
|
|
[ 1209141984, 1209141984), 105, 4)]], bad data!
|
|
dtype= [ f0 , <i4 ), f1 , <i4 )] ) i.e.
|
|
it seems like if array protocol is not being applied correctly when the sequence is made of objects that are not numpy.","bug"
|
|
"6013","907","AttributeError list object has no attribute shape Trac 309).
|
|
Original ticket http projects.scipy.org numpy ticket 309 on 2006 10 03 by nilswagner01, assigned to unknown.
|
|
====================================================================== ERROR Ticket 160 Traceback most recent call last) File usr lib python2.4 site packages numpy core tests test regression.py , line 267, in check rec iterate [2,2.2, 2.0 ]],dtype=descr) File usr lib python2.4 site packages numpy core records.py , line 502, in array return fromarrays obj, dtype=dtype, shape=shape, kwds) File usr lib python2.4 site packages numpy core records.py , line 268, in fromarrays shape = arrayList[0].shape AttributeError list object has no attribute shape Ran 497 tests in 0.665s FAILED errors=1) >>> numpy.
|
|
version 1.0.dev3246","bug"
|
|
"6013","908","Typo in records.py Trac 310).
|
|
Original ticket http projects.scipy.org numpy ticket 310 on 2006 10 03 by FrancescAlted, assigned to unknown.
|
|
Hi, You should have introduced a typo recently in records.py.
|
|
The next is a patch for this usr local lib python2.5 site packages numpy core records.py 2006 10 03 17 27 22.000000000 0200 usr local lib python2.5 site packages numpy core records.py 2006 10 03 17 45 45.000000000 0200 291,7 291,7 if dtype is not None descr = sb.dtype dtype) names = dtype.names names = descr.names else parsed = format parser formats, names, titles, aligned, byteorder) names = parsed.
|
|
names","bug"
|
|
"6013","909","f77exec= and f90exec= are not honoured in f2py compile Trac 311).
|
|
Original ticket http projects.scipy.org numpy ticket 311 on 2006 10 04 by trac user Dianora, assigned to unknown.
|
|
patch included compiling Fortran sources Traceback most recent call last) File usr local bin f2py , line 6, in ?
|
|
f2py.main ) File usr local lib python2.4 site packages numpy f2py f2py2e.py , line 546, in main run compile ) File usr local lib python2.4 site packages numpy f2py f2py2e.py , line 533, in run compile setup ext modules = [ext]) File usr local lib python2.4 site packages numpy distutils core.py , line 174, in setup return old setup new attr) File usr local lib python2.4 distutils core.py , line 149, in setup dist.run commands ) File usr local lib python2.4 distutils dist.py , line 946, in run commands self.run command cmd) File usr local lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr local lib python2.4 distutils command build.py , line 112, in run self.run command cmd name) File usr local lib python2.4 distutils cmd.py , line 333, in run command self.distribution.run command command) File usr local lib python2.4 distutils dist.py , line 966, in run command cmd obj.run ) File usr local lib python2.4 site packages numpy distutils command build ext.py , line 113, in run self.build extensions ) File usr local lib python2.4 distutils command build ext.py , line 405, in build extensions self.build extension ext) File usr local lib python2.4 site packages numpy distutils command build ext.py , line 252, in build extension f objects = self.fcompiler.compile f sources, AttributeError NoneType object has no attribute compile numpy distutils fcompiler init .py.orig Tue Oct 3 20 39 58 2006 numpy distutils fcompiler init .py Tue Oct 3 22 43 18 2006 268,11 268,7 f90 = self.
|
|
get cmd compiler f90 , F90 , conf, f90exec )) Temporarily setting f77,f90 compilers so that version cmd can use their executables.
|
|
if f77 self.set executables compiler f77=[f77]) if f90 self.set executables compiler f90=[f90]) If you do that, f77exec and f90exec break db Must set version cmd before others as self.get flags methods may call self.get version.
|
|
vers cmd = self.
|
|
get cmd self.get version cmd) This fixes f77exec and f90exec for me, but possibly breaks something else.
|
|
I d like someone with more python clue than I have to verify this is correct.","bug"
|
|
"6013","910","numpy.rec.array doesn t create a copy of buffer passed Trac 312).
|
|
Original ticket http projects.scipy.org numpy ticket 312 on 2006 10 04 by FrancescAlted, assigned to unknown.
|
|
I don t know if the next behaviour is expected In [38] ra=numpy.array [ 1,2.3)], dtype= i2,f8 ) In [39] ra2=numpy.rec.array ra, dtype= i2,f8 ) In [40] ra[ f0 ] Out[40] array [1], dtype=int16) In [41] ra2[ f0 ] = 2 In [42] ra[ f0 ] Out[42] array [2], dtype=int16) i.e.
|
|
numpy.rec.array doesn t make a copy of the original array.
|
|
I find this a bit counter intuitive, because it seems to me that all the array ) factories should deliver a copy of the buffer passed.
|
|
In case this is the expected behaviour, I d say that adding a copy parameter in numpy.rec.array would be convenient.","bug"
|
|
"6013","911","Creating a recarray from columns with different shapes fails Trac 313).
|
|
Original ticket http projects.scipy.org numpy ticket 313 on 2006 10 04 by FrancescAlted, assigned to unknown.
|
|
The next works well In [83] numpy.rec.fromarrays [ 1,2), red , green )], dtype=[ x , i4 ), color , a5 )]) Out[83] recarray [ 1, red ), 2, green )], dtype=[ x , <i4 ), color , |S5 )]) but, when trying to put together columns with different shapes, an error occurs In [84] numpy.rec.fromarrays [ 1,2), 2,3)), red , green )], dtype=[ x , 2,)i4 ), color , a5 )]) exceptions.ValueError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> usr lib python2.4 site packages numpy core records.py in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 288 for obj in arrayList 289 if obj.shape != shape > 290 raise ValueError, array has different shape 291 292 if dtype is not None ValueError array has different shape Note how the dtype in second case is adequate for columns with different shape.","bug"
|
|
"6013","912","array factory aborts when passed a list with a void type inside Trac 314).
|
|
Original ticket http projects.scipy.org numpy ticket 314 on 2006 10 04 by FrancescAlted, assigned to unknown.
|
|
The next exposes the problem In [1] import numpy In [2] rows=[ 0,)] In [3] descr=[ x , i4 )] In [4] numpy.array [rows[0]], dtype=descr) Out[4] array [ 0,)], dtype=[ x , <i4 )]) In [5] scra = numpy.array rows, dtype=descr)[0] In [6] numpy.array scra, dtype=descr) Out[6] array 0,), dtype=[ x , <i4 )]) So far so good, but....
|
|
In [7] numpy.array [scra], dtype=descr) glibc detected double free or corruption fasttop) 0x081e46d0 Avortat Using numpy 1.0.dev3249","bug"
|
|
"6013","913","Inconsistencies with sequence parameters Trac 315).
|
|
Original ticket http projects.scipy.org numpy ticket 315 on 2006 10 04 by FrancescAlted, assigned to unknown.
|
|
Hi, Is the next behavior intended?
|
|
In [18] numpy.array [ aaa , x )], dtype= S3,S1 ) Out[18] array [ aaa , x )], Good dtype=[ f0 , |S3 ), f1 , |S1 )]) In [19] numpy.array [ aaa , x ], dtype= S3,S1 ) Out[19] array [ aaa , \xb7 ), x[, , \x08 )], uh?
|
|
dtype=[ f0 , |S3 ), f1 , |S1 )]) In [20] numpy.array [ aaa , x ,24], dtype= S3,S1,i4 ) exceptions.TypeError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> TypeError expected a readable buffer object i.e.
|
|
numpy doesn t complain when dtype only contains string types but, still, gives bad results).
|
|
Furthermore, the above seems to work with rec.array In [22] numpy.rec.array [ aaa , x ], dtype= S3,S1 ) Out[22] recarray aaa , x ), dtype=[ f0 , |S3 ), f1 , |S1 )]) In [23] numpy.rec.array [ aaa , x ,24], dtype= S3,S1,i4 ) Out[23] recarray aaa , x , 24), dtype=[ f0 , |S3 ), f1 , |S1 ), f2 , <i4 )]) Should I understand that rec.array is preferred for getting recarrays from lists?
|
|
Thanks","bug"
|
|
"6013","914","blackmanharris, nuttall and flattop windows Trac 316).
|
|
Original ticket http projects.scipy.org numpy ticket 316 on 2006 10 05 by trac user gregwillden, assigned to unknown.
|
|
I added the Blackman Harris, Nuttall and Flat Top window functions to function base.py.
|
|
[[BR]] I also added See also sections to the end of all window docstrings.","enhancement"
|
|
"6013","915","Typo in records.py Trac 317).
|
|
Original ticket http projects.scipy.org numpy ticket 317 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
numpy core records.py 2006 10 05 10 05 01.000000000 0200 usr lib python2.4 site packages numpy core records.py 2006 10 05 12 16 04.000000000 0200 514,7 514,7 elif isinstance obj, recarray) copied = 0 if dtype is not None and new.dtype != dtype) if dtype is not None and obj.dtype != dtype) new = obj.astype dtype) copied = 1 else Cheers,","bug"
|
|
"6013","916","Easy way to increment the nesting level in recarrays Trac 318).
|
|
Original ticket http projects.scipy.org numpy ticket 318 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
Hi, I d like to see an easy way to increment the level of nesting in recarrays.
|
|
For me, a good way would be that the next could work In [33] nr=numpy.rec.array a 8, dtype=[ x , i4 ), y , i4 )]) In [34] nr Out[34] recarray [ 1633771873, 1633771873)], dtype=[ x , <i4 ), y , <i4 )]) In [35] numpy.rec.array nr, dtype=[ x , i4 ), y , i4 )]) Out[35] recarray [ 1633771873, 1633771873)], dtype=[ x , <i4 ), y , <i4 )]) Fine until here, but In [36] numpy.rec.array nr, dtype=[ new ,[ x , i4 ), y , i4 )])]) exceptions.ValueError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> usr lib python2.4 site packages numpy core records.py in array obj, dtype, shape, offset, strides, formats, names, titles, aligned, byteorder, copy) 516 copied = 0 517 if dtype is not None and obj.dtype != dtype) > 518 new = obj.astype dtype) 519 copied = 1 520 else ValueError size of tuple must match number of fields.
|
|
i.e.
|
|
in my opinion, nr is a buffer that is compatible with both [ x , i4 ), y , i4 )] and [ new ,[ x , i4 ), y , i4 )])] datatypes, and there wouldn t be any harm in allowing this.
|
|
I understand that allowing this in numpy.array would a bit controversial, but at least allowing it in numpy.rec.array would not create any secondary effect and this feature can be handy in some situations.","enhancement"
|
|
"6013","917","Setting a column in a recarray from another recarray corrupts data Trac 319).
