merge upstream

This commit is contained in:
aholic 2014-01-12 18:12:22 +08:00
commit 680399efdc
54 changed files with 259357 additions and 278 deletions

1
.gitignore vendored
View File

@ -14,3 +14,4 @@ prior.gbk
tmp
t.*
*.pid
build

View File

@ -2,16 +2,22 @@ PROJECT(CPPJIEBA)
CMAKE_MINIMUM_REQUIRED (VERSION 2.8)
SET(CMAKE_INSTALL_PREFIX /usr)
ADD_DEFINITIONS(-std=c++0x -O3)
if (CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT)
set (CMAKE_INSTALL_PREFIX "/usr" CACHE PATH "default install path" FORCE )
endif()
ADD_DEFINITIONS(-std=c++0x -O3 -Wall)
IF (DEFINED ENC)
ADD_DEFINITIONS(-DCPPJIEBA_${ENC})
ENDIF()
#ADD_DEFINITIONS(-DNO_FILTER)
ADD_SUBDIRECTORY(src)
ADD_SUBDIRECTORY(dicts)
ADD_SUBDIRECTORY(scripts)
ADD_SUBDIRECTORY(dict)
if (!${APPLE})
ADD_SUBDIRECTORY(script)
ADD_SUBDIRECTORY(conf)
endif()
ADD_SUBDIRECTORY(test)
ENABLE_TESTING()

View File

@ -28,7 +28,7 @@ sudo make install
#### 测试
```sh
cd build && ./test/segment.demo
make test
```
### 启动服务
@ -165,7 +165,7 @@ MPSegment.hpp
HMMSegment.hpp
是根据HMM模型来进行分词主要算法思路是根据(B,E,M,S)四个状态来代表每个字的隐藏状态。
HMM模型由dicts/下面的`hmm_model.utf8`提供。
HMM模型由dict/下面的`hmm_model.utf8`提供。
分词算法即viterbi算法。
FullSegment.hpp
@ -173,7 +173,7 @@ FullSegment.hpp
#### TransCode模块
TransCode.cpp/TransCode.h 负责转换编码类型将utf8和gbk转换成`uint16_t`类型,也负责逆转换。
TransCode.hpp 负责转换编码类型将utf8和gbk转换成`uint16_t`类型,也负责逆转换。
### src/Husky

View File

@ -10,10 +10,10 @@ thread_num=4
daemonize=true
#dict path
dict_path=/usr/share/CppJieba/dicts/jieba.dict.utf8
dict_path=/usr/share/CppJieba/dict/jieba.dict.utf8
#model path
model_path=/usr/share/CppJieba/dicts/hmm_model.utf8
model_path=/usr/share/CppJieba/dict/hmm_model.utf8
#pid file
pid_file=/var/run/CppJieba/cjserver.pid

View File

@ -1 +1 @@
INSTALL(FILES hmm_model.utf8 jieba.dict.utf8 DESTINATION share/CppJieba/dicts)
INSTALL(FILES hmm_model.utf8 jieba.dict.utf8 DESTINATION share/CppJieba/dict)

258826
dict/idf.utf8 Normal file

File diff suppressed because it is too large Load Diff

5
script/cjseg.sh Executable file
View File

@ -0,0 +1,5 @@
if [ $# -lt 1 ]; then
echo "usage: $0 <file>"
exit 1
fi
cjsegment --dictpath /usr/share/CppJieba/dict/jieba.dict.utf8 --modelpath /usr/share/CppJieba/dict/hmm_model.utf8 $1

View File

@ -1,5 +0,0 @@
if [ $# -lt 1 ]; then
echo "usage: $0 <file>"
exit 1
fi
cjsegment --dictpath /usr/share/CppJieba/dicts/jieba.dict.utf8 --modelpath /usr/share/CppJieba/dicts/hmm_model.utf8 $1

View File

@ -1,13 +1,15 @@
SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR}/bin)
SET(LIBRARY_OUTPUT_PATH ${PROJECT_BINARY_DIR}/lib)
INCLUDE_DIRECTORIES(${PROJECT_SOURCE_DIR}/src)
ADD_EXECUTABLE(cjsegment segment.cpp)
ADD_EXECUTABLE(cjserver server.cpp)
TARGET_LINK_LIBRARIES(cjserver pthread)
INSTALL(TARGETS cjsegment RUNTIME DESTINATION bin)
INSTALL(TARGETS cjserver RUNTIME DESTINATION bin)
INSTALL(FILES ChineseFilter.hpp HMMSegment.hpp MPSegment.hpp Trie.hpp ISegment.hpp MixSegment.hpp SegmentBase.hpp TransCode.hpp DESTINATION include/CppJieba)
INSTALL(FILES HMMSegment.hpp MPSegment.hpp Trie.hpp TrieManager.hpp ISegment.hpp MixSegment.hpp SegmentBase.hpp TransCode.hpp KeywordExtractor.hpp DESTINATION include/CppJieba)
ADD_SUBDIRECTORY(Husky)
ADD_SUBDIRECTORY(Limonp)

View File

@ -1,49 +0,0 @@
#ifndef CPPJIEBA_CHINESEFILTER_H
#define CPPJIEBA_CHINESEFILTER_H
#include "TransCode.hpp"
namespace CppJieba
{
/*
* if char is ascii, count the ascii string's length and return 0;
* else count the nonascii string's length and return 1;
* if errors, return -1;
* */
inline int filterAscii(const char* str, uint len, uint& resLen)
{
if(!str || !len)
{
return -1;
}
char x = 0x80;
int resFlag = (str[0] & x ? 1 : 0);
resLen = 0;
if(!resFlag)
{
while(resLen < len && !(str[resLen] & x))
{
resLen ++;
}
}
else
{
while(resLen < len && (str[resLen] & x))
{
#ifdef CPPJIEBA_GBK
resLen += 2;
#else
resLen ++;
#endif
}
}
if(resLen > len)
{
return -1;
}
return resFlag;
}
}
#endif

