bool cut(Unicode::const_iterator begin , Unicode::const_iterator end, vector<TrieNodeInfo>& segWordInfos)const
{
if(!_getInitFlag())
{
LogError("not inited.");
return false;
}
SegmentContext segContext;
//calc DAG
if(!_calcDAG(begin, end, segContext))
{
LogError("_calcDAG failed.");
return false;
}
if(!_calcDP(segContext))
{
LogError("_calcDP failed.");
return false;
}
if(!_cut(segContext, segWordInfos))
{
LogError("_cut failed.");
return false;
}
return true;
}
virtual bool cut(Unicode::const_iterator begin, Unicode::const_iterator end, vector<string>& res)const
{
if(!_getInitFlag())
{
LogError("not inited.");
return false;
}
vector<TrieNodeInfo> segWordInfos;
if(!cut(begin, end, segWordInfos))
{
return false;
}
string tmp;
for(uint i = 0; i < segWordInfos.size(); i++)
{
if(TransCode::encode(segWordInfos[i].word, tmp))
{
res.push_back(tmp);
}
else
{
LogError("encode failed.");
}
}
return true;
}
bool Trie::find(const Unicode& unico, vector<pair<uint, const TrieNodeInfo*> >& res)const
{
if(!_getInitFlag())
{
LogFatal("trie not initted!");
return false;
}
TrieNode* p = _root;
//for(Unicode::const_iterator it = begin; it != end; it++)
for(uint i = 0; i < unico.size(); i++)
{
if(p->hmap.find(unico[i]) == p-> hmap.end())
{
break;
}
p = p->hmap[unico[i]];
if(p->isLeaf)
{
uint pos = p->nodeInfoVecPos;
if(pos < _nodeInfoVec.size())
{
res.push_back(make_pair(i, &_nodeInfoVec[pos]));
}
else
{
LogFatal("node's nodeInfoVecPos is out of _nodeInfoVec's range");
return false;
}
}
}
return !res.empty();
}
bool Trie::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)
{
LogError("_trieInsert failed.");
return false;
}
res = _countWeight();
if(!res)
{
LogError("_countWeight failed.");
return false;
}
return true;
}
virtual bool dispose()
{
if(!_getInitFlag())
{
return true;
}
_trie.dispose();
_setInitFlag(false);
return true;
}
bool MixSegment::dispose()
{
if(!_getInitFlag())
{
return true;
}
_mpSeg.dispose();
_hmmSeg.dispose();
_setInitFlag(false);
return true;
}
bool Trie::insert(const TrieNodeInfo& nodeInfo)
{
if(!_getInitFlag())
{
LogFatal("not initted!");
return false;
}
const Unicode& uintVec = nodeInfo.word;
TrieNode* p = _root;
for(uint i = 0; i < uintVec.size(); i++)
{
uint16_t cu = uintVec[i];
if(NULL == p)
{
return false;
}
if(p->hmap.end() == p->hmap.find(cu))
{
TrieNode * next = NULL;
try
{
next = new TrieNode;
}
catch(const bad_alloc& e)
{
return false;
}
p->hmap[cu] = next;
p = next;
}
else
{
p = p->hmap[cu];
}
}
if(NULL == p)
{
return false;
}
if(p->isLeaf)
{
LogError("this node already inserted");
return false;
}
p->isLeaf = true;
_nodeInfoVec.push_back(nodeInfo);
p->nodeInfoVecPos = _nodeInfoVec.size() - 1;
return true;
}
const TrieNodeInfo* Trie::findPrefix(const string& str)const
{
if(!_getInitFlag())
{
LogFatal("trie not initted!");
return NULL;
}
Unicode uintVec;
if(!TransCode::decode(str, uintVec))
{
LogError("TransCode::decode failed.");
return NULL;
}
//find
TrieNode* p = _root;
uint pos = 0;
uint16_t chUni = 0;
const TrieNodeInfo * res = NULL;
for(uint i = 0; i < uintVec.size(); i++)
{
chUni = uintVec[i];
if(p->isLeaf)
{
pos = p->nodeInfoVecPos;
if(pos >= _nodeInfoVec.size())
{
LogFatal("node's nodeInfoVecPos is out of _nodeInfoVec's range");
return NULL;
}
res = &(_nodeInfoVec[pos]);
}
if(p->hmap.find(chUni) == p->hmap.end())
{