|
|
Original ticket http projects.scipy.org numpy ticket 319 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
The next code snipped shows the problem sorry but I haven t able to simplify it more than it is now) import numpy rows = [ [3, 2], 6j, 6.0, nn , [6j, 4j], [6.0, 4.0], [1L, 2L]), NN , 8), cc , NN , 6j), [[6.0, 4.0], [6.0, 4.0]], 8), [4, 3], 7j, 7.0, oo , [7j, 5j], [7.0, 5.0], [2L, 1L]), OO , 9), dd , OO , 7j), [[7.0, 5.0], [7.0, 5.0]], 9)] descr1 = [ x , <i4 , 2,)), Info , [ value , <c16 ), y2 , <f8 ), Info2 , [ name , |S2 ), value , <c16 , 2,)), y3 , <f8 , 2,)), z3 , <u4 , 2,))]), name , |S2 ), z2 , |u1 )]), color , |S2 ), info , [ Name , |S2 ), Value , <c16 )]), y , <f8 , 2, 2)), z , |u1 )] descr2 = [ value , 1c16 ), y2 , 1f8 ), Info2 , [ name , 1a2 ), value , 2,)c16 ), y3 , 2,)f8 ), z3 , 2,)u4 )]), name , 1a2 ), z2 , 1u1 )] ra1 = numpy.array rows, dtype=descr1) ra4 = ra1.copy ) print ra1 orig) > , ra1 ra2 = ra1[ Info ] ra3 = numpy.rec.array ra2, dtype=descr2) ra1[ Info ] = ra2 this works ra1[ Info ] = ra3 this don t print ra1 final) > , ra1 assert numpy.all ra1 == ra4) the output for my machine is ra1 orig) > [ [3, 2], 6j, 6.0, nn , [6j, 4j], [6.0, 4.0], [1L, 2L]), NN , 8), cc , NN , 6j), [[6.0, 4.0], [6.0, 4.0]], 8) [4, 3], 7j, 7.0, oo , [7j, 5j], [7.0, 5.0], [2L, 1L]), OO , 9), dd , OO , 7j), [[7.0, 5.0], [7.0, 5.0]], 9)] ra1 final) > [ [3, 2], 6j, 6.0, nn , [6j, 4j], [6.0, 4.0], [1L, 2L]), NN , 8), , , 6j), [[6.0, 4.0], [6.0, 4.0]], 8) [4, 3], 7j, 7.0, oo , [7j, 5j], [7.0, 5.0], [2L, 1L]), OO , 9), dd , OO , 7j), [[7.0, 5.0], [7.0, 5.0]], 9)] Traceback most recent call last) File prova4.py , line 33, in ?
|
|
assert numpy.all ra1 == ra4) AssertionError it seems that there is a bizarre effect in creating the recarray using rec.array.","bug"
|
|
"6013","918","Another typo in records.py Trac 320).
|
|
Original ticket http projects.scipy.org numpy ticket 320 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
numpy core records.py 2006 10 05 16 11 39.000000000 0200 usr lib python2.4 site packages numpy core records.py 2006 10 05 16 45 08.000000000 0200 306,7 306,7 for k, obj in enumerate arrayList) nn = len descr[k].shape) if obj.shape[nn ] != shape raise ValueError, array shape mismatch in array , k raise ValueError, array shape mismatch in array s k array = recarray shape, descr)","bug"
|
|
"6013","919","Semantic difference in formats and dtype parameters in rec.fromarrays Trac 321).
|
|
Original ticket http projects.scipy.org numpy ticket 321 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
Maybe this happens in other contexts as well, but I noticed it first in rec.fromarrays In [26] numpy.rec.fromarrays [[[0,1],[1,2]]], formats= 2,)i4 ) Out[26] recarray [ array [0, 1]),), array [1, 2]),)], dtype=[ f1 , <i4 , 2,))]) Fine, but...
|
|
In [27] numpy.rec.fromarrays [[[0,1],[1,2]]], dtype= 2,)i4 ) exceptions.ValueError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> usr lib python2.4 site packages numpy core records.py in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 296 print descr > , descr, type descr) 297 > 298 Determine shape from data type.
|
|
299 if len descr) != len arrayList) 300 raise ValueError, mismatch between the number of fields \ ValueError mismatch between the number of fields and the number of arrays","bug"
|
|
"6013","920","Shape mismatch error in rec.fromarray ) constructor Trac 322).
|
|
Original ticket http projects.scipy.org numpy ticket 322 on 2006 10 05 by FrancescAlted, assigned to unknown.
|
|
A strange error follows.
|
|
The next works just fine In [102] x1=numpy.array [[1,2],[3,4]]) In [103] x2=numpy.array [ a , dd ]) In [104] x3=numpy.array [1.1,2]) In [105] numpy.rec.fromarrays [x1,x2,x3], formats= 2,)i4,a3,f8 ) Out[105] recarray [ array [1, 2]), a , 1.1000000000000001), array [3, 4]), dd , 2.0)], dtype=[ f1 , <i4 , 2,)), f2 , |S3 ), f3 , <f8 )]) But In [109] x1=numpy.array [[1,2],[3,4],[5,6]]) In [110] x2=numpy.array [ a , dd , xyz ]) In [111] x3=numpy.array [1.1,2,3]) In [112] numpy.rec.fromarrays [x1,x2,x3], formats= 2,)i4,a3,f8 ) exceptions.ValueError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> usr lib python2.4 site packages numpy core records.py in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 307 308 for k, obj in enumerate arrayList) > 309 nn = len descr[k].shape) 310 if obj.shape[nn ] != shape 311 raise ValueError, array shape mismatch in array s k ValueError array shape mismatch in array 1","bug"
|
|
"6013","921","roots polynomial) drops zero roots if only one non zero coeff in polynomial Trac 323).
|
|
Original ticket http projects.scipy.org numpy ticket 323 on 2006 10 05 by trac user RyanGutenkunst, assigned to unknown.
|
|
roots polynomial) drops zero roots if only one non zero coeff in polynomial Example roots of x {2} == 0) >>> numpy.roots 1,0,0)) array [], dtype=float64) Also, roots fails if all coeffs are 0 >>> numpy.roots 0,0)) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File afs msc.cornell.edu home jsethna ryang installed lib python2.4 site packages numpy lib polynomial.py , line 96, in roots trailing zeros = len p) non zero[ 1] 1 IndexError index out of bounds The attached patch fixes both issues.
|
|
For the all zero polynomial, an empty array is returned.
|
|
This is what matlab does.)","bug"
|
|
"6013","922","ndmin ignored in array constructor Trac 324).
|
|
Original ticket http projects.scipy.org numpy ticket 324 on 2006 10 06 by stefanv, assigned to unknown.
|
|
x = N.array [1],dtype=N.float64) N.array x,dtype=N.float64,ndmin=2) should give a 2 dimensional array, but doesn t.","bug"
|
|
"6013","923","memory error using vectorize Trac 325).
|
|
Original ticket http projects.scipy.org numpy ticket 325 on 2006 10 06 by stefanv, assigned to unknown.
|
|
The following code causes python and valgrind!)
|
|
to segfault from numpy import vectorize, zeros vt = vectorize lambda args args) Removing either of the following lines cures the segfault vt zeros 1,2,1)), zeros 2,1,1)), zeros 1,1,2))) vt zeros 1,2,1)), zeros 2,1,1)), zeros 1,1,2)), zeros 2,2))) with gdb pronouncing Program received signal SIGSEGV, Segmentation fault.
|
|
[Switching to Thread 1210120528 LWP 11564)] DOUBLE setitem op=0x0, ov=0x824b820 xefxbfxbd1xefxbfxbdxefxbfxbdxefxbfxbd1xefxbfxbdxefxbfxbd\002 , ap=0x8226d28) at numpy core src arraytypes.inc.src 129 129 if PyArray IsScalar op, kind )) {","bug"
|
|
"6013","924","recarray transpose ) Trac 326).
|
|
Original ticket http projects.scipy.org numpy ticket 326 on 2006 10 06 by trac user parejkoj, assigned to unknown.
|
|
The following code should display a different array from the original transpose is not working on recarrays containing records).
|
|
type result[0]) is numpy.core.records.record.
|
|
import numpy.core.records as rec rowData = [ 0,123.4), 1,234.5), 2,345.6)] formats = i4,i8 names = i,j result=rec.fromrecords array rowData),names=names,formats=formats,shape=len rowData)) print result.transpose )","bug"
|
|
"6013","925","Memory error in ufuncobject Trac 327).
|
|
Original ticket http projects.scipy.org numpy ticket 327 on 2006 10 07 by stefanv, assigned to unknown.
|
|
Upon running from numpy import arange, add data = arange 5) data = add.outer data,data) Valgrind reports the following error ==4854== Conditional jump or move depends on uninitialised value s) ==4854== at 0x45541B1 construct loop ufuncobject.c 1181) ==4854== by 0x4554AEC PyUFunc GenericFunction ufuncobject.c 1475) ==4854== by 0x4555BD3 ufunc generic call ufuncobject.c 2786) ==4854== by 0x4556727 ufunc outer ufuncobject.c 3379) ==4854== by 0x80B8BD0 PyEval EvalFrame in usr bin python2.4) ==4854== by 0x80B9FC8 PyEval EvalCodeEx in usr bin python2.4) ==4854== by 0x80BA036 PyEval EvalCode in usr bin python2.4) ==4854== by 0x80DD569 PyRun FileExFlags in usr bin python2.4) ==4854== by 0x80DD756 PyRun SimpleFileExFlags in usr bin python2.4) ==4854== by 0x8055C81 Py Main in usr bin python2.4) which looks like it is caused by the new axis optimisation code.","bug"
|
|
"6013","926","numpy.insert ) won t insert at the end of an array Trac 328).
|
|
Original ticket http projects.scipy.org numpy ticket 328 on 2006 10 07 by trac user vomjom, assigned to unknown.
|
|
Here s a log >>> import numpy >>> a = numpy.array [1,2,4]) >>> numpy.insert a, 3, 5) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
File usr lib python2.4 site packages numpy lib function base.py , line 1290, in insert raise ValueError, index d) out of range 0<=index<= d) \ ValueError index 3) out of range 0<=index<=3) in dimension 0 3 certainly is within that range )","bug"
|
|
"6013","927","Memory error on array deallocation 64 bit platforms) Trac 329).
|
|
Original ticket http projects.scipy.org numpy ticket 329 on 2006 10 09 by stefanv, assigned to teoliphant.
|
|
While running import numpy as N N.eye 128)) on a 64 bit machine, Valgrind complained ==18199== Invalid read of size 4 ==18199== at 0x43CFC6 PyObject Free in usr bin python2.4) ==18199== by 0x582179D arraymultiter dealloc arrayobject.c 10352) ==18199== by 0x584AA52 broadcast cast arrayobject.c 7510) ==18199== by 0x584D10C PyArray FromArray arrayobject.c 7851) ==18199== by 0x5842C05 PyArray FromAny arrayobject.c 8257) ==18199== by 0x599A8A7 construct loop ufuncobject.c 856) ==18199== by 0x599B5BF PyUFunc GenericFunction ufuncobject.c 1475) ==18199== by 0x599C6BF ufunc generic call ufuncobject.c 2786) ==18199== by 0x41448F PyObject Call in usr bin python2.4) ==18199== by 0x475DE4 PyEval EvalFrame in usr bin python2.4) ==18199== Address 0x6F31020 is 56 bytes inside a block of size 1,344 free d ==18199== at 0x4A1A5B3 free vg replace malloc.c 235) ==18199== by 0x582179D arraymultiter dealloc arrayobject.c 10352) ==18199== by 0x584AA52 broadcast cast arrayobject.c 7510) ==18199== by 0x584D10C PyArray FromArray arrayobject.c 7851) ==18199== by 0x5842C05 PyArray FromAny arrayobject.c 8257) ==18199== by 0x599A8A7 construct loop ufuncobject.c 856) ==18199== by 0x599B5BF PyUFunc GenericFunction ufuncobject.c 1475) ==18199== by 0x599C6BF ufunc generic call ufuncobject.c 2786) ==18199== by 0x41448F PyObject Call in usr bin python2.4) ==18199== by 0x475DE4 PyEval EvalFrame in usr bin python2.4) unhandled opc aux = 0x7 first opcode == 0xDD vex amd64 >IR unhandled instruction bytes 0xDD 0x7C 0x24 0xFC This doesn t happen for .. Numpy version is r3294.","bug"
|
|
"6013","928","float|int)<bits> 0).imag causes segfault Trac 330).