View File

@ -71,7 +71,7 @@ namespace CppJieba
for (vector<pair<uint, const TrieNodeInfo*> >::const_iterator itr = tRes.begin(); itr != tRes.end(); itr++)
{
wordLen = itr->second->word.size();
if (wordLen >= 2 || tRes.size() == 1 && maxIdx <= uIdx)
if (wordLen >= 2 || (tRes.size() == 1 && maxIdx <= uIdx))
{
res.push_back(itr->second->word);
}

View File

@ -46,6 +46,7 @@ namespace CppJieba
{
if(_getInitFlag())
{
LogError("inited already.");
return false;
}
memset(_startProb, 0, sizeof(_startProb));
@ -58,7 +59,13 @@ namespace CppJieba
_emitProbVec.push_back(&_emitProbE);
_emitProbVec.push_back(&_emitProbM);
_emitProbVec.push_back(&_emitProbS);
return _setInitFlag(_loadModel(filePath.c_str()));
if(!_setInitFlag(_loadModel(filePath.c_str())))
{
LogError("_loadModel(%s) failed.", filePath.c_str());
return false;
}
LogInfo("HMMSegment init(%s) ok.", filePath.c_str());
return true;
}
public:
using SegmentBase::cut;
@ -198,7 +205,7 @@ namespace CppJieba
}
bool _loadModel(const char* const filePath)
{
LogInfo("loadModel [%s] start ...", filePath);
LogDebug("loadModel [%s] start ...", filePath);
ifstream ifile(filePath);
string line;
vector<string> tmp;
@ -208,7 +215,7 @@ namespace CppJieba
{
return false;
}
splitStr(line, tmp, " ");
split(line, tmp, " ");
if(tmp.size() != STATUS_SUM)
{
LogError("start_p illegal");
@ -227,7 +234,7 @@ namespace CppJieba
{
return false;
}
splitStr(line, tmp, " ");
split(line, tmp, " ");
if(tmp.size() != STATUS_SUM)
{
LogError("trans_p illegal");
@ -264,7 +271,7 @@ namespace CppJieba
return false;
}
LogInfo("loadModel [%s] end.", filePath);
LogDebug("loadModel [%s] end.", filePath);
return true;
}
@ -277,7 +284,7 @@ namespace CppJieba
{
continue;
}
if(strStartsWith(line, "#"))
if(startsWith(line, "#"))
{
continue;
}
@ -293,10 +300,10 @@ namespace CppJieba
}
vector<string> tmp, tmp2;
uint16_t unico = 0;
splitStr(line, tmp, ",");
split(line, tmp, ",");
for(uint i = 0; i < tmp.size(); i++)
{
splitStr(tmp[i], tmp2, ":");
split(tmp[i], tmp2, ":");
if(2 != tmp2.size())
{
LogError("_emitProb illegal.");

View File

@ -1,4 +1,5 @@
SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR}/bin)
SET(LIBRARY_OUTPUT_PATH ${PROJECT_BINARY_DIR}/lib)
INSTALL(FILES HttpReqInfo.hpp ServerFrame.hpp ThreadManager.hpp DESTINATION include/CppJieba/Husky)
FILE(GLOB SRCS *.hpp)
INSTALL(FILES ${SRCS} DESTINATION include/CppJieba/Husky)

View File

@ -3,7 +3,7 @@
#include <iostream>
#include <string>
#include "../Limonp/logger.hpp"
#include "Limonp/logger.hpp"
namespace Husky
{
@ -88,7 +88,7 @@ namespace Husky
}
string firstline(headerStr, lpos, rpos - lpos);
trim(firstline);
if(!splitStr(firstline, buf, " ") || 3 != buf.size())
if(!split(firstline, buf, " ") || 3 != buf.size())
{
LogFatal("parse header first line failed.");
return false;

View File

@ -29,7 +29,7 @@ namespace Husky
using namespace Limonp;
typedef int SOCKET;
const struct timeval SOCKET_TIMEOUT = {2, 0};
const char* const RESPONSE_FORMAT = "HTTP/1.1 200 OK\r\nConnection: close\r\nServer: FrameServer/1.0.0\r\nContent-Type: text/json; charset=%s\r\nContent-Length: %d\r\n\r\n";
const char* const RESPONSE_FORMAT = "HTTP/1.1 200 OK\r\nConnection: close\r\nServer: HuskyServer/1.0.0\r\nContent-Type: text/json; charset=%s\r\nContent-Length: %d\r\n\r\n";
const char* const RESPONSE_CHARSET_UTF8 = "UTF-8";
const char* const RESPONSE_CHARSET_GB2312 = "GB2312";
const char* const CLIENT_IP_K = "CLIENT_IP";
@ -53,13 +53,13 @@ namespace Husky
bool * pShutdown;
};
class ServerFrame//: public IWorkHandler
class HuskyServer
{
private:
pthread_mutex_t m_pmAccept;
bool m_bShutdown;
public:
ServerFrame(unsigned nPort, unsigned nThreadCount, IRequestHandler* pHandler)
explicit HuskyServer(unsigned nPort, unsigned nThreadCount, IRequestHandler* pHandler)
{
m_bShutdown = false;
m_nLsnPort = nPort;
@ -68,7 +68,7 @@ namespace Husky
assert(pHandler);
pthread_mutex_init(&m_pmAccept,NULL);
};
virtual ~ServerFrame(){pthread_mutex_destroy(&m_pmAccept);};
virtual ~HuskyServer(){pthread_mutex_destroy(&m_pmAccept);};
virtual bool init()
{
@ -292,8 +292,6 @@ namespace Husky
u_short m_nThreadCount;
SOCKET m_lsnSock;
IRequestHandler *m_pHandler;
//static bool m_bShutdown;
//static pthread_mutex_t m_pmAccept;
};

