third_codes/cppjieba
1
0
Fork 0
mirror of https://github.com/yanyiwu/cppjieba.git synced 2025-07-18 00:00:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1 from yanyiwu/patch-1

Browse Source 
Update TextRankExtractor.hpp: use yanyiwu's correction
This commit is contained in:
questionfish 2016-05-03 21:46:01 +08:00
commit 04c176de08
3 changed files with 52 additions and 49 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

82
include/cppjieba/TextRankExtractor.hpp
View File

@ -1,4 +1,4 @@
#ifndef CPPJIEBA_TEXTRANK_EXTRACTOR_H #ifndef CPPJIEBA_TEXTRANK_EXTRACTOR_H
#define CPPJIEBA_TEXTRANK_EXTRACTOR_H #define CPPJIEBA_TEXTRANK_EXTRACTOR_H
#include <cmath> #include <cmath>
@ -12,15 +12,18 @@ namespace cppjieba {
public: public:
typedef struct _Word {string word;vector<size_t> offsets;double weight;} Word; // struct Word typedef struct _Word {string word;vector<size_t> offsets;double weight;} Word; // struct Word
private: private:
typedef std::unordered_map<string,Word> WordMap; typedef std::map<string,Word> WordMap;
class WordGraph{ class WordGraph{
private: private:
typedef double Score; typedef double Score;
typedef string Node; typedef string Node;
typedef std::unordered_set<Node> NodeSet; typedef std::set<Node> NodeSet;
typedef std::unordered_map<Node,double> Edges;
typedef std::unordered_map<Node,Edges> Graph; typedef std::map<Node,double> Edges;
typedef std::map<Node,Edges> Graph;
//typedef std::unordered_map<Node,double> Edges;
//typedef std::unordered_map<Node,Edges> Graph;
double d; double d;
Graph graph; Graph graph;
@ -82,40 +85,40 @@ namespace cppjieba {
} }
}; };
public: public:
TextRankExtractor(const string& dictPath, TextRankExtractor(const string& dictPath,
const string& hmmFilePath, const string& hmmFilePath,
const string& stopWordPath, const string& stopWordPath,
const string& userDict = "") const string& userDict = "")
: segment_(dictPath, hmmFilePath, userDict) { : segment_(dictPath, hmmFilePath, userDict) {
LoadStopWordDict(stopWordPath); LoadStopWordDict(stopWordPath);
} }
TextRankExtractor(const DictTrie* dictTrie, TextRankExtractor(const DictTrie* dictTrie,
const HMMModel* model, const HMMModel* model,
const string& stopWordPath) const string& stopWordPath)
: segment_(dictTrie, model) { : segment_(dictTrie, model) {
LoadStopWordDict(stopWordPath); LoadStopWordDict(stopWordPath);
} }
TextRankExtractor(const Jieba& jieba, const string& stopWordPath) : segment_(jieba.GetDictTrie(), jieba.GetHMMModel()) { TextRankExtractor(const Jieba& jieba, const string& stopWordPath) : segment_(jieba.GetDictTrie(), jieba.GetHMMModel()) {
LoadStopWordDict(stopWordPath); LoadStopWordDict(stopWordPath);
} }
~TextRankExtractor() { ~TextRankExtractor() {
} }
void Extract(const string& sentence, vector<string>& keywords, size_t topN) const { void Extract(const string& sentence, vector<string>& keywords, size_t topN) const {
vector<Word> topWords; vector<Word> topWords;
Extract(sentence, topWords, topN); Extract(sentence, topWords, topN);
for (size_t i = 0; i < topWords.size(); i++) { for (size_t i = 0; i < topWords.size(); i++) {
keywords.push_back(topWords[i].word); keywords.push_back(topWords[i].word);
} }
} }
void Extract(const string& sentence, vector<pair<string, double> >& keywords, size_t topN) const { void Extract(const string& sentence, vector<pair<string, double> >& keywords, size_t topN) const {
vector<Word> topWords; vector<Word> topWords;
Extract(sentence, topWords, topN); Extract(sentence, topWords, topN);
for (size_t i = 0; i < topWords.size(); i++) { for (size_t i = 0; i < topWords.size(); i++) {
keywords.push_back(pair<string, double>(topWords[i].word, topWords[i].weight)); keywords.push_back(pair<string, double>(topWords[i].word, topWords[i].weight));
} }
} }
void Extract(const string& sentence, vector<Word>& keywords, size_t topN, size_t span=5,size_t rankTime=10) const { void Extract(const string& sentence, vector<Word>& keywords, size_t topN, size_t span=5,size_t rankTime=10) const {
@ -176,13 +179,6 @@ namespace cppjieba {
return false; return false;
} }
static void sortMapValue(WordMap &map,vector<Word>& result,size_t topN){
for(auto i=map.begin();i!=map.end();i++){
result.push_back(i->second);
}
partial_sort(result.begin(),result.begin()+topN,result.end(),Compare);
}
static bool Compare(const Word &x,const Word &y){ static bool Compare(const Word &x,const Word &y){
return x.weight > y.weight; return x.weight > y.weight;
} }
@ -190,9 +186,9 @@ namespace cppjieba {
MixSegment segment_; MixSegment segment_;
unordered_set<string> stopWords_; unordered_set<string> stopWords_;
}; };
inline ostream& operator << (ostream& os, const TextRankExtractor::Word& word) { inline ostream& operator << (ostream& os, const TextRankExtractor::Word& word) {
return os << word.word << '|' << word.offsets << '|' << word.weight; return os << word.word << '|' << word.offsets << '|' << word.weight;
} }
} // namespace cppjieba } // namespace cppjieba