|
|
Original ticket http projects.scipy.org numpy ticket 330 on 2006 10 10 by pearu, assigned to teoliphant.","bug"
|
|
"6013","929","Empty strings not allowed in adaptable string data types Trac 331).
|
|
Original ticket http projects.scipy.org numpy ticket 331 on 2006 10 11 by FrancescAlted, assigned to unknown.
|
|
Why it is not possible to create arrays with adaptable size coming from empty strings?
|
|
In [27] numpy.array , dtype= S ) exceptions.ValueError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> ValueError Empty data type I think that in this case it would be nice to allow this by returning a S1 dtype i.e.
|
|
the minimum length that is able to keep an empty string).
|
|
That would be the equivalent to In [29] numpy.array , dtype= S1 ) Out[29] array , dtype= |S1 )","enhancement"
|
|
"6013","930","Proper .name in nested types Trac 332).
|
|
Original ticket http projects.scipy.org numpy ticket 332 on 2006 10 11 by FrancescAlted, assigned to unknown.
|
|
Hi, I d find rather useful to be able to access to the proper name of a nested type in this context In [106] nesteddtype=numpy.dtype [ info , [ Name , 1a2 ), Value , 1c16 )]), y , 2, 2)f8 )]) In [107] nesteddtype[ info ] Out[107] dtype [ Name , |S2 ), Value , <c16 )]) In [108] nesteddtype[ info ].name Out[108] void144 i.e.
|
|
I d expect the output for 108 to be info instead of void144 .
|
|
Would this be difficult to implement?
|
|
Thanks","enhancement"
|
|
"6013","931","Improve error message when making a structured array from an invalid input Trac 333).
|
|
Original ticket http projects.scipy.org numpy ticket 333 on 2006 10 11 by FrancescAlted, assigned to unknown.
|
|
The next exposes the problem In [142] dt1=numpy.dtype i4 ) In [143] numpy.array 0, dtype=dt1.type) Out[143] array 0) good, but In [144] dt2=numpy.dtype i4 , 2,))) In [145] numpy.array 0, dtype=dt2.type) exceptions.MemoryError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> MemoryError Moreover In [147] numpy.array [0], dtype=dt1.type) Out[147] array [0]) good, but In [148] numpy.array [0], dtype=dt2.type) exceptions.TypeError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> TypeError expected a readable buffer object I d say that the .type is an scalar type that should be independent of the shape of the base dtype.","enhancement"
|
|
"6013","932","Inconsistency in dtypes with scalar shape Trac 334).
|
|
Original ticket http projects.scipy.org numpy ticket 334 on 2006 10 11 by FrancescAlted, assigned to unknown.
|
|
Hi, Is the next expected?
|
|
In [226] numpy.dtype i4 , ))) Out[226] dtype int32 , ))) In [227] numpy.dtype i4 ) Out[227] dtype int32 ) In [230] numpy.dtype i4 , ))).type Out[230] <type numpy.void > In [231] numpy.dtype i4 ).type Out[231] <type numpy.int32 > in other words, why this difference between numpy.dtype i4 , ))) and numpy.dtype i4 )?.
|
|
I d say that it is artificiuos, as per In [228] numpy.dtype i4 ).shape Out[228] ) In [229] numpy.dtype i4 , ))).shape Out[229] ) In [232] numpy.array [0], numpy.dtype i4 )) Out[232] array [0]) In [233] numpy.array [0], numpy.dtype i4 , )))) Out[233] array [0]) In [234] numpy.array [0], numpy.dtype i4 )) == numpy.array [0], numpy.dtype i4 , )))) Out[234] array [True], dtype=bool)","bug"
|
|
"6013","933",")type not allowed in creating dtypes Trac 335).
|
|
Original ticket http projects.scipy.org numpy ticket 335 on 2006 10 12 by FrancescAlted, assigned to unknown.
|
|
Hi, In [12] numpy.dtype [ col1 , 1,)i4 )]) Out[12] dtype [ col1 , <i4 , 1,))]) In [13] numpy.dtype [ col1 , 1i4 )]) Out[13] dtype [ col1 , <i4 )]) In [14] numpy.dtype [ col1 , )i4 )]) )i4 exceptions.TypeError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> TypeError data type not understood The statement 14 used to work until last night, so this might be a side effect of the changeset 3309.","bug"
|
|
"6013","934","Creating new datatypes from existing ones Trac 336).
|
|
Original ticket http projects.scipy.org numpy ticket 336 on 2006 10 12 by FrancescAlted, assigned to unknown.
|
|
Hi, One thing that I believe would be quite useful at least, it is for me), is to be able to create datatypes from pieces of other datatypes.
|
|
An example would clarify this.
|
|
Let s suppose that we have a generic nested dtype In [35] nt=numpy.dtype [ info , [ Name , 1a2 ), Value , 1c16 )]), .... x , 2, 2)f8 ), .... y , 2, 2)f8 )]) Now, I want to create a new datatype but using just cols info and y i.e.
|
|
remove the x col).
|
|
So, it would be nice if the next or something equivalent) would work In [40] numpy.dtype [nt[ info ], nt[ y ]]) exceptions.TypeError Traceback most recent call last) home faltet python.nobackup numpy <ipython console> TypeError data type not understood Note that, for better reproducing the original column names, the dtype columns should keep their original name see ticket 332).
|
|
Thanks,","enhancement"
|
|
"6013","935","Compiler warnings on 64 bit Linux Trac 337).
|
|
Original ticket http projects.scipy.org numpy ticket 337 on 2006 10 12 by alberts, assigned to unknown.
|
|
When compiling NumPy r3317 on 64 bit Linux with GCC 4.0.3, the following warnings are generated gcc numpy core src multiarraymodule.c In file included from numpy core src multiarraymodule.c 63 numpy core src arrayobject.c In function iter ass subscript numpy core src arrayobject.c 9285 warning format d expects type int , but argument 3 has type npy intp numpy core src scalartypes.inc.src In function gentype imag get numpy core src scalartypes.inc.src 834 warning typecode may be used uninitialized in this function gcc build src.linux x86 64 2.4 numpy core src umathmodule.c In file included from numpy core src umathmodule.c.src 2056 numpy core src ufuncobject.c In function PyUFunc GetPyValues numpy core src ufuncobject.c 795 warning format ld expects type long int , but argument 4 has type int numpy core src ufuncobject.c 795 warning format ld expects type long int , but argument 5 has type int","bug"
|
|
"6013","936","Valgrind warning when calling scipy.interpolate.interp1d Trac 338).
|
|
Original ticket http projects.scipy.org numpy ticket 338 on 2006 10 12 by alberts, assigned to unknown.
|
|
When running the following code !python import numpy as np import scipy as sp from scipy.interpolate import interp1d print np.
|
|
version print sp.
|
|
version x10 = np.arange 10.)
|
|
y210 = np.arange 20.
|
|
).reshape 2, 10)) interp210 = interp1d x10, y210) interp210 1.5) Valgrind emits the following warning on 32 bit Linux ==1076== Conditional jump or move depends on uninitialised value s) ==1076== at 0x4433E14 PyArray MapIterReset arrayobject.c 9684) ==1076== by 0x44679F3 array ass sub arrayobject.c 2442) ==1076== by 0x4A1EA6B PyObject SetItem in usr lib libpython2.4.so.1.0) ==1076== by 0x4A7D92C PyEval EvalFrame in usr lib libpython2.4.so.1.0) ==1076== by 0x4A81757 PyEval EvalCodeEx in usr lib libpython2.4.so.1.0) ...
|
|
Similar warning on 64 bit Linux ==31596== Conditional jump or move depends on uninitialised value s) ==31596== at 0x571ACD0 PyArray MapIterReset arrayobject.c 9650) ==31596== by 0x575CED0 array ass sub arrayobject.c 2433) ==31596== by 0x474080 PyEval EvalFrame in usr bin python2.4) ==31596== by 0x4783FE PyEval EvalCodeEx in usr bin python2.4) ...
|
|
Tested with Valgrind 3.1.0 and 3.2.1.
|
|
Used !NumPy 1.0.dev3217 and !SciPy 0.5.2.dev2254 on 64 bit system.
|
|
Used !NumPy 1.0.dev3315 and !SciPy 0.5.2.dev2241 on 32 bit system.","bug"
|
|
"6013","937","nd grid with 1 sample Trac 339).
|
|
Original ticket http projects.scipy.org numpy ticket 339 on 2006 10 12 by trac user guimond, assigned to unknown.
|
|
hi.
|
|
i ve made modifications to index tricks to enable the use of nd grid with a number of samples equal to 1, which currently doesn t work because of a division by num samples 1).
|
|
It is then possible to call mgrid[ 0 0 1j ] which returns array [ 0.]).
|
|
this is against index tricks rev 3224. see attached file","enhancement"
|
|
"6013","938","ndpointer zeros part of 64 bit address Trac 340).
|
|
Original ticket http projects.scipy.org numpy ticket 340 on 2006 10 12 by alberts, assigned to unknown.
|
|
SConstruct env = Environment ) env.Replace CCFLAGS=[ O0 , ggdb , Wall , ansi , pedantic ]) env.SharedLibrary spfuncs ,[ spfuncs.c ]) C code !c include <stdio.h> void nnz double ary) { printf ary = p\n , void )ary); } Python code !python import numpy as N from ctypes import from numpy.ctypeslib import ndpointer libspfuncs = N.ctypeslib.load library libspfuncs , file ) libspfuncs.nnz.restype = None A = N.eye 128)) print A.ctypes.data as c void p) print hex A. array interface [ data ][0]) libspfuncs.nnz.argtypes = [POINTER c double)] libspfuncs.nnz A.ctypes.data as POINTER c double))) libspfuncs.nnz.argtypes = [ndpointer dtype = N.float64)] libspfuncs.nnz A) Output on 64 bit Linux with ctypes 1.0.0 and !NumPy r3317 c void p 46912559644688) 0x2aaaae740010 ary = 0x2aaaae740010 ary = 0xae740010","bug"
|
|
"6013","939","Inconsistent behavior of comparison operators on character arrays, created with c typecode Trac 341).
|
|
Original ticket http projects.scipy.org numpy ticket 341 on 2006 10 13 by trac user gkoczyk, assigned to unknown.
|
|
Arrays created using old Numeric c typecode, seem to have |S1 datatype, but raise errors on comparison with Python strings or sequences.
|
|
>>> numpy.array [ X ], c ) array [ X ], dtype= |S1 ) >>> numpy.array [ X ], c ) == X Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError invalid string data types in comparison >>> numpy.array [ X , X ], c ) < [ X ] Traceback most recent call last) File <stdin> , line 1, in ?
|
|
TypeError invalid string data types in comparison On a sidenote, functor objects implementing comparisons on |S1 arrays raise NotImplemented >>> numpy.equal numpy.array [ X , X ], c ) , numpy.array [ X , X ], c ) ) NotImplemented >>> numpy.equal numpy.array [ X , X ], c ) , numpy.array [ X , X ], |S1 ) ) NotImplemented >>> numpy.equal numpy.array [ X , X ], |S1 ) , numpy.array [ X , X ], |S1 ) ) NotImplemented","bug"
|
|
"6013","940","Inconsistent behavior of string array constructors Trac 342).
|
|
Original ticket http projects.scipy.org numpy ticket 342 on 2006 10 13 by trac user gkoczyk, assigned to unknown.
|
|
If a single string is given as one of the columns of an n x m array, it seems to be silently used to construct the remainder of entries along the same dimension.