View File

@ -5,8 +5,6 @@
#include <vector>
#include <map>
#define INFINITE 0
namespace Husky
{
using namespace std;
@ -14,48 +12,44 @@ namespace Husky
class ThreadManager
{
private:
typedef int HANDLE;
typedef int DWORD;
typedef pthread_t HANDLE;
typedef void *(* PThreadFunc)(void* param);
public:
ThreadManager(){;}
~ThreadManager(){}
unsigned int HandleCount(){return m_vecHandle.size();}
size_t HandleCount(){return _handles.size();}
void clear()
{
m_vecHandle.clear();
_handles.clear();
}
HANDLE CreateThread( PThreadFunc pFunc,void *pPara)
int CreateThread( PThreadFunc pFunc,void *pPara)
{
pthread_t pt;
int nErrorCode=pthread_create(&pt,NULL,pFunc,pPara);
if(nErrorCode!=0)
int nErrorCode = pthread_create(&pt,NULL,pFunc,pPara);
if(nErrorCode != 0)
return nErrorCode;
m_vecHandle.push_back(pt); //加入线程列表 为WaitForMultipleObjects准备
_handles.push_back(pt);
return nErrorCode;
}
//hThread (thread handler) : 为0时为默认最后一个加入管理器的线程句柄
//dwMilliseconds等待时间 : 单位毫秒,默认值无穷时间
//return value : -1句柄无效其他值 WaitForSingleObject函数的返回值
DWORD Wait(HANDLE hThread=0,DWORD dwMilliseconds=INFINITE )
int Wait(HANDLE hThread = 0)
{
if( hThread==0)//最后一个加入的线程
if( hThread == 0)//the last handle
{
if(!m_vecHandle.empty())
if(!_handles.empty())
{
return pthread_join(m_vecHandle.back(),NULL);
return pthread_join(_handles.back(),NULL);
}
else
return -1;
}
else
{
if (find(m_vecHandle.begin(),m_vecHandle.end(),hThread)==m_vecHandle.end())//不存在此句柄
if (find(_handles.begin(),_handles.end(),hThread) == _handles.end())
{
return -1;
}
@ -65,31 +59,26 @@ namespace Husky
}
//等待所有线程执行完毕
//bWaitAll是否所有线程 : 默认值1等待所有线程,0有任何线程结束此函数返回
//dwMilliseconds : 单位毫秒,默认值无穷时间
//return value : -1没有任何句柄其他值 WaitForMultipleObjects函数的返回值
DWORD WaitMultipleThread( bool bWaitAll=1,DWORD dwMilliseconds=INFINITE)
int WaitMultipleThread()
{
if (m_vecHandle.empty())
if (_handles.empty())
return -1;
int nErrorcode;
for (uint i=0;i<m_vecHandle.size();++i)
for (uint i = 0; i < _handles.size(); i++)
{
nErrorcode=pthread_join(m_vecHandle[i], NULL);
if (nErrorcode!=0)
nErrorcode = pthread_join(_handles[i], NULL);
if (nErrorcode != 0)
return nErrorcode;
}
return 0;
}
private:
vector<pthread_t> m_vecHandle;
vector<pthread_t> _handles;
private:
ThreadManager(const ThreadManager&){;}// copy forbidden
void operator=(const ThreadManager &){}// copy forbidden
void operator = (const ThreadManager &){}// copy forbidden
};
}

129
src/KeywordExtractor.hpp Normal file
View File

@ -0,0 +1,129 @@
#ifndef CPPJIEBA_KEYWORD_EXTRACTOR_H
#define CPPJIEBA_KEYWORD_EXTRACTOR_H
#include "MPSegment.hpp"
#include <cmath>
#define MIN(X,Y) ((X) < (Y) ? (X) : (Y))
namespace CppJieba
{
using namespace Limonp;
//struct KeyWordInfo
//{
// string word;
// double tfidf;
//};
//inline ostream& operator << (ostream& os, const KeyWordInfo & keyword)
//{
// return os << keyword.word << "," << keyword.idf;
//}
class KeywordExtractor
{
private:
MPSegment _segment;
private:
unordered_map<string, double> _idfMap;
protected:
bool _isInited;
bool _getInitFlag()const{return _isInited;};
bool _setInitFlag(bool flag){return _isInited = flag;};
public:
operator bool(){return _getInitFlag();};
public:
KeywordExtractor(){_setInitFlag(false);};
explicit KeywordExtractor(const string& dictPath, const string& idfPath){_setInitFlag(init(dictPath, idfPath));};
~KeywordExtractor(){};
public:
bool init(const string& dictPath, const string& idfPath)
{
ifstream ifs(idfPath.c_str());
if(!ifs)
{
LogError("open %s failed.", idfPath.c_str());
return false;
}
string line ;
vector<string> buf;
for(uint lineno = 0; getline(ifs, line); lineno++)
{
buf.clear();
if(line.empty())
{
LogError("line[%d] empty. skipped.", lineno);
continue;
}
if(!split(line, buf, " ") || buf.size() != 2)
{
LogError("line %d [%s] illegal. skipped.", lineno, line.c_str());
continue;
}
_idfMap[buf[0]] = atof(buf[1].c_str());
}
return _setInitFlag(_segment.init(dictPath));
};
public:
bool extract(const string& str, vector<string>& keywords, uint topN) const
{
assert(_getInitFlag());
vector<pair<string, double> > topWords;
if(!extract(str, topWords, topN))
{
return false;
}
for(uint i = 0; i < topWords.size(); i++)
{
keywords.push_back(topWords[i].first);
}
return true;
}
bool extract(const string& str, vector<pair<string, double> >& keywords, uint topN) const
{
vector<string> words;
if(!_segment.cut(str, words))
{
LogError("segment cut(%s) failed.", str.c_str());
return false;
}
unordered_map<string, double> wordmap;
for(uint i = 0; i < words.size(); i ++)
{
wordmap[ words[i] ] += 1.0;
}
for(unordered_map<string, double>::iterator itr = wordmap.begin(); itr != wordmap.end();)
{
unordered_map<string, double>::const_iterator cit = _idfMap.find(itr->first);
if(cit != _idfMap.end())
{
itr->second *= cit->second;
itr ++;
}
else
{
itr = wordmap.erase(itr);
}
}
keywords.resize(MIN(topN, wordmap.size()));
partial_sort_copy(wordmap.begin(), wordmap.end(), keywords.begin(), keywords.end(), _cmp);
return true;
}
private:
static bool _cmp(const pair<string, uint>& lhs, const pair<string, uint>& rhs)
{
return lhs.second > rhs.second;
}
};
}
#endif

