diff --git a/match_program/src/main/java/com/ossean/match/lucene/LuceneIndex.java b/match_program/src/main/java/com/ossean/match/lucene/LuceneIndex.java index a381f3030..1b3d6658c 100644 --- a/match_program/src/main/java/com/ossean/match/lucene/LuceneIndex.java +++ b/match_program/src/main/java/com/ossean/match/lucene/LuceneIndex.java @@ -197,12 +197,12 @@ public class LuceneIndex { List memos = memoDao.getMemoInfo(startId, endId); for(RelativeMemo memo : memos) { lastMemoId = memo.getId(); - if (memo.getSubject() == null) { + if (memo.getTitle() == null) { continue; } Document doc = new Document(); doc.add(new StringField(memoIdFieldName, String.valueOf(lastMemoId), Store.YES)); - doc.add(new TextField(titleFieldName, memo.getSubject(), Store.NO)); + doc.add(new TextField(titleFieldName, memo.getTitle(), Store.NO)); String memoTagsString = memo.getTags(); if (memoTagsString == null) { memoTagsString = ""; @@ -253,12 +253,12 @@ public class LuceneIndex { int rsId = -1; for(RelativeMemo memo : memos) { rsId = memo.getId(); - if (memo.getSubject() == null) { + if (memo.getTitle() == null) { continue; } Document doc = new Document(); doc.add(new StringField(memoIdFieldName, String.valueOf(rsId), Store.YES)); - doc.add(new TextField(titleFieldName, memo.getSubject(), Store.NO)); + doc.add(new TextField(titleFieldName, memo.getTitle(), Store.NO)); String memoTagsString = memo.getTags(); if (memoTagsString == null) { memoTagsString = ""; diff --git a/match_program/src/main/java/com/ossean/match/matchprocess/Match.java b/match_program/src/main/java/com/ossean/match/matchprocess/Match.java index 5b3f8c4bc..01fde04d9 100644 --- a/match_program/src/main/java/com/ossean/match/matchprocess/Match.java +++ b/match_program/src/main/java/com/ossean/match/matchprocess/Match.java @@ -207,39 +207,46 @@ public class Match { * @param osp_id * @return */ - public static String getTargetTable(int osp_id){ + public static String getTargetTable(int ospId){ String targetTableName = ""; - if (osp_id < 500) { - targetTableName = "relative_memo_to_open_source_projects_1"; - } - else if (osp_id >= 500 && osp_id < 1000) { - targetTableName = "relative_memo_to_open_source_projects_2"; - } - else if (osp_id >= 1000 && osp_id < 1500) { - targetTableName = "relative_memo_to_open_source_projects_3"; - } - else if (osp_id >= 1500 && osp_id < 2000) { - targetTableName = "relative_memo_to_open_source_projects_4"; - } - else if (osp_id >= 2000 && osp_id < 3000) { - targetTableName = "relative_memo_to_open_source_projects_5"; - } - else if (osp_id >= 3000 && osp_id < 5000) { - targetTableName = "relative_memo_to_open_source_projects_6"; - } - else if (osp_id >= 5000 && osp_id < 7500) { - targetTableName = "relative_memo_to_open_source_projects_7"; - } - else if (osp_id >= 7500 && osp_id < 10000) { - targetTableName = "relative_memo_to_open_source_projects_8"; - } - else if (osp_id >= 10000 && osp_id < 310000) { - int a = 7 + osp_id/5000; - targetTableName = "relative_memo_to_open_source_projects_" + a; - } - else if (osp_id >= 310000) { + if(ospId >= 770000){ targetTableName = "relative_memo_to_open_source_projects_70"; } + else{ + int a = 1 + ospId/11000; + targetTableName = "relative_memo_to_open_source_projects_" + a; + } +// if (osp_id < 500) { +// targetTableName = "relative_memo_to_open_source_projects_1"; +// } +// else if (osp_id >= 500 && osp_id < 1000) { +// targetTableName = "relative_memo_to_open_source_projects_2"; +// } +// else if (osp_id >= 1000 && osp_id < 1500) { +// targetTableName = "relative_memo_to_open_source_projects_3"; +// } +// else if (osp_id >= 1500 && osp_id < 2000) { +// targetTableName = "relative_memo_to_open_source_projects_4"; +// } +// else if (osp_id >= 2000 && osp_id < 3000) { +// targetTableName = "relative_memo_to_open_source_projects_5"; +// } +// else if (osp_id >= 3000 && osp_id < 5000) { +// targetTableName = "relative_memo_to_open_source_projects_6"; +// } +// else if (osp_id >= 5000 && osp_id < 7500) { +// targetTableName = "relative_memo_to_open_source_projects_7"; +// } +// else if (osp_id >= 7500 && osp_id < 10000) { +// targetTableName = "relative_memo_to_open_source_projects_8"; +// } +// else if (osp_id >= 10000 && osp_id < 310000) { +// int a = 7 + osp_id/5000; +// targetTableName = "relative_memo_to_open_source_projects_" + a; +// } +// else if (osp_id >= 310000) { +// targetTableName = "relative_memo_to_open_source_projects_70"; +// } return