|
|
>>> numpy.array [ [ X ], [ X , X , X ] ], |S1 ) array [[ X , X , X ], [ X , X , X ]], dtype= |S1 ) If more strings are supplied, but still not enough to match dimension length, ValueError is raised correctly.
|
|
>>> numpy.array [ [ X , X ], [ X , X , X ] ], |S1 ) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
ValueError setting an array element with a sequence","bug"
|
|
"6013","941","numpy.dtype dtype, int) should raise an error Trac 343).
|
|
Original ticket http projects.scipy.org numpy ticket 343 on 2006 10 13 by FrancescAlted, assigned to unknown.
|
|
Hi, The next is confusing to me In [37] numpy.dtype uint32 , 342) == numpy.dtype uint32 ) Out[37] True I d either forbid passing two parameters to the dtype factory or make the next statements equivalent In [38] numpy.dtype uint32 , 342) == numpy.dtype uint32 ,342)) Out[38] True It currently returns False","bug"
|
|
"6013","942","Normalize shapes in dtype representation Trac 344).
|
|
Original ticket http projects.scipy.org numpy ticket 344 on 2006 10 13 by FrancescAlted, assigned to unknown.
|
|
It would be nice if shapes in repr dtype) could be normalized to use always tuples.
|
|
For example In [63] dt1=numpy.dtype uint32 , 2)) In [64] dt2=numpy.dtype uint32 , 2,))) In [65] dt1 == dt2 Out[65] True In [66] dt1.shape Out[66] 2,) In [67] dt2.shape Out[67] 2,) But....
|
|
In [68] dt1 Out[68] dtype uint32 ,2)) In [69] dt2 Out[69] dtype uint32 , 2,))) for an inexpert user, repr of dt1 and dt2 may lead him to think that they are different dtypes and they are not.","enhancement"
|
|
"6013","943","python 2.5 syntax in test errstate.py breaks installation Trac 345).
|
|
Original ticket http projects.scipy.org numpy ticket 345 on 2006 10 14 by astraw, assigned to unknown.
|
|
The Python2.5 syntax with statements) breaks the byte compilation step of installation.
|
|
This kills installation on Debian as probably other) systems.
|
|
From the commented out code in the file, it s clear that someone knows about this issue, but in the runup to 1.0, I thought I d make sure the latest SVN trunk is working and came across this build gotcha.
|
|
Patch attached.","bug"
|
|
"6013","944","Invalid promotion of numpy scalar to float64 Trac 346).
|
|
Original ticket http projects.scipy.org numpy ticket 346 on 2006 10 14 by charris, assigned to unknown.
|
|
In [95] a = np.float32 1) In [96] a.dtype Out[96] dtype float32 ) In [97] a = 2 In [98] a.dtype Out[98] dtype float64 ) Result type of a should still be float32.
|
|
This also holds for =, =, and =.","bug"
|
|
"6013","945","qr only works for square matrices Trac 347).
|
|
Original ticket http projects.scipy.org numpy ticket 347 on 2006 10 16 by trac user bartvde, assigned to charris.
|
|
The qr method doesn t work for non square matrices.
|
|
In addition, there is no unit test for the qr method.
|
|
The scipy.linalg.qr method works fine.
|
|
In [15] a=numpy.random.rand 6,5) In [16] q,r=numpy.linalg.qr a) exceptions.ValueError Traceback most recent call last) home bartvde Documents ADI python <ipython console> home bartvde Local lib python numpy linalg linalg.py in qr a, mode) 307 r = zeros mn,n), result t) 308 for i in range mn) > 309 r[i, i ] = atemp[i, i ] 310 311 r mode, that is, calculate only r ValueError shape mismatch objects cannot be broadcast to a single shape","bug"
|
|
"6013","946","Creating recarrays from numpy scalar voids fails Trac 348).
|
|
Original ticket http projects.scipy.org numpy ticket 348 on 2006 10 16 by FrancescAlted, assigned to unknown.
|
|
Hi, Perhaps we ve discussed this already, but I find the next a bit inconsistent.
|
|
The next works consistently In [87] numpy.rec.array 1, 2L), dtype=descr) Out[87] recarray 1, 2L), dtype=[ x , |u1 ), y , <u4 )]) In [88] numpy.rec.array [ 1, 2L)], dtype=descr) Out[88] recarray [ 1, 2L)], dtype=[ x , |u1 ), y , <u4 )]) i.e.
|
|
a recarray with shape ) or 1,) is returned depending on the shape of the input.
|
|
However In [89] numpy.rec.array numpy.array 1, 2L), dtype=descr), dtype=descr) Out[89] recarray 1, 2L), dtype=[ x , |u1 ), y , <u4 )]) Equals to 87, but...
|
|
In [90] numpy.rec.array [numpy.array 1, 2L), dtype=descr)], dtype=descr) exceptions.ValueError Traceback most recent call last) home faltet <ipython console> usr lib python2.4 site packages numpy core records.py in array obj, dtype, shape, offset, strides, formats, names, titles, aligned, byteorder, copy) 512 return fromrecords obj, dtype=dtype, shape=shape, kwds) 513 else > 514 return fromarrays obj, dtype=dtype, shape=shape, kwds) 515 516 elif isinstance obj, recarray) usr lib python2.4 site packages numpy core records.py in fromarrays arrayList, dtype, shape, formats, names, titles, aligned, byteorder) 296 Determine shape from data type.
|
|
297 if len descr) != len arrayList) > 298 raise ValueError, mismatch between the number of fields \ 299 and the number of arrays 300 ValueError mismatch between the number of fields and the number of arrays for me, the output for 90 should be similar to 88.","bug"
|
|
"6013","947","Improve unit tests in linalg Trac 349).
|
|
Original ticket http projects.scipy.org numpy ticket 349 on 2006 10 16 by charris, assigned to unknown.
|
|
None of the unit tests check results for non square arrays.","bug"
|
|
"6013","948","Poor wording of exception string invalid encountered in sqrt Trac 350).
|
|
Original ticket http projects.scipy.org numpy ticket 350 on 2006 10 17 by edschofield, assigned to unknown.
|
|
This exception string is poorly worded >>> import numpy >>> numpy.seterr invalid= raise ) >>> numpy.sqrt 1) Traceback most recent call last) File <stdin> , line 1, in ?
|
|
FloatingPointError invalid encountered in sqrt Perhaps invalid value encountered would be better or, better, negative value encountered if this is only raised with negative values.","bug"
|
|
"6013","949","set string function None) for reset of array repr Trac 351).
|
|
Original ticket http projects.scipy.org numpy ticket 351 on 2006 10 17 by trac user mackeith, assigned to unknown.
|
|
Currently it is not possible to reset the repr and str methods of the numpy array to use the built in functions array repr and array str.
|
|
These are set in numeric.py to use Python functions, and cannot be reset to the built in c functions.
|
|
I propose that set string function None, 0 1) will reset the str repr function to the built in version.
|
|
The changes to arrayobject.c and multiarraymodule.c are below arrayobject.c Tue Oct 17 17 53 35 2006 4132,4148 if repr) { Dispose of previous callback Py XDECREF PyArray ReprFunction); !
|
|
Add a reference to new callback !
|
|
Py XINCREF op); !
|
|
Remember new callback !
|
|
PyArray ReprFunction = op; } else { Dispose of previous callback Py XDECREF PyArray StrFunction); !
|
|
Add a reference to new callback !
|
|
Py XINCREF op); !
|
|
Remember new callback !
|
|
PyArray StrFunction = op; } } 4132,4162 if repr) { Dispose of previous callback Py XDECREF PyArray ReprFunction); !
|
|
if op == Py None) !
|
|
{ !
|
|
PyArray ReprFunction = NULL; ! }
|
|
!
|
|
else !
|
|
{ !
|
|
Add a reference to new callback !
|
|
Py XINCREF op); !
|
|
Remember new callback !
|
|
PyArray ReprFunction = op; ! }
|
|
} else { Dispose of previous callback Py XDECREF PyArray StrFunction); !
|
|
if op == Py None) !
|
|
{ !
|
|
PyArray StrFunction = NULL; ! }
|
|
!
|
|
else !
|
|
{ !
|
|
Add a reference to new callback !
|
|
Py XINCREF op); !
|
|
Remember new callback !
|
|
PyArray StrFunction = op; ! }
|
|
} } multiarraymodule.c Tue Oct 17 14 08 51 2006 6376,6382 if !PyArg ParseTupleAndKeywords args, kwds, O|i , kwlist, op, repr)) return NULL; !
|
|
if !PyCallable Check op)) { PyErr SetString PyExc TypeError, Argument must be callable.
|
|
); return NULL; 6376,6382 if !PyArg ParseTupleAndKeywords args, kwds, O|i , kwlist, op, repr)) return NULL; !
|
|
if !PyCallable Check op) op != Py None) { PyErr SetString PyExc TypeError, Argument must be callable or None.
|
|
); return NULL;}}}","enhancement"
|
|
"6013","950",".repeat on slice gets wrong slice Trac 352).
|
|
Original ticket http projects.scipy.org numpy ticket 352 on 2006 10 18 by trac user mrovner, assigned to unknown.
|
|
I got strange discrepance between 2.4 0.9.8 and 2.5 1.0rc2 model lib pool 0 !5019 python2.5 Python 2.5 r25 51908, Oct 17 2006, 16 16 21) [GCC 3.2.3 20030502 Red Hat Linux 3.2.3 14)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> a=arange 12).reshape 4,3) >>> a array [[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) >>> a[ ,2] array [ 2, 5, 8, 11]) >>> a[ ,2].repeat 3) array [2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5]) >>> import numpy >>> numpy.
|
|
version 1.0rc2 >>> Old version looks right Python 2.4.2 5, Jun 2 2006, 18 33 20) [GCC 3.2.3 20030502 Red Hat Linux 3.2.3 20)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> from numpy import >>> a=arange 12).reshape 4,3) >>> a array [[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) >>> a[ ,2] array [ 2, 5, 8, 11]) >>> a[ ,2].repeat 3) array [ 2, 2, 2, 5, 5, 5, 8, 8, 8, 11, 11, 11]) >>> import numpy >>> numpy.
|
|
version 0.9.8 >>>","bug"
|
|
"6013","951","Creating 0 dim arrays from an empty scalar fails Trac 353).
|
|
Original ticket http projects.scipy.org numpy ticket 353 on 2006 10 18 by FrancescAlted, assigned to unknown.
|
|
Hi, I don t know if this is related with ticket 331, but I think the next should be a bug In [71] == numpy.string ) Out[71] True In [72] numpy.array ) Out[72] array , dtype= |S1 ) so far so good, but...
|
|
In [73] numpy.array numpy.string )) exceptions.ValueError Traceback most recent call last) home faltet <ipython console> ValueError Empty data type","bug"
|
|
"6013","952","Possible inconsistency in 0 dim and scalar empty array types Trac 354).
|
|
Original ticket http projects.scipy.org numpy ticket 354 on 2006 10 18 by FrancescAlted, assigned to unknown.
|
|
Do you think that the next is inconsistent?
|
|
In [79] numpy.array ).dtype Out[79] dtype |S1 ) In [80] numpy.array )[ )].dtype Out[80] dtype |S0 ) In [81] numpy.string ).dtype Out[81] dtype |S0 ) I d tend to think that it is.
|
|
Regards","bug"
|
|
"6013","953","Valgrind warning on arrayobject.c Trac 355).
|
|
Original ticket http projects.scipy.org numpy ticket 355 on 2006 10 18 by FrancescAlted, assigned to unknown.