View File

@ -22,9 +22,9 @@ namespace Limonp
for(int i = 0; i < argc; i++)
{
if(strStartsWith(argv[i], "-"))
if(startsWith(argv[i], "-"))
{
if(i + 1 < argc && !strStartsWith(argv[i + 1], "-"))
if(i + 1 < argc && !startsWith(argv[i + 1], "-"))
{
_mpss[argv[i]] = argv[i+1];
i++;

View File

@ -18,7 +18,17 @@ namespace Limonp
class Config
{
public:
bool loadFile(const char * const filePath)
Config(const char * const filePath)
{
_loadFile(filePath);
}
public:
operator bool ()
{
return !_map.empty();
}
private:
bool _loadFile(const char * const filePath)
{
ifstream ifs(filePath);
if(!ifs)
@ -33,12 +43,12 @@ namespace Limonp
{
lineno ++;
trim(line);
if(line.empty() || strStartsWith(line, "#"))
if(line.empty() || startsWith(line, "#"))
{
continue;
}
vecBuf.clear();
if(!splitStr(line, vecBuf, "=") || 2 != vecBuf.size())
if(!split(line, vecBuf, "=") || 2 != vecBuf.size())
{
LogFatal("line[%d:%s] is illegal.", lineno, line.c_str());
return false;
@ -57,6 +67,7 @@ namespace Limonp
ifs.close();
return true;
}
public:
bool get(const string& key, string& value) const
{
map<string, string>::const_iterator it = _map.find(key);
@ -73,7 +84,7 @@ namespace Limonp
friend ostream& operator << (ostream& os, const Config& config);
};
ostream& operator << (ostream& os, const Config& config)
inline ostream& operator << (ostream& os, const Config& config)
{
return os << config._map;
}

View File

@ -11,6 +11,7 @@
#include <string>
#include <stdio.h>
#include <stdarg.h>
#include <cassert>
#include "io_functs.hpp"
#include "str_functs.hpp"
@ -23,6 +24,7 @@
#define LogFatal(fmt, ...) Logger::LoggingF(LL_FATAL, FILE_BASENAME, __LINE__, fmt, ## __VA_ARGS__)
namespace Limonp
{
using namespace std;
@ -36,16 +38,11 @@ namespace Limonp
public:
static bool Logging(uint level, const string& msg, const char* fileName, int lineNo)
{
if(level > LL_FATAL)
{
cerr<<"level's value is out of range"<<endl;
return false;
}
assert(level <= LL_FATAL);
char buf[CSTR_BUFFER_SIZE];
time_t timeNow;
time(&timeNow);
size_t ret = strftime(buf, sizeof(buf), LOG_TIME_FORMAT, localtime(&timeNow));
if(0 == ret)
if(!strftime(buf, sizeof(buf), LOG_TIME_FORMAT, localtime(&timeNow)))
{
fprintf(stderr, "stftime failed.\n");
return false;
@ -55,6 +52,9 @@ namespace Limonp
}
static bool LoggingF(uint level, const char* fileName, int lineNo, const string& fmt, ...)
{
#ifdef LOGGER_LEVEL
if(level < LOGGER_LEVEL) return true;
#endif
int size = 256;
string msg;
va_list ap;

View File

@ -34,7 +34,7 @@
namespace Limonp
{
#pragma region MD5 defines
//#pragma region MD5 defines
// Constants for MD5Transform routine.
#define S11 7
#define S12 12
@ -85,7 +85,7 @@ namespace Limonp
(a) = ROTATE_LEFT ((a), (s)); \
(a) += (b); \
}
#pragma endregion
//#pragma endregion
typedef unsigned char BYTE ;
@ -115,7 +115,7 @@ private:
unsigned char buffer[64]; /* input buffer */
} context ;
#pragma region static helper functions
//#pragma region static helper functions
// The core of the MD5 algorithm is here.
// MD5 basic transformation. Transforms state based on block.
static void MD5Transform( UINT4 state[4], unsigned char block[64] )
@ -229,7 +229,7 @@ private:
output[i] = ((UINT4)input[j]) | (((UINT4)input[j+1]) << 8) |
(((UINT4)input[j+2]) << 16) | (((UINT4)input[j+3]) << 24);
}
#pragma endregion
//#pragma endregion
public:
@ -354,7 +354,7 @@ public:
}
else
{
while( len = fread( buffer, 1, 1024, file ) )
while( (len = fread( buffer, 1, 1024, file )) )
Update( buffer, len ) ;
Final();