1
test/unittest/CMakeLists.txt
View File

@ -13,6 +13,7 @@ ADD_EXECUTABLE(test.run
pos_tagger_test.cpp pos_tagger_test.cpp
jieba_test.cpp jieba_test.cpp
pre_filter_test.cpp pre_filter_test.cpp
textrank_test.cpp
) )
TARGET_LINK_LIBRARIES(test.run gtest pthread) TARGET_LINK_LIBRARIES(test.run gtest pthread)

18
test/unittest/textrank_test.cpp
View File

@ -24,14 +24,16 @@ TEST(TextRankExtractorTest, Test1) {
vector<pair<string, double> > words; vector<pair<string, double> > words;
Extractor.Extract(s, words, topN); Extractor.Extract(s, words, topN);
res << words; res << words;
ASSERT_EQ(res, "[\"世界:1\", \"你好:0.514286\"]"); //ASSERT_EQ(res, "[\"世界:1\", \"你好:0.514286\"]");
ASSERT_EQ(res, "[\"\xE4\xB8\x96\xE7\x95\x8C:1\", \"\xE4\xBD\xA0\xE5\xA5\xBD:0.519787\"]");
} }
{ {
vector<TextRankExtractor::Word> words; vector<TextRankExtractor::Word> words;
Extractor.Extract(s, words, topN); Extractor.Extract(s, words, topN);
res << words; res << words;
ASSERT_EQ(res, "[\"世界|[\"6\", \"12\"]|1\", \"你好|[\"0\"]|0.514286\"]"); //ASSERT_EQ(res, "[\"世界|[\"6\", \"12\"]|1\", \"你好|[\"0\"]|0.514286\"]");
ASSERT_EQ(res, "[\"\xE4\xB8\x96\xE7\x95\x8C|[\"6\", \"12\"]|1\", \"\xE4\xBD\xA0\xE5\xA5\xBD|[\"0\"]|0.519787\"]");
} }
} }
@ -42,7 +44,8 @@ TEST(TextRankExtractorTest, Test1) {
size_t topN = 5; size_t topN = 5;
Extractor.Extract(s, wordweights, topN); Extractor.Extract(s, wordweights, topN);
res << wordweights; res << wordweights;
ASSERT_EQ(res, "[\"\xE4\xB8\x93\xE4\xB8\x9A|[\"36\"]|1\", \"CEO|[\"94\"]|0.953149\", \"\xE6\x89\x8B\xE6\x89\xB6\xE6\x8B\x96\xE6\x8B\x89\xE6\x9C\xBA|[\"21\"]|0.794203\", \"\xE5\xBD\x93\xE4\xB8\x8A|[\"87\"]|0.78716\", \"\xE8\xB5\xB0\xE4\xB8\x8A|[\"100\"]|0.767636\"]"); ASSERT_EQ(res, "[\"\xE4\xB8\x93\xE4\xB8\x9A|[\"36\"]|1\", \"CEO|[\"94\"]|0.95375\", \"\xE6\x89\x8B\xE6\x89\xB6\xE6\x8B\x96\xE6\x8B\x89\xE6\x9C\xBA|[\"21\"]|0.801701\", \"\xE5\xBD\x93\xE4\xB8\x8A|[\"87\"]|0.798968\", \"\xE8\xB5\xB0\xE4\xB8\x8A|[\"100\"]|0.775505\"]");