|
|
Hi, I m trying to run the pytables tests against valgrind and so far, I ve got the next warning ==6038== Conditional jump or move depends on uninitialised value s) ==6038== at 0x49A1D1F PyArray Scalar arrayobject.c 1329) ==6038== by 0x49A3B2E array item nice arrayobject.c 2001) ==6038== by 0x49A65E4 array subscript nice arrayobject.c 2931) ==6038== by 0x80B7394 PyEval EvalFrame in usr bin python2.4) This was using numpy 1.0.dev3341 , so the affected revision is source trunk numpy core src arrayobject.c 3341 I have had a look at the offending line, but it seems ok to me.
|
|
If you think that these kind of errors are not worth to note, please, tell me and I won t file anymore unless they would be really harmful).","bug"
|
|
"6013","954","array.reshape None) segfaults Trac 356).
|
|
Original ticket http projects.scipy.org numpy ticket 356 on 2006 10 19 by huard, assigned to unknown.
|
|
a.reshape None) a.resize None) crash the shell.
|
|
1.0.dev3327","bug"
|
|
"6013","955","Memory leak iterating over arange Trac 357).
|
|
Original ticket http projects.scipy.org numpy ticket 357 on 2006 10 20 by trac user tlbtlbtlb, assigned to unknown.
|
|
Dereferencing the result of arange seems to leak memory, about 80 bytes each time.
|
|
In the attached example code, either function will run out of memory.
|
|
I run it with a 30 MB limit ulimit d 30000 ; .
|
|
t nump leak.py) so you can see it run out without hosing your machine, but these will rapidly use up any amount of memory.
|
|
I m using 1.0rc2 from subversion, built on FreeBSD 5.4 with Python 2.4.3","bug"
|
|
"6013","956","Reshape makes unnecessary copies Trac 358).
|
|
Original ticket http projects.scipy.org numpy ticket 358 on 2006 10 20 by trac user peridot, assigned to unknown.
|
|
Reshape s rule for detecting when it can return a view is too conservative.
|
|
This results in unnecessary copying of data, and it restricts the utility of assigning to the shape attribute which never copies).
|
|
Consider the following session In [3] from numpy import In [4] z = zeros 4,4))[ ,1] In [5] z Out[5] array [ 0., 0., 0., 0.])
|
|
In [6] r = z.reshape 2,2)) In [7] r[0]=1 In [8] z Out[8] array [ 0., 0., 0., 0.])
|
|
The elements of z are not contiguous, but they are a linear array spaced by 4 sizeof float); there s no reason this can t be reshaped with a view.
|
|
Attached is a patch that recognizes all situations where a view can be returned.
|
|
It passes all tests in numpy, and a few of my own devising although it could use a more thorough and organized test suite).
|
|
The new code is only called when a copy would have been made under the current rules, so it should not affect the speed of easy reshapes.
|
|
It does correctly handle the easy reshapes, so the code could perhaps be simplified by removing some special case detection.","bug"
|
|
"6013","957","Compiler warning in ufuncobject.c Trac 359).
|
|
Original ticket http projects.scipy.org numpy ticket 359 on 2006 10 20 by alberts, assigned to unknown.
|
|
Compiled with GCC 4.1.1 on 32 bit Linux numpy core src ufuncobject.c 699 warning i may be used uninitialized in this function Checked with r3374.","bug"
|
|
"6013","958","Valgrind errors Trac 360).
|
|
Original ticket http projects.scipy.org numpy ticket 360 on 2006 10 20 by alberts, assigned to unknown.
|
|
Valgrind warnings when running the test suite with r3374.
|
|
Warnings pop up when running the regression test for ticket 852 and when running test ma.check testArithmetic in numpy core tests test ma.py.
|
|
Valgrind says ==2256== ==2256== Use of uninitialised value of size 4 ==2256== at 0x4A42EDD PyDict Next in usr lib libpython2.4.so.1.0) ==2256== by 0x45CEBC1 construct loop ufuncobject.c 1615) ==2256== by 0x45D03A2 PyUFunc GenericFunction ufuncobject.c 1735) ==2256== by 0x45D14CA ufunc generic call ufuncobject.c 3047) ==2256== by 0x4A1B586 PyObject Call in usr lib libpython2.4.so.1.0) ...
|
|
Tested with Valgrind 3.1.0.
|
|
If any developer s) need help in getting Valgrind going on their systems, I would be more than happy to lend a hand.","bug"
|
|
"6013","959","f2py SyntaxError invalid syntax Trac 361).
|
|
Original ticket http projects.scipy.org numpy ticket 361 on 2006 10 21 by nilswagner01, assigned to unknown.
|
|
python setup.py install yields syntax errors byte compiling usr lib python2.4 site packages numpy f2py lib parser base classes.py to base classes.pyc File usr lib python2.4 site packages numpy f2py lib parser base classes.py , line 369 return self.is allocatable ) or self.is pointer ) SyntaxError invalid syntax byte compiling usr lib python2.4 site packages numpy f2py lib parser expressions.py to expressions.pyc File usr lib python2.4 site packages numpy f2py lib parser expressions.py , line 142 def compose patterns pattern list, names join= ) SyntaxError invalid syntax byte compiling usr lib python2.4 site packages numpy f2py lib parser test expressions.py to test expressions.pyc","bug"
|
|
"6013","960","UFUNC CHECK STATUS doesn t correctly clear the fp flags on AIX Trac 362).
|
|
Original ticket http projects.scipy.org numpy ticket 362 on 2006 10 22 by trac user bgranger, assigned to unknown.
|
|
On AIX, UFUNC CHECK STATUS in ufuncobject.h) always return 8 indicating that an INVALID floating point error has occured.
|
|
This is happening because the fp flags are not being cleared.
|
|
The current code tries to clear the flags with the call fp clr flag FP DIV BY ZERO | FP OVERFLOW | FP UNDERFLOW | FP INVALID); But this doesn t work as expected.
|
|
Instead, after this call is made the FP INVALID flag is set to true so all further calls to UFUNC CHECK STATUS return 8.
|
|
The correct way to clear the fp flags on AIX is fp swap flag 0); the AIX documentation http publib.boulder.ibm.com infocenter pseries v5r3 index.jsp?topic= com.ibm.aix.basetechref doc basetrf1 fp clr flag.htm Makes only a vague reference to this being necessary, but apparently it is.
|
|
Could we fix this before 1.0?
|
|
Thanks!!!
|
|
!","bug"
|
|
"6013","961","failure to install numpy Trac 363).
|
|
Original ticket http projects.scipy.org numpy ticket 363 on 2006 10 23 by trac user mbennun, assigned to unknown.
|
|
Hello We have tried to install numpy and failed.
|
|
We have python2.4 and were able to successfully install f2py and scipy distutils.
|
|
The output of the python setup.py install command is attached below.
|
|
Thank you for your help.
|
|
Michal Ben Nun Running from numpy source directory.
|
|
F2PY Version 2 3364 blas opt info blas mkl info libraries mkl,vml,guide not found in usr1 mbennun work NSFCyber python intel lib libraries mkl,vml,guide not found in usr local lib libraries mkl,vml,guide not found in usr lib NOT AVAILABLE atlas blas threads info Setting PTATLAS=ATLAS libraries lapack,blas not found in usr1 mbennun work NSFCyber python intel lib Setting PTATLAS=ATLAS Setting PTATLAS=ATLAS FOUND libraries = [ lapack , blas ] library dirs = [ usr local lib ] language = c include dirs = [ usr local include atlas ] Could not locate executable f95 customize GnuFCompiler customize GnuFCompiler customize GnuFCompiler using config compiling configtest.c This file is generated from numpy distutils system info.py void ATL buildinfo void); int main void) { ATL buildinfo ); return 0; } C compiler gcc pthread fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes fPIC compile options c gcc configtest.c tmp ccSQrFKJ.s Assembler messages tmp ccSQrFKJ.s 20 Error suffix or operands invalid for `push tmp ccSQrFKJ.s 22 Error suffix or operands invalid for `push tmp ccSQrFKJ.s 22 Error suffix or operands invalid for `push tmp ccSQrFKJ.s 31 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
tmp ccSQrFKJ.s Assembler messages tmp ccSQrFKJ.s 20 Error suffix or operands invalid for `push tmp ccSQrFKJ.s 22 Error suffix or operands invalid for `push tmp ccSQrFKJ.s 31 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
failure.
|
|
removing configtest.c configtest.o Status 255 Output FOUND libraries = [ lapack , blas ] library dirs = [ usr local lib ] language = c define macros = [ NO ATLAS INFO , 2)] include dirs = [ usr local include atlas ] lapack opt info lapack mkl info mkl info libraries mkl,vml,guide not found in usr1 mbennun work NSFCyber python intel lib libraries mkl,vml,guide not found in usr local lib libraries mkl,vml,guide not found in usr lib NOT AVAILABLE NOT AVAILABLE atlas threads info Setting PTATLAS=ATLAS libraries lapack,blas not found in usr1 mbennun work NSFCyber python intel lib libraries lapack atlas not found in usr1 mbennun work NSFCyber python intel lib libraries lapack atlas not found in usr local lib numpy.distutils.system info.atlas threads info Setting PTATLAS=ATLAS usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils system info.py 934 UserWarning Lapack library from ATLAS) is probably incomplete size of usr local lib liblapack.a is 3887k expected >4000k) Follow the instructions in the KNOWN PROBLEMS section of the file numpy INSTALL.txt.
|
|
warnings.warn message) Setting PTATLAS=ATLAS FOUND libraries = [ lapack , lapack , blas ] library dirs = [ usr local lib ] language = c include dirs = [ usr local include atlas ] customize GnuFCompiler customize GnuFCompiler customize GnuFCompiler using config compiling configtest.c This file is generated from numpy distutils system info.py void ATL buildinfo void); int main void) { ATL buildinfo ); return 0; } C compiler gcc pthread fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes fPIC compile options c gcc configtest.c tmp ccYe889d.s Assembler messages tmp ccYe889d.s 20 Error suffix or operands invalid for `push tmp ccYe889d.s 22 Error suffix or operands invalid for `push tmp ccYe889d.s 31 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
tmp ccYe889d.s Assembler messages tmp ccYe889d.s 20 Error suffix or operands invalid for `push tmp ccYe889d.s 22 Error suffix or operands invalid for `push tmp ccYe889d.s 31 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
failure.
|
|
removing configtest.c configtest.o Status 255 Output FOUND libraries = [ lapack , lapack , blas ] library dirs = [ usr local lib ] language = c define macros = [ NO ATLAS INFO , 2)] include dirs = [ usr local include atlas ] running install running build running config fc running build src building py modules sources building extension numpy.core.multiarray sources Generating build src.linux x86 64 2.4 numpy core config.h customize GnuFCompiler customize GnuFCompiler customize GnuFCompiler using config C compiler gcc pthread fno strict aliasing DNDEBUG g O3 Wall Wstrict prototypes fPIC compile options I usr1 mbennun work NSFCyber python intel include python2.4 Inumpy core src Inum py core include I usr1 mbennun work NSFCyber python intel include python2.4 c gcc configtest.c tmp ccaSumWn.s Assembler messages tmp ccaSumWn.s 59 Error suffix or operands invalid for `push tmp ccaSumWn.s 61 Error suffix or operands invalid for `push tmp ccaSumWn.s 65 Error suffix or operands invalid for `push tmp ccaSumWn.s 67 Error suffix or operands invalid for `push tmp ccaSumWn.s 70 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
tmp ccaSumWn.s Assembler messages tmp ccaSumWn.s 59 Error suffix or operands invalid for `push tmp ccaSumWn.s 61 Error suffix or operands invalid for `push tmp ccaSumWn.s 65 Error suffix or operands invalid for `push tmp ccaSumWn.s 67 Error suffix or operands invalid for `push tmp ccaSumWn.s 70 Internal error, aborting at .. .. gas config tc i386.c line 3501 in output imm Please report this bug.