View File

@ -72,28 +72,6 @@ namespace Limonp
}
}
//inline bool joinStr(const vector<string>& src, string& dest, const string& connectorStr)
//{
// if(src.empty())
// {
// return false;
// }
// for(uint i = 0; i < src.size() - 1; i++)
// {
// dest += src[i];
// dest += connectorStr;
// }
// dest += src[src.size() - 1];
// return true;
//}
//inline string joinStr(const vector<string>& source, const string& connector)
//{
// string res;
// joinStr(source, res, connector);
// return res;
//}
template<class T>
void join(T begin, T end, string& res, const string& connector)
{
@ -122,7 +100,7 @@ namespace Limonp
inline bool splitStr(const string& src, vector<string>& res, const string& pattern)
inline bool split(const string& src, vector<string>& res, const string& pattern)
{
if(src.empty())
{
@ -181,20 +159,9 @@ namespace Limonp
}
inline uint16_t twocharToUint16(char high, char low)
{
return (((uint16_t(high) & 0x00ff ) << 8) | (uint16_t(low) & 0x00ff));
}
inline pair<char, char> uint16ToChar2(uint16_t in)
{
pair<char, char> res;
res.first = (in>>8) & 0x00ff; //high
res.second = (in) & 0x00ff; //low
return res;
}
inline bool strStartsWith(const string& str, const string& prefix)
inline bool startsWith(const string& str, const string& prefix)
{
//return str.substr(0, prefix.size()) == prefix;
if(prefix.length() > str.length())
@ -204,7 +171,7 @@ namespace Limonp
return 0 == str.compare(0, prefix.length(), prefix);
}
inline bool strEndsWith(const string& str, const string& suffix)
inline bool endsWith(const string& str, const string& suffix)
{
if(suffix.length() > str.length())
{
@ -218,13 +185,19 @@ namespace Limonp
return str.find(ch) != string::npos;
}
inline uint16_t twocharToUint16(char high, char low)
{
return (((uint16_t(high) & 0x00ff ) << 8) | (uint16_t(low) & 0x00ff));
}
inline bool utf8ToUnicode(const char * const str, uint len, vector<uint16_t>& vec)
{
char ch1, ch2;
if(!str)
{
return false;
}
char ch1, ch2;
uint16_t tmp;
vec.clear();
for(uint i = 0;i < len;)
{
@ -237,14 +210,16 @@ namespace Limonp
{
ch1 = (str[i] >> 2) & 0x07;
ch2 = (str[i+1] & 0x3f) | ((str[i] & 0x03) << 6 );
vec.push_back(twocharToUint16(ch1, ch2));
tmp = (((uint16_t(ch1) & 0x00ff ) << 8) | (uint16_t(ch2) & 0x00ff));
vec.push_back(tmp);
i += 2;
}
else if((unsigned char)str[i] <= 0xef && i + 2 < len)
{
ch1 = (str[i] << 4) | ((str[i+1] >> 2) & 0x0f );
ch2 = ((str[i+1]<<6) & 0xc0) | (str[i+2] & 0x3f);
vec.push_back(twocharToUint16(ch1, ch2));
tmp = (((uint16_t(ch1) & 0x00ff ) << 8) | (uint16_t(ch2) & 0x00ff));
vec.push_back(tmp);
i += 3;
}
else
@ -310,7 +285,8 @@ namespace Limonp
{
if(i + 1 < len) //&& (str[i+1] & 0x80))
{
vec.push_back(twocharToUint16(str[i], str[i + 1]));
uint16_t tmp = (((uint16_t(str[i]) & 0x00ff ) << 8) | (uint16_t(str[i+1]) & 0x00ff));
vec.push_back(tmp);
i += 2;
}
else
@ -321,11 +297,20 @@ namespace Limonp
}
return true;
}
inline bool gbkTrans(const string& str, vector<uint16_t>& vec)
{
return gbkTrans(str.c_str(), str.size(), vec);
}
//inline pair<char, char> uint16ToChar2(uint16_t in)
//{
// pair<char, char> res;
// res.first = (in>>8) & 0x00ff; //high
// res.second = (in) & 0x00ff; //low
// return res;
//}
inline bool gbkTrans(vector<uint16_t>::const_iterator begin, vector<uint16_t>::const_iterator end, string& res)
{
if(begin >= end)
@ -333,18 +318,21 @@ namespace Limonp
return false;
}
res.clear();
pair<char, char> pa;
//pair<char, char> pa;
char first, second;
while(begin != end)
{
pa = uint16ToChar2(*begin);
if(pa.first & 0x80)
//pa = uint16ToChar2(*begin);
first = ((*begin)>>8) & 0x00ff;
second = (*begin) & 0x00ff;
if(first & 0x80)
{
res += pa.first;
res += pa.second;
res += first;
res += second;
}
else
{
res += pa.second;
res += second;
}
begin++;
}

View File

@ -32,7 +32,7 @@ namespace CppJieba
class MPSegment: public SegmentBase
{
private:
protected:
Trie* _trie;
public:
@ -56,6 +56,7 @@ namespace CppJieba
LogError("get a NULL pointor form getTrie(\"%s\").", dictPath.c_str());
return false;
}
LogInfo("MPSegment init(%s) ok", dictPath.c_str());
return _setInitFlag(true);
}
public:

View File

@ -15,9 +15,9 @@ namespace CppJieba
HMMSegment _hmmSeg;
public:
MixSegment(){_setInitFlag(false);};
explicit MixSegment(const string& mpSegDict, const string& hmmSegDict): _mpSeg(mpSegDict), _hmmSeg(hmmSegDict)
explicit MixSegment(const string& mpSegDict, const string& hmmSegDict)
{
_setInitFlag(_mpSeg && _hmmSeg);
_setInitFlag(init(mpSegDict, hmmSegDict));
}
virtual ~MixSegment(){}
public:
@ -38,6 +38,7 @@ namespace CppJieba
LogError("_hmmSeg init");
return false;
}
LogInfo("MixSegment init(%s, %s)", mpSegDict.c_str(), hmmSegDict.c_str());
return _setInitFlag(true);
}
public:

View File

@ -1,10 +1,11 @@
#ifndef CPPJIEBA_SEGMENTBASE_H
#define CPPJIEBA_SEGMENTBASE_H
#include "ISegment.hpp"
#include "ChineseFilter.hpp"
#include "Limonp/str_functs.hpp"
#include "TransCode.hpp"
#include "Limonp/logger.hpp"
#include "ISegment.hpp"
#include <cassert>
namespace CppJieba
{
@ -18,9 +19,10 @@ namespace CppJieba
bool _isInited;
bool _getInitFlag()const{return _isInited;};
bool _setInitFlag(bool flag){return _isInited = flag;};
public:
operator bool(){return _getInitFlag();};
public:
virtual bool cut(Unicode::const_iterator begin, Unicode::const_iterator end, vector<string>& res)const = 0;
virtual bool cut(const string& str, vector<string>& res)const
{
@ -74,6 +76,46 @@ namespace CppJieba
return true;
#endif
}
public:
/*
* if char is ascii, count the ascii string's length and return 0;
* else count the nonascii string's length and return 1;
* if errors, return -1;
* */
static int filterAscii(const char* str, uint len, uint& resLen)
{
if(!str || !len)
{
return -1;
}
char x = 0x80;
int resFlag = (str[0] & x ? 1 : 0);
resLen = 0;
if(!resFlag)
{
while(resLen < len && !(str[resLen] & x))
{
resLen ++;
}
}
else
{
while(resLen < len && (str[resLen] & x))
{
#ifdef CPPJIEBA_GBK
resLen += 2;
#else
resLen ++;
#endif
}
}
if(resLen > len)
{
return -1;
}
return resFlag;
}
};
}