// ASSERT_EQ(res, "[\"\xE4\xB8\x93\xE4\xB8\x9A|[\"36\"]|1\", \"CEO|[\"94\"]|0.953149\", \"\xE6\x89\x8B\xE6\x89\xB6\xE6\x8B\x96\xE6\x8B\x89\xE6\x9C\xBA|[\"21\"]|0.794203\", \"\xE5\xBD\x93\xE4\xB8\x8A|[\"87\"]|0.78716\", \"\xE8\xB5\xB0\xE4\xB8\x8A|[\"100\"]|0.767636\"]");
} }
{ {
@ -52,7 +55,8 @@ TEST(TextRankExtractorTest, Test1) {
size_t topN = 5; size_t topN = 5;
Extractor.Extract(s, wordweights, topN); Extractor.Extract(s, wordweights, topN);
res << wordweights; res << wordweights;
ASSERT_EQ(res, "[\"iPhone6|[\"6\"]|1\", \"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|0.996126\"]"); ASSERT_EQ(res, "[\"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|1\", \"iPhone6|[\"6\"]|0.996126\"]");
//ASSERT_EQ(res, "[\"iPhone6|[\"6\"]|1\", \"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|0.996126\"]");
} }
} }
@ -70,7 +74,8 @@ TEST(TextRankExtractorTest, Test2) {
size_t topN = 5; size_t topN = 5;
Extractor.Extract(s, wordweights, topN); Extractor.Extract(s, wordweights, topN);
res << wordweights; res << wordweights;
ASSERT_EQ(res, "[\"\xE4\xBC\x98\xE7\xA7\x80|[\"6\"]|1\", \"\xE6\xAF\x95\xE4\xB8\x9A\xE7\x94\x9F|[\"12\"]|0.996685\", \"\xE8\x93\x9D\xE7\xBF\x94|[\"0\"]|0.992994\"]"); ASSERT_EQ(res, "[\"蓝翔|[\"0\"]|1\", \"毕业生|[\"12\"]|0.996685\", \"优秀|[\"6\"]|0.992994\"]");
//ASSERT_EQ(res, "[\"\xE4\xBC\x98\xE7\xA7\x80|[\"6\"]|1\", \"\xE6\xAF\x95\xE4\xB8\x9A\xE7\x94\x9F|[\"12\"]|0.996685\", \"\xE8\x93\x9D\xE7\xBF\x94|[\"0\"]|0.992994\"]");
} }
{ {
@ -80,6 +85,7 @@ TEST(TextRankExtractorTest, Test2) {
size_t topN = 5; size_t topN = 5;
Extractor.Extract(s, wordweights, topN); Extractor.Extract(s, wordweights, topN);
res << wordweights; res << wordweights;
ASSERT_EQ(res, "[\"iPhone6|[\"6\"]|1\", \"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|0.996126\"]"); //ASSERT_EQ(res, "[\"iPhone6|[\"6\"]|1\", \"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|0.996126\"]");
ASSERT_EQ(res, "[\"\xE4\xB8\x80\xE9\x83\xA8|[\"0\"]|1\", \"iPhone6|[\"6\"]|0.996126\"]");
} }
} }