|
|
failure.
|
|
removing configtest.c configtest.o Traceback most recent call last) File setup.py , line 89, in ?
|
|
setup package ) File setup.py , line 82, in setup package configuration=configuration ) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils core.py , line 174, in setup return old setup new attr) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils core.py , line 149, in setup dist.run commands ) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils dist.py , line 946, in run c ommands self.run command cmd) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils dist.py , line 966, in run c ommand cmd obj.run ) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils command install.py , line 16, in run r = old install.run self) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils command install.py , line 50 6, in run self.run command build ) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils cmd.py , line 333, in run co mmand self.distribution.run command command) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils dist.py , line 966, in run c ommand cmd obj.run ) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils command build.py , line 112, in run self.run command cmd name) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils cmd.py , line 333, in run co mmand self.distribution.run command command) File usr1 mbennun work NSFCyber python intel lib python2.4 distutils dist.py , line 966, in run c ommand cmd obj.run ) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils command build src.py , line 87, in r un self.build sources ) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils command build src.py , line 106, in build sources self.build extension sources ext) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils command build src.py , line 212, in build extension sources sources = self.generate sources sources, ext) File usr1 mbennun work NSFCyber numpy 1.0rc3 numpy distutils command build src.py , line 270, in generate sources source = func extension, build dir) File numpy core setup.py , line 50, in generate config h raise ERROR Failed to test configuration ERROR Failed to test configuration","bug"
|
|
"6013","962","numpy.core.tests.test defmatrix.test matrix return test failure Trac 364).
|
|
Original ticket http projects.scipy.org numpy ticket 364 on 2006 10 23 by alberts, assigned to unknown.
|
|
r3386 seems to have broken numpy.core.tests.test defmatrix.test matrix return.
|
|
====================================================================== FAIL check instance methods numpy.core.tests.test defmatrix.test matrix return) Traceback most recent call last) File usr lib python2.4 site packages numpy core tests test defmatrix.py , line 167, in check instance methods assert type b) is matrix, s attrib AssertionError getA1","bug"
|
|
"6013","963","docstring for modes of multiarray.correlate Trac 365).
|
|
Original ticket http projects.scipy.org numpy ticket 365 on 2006 10 25 by cournape, assigned to unknown.
|
|
It would be good to have a short docstring for numpy.correlate right now, there is no information on the modes, you have to look for them in scipy.signal.correlate...","enhancement"
|
|
"6013","964","Warnings from Intel C compiler 9.1 Trac 366).
|
|
Original ticket http projects.scipy.org numpy ticket 366 on 2006 10 27 by alberts, assigned to unknown.
|
|
The Intel C Compiler turned up a few warnings that should probably be looked at.
|
|
Most of them look harmless, but maybe one or two could indicate the presence of a bug.","bug"
|
|
"6013","965","Startup errors in latest build variable as Trac 367).
|
|
Original ticket http projects.scipy.org numpy ticket 367 on 2006 10 28 by trac user melson, assigned to unknown.
|
|
Using Python revision 52498 and numpy revision 3413 I get an error when importing numpy me legion>python setup.py build Work matplotlib svn Traceback most recent call last) File setup.py , line 118, in <module> import numpy File home me Work lib python2.6 site packages numpy init .py , line 36, in <module> import core File home me Work lib python2.6 site packages numpy core init .py , line 10, in <module> from numeric import File home me Work lib python2.6 site packages numpy core numeric.py , line 293 as = a.shape SyntaxError invalid syntax It looks like the problem is the use of as as a variable name.
|
|
I ve attached a patch changing it to a s .","bug"
|
|
"6013","966","Patch to remove 2 instances of as as a variable name Trac 368).
|
|
Original ticket http projects.scipy.org numpy ticket 368 on 2006 10 29 by trac user melson, assigned to unknown.
|
|
numpy lib shape base.py and numpy f2py crackfortran.py both contain variables named as , a reserved keyword.
|
|
The first is preventing matplotlib from installing, the second isn t causing an error for me but some code might break.
|
|
Attached patch renames both to as .","bug"
|
|
"6013","967","One point windows invalid Trac 369).
|
|
Original ticket http projects.scipy.org numpy ticket 369 on 2006 10 30 by alberts, assigned to unknown.
|
|
The blackman, bartlett, hanning and hamming windows return invalid values for 1 point windows, e.g.
|
|
In [4] N.hamming 1) Warning invalid value encountered in divide Out[4] Warning invalid value encountered in absolute array [ 1.
|
|
IND]) MATLAB returns >> hamming 1) ans = 1 Having these functions return valid 1 point windows is useful when writing unit tests for other code.","bug"
|
|
"6013","968","oldnumeric.Int32 mapping on 32 bit platforms seems wrong Trac 370).
|
|
Original ticket http projects.scipy.org numpy ticket 370 on 2006 11 02 by FrancescAlted, assigned to unknown.
|
|
Hi, I think that, for 32 bit platforms, oldnumeric.Int32 should be mapped to a C long, instead as a C int as it is now.
|
|
This is to be consistent with In [16] numpy.zeros 1,), numpy.int32).dtype.num Out[16] 7 In [17] numpy.zeros 1,), numpy.sctypeNA[ Int32 ]).dtype.num Out[17] 7 In [18] numpy.zeros 1,), oldnumeric.int32).dtype.num Out[18] 7 Fine until here, but In [19] numpy.zeros 1,), oldnumeric.Int32).dtype.num Out[19] 5 !!
|
|
This can effectively be a source of problems with packages that use oldnumeric.
|
|
scipy.ndimage is one of them that gets affected by this.
|
|
See http aspn.activestate.com ASPN Mail Message scipy user 3321095 Regards,","bug"
|
|
"6013","969","Potential memory leak with numpy 1.0 Trac 371).
|
|
Original ticket http projects.scipy.org numpy ticket 371 on 2006 11 03 by trac user tepperly, assigned to unknown.
|
|
compiled python 2.5 with .
|
|
configure enable shared without pymalloc, and I defined define d Py USING MEMORY DEBUGGER in Python 2.5 Objects obmalloc.c If I run python c import numpy , I see some unfreed pointers in the numpy code.
|
|
Here is the valgrind output for the potential leak ==7786== 60 bytes in 10 blocks are still reachable in loss record 13 of 48 ==7786== at 0x401A662 malloc vg replace malloc.c 149) ==7786== by 0x49C807F PyArray NewFromDescr arrayobject.c 5361) ==7786== by 0x49C54D3 PyArray FromScalar scalartypes.inc.src 187) ==7786== by 0x49CA561 PyArray FromAny arrayobject.c 8296) ==7786== by 0x49CA912 PyArray CheckFromAny arrayobject.c 8476) ==7786== by 0x4A04ED1 array fromobject multiarraymodule.c 5392) ==7786== by 0x407224D PyCFunction Call methodobject.c 108) ==7786== by 0x40B7CE4 call function ceval.c 3566) ==7786== by 0x40B5D3E PyEval EvalFrameEx ceval.c 2269) ==7786== by 0x40B6729 PyEval EvalCodeEx ceval.c 2833) ==7786== by 0x40B7E1E fast function ceval.c 3662) ==7786== by 0x40B7A38 call function ceval.c 3587) This looks like a potential leak.
|
|
The whole valgrind output log is in the attached file memleak.txt.
|
|
It also has a bigger stack trace.","bug"
|
|
"6013","970","Shifted numbering in type names in records.py module Trac 372).
|
|
Original ticket http projects.scipy.org numpy ticket 372 on 2006 11 04 by FrancescAlted, assigned to unknown.
|
|
When you create recarrays with records.py and not specify the dtype argument explicitely, the names are generated with 1 as base i.e.
|
|
f1 , f2 , ...) instead of using 0 i.e.
|
|
f0 , f1 ,...) which is the default in !NumPy.
|
|
The next exposes the issue In [7] numpy.array [ 1,2), 3,4)], i4,i4 ) Out[7] array [ 1, 2), 3, 4)], dtype=[ f0 , <i4 ), f1 , <i4 )]) In [8] numpy.rec.array [ 1,2), 3,4)], i4,i4 ) Out[8] recarray [ 1, 2), 3, 4)], dtype=[ f0 , <i4 ), f1 , <i4 )]) In [9] numpy.rec.array [ 1,2), 3,4)]) Out[9] recarray [ 1, 2), 3, 4)], dtype=[ f1 , <i4 ), f2 , <i4 )]) !
|
|
In [10] numpy.rec.fromarrays [ 1,2), 3,4)], i4,i4 ) Out[10] recarray [ 1, 3), 2, 4)], dtype=[ f0 , <i4 ), f1 , <i4 )]) In [11] numpy.rec.fromarrays [ 1,2), 3,4)]) Out[11] recarray [ 1, 3), 2, 4)], dtype=[ f1 , <i4 ), f2 , <i4 )]) !
|
|
The next patch is a cure for this numpy core records.py revision 3428) numpy core records.py working copy) 82,11 82,11 else self.
|
|
names = [] if the names are not specified, they will be assigned as f1, f2,... if not enough names are specified, they will be assigned as f[n 1], f[n 2],... etc.
|
|
where n is the number of specified names... self.
|
|
names = [ f d i for i in range len self.
|
|
names) 1, self.
|
|
nfields 1)] if the names are not specified, they will be assigned as f0, f1,... if not enough names are specified, they will be assigned as f[n], f[n 1],... etc.
|
|
where n is the number of specified names... self.
|
|
names = [ f d i for i in range len self.
|
|
names), self.
|
|
nfields)] check for redundant names dup = find duplicate self.
|
|
names) if dup","bug"
|
|
"6013","971","Yet another valgrind warning Trac 373).
|
|
Original ticket http projects.scipy.org numpy ticket 373 on 2006 11 04 by FrancescAlted, assigned to unknown.
|
|
Hi, Valgrind is detecting this in some tests of mine ==12109== Invalid read of size 4 ==12109== at 0x446374 PyObject Free obmalloc.c 919) ==12109== by 0x6010F87 array dealloc arrayobject.c 1930) ==12109== by 0x4CD67B frame dealloc frameobject.c 416) ==12109== by 0x4833ED PyEval EvalFrameEx ceval.c 3654) ==12109== by 0x4833CA PyEval EvalFrameEx ceval.c 3652) ==12109== by 0x4845B0 PyEval EvalCodeEx ceval.c 2833) [snip] ==12109== Conditional jump or move depends on uninitialised value s) ==12109== at 0x44637D PyObject Free obmalloc.c 919) ==12109== by 0x6010F87 array dealloc arrayobject.c 1930) ==12109== by 0x6010F87 array dealloc arrayobject.c 1930) ==12109== by 0x4CD67B frame dealloc frameobject.c 416) ==12109== by 0x4833ED PyEval EvalFrameEx ceval.c 3654) ==12109== by 0x4845B0 PyEval EvalCodeEx ceval.c 2833) ==12109== by 0x4CF0D3 function call funcobject.c 517) ==12109== by 0x415C32 PyObject Call abstract.c 1860) ==12109== by 0x4817F8 PyEval EvalFrameEx ceval.c 3846) ==12109== by 0x4845B0 PyEval EvalCodeEx ceval.c 2833) ==12109== by 0x4CF0D3 function call funcobject.c 517) ==12109== by 0x415C32 PyObject Call abstract.c 1860) ==12109== ==12109== Use of uninitialised value of size 8 ==12109== at 0x446396 PyObject Free obmalloc.c 919) ==12109== by 0x6010F87 array dealloc arrayobject.c 1930) ==12109== by 0x6010F87 array dealloc arrayobject.c 1930) ==12109== by 0x4CD67B frame dealloc frameobject.c 416) ==12109== by 0x4833ED PyEval EvalFrameEx ceval.c 3654) ==12109== by 0x4845B0 PyEval EvalCodeEx ceval.c 2833) Is self.base being always initialized properly?