View File

@ -1,32 +0,0 @@
#ifndef CPPJIEBA_TFIDF_H
#define CPPJIEBA_TFIDF_H
#include "MPSegment.hpp"
namespace CppJieba
{
using namespace Limonp;
class TfIdfKeyWord
{
private:
MPSegment _segment;
public:
TfIdfKeyWord(const char* dictFile): _segment(dictFile){};
~TfIdfKeyWord(){};
public:
bool init(){return _segment.init();};
bool dispose(){return _segment.dispose();};
public:
bool extract(const string& str, vector<string>& words, uint topN)
{
return _segment.cut(words);
return true;
}
};
}
#endif

View File

@ -130,26 +130,13 @@ namespace CppJieba
}
bool loadDict(const char * const filePath)
{
if(!_getInitFlag())
{
LogError("not initted.");
return false;
}
if(!checkFileExist(filePath))
{
LogError("cann't find fiel[%s].",filePath);
return false;
}
bool res = false;
res = _trieInsert(filePath);
if(!res)
assert(_getInitFlag());
if(!_trieInsert(filePath))
{
LogError("_trieInsert failed.");
return false;
}
res = _countWeight();
if(!res)
if(!_countWeight())
{
LogError("_countWeight failed.");
return false;
@ -339,22 +326,30 @@ namespace CppJieba
private:
bool _trieInsert(const char * const filePath)
{
ifstream ifile(filePath);
ifstream ifs(filePath);
if(!ifs)
{
LogError("open %s failed.", filePath);
return false;
}
string line;
vector<string> vecBuf;
TrieNodeInfo nodeInfo;
while(getline(ifile, line))
size_t lineno = 0;
while(getline(ifs, line))
{
vecBuf.clear();
splitStr(line, vecBuf, " ");
lineno ++;
split(line, vecBuf, " ");
if(3 < vecBuf.size())
{
LogError("line[%s] illegal.", line.c_str());
LogError("line[%u:%s] illegal.", lineno, line.c_str());
return false;
}
if(!TransCode::decode(vecBuf[0], nodeInfo.word))
{
LogError("line[%u:%s] illegal.", lineno, line.c_str());
return false;
}
nodeInfo.freq = atoi(vecBuf[1].c_str());

View File

@ -23,15 +23,13 @@ namespace CppJieba
LogError("error when getting md5 for file '%s'", dictpath);
return NULL;
}
LogInfo("md5 for file '%s': %s", dictpath, md5.c_str());
if (_tries.find(md5) != _tries.end())
{
LogInfo("find a exits trie for md5: '%s'", md5.c_str());
return _tries[md5.c_str()];
}
LogInfo("create a new trie for md5: '%s'", md5.c_str());
//LogDebug("create a new trie for md5: '%s'", md5.c_str());
Trie* trie = NULL;
try
{
@ -54,15 +52,14 @@ namespace CppJieba
return NULL;
}
LogInfo("trie->loadDict(%s) start...", dictpath);
if (!trie->loadDict(dictpath))
{
LogError("trie->loadDict(%s) failed...", dictpath);
return NULL;
}
LogInfo("trie->loadDict end...");
_tries[md5.c_str()] = trie;
LogDebug("trie->loadDict(%s)", dictpath);
return trie;
}

View File

@ -45,11 +45,11 @@ int main(int argc, char ** argv)
<<"\t--modelpath\tsee example\n"
<<"\t--maxlen\tspecify the granularity of cut used in cutQuery. \n\t\t\tIf not specified, the default is 3\n"
<<"example:\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dicts/jieba.dict.utf8 --algorithm cutDAG\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dicts/jieba.dict.utf8 --algorithm cutFull\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --modelpath ../dicts/hmm_model.utf8 --algorithm cutHMM\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dicts/jieba.dict.utf8 --modelpath ../dicts/hmm_model.utf8 --algorithm cutMix\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dicts/jieba.dict.utf8 --modelpath ../dicts/hmm_model.utf8 --algorithm cutQuery --maxlen 3\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dict/jieba.dict.utf8 --algorithm cutDAG\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dict/jieba.dict.utf8 --algorithm cutFull\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --modelpath ../dict/hmm_model.utf8 --algorithm cutHMM\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dict/jieba.dict.utf8 --modelpath ../dict/hmm_model.utf8 --algorithm cutMix\n"
<<"\t"<<argv[0]<<" ../test/testdata/testlines.utf8 --dictpath ../dict/jieba.dict.utf8 --modelpath ../dict/hmm_model.utf8 --algorithm cutQuery --maxlen 3\n"
<<endl;
return EXIT_FAILURE;

View File

@ -4,7 +4,7 @@
#include <ctype.h>
#include <string.h>
#include "Limonp/Config.hpp"
#include "Husky/ServerFrame.hpp"
#include "Husky/HuskyServer.hpp"
#include "MPSegment.hpp"
#include "HMMSegment.hpp"
#include "MixSegment.hpp"
@ -38,8 +38,8 @@ bool run(int argc, char** argv)
{
return false;
}
Config conf;
if(!conf.loadFile(argv[1]))
Config conf(argv[1]);
if(!conf)
{
return false;
}
@ -90,7 +90,7 @@ bool run(int argc, char** argv)
}
ReqHandler reqHandler(dictPath, modelPath);
ServerFrame sf(port, threadNum, &reqHandler);
HuskyServer sf(port, threadNum, &reqHandler);
return sf.init() && sf.run();
}

View File

@ -22,28 +22,19 @@ void cut(const ISegment * seg, const char * const filePath, size_t times = 10)
for(uint i = 0; i < times; i ++)
{
LogInfo("times[%u]", i);
//ifile.seekg(0);
//while(getline(ifile, line))
//{
// if(!line.empty())
// {
res.clear();
seg->cut(doc, res);
//print(res);
//cout<<join(res.begin(), res.end(),"/")<<endl;
// }
//}
res.clear();
seg->cut(doc, res);
}
}
int main(int argc, char ** argv)
{
{
MixSegment seg("../dicts/jieba.dict.utf8", "../dicts/hmm_model.utf8");
MixSegment seg("../dict/jieba.dict.utf8", "../dict/hmm_model.utf8");
if(!seg)
{
cout<<"seg init failed."<<endl;
return false;
return EXIT_FAILURE;
}
cut(&seg, "../test/testdata/weicheng.utf8");
}

