Update matrix.cpp

src/gausskernel/dbmind/db4ai/executor/matrix.cpp

@@ -33,6 +33,11 @@ calculation, and for a square matrix, determinant calculation is also required.*
 
 #define MATRIX_LIMITED_OUTPUT 30
 
+
+/*Function: void matrix_ Init_ Random_ Gaussian
+Formal parameters: (Matrix * matrix, int rows, int columns, float8 mu, float8 sigma, int seed)
+Return: None
+Production random Gaussian matrix*/
 // using Box-Muller implementation
 void matrix_init_random_gaussian(Matrix *matrix, int rows, int columns, float8 mu, float8 sigma, int seed)
 {
@@ -60,12 +65,20 @@ void matrix_init_random_gaussian(Matrix *matrix, int rows, int columns, float8 m
     }
 }
 
+/*Function: matrix_ Init_ Kernel_ Gaussian
+Formal parameters: (int features, int components, float8 gamma, int seed, Matrix * weights, Matrix * offsets)
+Return: None
+Initialize a matrix using a specified number*/
 void matrix_init_kernel_gaussian(int features, int components, float8 gamma, int seed, Matrix *weights, Matrix *offsets)
 {
     matrix_init_random_gaussian(weights, features, components, 0.0, sqrt(2.0 * gamma), seed);
     matrix_init_random_uniform(offsets, components, 1, 0.0, 2.0 * M_PI, seed+1);
 }
 
+/*Function: matrix_ Transform_ Kernel_ Gaussian
+Formal parameters: (const Matrix * input, const Matrix * weights, const Matrix * offsets, Matrix * output)
+Return: None
+Matrix transpose*/
 void matrix_transform_kernel_gaussian(const Matrix *input, const Matrix *weights, const Matrix *offsets, Matrix *output)
 {
     int components = weights->columns;
@@ -91,6 +104,10 @@ void matrix_transform_kernel_gaussian(const Matrix *input, const Matrix *weights
     matrix_mult_scalar(output, sqrt(2.0 / components));
 }
 
+/*Function: matrix_ Init_ Random_ Uniform
+Formal parameters: (Matrix * matrix, int rows, int columns, float8 min, float8 max, int seed)
+Return: None
+Initializing a matrix using random floating-point numbers*/
 void matrix_init_random_uniform(Matrix *matrix, int rows, int columns, float8 min, float8 max, int seed)
 {
     Assert(min < max);
@@ -109,7 +126,10 @@ void matrix_init_random_uniform(Matrix *matrix, int rows, int columns, float8 mi
         *pd++ = min + range * u;
     }
 }
-
+/*Function: matrix_ Init_ Random_ Bernoulli
+Formal parameters: (Matrix * matrix, int rows, int columns, float8 p, float8 min, float8 max, int seed)
+Return: None
+Generate Random Bernoulli Matrix*/
 void matrix_init_random_bernoulli(Matrix *matrix, int rows, int columns, float8 p, float8 min, float8 max, int seed)
 {
     matrix_init(matrix, rows, columns);
@@ -126,6 +146,11 @@ void matrix_init_random_bernoulli(Matrix *matrix, int rows, int columns, float8
     }
 }
 
+/*Function: matrix_ Init_ Kernel_ Polynomial
+Formal parameters: (int features, int components, int degree, float8 coef0, int seed, Matrix * weights,
+Matrix * coefs)
+Return: int*
+Initialize a polynomial matrix using a specified number*/
 int *matrix_init_kernel_polynomial(int features, int components, int degree, float8 coef0, int seed, Matrix *weights,
                                    Matrix *coefs)
 {
@@ -160,6 +185,11 @@ int *matrix_init_kernel_polynomial(int features, int components, int degree, flo
     return pcomponents;
 }
 
+/*Function: matrix_ Transform_ Kernel_ Polynomial
+Formal parameters: (const Matrix * input, int ncomponents, int * components, const Matrix * weights,
+Const Matrix * coefficients, Matrix * output)
+Return: None
+Polynomial matrix transpose*/
 void matrix_transform_kernel_polynomial(const Matrix *input, int ncomponents, int *components, const Matrix *weights,
                                         const Matrix *coefficients, Matrix *output)
 {
@@ -185,6 +215,10 @@ void matrix_transform_kernel_polynomial(const Matrix *input, int ncomponents, in
     matrix_mult_scalar(output, sqrt(1.0 / output->rows));
 }
 
+/*Function: matrix_ Mult
+Formal parameters: (const Matrix * matrix1, const Matrix * matrix2, Matrix * result)
+Return value: None
+matrix multiplication*/
 void matrix_mult(const Matrix *matrix1, const Matrix *matrix2, Matrix *result)
 {
     Assert(matrix1 != nullptr);
@@ -213,6 +247,10 @@ void matrix_mult(const Matrix *matrix1, const Matrix *matrix2, Matrix *result)
     }
 }
 
+/*Function: matrix_ Print
+Formal parameters: (const Matrix * matrix, StringInfo buf, bool full)
+Return value: None
+Print Matrix*/
 void matrix_print(const Matrix *matrix, StringInfo buf, bool full)
 {
     Assert(matrix != nullptr);
@@ -253,6 +291,24 @@ void matrix_print(const Matrix *matrix, StringInfo buf, bool full)
     appendStringInfoChar(buf, ']');
 }
 
+/*Function: elog_ Matrix
+Formal parameters: (int level, const char * msg, const matrix * matrix)
+Return value: None
+Matrix error*/
+/*elog is an old mode that can be equivalent to the ereport mode. 
+You can see that it provides level and the error level is the same, 
+but it does not provide errcode. As mentioned earlier, the default 
+errcode is provided based on the severity level. Then the message 
+is passed through an auxiliary function errmsg_ Internal() goes to 
+show it, and the process is different from the errmsg in ereport mentioned
+earlier. errmsg() is set according to regional settings, such as it can be
+translated into the language of the corresponding country, such as Chinese.
+In fact, errmsg_ Internal() is a language that is not limited by translation and
+can automatically print out the original language.
+Why should we keep this old pattern? Because it is concise enough, when
+there are some internal errors, such as internal errors in the PG kernel, these
+errors are not actually displayed to the user and are not of interest to the user.
+This concise mode can be used for printing, which is very convenient and has been preserved.*/
 void elog_matrix(int elevel, const char *msg, const Matrix *matrix)
 {
     if (is_errmodule_enable(elevel, MOD_DB4AI)) {