|
|
This is using pure !NumPy i.e.
|
|
without doing conversions with array protocol).
|
|
I m using r3428, so the affected line should be seen in source trunk numpy core src arrayobject.c 3428 Sorry, but again, this come without a code snippet","bug"
|
|
"6013","972","random.shuffle breaks for certain shapes Trac 374).
|
|
Original ticket http projects.scipy.org numpy ticket 374 on 2006 11 05 by stefanv, assigned to rkern.
|
|
As reported on the mailing list by Yves Frederix In [1] import numpy as N In [2] a = N.arange 5).reshape 5,1)) In [3] a Out[3] array [[0], [1], [2], [3], [4]]) In [4] N.random.shuffle a) In [5] a Out[5] array [[0], [1], [1], [3], [3]])","bug"
|
|
"6013","973","Numpy.arange ) gives confusing error message Trac 375).
|
|
Original ticket http projects.scipy.org numpy ticket 375 on 2006 11 08 by trac user gpk2, assigned to unknown.
|
|
if x is a Numpy array, then Num.arange x.shape) reports TypeError unsupported operand type s) for tuple and int This is confusing.
|
|
What it should report is that arange needs integer argument s), rather than a tuple.
|
|
Normally, this would not be a big deal, but it s a very easy error to make, and one that is not particularly easy to see when one is debugging.","bug"
|
|
"6013","974","ma.average destroys mask Trac 376).
|
|
Original ticket http projects.scipy.org numpy ticket 376 on 2006 11 08 by trac user roberto, assigned to unknown.
|
|
The function ma.average does not preserve the mask >>> from numpy import >>> a = ma.masked values arange 8), 4) >>> a.shape = 2,2,2) >>> assert a.shape == a.mask.shape >>> b = ma.average a, axis=0) >>> assert b.shape == b.mask.shape Traceback most recent call last) File <stdin> , line 1, in ?
|
|
AssertionError >>> print b.shape, b.mask.shape 2, 2) )","enhancement"
|
|
"6013","975","Compiler warnings in lapack lite.c Trac 377).
|
|
Original ticket http projects.scipy.org numpy ticket 377 on 2006 11 11 by alberts, assigned to unknown.
|
|
Warning from GCC 4.0.3 when building lapack litemodule.c as of r3432 numpy linalg lapack litemodule.c In function lapack lite dgesdd numpy linalg lapack litemodule.c 462 warning label defaul defined but not used numpy linalg lapack litemodule.c 446 warning nwork may be used uninitialized in this function","bug"
|
|
"6013","976","NumPy or Python crashing with segfault Trac 378).
|
|
Original ticket http projects.scipy.org numpy ticket 378 on 2006 11 11 by trac user tecki, assigned to unknown.
|
|
I just ran the following little python script import numpy a = numpy.arange 100 100) a = a.reshape 100, 100) b = a def p x, y) print x, y return x v = numpy.vectorize p) print v a,b) and I god a segmentation fault.
|
|
I considered this strange, since there should be no segfaults using python, but exceptions, no?
|
|
OK, I am using slackware linux on an i386, tried their packages python 2.4.1, python 2.4.3 and a self compiled python 2.5, all with numpy 1.0.
|
|
Since the python version doesn t matter, I guessed it is the numpy package which makes problems.","bug"
|
|
"6013","977","different error when slicing with Python 2.4 and 2.5 Trac 379).
|
|
Original ticket http projects.scipy.org numpy ticket 379 on 2006 11 13 by bthyreau, assigned to unknown.
|
|
Hi, I happened to get an error when slicing with an array.
|
|
from numpy import names = arange 8) v = array [8]) names[ v] Python 2.4 outputs `array [0, 1, 2, 3, 4, 5, 6, 7])` while Python 2.5 outputs `TypeError only 0 d integerarrays can be converted to an index` When `v = array [7, 8])`, Python 2.4 outputs `IndexError invalid slice` This was tested using various >1.0 Numpy.
|
|
On my case 1.0, 1.0.1.dev3416, 1.0.1.dev3432...
|
|
Since slicing with an 1 element array is wrong anyway, this shouldn t occurs frequently, but might break badly written code.
|
|
Thanks","bug"
|
|
"6013","978","broadcasting problem?
|
|
differs from = and differs from = Trac 380).
|
|
Original ticket http projects.scipy.org numpy ticket 380 on 2006 11 14 by trac user holistone, assigned to unknown.
|
|
!python a = ones 3,2)) print a a[0] [[ 0.
|
|
0.]
|
|
[ 0.
|
|
0.]
|
|
[ 0.
|
|
0.]]
|
|
a = a[0] print a [[ 0.
|
|
0.]
|
|
[ 1.
|
|
1.]
|
|
[ 1.
|
|
1.]]
|
|
a = ones 3,2)) a a[0] array [[ 2., 2.
|
|
], [ 2., 2.
|
|
], [ 2., 2.]])
|
|
a = a[0] a array [[ 2., 2.
|
|
], [ 3., 3.
|
|
], [ 3., 3.]])
|
|
I expected the same result in both cases.
|
|
Is this the intended behavior?","bug"
|
|
"6013","979","getattr True , imag ) and getattr False , imag ) cause a seg fault Trac 381).
|
|
Original ticket http projects.scipy.org numpy ticket 381 on 2006 11 15 by trac user davebaum, assigned to unknown.
|
|
The following code causes a core dump import numpy getattr numpy.False , imag ) I have tried this on a 1.0b5 build on Mac OSX and 1.0rc1 build on Linux.","bug"
|
|
"6013","980","Huge file in SVN repositori Trac 382).
|
|
Original ticket http projects.scipy.org numpy ticket 382 on 2006 11 16 by FrancescAlted, assigned to unknown.
|
|
Hi, From some time ago, I m detecting a very huge file in the !NumPy repository trunk).
|
|
The file is ll trunk numpy f2py lib research total 4585 rw r r 1 faltet users 4687328 2006 11 16 20 24 04 007.pdf It seems that it is the draft for a forthcoming version of Fortran.
|
|
Given that it doesn t seem relevant for NumPy to work, I d remove it from the repository.","bug"
|
|
"6013","981","PyArray GETPTR C API macros defect Trac 383).
|
|
Original ticket http projects.scipy.org numpy ticket 383 on 2006 11 19 by trac user yairchu, assigned to charris.
|
|
missing ) for the i of PyArray GETPTR1 obj, i) currently = void ) PyArray BYTES obj) i PyArray STRIDES obj)[0]) should turn i to i) might create hard to find bugs like it did for me )","bug"
|
|
"6013","982","numpy.fft.fft segfaults when called with null of negative argument for size Trac 384).
|
|
Original ticket http projects.scipy.org numpy ticket 384 on 2006 11 20 by cournape, assigned to unknown.
|
|
I made a mistake in my code by giving an axis argument instead of the expected size argument in numpy.fft.fft this resulted in segfault.
|
|
!python from numpy import fft fft random.randn 10, 2), 0) >segfaults !python from numpy import fft fft random.randn 10, 2), 1) >100 CPU, and fails with floating point excetpion I understand the arguments does not make any sense, but would it be possible to detect this ?","bug"
|
|
"6013","983","distutils has invalid flag for nagware fortran compiler version 5 Trac 385).
|
|
Original ticket http projects.scipy.org numpy ticket 385 on 2006 11 21 by trac user jgraham, assigned to unknown.
|
|
On upgrading from version 4.something I think) to version 5.1 of the nagware compiler for Linux, f2py no longer works as the ` target=native` option is no longer recognised; the last useful error message before failure is `Option error Unrecognised argument to target=` I think just removing this option in distutils fcompiler nag.py is OK for version 5 but I don t know what effect this will have on earlier versions.","bug"
|
|
"6013","984","import error Trac 386).
|
|
Original ticket http projects.scipy.org numpy ticket 386 on 2006 11 22 by astraw, assigned to stefanv.
|
|
With revision 3457 current as of now), the script from numpy.numarray.linear algebra import solve linear equations as solve from numpy.oldnumeric.linear algebra import results in the traceback Traceback most recent call last) File NUMPY IMPORT BUG.py , line 2, in ?
|
|
from numpy.oldnumeric.linear algebra import AttributeError module object has no attribute qr decomposition I will debug this and submit a patch as I have time.
|
|
If someone beats me to it, great.","bug"
|
|
"6013","985","Small polynomial bug with patch suggestion) Trac 387).
|
|
Original ticket http projects.scipy.org numpy ticket 387 on 2006 11 22 by trac user clovisgo, assigned to unknown.
|
|
Subject Small bug in polynomial There seems to exist a small bug in the numpy.polynomial functions Consider the following script import numpy poly1 = numpy.poly1d [1.0]) poly2 = numpy.poly1d [1.0, 1.0]) poly3 = numpy.poly1d [2.0,1.0]) print Poly1 is print poly1 print type poly1) print Poly2 is print poly2 print type poly2) print Poly3 is print poly3 print type poly3) print Poly1 Poly2 = auxvar = numpy.polyadd poly1,poly2) print auxvar print type auxvar) print Poly2 Poly3 = auxvar = numpy.polyadd poly2,poly3) print auxvar print type auxvar) Raw output is given below Poly1 is 1 <class numpy.lib.polynomial.poly1d > Poly2 is 1 x 1 <class numpy.lib.polynomial.poly1d > Poly3 is 2 x 1 <class numpy.lib.polynomial.poly1d > Poly1 Poly2 = 1 x 2 <class numpy.lib.polynomial.poly1d > Poly2 Poly3 = [ 3.
|
|
2.]
|
|
<type numpy.ndarray > Comments.
|
|
The last test Poly2 Poly3) is bad!
|
|
The result shoud be 3 x 2 and the type should be <class numpy.lib.polynomial.poly1d > and not <type numpy.ndarray > The same problem exists with polysub.
|
|
The solution seems to be simple.
|
|
The original code for the polyadd function with line numbers) is 1 def polyadd a1, a2) 2 Adds two polynomials represented as sequences 3 4 truepoly = isinstance a1, poly1d) or isinstance a2, poly1d)) 5 a1 = atleast 1d a1) 6 a2 = atleast 1d a2) 7 diff = len a2) len a1) 8 if diff == 0 9 return a1 a2 10 elif diff > 0 11 zr = NX.zeros diff, a1.dtype) 12 val = NX.concatenate zr, a1)) a2 13 else 14 zr = NX.zeros abs diff), a2.dtype) 15 val = a1 NX.concatenate zr, a2)) 16 if truepoly 17 val = poly1d val) 18 return val In order to correct the problem is sufficient to replace the original line 9 by 9 val = a1 a2 If this change is made, the return value will be an array if input variables are arrays) or a poly1d if one of the input variables are polynomials).
|
|
A similar change should be made for the polysub function.
|
|
Clovis","bug"
|
|
"6013","986","Doc error for std in http scipy.org Numpy Example List std Trac 388).
|
|
Original ticket http projects.scipy.org numpy ticket 388 on 2006 11 22 by trac user g2boojum, assigned to unknown.
|
|
According to the example in http scipy.org Numpy Example List std, array.std ) uses an N 1 normalization and that s what the example shows).
|
|
With numpy 1.0 on my machine In [22] a = array [1., 2., 7.])