View File

@ -23,8 +23,8 @@ void cut(const ISegment * seg, const char * const filePath)
}
const char * const TEST_FILE = "../test/testdata/testlines.utf8";
const char * const JIEBA_DICT_FILE = "../dicts/jieba.dict.utf8";
const char * const HMM_DICT_FILE = "../dicts/hmm_model.utf8";
const char * const JIEBA_DICT_FILE = "../dict/jieba.dict.utf8";
const char * const HMM_DICT_FILE = "../dict/hmm_model.utf8";
int main(int argc, char ** argv)
{

View File

@ -8,8 +8,8 @@
using namespace Husky;
using namespace CppJieba;
const char * const DEFAULT_DICTPATH = "../dicts/jieba.dict.utf8";
const char * const DEFAULT_MODELPATH = "../dicts/hmm_model.utf8";
const char * const DEFAULT_DICTPATH = "../dict/jieba.dict.utf8";
const char * const DEFAULT_MODELPATH = "../dict/hmm_model.utf8";
class ServerDemo: public IRequestHandler
{

91
test/servertest/load_test.py Executable file
View File

@ -0,0 +1,91 @@
#!/usr/bin/python
# coding:utf-8
import time
import urllib2
import threading
from Queue import Queue
from time import sleep
import sys
# 性能测试页面
#PERF_TEST_URL = "http://10.2.66.38/?yyid=-1&suv=1309231700203264&callback=xxxxx"
URLS = [line for line in open("../testdata/load_test.urls", "r")]
# 配置:压力测试
THREAD_NUM = 10 # 并发线程总数
ONE_WORKER_NUM = 500 # 每个线程的循环次数
LOOP_SLEEP = 0.01 # 每次请求时间间隔(秒)
# 配置:模拟运行状态
#THREAD_NUM = 10 # 并发线程总数
#ONE_WORKER_NUM = 10 # 每个线程的循环次数
#LOOP_SLEEP = 0 # 每次请求时间间隔(秒)
# 出错数
ERROR_NUM = 0
#具体的处理函数,负责处理单个任务
def doWork(index, url):
t = threading.currentThread()
#print "["+t.name+" "+str(index)+"] "+PERF_TEST_URL
try:
html = urllib2.urlopen(url).read()
except urllib2.URLError, e:
print "["+t.name+" "+str(index)+"] "
print e
global ERROR_NUM
ERROR_NUM += 1
#这个是工作进程,负责不断从队列取数据并处理
def working():
t = threading.currentThread()
print "["+t.name+"] Sub Thread Begin"
i = 0
while i < ONE_WORKER_NUM:
i += 1
doWork(i, URLS[i % len(URLS)])
sleep(LOOP_SLEEP)
print "["+t.name+"] Sub Thread End"
def main():
#doWork(0)
#return
t1 = time.time()
Threads = []
# 创建线程
for i in range(THREAD_NUM):
t = threading.Thread(target=working, name="T"+str(i))
t.setDaemon(True)
Threads.append(t)
for t in Threads:
t.start()
for t in Threads:
t.join()
print "main thread end"
t2 = time.time()
print "========================================"
#print "URL:", PERF_TEST_URL
print "任务数量:", THREAD_NUM, "*", ONE_WORKER_NUM, "=", THREAD_NUM*ONE_WORKER_NUM
print "总耗时(秒):", t2-t1
print "每次请求耗时(秒):", (t2-t1) / (THREAD_NUM*ONE_WORKER_NUM)
print "每秒承载请求数:", 1 / ((t2-t1) / (THREAD_NUM*ONE_WORKER_NUM))
print "错误数量:", ERROR_NUM
if __name__ == "__main__":
main()

11
test/servertest/run_curl.sh Executable file
View File

@ -0,0 +1,11 @@
CURL_RES=../testdata/curl.res
TMP=curl.res.tmp
curl -s "http://127.0.0.1:11200/?key=南京市长江大桥" >> $TMP
if diff $TMP $CURL_RES >> /dev/null
then
echo "ok";
else
echo "failed."
fi
rm $TMP

1
test/testdata/curl.res vendored Normal file
View File

@ -0,0 +1 @@
["南京市", "长江大桥"]

1
test/testdata/load_test.urls vendored Normal file
View File

@ -0,0 +1 @@
http://127.0.0.1:11200/?key=南京市长江大桥

View File

@ -3,9 +3,11 @@ SET(LIBRARY_OUTPUT_PATH ${PROJECT_BINARY_DIR}/test/lib)
SET(GTEST_ROOT_DIR gtest-1.6.0)
ADD_DEFINITIONS(-DLOGGER_LEVEL=LL_WARN)
INCLUDE_DIRECTORIES(${GTEST_ROOT_DIR} ${GTEST_ROOT_DIR}/include ${PROJECT_SOURCE_DIR})
ADD_LIBRARY(gtest STATIC ${GTEST_ROOT_DIR}/src/gtest-all.cc)
ADD_EXECUTABLE(test.run gtest_main.cc TChineseFilter.cpp TMixSegment.cpp TMPSegment.cpp THMMSegment.cpp TTrie.cpp TFullSegment.cpp TQuerySegment.cpp TTrieManager.cpp TMd5.cpp)
FILE(GLOB SRCFILES *.cpp)
ADD_EXECUTABLE(test.run ${SRCFILES})
TARGET_LINK_LIBRARIES(gtest pthread)
TARGET_LINK_LIBRARIES(test.run gtest pthread)

View File

@ -5,7 +5,7 @@ using namespace CppJieba;
TEST(FullSegment, Test1)
{
FullSegment segment("../dicts/jieba.dict.utf8");
FullSegment segment("../dict/jieba.dict.utf8");
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"", "来自", "北京", "北京邮电", "北京邮电大学", "邮电", "邮电大学", "电大", "大学", "", "", "", " ", "学号", " 123456"};
vector<string> words;