targetTableName; //return "relative_memo_to_open_source_projects"; } @@ -268,13 +275,13 @@ public class Match { String memoTitle = ""; String memoTags = ""; memoId = memo.getId(); - if (memo.getSubject() == null || memo.getSubject().length() == 0) { + if (memo.getTitle() == null || memo.getTitle().length() == 0) { continue; } if (memo.getTags() != null && memo.getTags().length() > 0) { memoTags = memo.getTags().toLowerCase(); } - memoTitle = memo.getSubject().toLowerCase(); + memoTitle = memo.getTitle().toLowerCase(); if (memoTitle.length() == 0) { continue; } diff --git a/match_program/src/main/java/com/ossean/match/model/Project.java b/match_program/src/main/java/com/ossean/match/model/Project.java index 8e20e1c8b..d95beb43b 100644 --- a/match_program/src/main/java/com/ossean/match/model/Project.java +++ b/match_program/src/main/java/com/ossean/match/model/Project.java @@ -1,7 +1,5 @@ package com.ossean.match.model; -import java.util.HashMap; - public class Project { private int id; private String url; @@ -12,7 +10,6 @@ public class Project { private double rank_star = -1.0; private String tags; private String source; - private HashMap urls; @Override public String toString() { // TODO Auto-generated method stub @@ -43,9 +40,6 @@ public class Project { public void setSynonyms(String synonyms) { this.synonyms = synonyms; } - public void setUrls(HashMap urls) { - this.urls = urls; - } public int getRank(){ return rank; diff --git a/match_program/src/main/java/com/ossean/match/model/RelativeMemo.java b/match_program/src/main/java/com/ossean/match/model/RelativeMemo.java index 5ce559da1..3abdf856d 100644 --- a/match_program/src/main/java/com/ossean/match/model/RelativeMemo.java +++ b/match_program/src/main/java/com/ossean/match/model/RelativeMemo.java @@ -3,51 +3,32 @@ package com.ossean.match.model; public class RelativeMemo { private int id; - private int osp_id; - private String subject; + private String title; private String content; - private String author; - private int replies_num; - private int lock; - private int sticky; private String created_time; private String updated_time; - private String url; - private int view_num_crawled; - private int vote_up_num; - private int collection_num; - private String abstractText; - private String memo_type; - private String source; - private String category; - private int view_num_trustie; - private int author_id; - private int parent_id; - private int last_reply_id; - private int is_quote; - private String username; - private String userhomeurl; - private String crawled_time; - private String author_url; - private String url_md5; + private String type; private String tags; + private String source; + private String url; + private String url_md5; + private String author; + private String author_url; + private int view_num; + private int review_num; + private int view_num_ossean; + private String extracted_time; public int getId() { return id; } public void setId(int id) { this.id = id; } - public int getOsp_id() { - return osp_id; + public String getTitle() { + return title; } - public void setOsp_id(int osp_id) { - this.osp_id = osp_id; - } - public String getSubject() { - return subject; - } - public void setSubject(String subject) { - this.subject = subject; + public void setTitle(String title) { + this.title = title; } public String getContent() { return content; @@ -55,30 +36,6 @@ public class RelativeMemo { public void setContent(String content) { this.content = content; } - public String getAuthor() { - return author; - } - public void setAuthor(String author) { - this.author = author; - } - public int getReplies_num() { - return replies_num; - } - public void setReplies_num(int replies_num) { - this.replies_num = replies_num; - } - public int getLock() { - return lock; - } - public void setLock(int lock) { - this.lock = lock; - } - public int getSticky() { - return sticky; - } - public void setSticky(int sticky) { - this.sticky = sticky; - } public String getCreated_time() { return created_time; } @@ -91,113 +48,11 @@ public class RelativeMemo { public void setUpdated_time(String updated_time) { this.updated_time = updated_time; } - public String getUrl() { - return url; + public String getType() { + return type; } - public void setUrl(String url) { - this.url = url; - } - public int getView_num_crawled() { - return view_num_crawled; - } - public void setView_num_crawled(int view_num_crawled) { - this.view_num_crawled = view_num_crawled; - } - public int getVote_up_num() { - return vote_up_num; - } - public void setVote_up_num(int vote_up_num) { - this.vote_up_num = vote_up_num; - } - public int getCollection_num() { - return collection_num; - } - public void setCollection_num(int collection_num) { - this.collection_num = collection_num; - } - public String getAbstractText() { - return abstractText; - } - public void setAbstractText(String abstractText) { - this.abstractText = abstractText; - } - public String getMemo_type() { - return memo_type; - } - public void setMemo_type(String memo_type) { - this.memo_type = memo_type; - } - public String getSource() { - return source; - } - public void setSource(String source) { - this.source = source; - } - public String getCategory() { - return category; - } - public void setCategory(String category) { - this.category = category; - } - public int getView_num_trustie() { - return view_num_trustie; - } - public void setView_num_trustie(int view_num_trustie) { - this.view_num_trustie = view_num_trustie; - } - public int getAuthor_id() { - return author_id; - } - public void setAuthor_id(int author_id) { - this.author_id = author_id; - } - public int getParent_id() { - return parent_id; - } - public void setParent_id(int parent_id) { - this.parent_id = parent_id; - } - public int getLast_reply_id() { - return last_reply_id; - } - public void setLast_reply_id(int last_reply_id) { - this.last_reply_id = last_reply_id; - } - public int getIs_quote() { - return is_quote; - } - public void setIs_quote(int is_quote) { - this.is_quote = is_quote; - } - public String getUsername() { - return username; - } - public void setUsername(String username) { - this.username = username; - } - public String getUserhomeurl() { - return userhomeurl; - } - public void setUserhomeurl(String userhomeurl) { - this.userhomeurl = userhomeurl; - } - public String getCrawled_time() { - return crawled_time; - } - public void setCrawled_time(String crawled_time) { - this.crawled_time = crawled_time; - } - public String getAuthor_url() { - return author_url; - } - public void setAuthor_url(String author_url) { - this.author_url = author_url; - } - public String getUrl_md5() { - return url_md5; - } - public void setUrl_md5(String url_md5) { - this.url_md5 = url_md5; + public void setType(String type) { + this.type = type; } public String getTags() { return tags; @@ -205,5 +60,59 @@ public class RelativeMemo { public void setTags(String tags) { this.tags = tags; } - + public String getUrl() { + return url; + } + public void setUrl(String url) { + this.url = url; + } + public String getUrl_md5() { + return url_md5; + } + public void setUrl_md5(String url_md5) { + this.url_md5 = url_md5; + } + public String getAuthor() { + return author; + } + public void setAuthor(String author) { + this.author = author; + } + public String getAuthor_url() { + return author_url; + } + public void setAuthor_url(String author_url) { + this.author_url = author_url; + } + public String getSource() { + return source; + } + public void setSource(String source) { + this.source = source; + } + public int getView_num() { + return view_num; + } + public void setView_num(int view_num) { + this.view_num = view_num; + } + public int getReview_num() { + return review_num; + } + public void setReview_num(int review_num) { + this.review_num = review_num; + } + public int getView_num_ossean() { + return view_num_ossean; + } + public void setView_num_ossean(int view_num_ossean) { + this.view_num_ossean = view_num_ossean; + } + public String getExtracted_time() { + return extracted_time; + } + public void setExtracted_time(String extracted_time) { + this.extracted_time = extracted_time; + } + } diff --git a/match_program/src/main/java/com/ossean/match/utils/Normalizer.java b/match_program/src/main/java/com/ossean/match/utils/Normalizer.java index b23340635..95a8c6627 100644 --- a/match_program/src/main/java/com/ossean/match/utils/Normalizer.java +++ b/match_program/src/main/java/com/ossean/match/utils/Normalizer.java @@ -65,6 +65,8 @@ public class Normalizer { public static void main(String []args) { String content = "大声道Corvette123 *&%9&^&%&^%(*_asd's_12 L'Aur--ore, 1766-1775"; + int a = 1 + 11000/11000; + String targetTableName = "relative_memo_to_open_source_projects_" + a; //System.out.println(StringFilter(content)); String result = content.replaceAll("[^0-9a-zA-Z]"," "); System.out.println(result);