|
|
In [23] a.std ) Out[23] 2.62466929134 In [24] sqrt 1. a.mean )) 2 2. a.mean )) 2 7. a.mean )) 2) 3) Out[24] 2.6246692913372702 In [25] sqrt 1. a.mean )) 2 2. a.mean )) 2 7. a.mean )) 2) 2) Out[25] 3.214550253664318 < That s the result in the example Please fix the docs.","bug"
|
|
"6013","987","ones is not working for compound types Trac 389).
|
|
Original ticket http projects.scipy.org numpy ticket 389 on 2006 11 22 by FrancescAlted, assigned to unknown.
|
|
Hi, I don t know if the next is intended or not In [51] numpy.ones shape=2, dtype= i2 ) Out[51] array [1, 1], dtype=int16) So far so good.
|
|
However, In [52] numpy.zeros shape=2, dtype= i4,i2 ) Out[52] array [ 0, 0), 0, 0)], dtype=[ f0 , <i4 ), f1 , <i2 )]) In [53] numpy.ones shape=2, dtype= i4,i2 ) exceptions.TypeError Traceback most recent call last) home faltet <ipython console> usr lib python2.4 site packages numpy core numeric.py in ones shape, dtype, order) 587 588 a = empty shape, dtype, order) > 589 a.fill 1) 590 Above is faster now after addition of fast loops.
|
|
591 a = zeros shape, dtype, order) TypeError expected a readable buffer object I d say that, provided that ones ) works well for single types, and zeros also works for compound types, then ones ) is having some problem there.
|
|
Using 1.0.1.dev3428","bug"
|
|
"6013","988","repr ) on recarray failure Trac 390).
|
|
Original ticket http projects.scipy.org numpy ticket 390 on 2006 11 22 by astraw, assigned to unknown.
|
|
The script import numpy a= numpy.rec.array [ 1, 2, 3)],dtype=[ a , numpy.int8), b , numpy.int8), c , numpy.int8)]) x=repr a) from numpy import eval x) results in Traceback most recent call last) File NUMPY RECARRAY BUG.py , line 10, in ?
|
|
eval x) File <string> , line 2, in ?
|
|
File usr lib python2.4 site packages numpy core records.py , line 181, in new self = ndarray.
|
|
new subtype, shape, record, descr)) TypeError an integer is required Adding `x = x.replace recarray , rec.array )` before eval ) stops the exception.
|
|
I think this illustrates that rec.array s ` repr )` should print rec.array rather than recarray .
|
|
However, I m no recarray expert...","bug"
|
|
"6013","989","argsort default axis axis=None Trac 391).
|
|
Original ticket http projects.scipy.org numpy ticket 391 on 2006 11 28 by pierregm, assigned to unknown.
|
|
Folks, The default axis for `argsort` method function) is 1, when it s `None` for `argmin` and `argmax`.
|
|
Moreover, forcing the axis to `None` raises a `TypeError an integer is required` exception.
|
|
>>>import numpy as N >>>N.version.version 1.0.1.dev3460 >>>a=N.array [[1,3],[2,4]]) >>>N.argmin a), N.argmax a), 0,3) >>>N.argsort a) array [[0, 1], [0, 1]]) >>>N.argsort a,None) usr lib64 python2.4 site packages numpy core fromnumeric.py in argsort a, axis, kind) 192 except AttributeError 193 return wrapit a, argsort , axis, kind) > 194 return argsort axis, kind) 195 196 def argmax a, axis=None) TypeError an integer is required","bug"
|
|
"6013","990","histogram reverse indicies Trac 392).
|
|
Original ticket http projects.scipy.org numpy ticket 392 on 2006 11 30 by trac user donovandak, assigned to unknown.
|
|
Are there any plans to add reverse indexing capability as found in IDL s histogram function) to numpy.histogram ?
|
|
This capability tells you, for each bin in the histogram, the actual indices of the data which fell into that bin.
|
|
It s actually more useful than it may sound... check out the reverse indicies section of the page below if you are unfamiliar with the concept [http www.dfanning.com tips histogram tutorial.html]","enhancement"
|
|
"6013","991","numpy 1.0 import fails on Red Hat 9 RH9) blas related) Trac 393).
|
|
Original ticket http projects.scipy.org numpy ticket 393 on 2006 12 01 by trac user twegener, assigned to unknown.
|
|
Importing numpy 1.0) fails in Red Hat 9 Linux blas 3.0 20).
|
|
Compiled and run on same machine.)
|
|
The error output is shown below Python 2.5 r25 51908, Oct 17 2006, 11 51 06) [GCC 3.2.2 20030222 Red Hat Linux 3.2.2 5)] on linux2 Type help , copyright , credits or license for more information.
|
|
>>> import numpy Traceback most recent call last) File <stdin> , line 1, in <module> File radstd python site packages i686 linux py2.5 numpy init .py , line 40, in <module> import linalg File radstd python site packages i686 linux py2.5 numpy linalg init .py , line 4, in <module> from linalg import File radstd python site packages i686 linux py2.5 numpy linalg linalg.py , line 25, in <module> from numpy.linalg import lapack lite ImportError usr lib libblas.so.3 undefined symbol e wsfe This happens with Python 2.5, 2.4 and 2.3.
|
|
It works fine with Fedora 3 blas 3.0 26.fc3) and Fedora 2 blas 3.0 24.1)","bug"
|
|
"6013","992","Memory leak in .argsort ) method Trac 394).
|
|
Original ticket http projects.scipy.org numpy ticket 394 on 2006 12 04 by FrancescAlted, assigned to unknown.
|
|
It seems that a memory leak in recent !NumPy 1.0.1 has slipped in.
|
|
The next code snipped reproduces the problem !python from numpy import N=1000 M=10000 Toggle to this for valgrind N=10 M=100 for i in xrange N) a = arange M) b = a.argsort ) and see evolution of memory consumption with an appropriate tool.
|
|
Running the above against valgrind, I ve got ==10942== 4,480 400 direct, 4,080 indirect) bytes in 10 blocks are definitely l ost in loss record 36 of 44 ==10942== at 0x401C38B malloc vg replace malloc.c 149) ==10942== by 0x45A414C array alloc arrayobject.c 6671) ==10942== by 0x45A075B PyArray NewFromDescr arrayobject.c 5323) ==10942== by 0x45C706D PyArray ArangeObj multiarraymodule.c 6476) ==10942== by 0x45C738E array arange multiarraymodule.c 6532) ==10942== by 0x80B954A PyEval EvalFrame in usr bin python2.4) ==10942== by 0x80BA6B4 PyEval EvalCodeEx in usr bin python2.4) ==10942== by 0x80BA718 PyEval EvalCode in usr bin python2.4) ==10942== by 0x80DD046 PyRun FileExFlags in usr bin python2.4) ==10942== by 0x80DD243 PyRun SimpleFileExFlags in usr bin python2.4) ==10942== by 0x8055BA7 Py Main in usr bin python2.4) ==10942== by 0x8055031 main in usr bin python2.4) So, it seems that the arrays created by .arange ) method don t got freed.
|
|
An inspection of the latests commits in trunk pointed directly to the r3466 changeset.
|
|
Below is a fix Index numpy core src multiarraymodule.c =================================================================== numpy core src multiarraymodule.c revision 3466) numpy core src multiarraymodule.c working copy) 2490,6 2490,7 Determine if we should use new algorithm or not if op2 >descr >f >argsort[which] != NULL) { Py DECREF op); return new argsort op2, axis, which); } This patch seems to behave well with both !NumPy and !PyTables tests.","bug"
|
|
"6013","993","GCC compiler warning Trac 395).
|
|
Original ticket http projects.scipy.org numpy ticket 395 on 2006 12 05 by alberts, assigned to unknown.
|
|
Recent changes to !NumPy probably in r3475 and r3476) have caused the following compiler warning when building with GCC 4.1.1 on Fedora Core 6 compile options Ibuild src.linux i686 2.4 numpy core src Inumpy core include Ibuild src.linux i686 2.4 numpy core Inumpy core src Inumpy core include I usr include python2.4 c gcc numpy core src multiarraymodule.c numpy core src multiarraymodule.c In function PyArray ArgSort numpy core src multiarraymodule.c 2494 warning assignment from incompatible pointer type numpy core src multiarraymodule.c 2496 warning return from incompatible pointer type","bug"
|
|
"6013","994","poly1d hangs Trac 396).
|
|
Original ticket http projects.scipy.org numpy ticket 396 on 2006 12 05 by trac user joris, assigned to unknown.
|
|
from numpy import p = poly1d [nan,nan,1], r=0) p.r The last command makes python hang, not even CTRL C helps.
|
|
In practice you don t give of course deliberately [nan,nan,1] as coefficients, but earlier in the code something may have gone wrong, so that the coefficients are not what is meant.
|
|
I would expect then an ArithmeticError of poly1d rather than hanging.
|
|
Versions Python 2.5 Numpy 1.0","bug"
|
|
"6013","995","Misspelled True in array.byteswap Trac 397).
|
|
Original ticket http projects.scipy.org numpy ticket 397 on 2006 12 10 by trac user ahonkela, assigned to unknown.
|
|
A call to array.byteswap in numpy.oldnumeric.array constructor defined in numpy oldnumeric compat.py) has an invalid argument TRUE instead of True.
|
|
The same bug also appears in the docstring for array.byteswap.
|
|
A patch to fix this is attached.","bug"
|
|
"6013","996","Compatibility loader for old Numeric pickles Trac 398).
|
|
Original ticket http projects.scipy.org numpy ticket 398 on 2006 12 10 by trac user ahonkela, assigned to unknown.
|
|
A code snippet illustrating loading pickled arrays saved with Numeric using either numpy or Numeric is attached.
|
|
Somehow incorporating this behaviour to regular numpy or at least documenting the procedure somehow would be rather useful.","enhancement"
|
|
"6013","997","Short notation typecode size) for integers is not present in sctypeNA dict Trac 399).
|
|
Original ticket http projects.scipy.org numpy ticket 399 on 2006 12 11 by FrancescAlted, assigned to unknown.
|
|
Hi, sctypeNA does have entries for booleans, floats, complex short notation In [27] numpy.sctypeNA[ b1 ] Out[27] Bool In [28] numpy.sctypeNA[ f4 ] Out[28] Float32 In [29] numpy.sctypeNA[ c8 ] Out[29] Complex32 but not for integers In [30] numpy.sctypeNA[ i4 ] exceptions.KeyError Traceback most recent call last) home faltet <ipython console> KeyError i4 In [31] numpy.sctypeNA[ u4 ] exceptions.KeyError Traceback most recent call last) home faltet <ipython console> KeyError u4","bug"
|
|
"6013","998","C API access to fft for C scipy extension ?
|
|
Trac 400).
|
|
Original ticket http projects.scipy.org numpy ticket 400 on 2006 12 12 by cournape, assigned to unknown.
|
|
Is there any plan to expose the fft used by numpy to C extension, so that other C extension to numpy could use it ?","enhancement"
|
|
"6013","999","numpy.roots ) does not return roots if they are all zero.
|
|
Trac 401).
|
|
Original ticket http projects.scipy.org numpy ticket 401 on 2006 12 12 by trac user DaleD, assigned to unknown.
|
|
Python 2.4.3, Enthought Edition 1.0.0, on Windows, numpy.version.version= 0.9.9.2706 If all of the roots of a polynomial are zero, then an empty array is returned from roots ).
|
|
Example >>> from numpy import >>> roots array [1.0,0,1.0])) array [ 0.
|
|
1.j, 0.
|
|
1.j]) correct >>> roots array [1.0,0,0])) array [], dtype=int32) either incorrect or inconvenient >>> roots array [1.0,1.0,0])) array [ 1., 0.])
|
|
returns the zero root okay here >>> Should either document this behavior yuck) or return all the zero roots cool).","bug"
|