View File

@ -5,7 +5,7 @@ using namespace CppJieba;
TEST(HMMSegmentTest, Test1)
{
HMMSegment segment("../dicts/hmm_model.utf8");;
HMMSegment segment("../dict/hmm_model.utf8");;
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"我来", "自北京", "邮电大学", "", "", "", " ", "学号", " 123456"};
//string s;

View File

@ -0,0 +1,56 @@
#include "src/KeywordExtractor.hpp"
#include "gtest/gtest.h"
using namespace CppJieba;
TEST(KeywordExtractorTest, Test1)
{
KeywordExtractor extractor("../dict/jieba.dict.utf8", "../dict/idf.utf8");
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"北京邮电大学", "来自"};
vector<string> words;
ASSERT_TRUE(extractor);
ASSERT_TRUE(extractor.extract(str, words, 2));
ASSERT_EQ(words, vector<string>(res, res + sizeof(res)/sizeof(res[0])));
}
TEST(KeywordExtractorTest, Test2)
{
KeywordExtractor extractor("../dict/jieba.dict.utf8", "../dict/idf.utf8");
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"北京邮电大学", "来自"};
vector<string> words;
ASSERT_TRUE(extractor);
ASSERT_TRUE(extractor.extract(str, words, 9));
ASSERT_EQ(words, vector<string>(res, res + sizeof(res)/sizeof(res[0])));
}
TEST(KeywordExtractorTest, Test3)
{
ifstream ifs("../test/testdata/weicheng.utf8");
ASSERT_TRUE(!!ifs);
string str((istreambuf_iterator<char>(ifs)), (istreambuf_iterator<char>()));
KeywordExtractor extractor("../dict/jieba.dict.utf8", "../dict/idf.utf8");
const char* res[] = {"小姐", "孙小姐", "方鸿渐", "自己", "没有"};
vector<string> keywords;
extractor.extract(str, keywords, 5);
ASSERT_EQ(keywords, vector<string>(res, res + sizeof(res)/sizeof(res[0])));
}
TEST(KeywordExtractorTest, Test4)
{
ifstream ifs("../test/testdata/weicheng.utf8");
ASSERT_TRUE(!!ifs);
string str((istreambuf_iterator<char>(ifs)), (istreambuf_iterator<char>()));
KeywordExtractor extractor("../dict/jieba.dict.utf8", "../dict/idf.utf8");
//const char* res[] = {"小姐", "孙小姐", "方鸿渐", "自己", "没有"};
vector<pair<string,double> > keywords;
extractor.extract(str, keywords, 5);
//print(keywords);
string res;
res << keywords;
ASSERT_EQ(res, "[\"小姐:4268.75\", \"孙小姐:3789.41\", \"方鸿渐:3030.35\", \"自己:2300.54\", \"没有:2104.27\"]");
}

View File

@ -5,7 +5,7 @@ using namespace CppJieba;
TEST(MPSegmentTest, Test1)
{
MPSegment segment("../dicts/jieba.dict.utf8");;
MPSegment segment("../dict/jieba.dict.utf8");;
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"", "来自", "北京邮电大学", "","",""," ","","", " 123456"};
vector<string> words;

View File

@ -5,7 +5,7 @@ using namespace CppJieba;
TEST(MixSegmentTest, Test1)
{
MixSegment segment("../dicts/jieba.dict.utf8", "../dicts/hmm_model.utf8");;
MixSegment segment("../dict/jieba.dict.utf8", "../dict/hmm_model.utf8");;
const char* str = "我来自北京邮电大学。。。 学号 123456";
const char* res[] = {"", "来自", "北京邮电大学", "","",""," ","学号", " 123456"};
vector<string> words;

View File

@ -5,7 +5,7 @@ using namespace CppJieba;
TEST(QuerySegment, Test1)
{
QuerySegment segment("../dicts/jieba.dict.utf8", "../dicts/hmm_model.utf8", 3);
QuerySegment segment("../dict/jieba.dict.utf8", "../dict/hmm_model.utf8", 3);
const char* str = "小明硕士毕业于中国科学院计算所,后在日本京都大学深造";
const char* res[] = {"小明", "硕士", "毕业", "", "中国", "中国科学院", "科学", "科学院", "学院", "计算所", "", "", "", "日本", "日本京都大学", "京都", "京都大学", "大学", "深造"};
vector<string> words;

View File

@ -1,9 +1,9 @@
#include "src/ChineseFilter.hpp"
#include "src/SegmentBase.hpp"
#include "gtest/gtest.h"
using namespace CppJieba;
TEST(ChineseFilterTest, Test1)
TEST(SegmentBaseTest, Test1)
{
const char* str = "heheh你好...hh";
string s;
@ -16,9 +16,9 @@ TEST(ChineseFilterTest, Test1)
uint offset = 0;
while(offset < size)
{
uint len;
uint len = 0;
const char* t = str + offset;
int ret = filterAscii(t, size - offset, len);
SegmentBase::filterAscii(t, size - offset, len);
s.assign(t, len);
res.push_back(s);
//cout<<s<<","<<ret<<","<<len<<endl;

View File

@ -3,7 +3,7 @@
using namespace CppJieba;
static const char* const DICT_FILE = "../dicts/jieba.dict.utf8";
static const char* const DICT_FILE = "../dict/jieba.dict.utf8";
TEST(TrieTest, Test1)
{

View File

@ -496,7 +496,20 @@
# undef _TR1_FUNCTIONAL // Allows the user to #include
// <tr1/functional> if he chooses to.
# else
# if defined (__cplusplus) && __cplusplus > 199711L
# include <tuple>
namespace std {
namespace tr1 {
using std::tuple;
using std::tuple_element;
using std::get;
using std::tuple_size;
using std::make_tuple;
}
}
# else
# include <tr1/tuple> // NOLINT
# endif
# endif // !GTEST_HAS_RTTI && GTEST_GCC_VER_ < 40302
# else