bool init(const string &dictPath, const string &hmmFilePath, const string &idfPath, const string &stopWordPath)
{
_loadIdfDict(idfPath);
_loadStopWordDict(stopWordPath);
LIMONP_CHECK(_segment.init(dictPath, hmmFilePath));
return true;
};
bool init(const string& dictPath, const string& hmmFilePath, const string& charStatus, const string& startProb, const string& emitProb, const string& endProb, const string& transProb)
{
assert(!_getInitFlag());
_dictTrie.init(dictPath);
assert(_dictTrie);
return _setInitFlag(_segment.init(dictPath, hmmFilePath));
};
int main(int argc, char ** argv)
{
if(argc < 2)
{
cout<<"usage: \n\t"<<argv[0]<<" [options] <filename>\n"
<<"options:\n"
<<"\t--algorithm\tSupported methods are [cutDAG, cutHMM, cutMix] for now. \n\t\t\tIf not specified, the default is cutDAG\n"
<<"\t--dictpath\tsee example\n"
<<"\t--modelpath\tsee example\n"
<<"example:\n"
<<"\t"<<argv[0]<<" testlines.utf8 --dictpath dicts/jieba.dict.utf8\n"
<<"\t"<<argv[0]<<" testlines.utf8 --modelpath dicts/hmm_model.utf8 --algorithm cutHMM\n"
<<"\t"<<argv[0]<<" testlines.utf8 --dictpath dicts/jieba.dict.utf8 --modelpath dicts/hmm_model.utf8 --algorithm cutMix\n"
<<endl;
return EXIT_FAILURE;
}
ArgvContext arg(argc, argv);
string dictPath = arg["--dictpath"];
string modelPath = arg["--modelpath"];
string algorithm = arg["--algorithm"];
if("cutHMM" == algorithm)
{
HMMSegment seg;
if(!seg.init(modelPath.c_str()))
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, arg[1].c_str());
seg.dispose();
}
else if("cutMix" == algorithm)
{
MixSegment seg;
if(!seg.init(dictPath.c_str(), modelPath.c_str()))
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, arg[1].c_str());
seg.dispose();
}
else
{
MPSegment seg;
if(!seg.init(dictPath.c_str()))
{
cout<<"seg init failed."<<endl;
return false;
}
cut(&seg, arg[1].c_str());
seg.dispose();
}
return EXIT_SUCCESS;
}
void init(
const string& dictPath,
const string& hmmFilePath,
const string& userDictPath = ""
)
{
LIMONP_CHECK(_segment.init(dictPath, hmmFilePath, userDictPath));
_dictTrie = _segment.getDictTrie();
LIMONP_CHECK(_dictTrie);
};
bool init(const string& dict, const string& model, size_t maxWordLen)
{
if (_getInitFlag())
{
LogError("inited already.");
return false;
}
if (!_mixSeg.init(dict, model))
{
LogError("_mixSeg init");
return false;
}
if (!_fullSeg.init(dict))
{
LogError("_fullSeg init");
return false;
}
_maxWordLen = maxWordLen;
return _setInitFlag(true);
}
int main(int argc, char ** argv)
{
//demo
{
HMMSegment seg;
if(!seg.init("../dicts/hmm_model.utf8"))
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, "testlines.utf8");
seg.dispose();
}
{
MixSegment seg;
if(!seg.init("../dicts/jieba.dict.utf8", "../dicts/hmm_model.utf8"))
{
cout<<"seg init failed."<<endl;
return EXIT_FAILURE;
}
cut(&seg, "testlines.utf8");
cout<<"Cut type"<<endl;
cut_type(&seg, "testlines.utf8");
cout<<endl;
seg.dispose();
}
{
MPSegment seg;
if(!seg.init("../dicts/jieba.dict.utf8"))
{
cout<<"seg init failed."<<endl;
return false;
}
cut(&seg, "testlines.utf8");
seg.dispose();
}
return EXIT_SUCCESS;
}
bool init_c(const char * dict_path, const char * model_path)
{
return segment.init(dict_path, model_path);
}
bool init(const string& dictPath, const string& hmmFilePath, const string& idfPath, const string& stopWordPath)
{
_loadIdfDict(idfPath);
_loadStopWordDict(stopWordPath);
return _setInitFlag(_segment.init(dictPath, hmmFilePath));
};