modify test/segment_demo.cpp

This commit is contained in:
wyy 2014-04-25 22:09:55 +08:00
parent f8487fd9cf
commit 57ef504d9b
4 changed files with 62 additions and 67 deletions

View File

@ -1,6 +1,6 @@
SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR})
ADD_EXECUTABLE(segment.demo segment.cpp)
ADD_EXECUTABLE(segment.demo segment_demo.cpp)
ADD_EXECUTABLE(keyword.demo keyword_demo.cpp)
ADD_EXECUTABLE(tagging.demo tagging_demo.cpp)
ADD_EXECUTABLE(load_test load_test.cpp)

View File

@ -1,60 +0,0 @@
#include <iostream>
#include <fstream>
#include "../src/MPSegment.hpp"
#include "../src/HMMSegment.hpp"
#include "../src/MixSegment.hpp"
using namespace CppJieba;
void cut(const ISegment * seg, const char * const filePath)
{
ifstream ifile(filePath);
vector<string> res;
string line;
while(getline(ifile, line))
{
if(!line.empty())
{
res.clear();
seg->cut(line, res);
cout<<join(res.begin(), res.end(),"/")<<endl;
}
}
}
const char * const TEST_FILE = "../test/testdata/testlines.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)
{
//demo
{
MPSegment seg(JIEBA_DICT_FILE);
if(!seg)
{
cout<<"seg init failed."<<endl;
return false;
}
cut(&seg, TEST_FILE);
}
{
HMMSegment seg(HMM_DICT_FILE);
if(!seg)
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, TEST_FILE);
}
{
MixSegment seg(JIEBA_DICT_FILE, HMM_DICT_FILE);
if(!seg)
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, TEST_FILE);
}
return EXIT_SUCCESS;
}

59
test/segment_demo.cpp Normal file
View File

@ -0,0 +1,59 @@
#include <iostream>
#include <fstream>
#define LOGGER_LEVEL LL_WARN
#include "../src/MPSegment.hpp"
#include "../src/HMMSegment.hpp"
#include "../src/MixSegment.hpp"
using namespace CppJieba;
const char * const TEST_FILE = "../test/testdata/testlines.utf8";
const char * const JIEBA_DICT_FILE = "../dict/jieba.dict.utf8";
const char * const HMM_DICT_FILE = "../dict/hmm_model.utf8";
const char * const USER_DICT_FILE = "../test/testdata/userdict.utf8";
void cut(const ISegment& seg, const char * const filePath)
{
ifstream ifile(filePath);
vector<string> words;
string line;
string res;
while(getline(ifile, line))
{
if(!line.empty())
{
words.clear();
seg.cut(line, words);
join(words.begin(), words.end(), res, "/");
cout<< res <<endl;
}
}
}
int main(int argc, char ** argv)
{
{
printf("\e[32m%s\e[0m\n", "[demo] MPSegment"); // colorful
MPSegment seg(JIEBA_DICT_FILE);
cut(seg, TEST_FILE);
}
{
printf("\e[32m%s\e[0m\n", "[demo] HMMSegment"); // colorful
HMMSegment seg(HMM_DICT_FILE);
cut(seg, TEST_FILE);
}
{
printf("\e[32m%s\e[0m\n", "[demo] MixSegment"); // colorful
MixSegment seg(JIEBA_DICT_FILE, HMM_DICT_FILE);
cut(seg, TEST_FILE);
}
{
printf("\e[32m%s\e[0m\n", "[demo] MixSegment with UserDict"); // colorful
MixSegment seg(JIEBA_DICT_FILE, HMM_DICT_FILE, USER_DICT_FILE);
cut(seg, TEST_FILE);
}
return EXIT_SUCCESS;
}

View File

@ -1,11 +1,7 @@
我来到北京清华大学
他来到了网易杭研大厦
杭研
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
我来自北京邮电大学。。。 学号 091111xx。。。
来这里看看别人正在搜索什么吧
我来到南京市长江大桥
请在一米线外等候
人事处女干事
去医院做B超号123
B超 T恤
去医院做B超号123
令狐冲是云计算行业的专家