break;
}
else
{
p = p->hmap[chUni];
}
}
return res;
}
bool Trie::dispose()
{
if(!_getInitFlag())
{
return false;
}
bool ret = _deleteNode(_root);
if(!ret)
{
LogFatal("_deleteNode failed!");
return false;
}
_root = NULL;
_nodeInfoVec.clear();
_setInitFlag(false);
return ret;
}
bool MixSegment::init(const char* const mpSegDict, const char* const hmmSegDict)
{
if(_getInitFlag())
{
LogError("inited.");
return false;
}
if(!_mpSeg.init(mpSegDict))
{
LogError("_mpSeg init");
return false;
}
if(!_hmmSeg.init(hmmSegDict))
{
LogError("_hmmSeg init");
return false;
}
return _setInitFlag(true);
}
bool cut(const string& str, vector<TrieNodeInfo>& segWordInfos)const
{
if(!_getInitFlag())
{
LogError("not inited.");
return false;
}
if(str.empty())
{
return false;
}
Unicode sentence;
if(!TransCode::decode(str, sentence))
{
LogError("TransCode::decode failed.");
return false;
}
return cut(sentence.begin(), sentence.end(), segWordInfos);
}
virtual bool init()
{
if(_getInitFlag())
{
LogError("already inited before now.");
return false;
}
if(!_trie.init())
{
LogError("_trie.init failed.");
return false;
}
LogInfo("_trie.loadDict(%s) start...", _dictPath.c_str());
if(!_trie.loadDict(_dictPath.c_str()))
{
LogError("_trie.loadDict faield.");
return false;
}
LogInfo("_trie.loadDict end.");
return _setInitFlag(true);
}
const TrieNodeInfo* Trie::find(Unicode::const_iterator begin, Unicode::const_iterator end)const
{
if(!_getInitFlag())
{
LogFatal("trie not initted!");
return NULL;
}
if(begin >= end)
{
return NULL;
}
TrieNode* p = _root;
for(Unicode::const_iterator it = begin; it != end; it++)
{
uint16_t chUni = *it;
if(p->hmap.find(chUni) == p-> hmap.end())
{
return NULL;
}
else
{
p = p->hmap[chUni];
}
}
if(p->isLeaf)
{
uint pos = p->nodeInfoVecPos;
if(pos < _nodeInfoVec.size())
{
return &(_nodeInfoVec[pos]);
}
else
{
LogFatal("node's nodeInfoVecPos is out of _nodeInfoVec's range");
return NULL;
}
}
return NULL;
}
bool Trie::init()
{
if(_getInitFlag())
{
LogError("already initted!");
return false;
}
try
{
_root = new TrieNode;
}
catch(const bad_alloc& e)
{
return false;
}
if(NULL == _root)
{
return false;
}
_setInitFlag(true);
return true;
}
operator bool() const {return _getInitFlag();};
bool MixSegment::cut(Unicode::const_iterator begin, Unicode::const_iterator end, vector<string>& res)const
{
if(!_getInitFlag())
{
LogError("not inited.");
return false;
}
if(begin == end)
{
return false;
}
vector<TrieNodeInfo> infos;
if(!_mpSeg.cut(begin, end, infos))
{
LogError("mpSeg cutDAG failed.");
return false;
}
Unicode unico;
vector<Unicode> hmmRes;
string tmp;
for(uint i= 0; i < infos.size(); i++)
{
TransCode::encode(infos[i].word,tmp);
if(1 == infos[i].word.size())
{
unico.push_back(infos[i].word[0]);
}
else
{
if(!unico.empty())
{
hmmRes.clear();
if(!_hmmSeg.cut(unico.begin(), unico.end(), hmmRes))
{
LogError("_hmmSeg cut failed.");
return false;
}
for(uint j = 0; j < hmmRes.size(); j++)
{
TransCode::encode(hmmRes[j], tmp);
res.push_back(tmp);
}
}
unico.clear();
TransCode::encode(infos[i].word, tmp);
res.push_back(tmp);
}
}
if(!unico.empty())
{
hmmRes.clear();
if(!_hmmSeg.cut(unico.begin(), unico.end(), hmmRes))
{
LogError("_hmmSeg cut failed.");
return false;
}
for(uint j = 0; j < hmmRes.size(); j++)
{
TransCode::encode(hmmRes[j], tmp);
res.push_back(tmp);
}
}
return true;
}
