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

modify calcDAG try to speed up

Browse Source
This commit is contained in:
wyy 2013-12-07 07:45:06 -08:00
parent 32bafd78f0
commit 81c2d3caf1
2 changed files with 36 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

49
src/MPSegment.hpp
View File

@ -126,13 +126,9 @@ namespace CppJieba
return false;
}
SegmentContext segContext;
for(Unicode::const_iterator it = begin; it != end; it++)
{
segContext.push_back(SegmentChar(*it));
}
//calc DAG
if(!_calcDAG(segContext))
if(!_calcDAG(begin, end, segContext))
{
LogError("_calcDAG failed.");
return false;
@ -154,36 +150,47 @@ namespace CppJieba
}
private:
bool _calcDAG(SegmentContext& segContext)const
bool _calcDAG(Unicode::const_iterator begin, Unicode::const_iterator end, SegmentContext& segContext) const
{
if(segContext.empty())
if(begin >= end)
{
LogError("segContext empty.");
LogError("begin >= end.");
return false;
}
Unicode unicode;
vector<pair<uint, const TrieNodeInfo*> > vp;
for(uint i = 0; i < segContext.size(); i++)
for(Unicode::const_iterator it = begin; it != end; it++)
{
unicode.clear();
for(uint j = i ; j < segContext.size(); j++)
{
unicode.push_back(segContext[j].uniCh);
}
segContext.push_back(SegmentChar(*it));
}
vector<pair<uint, const TrieNodeInfo*> > vp;
//for(Unicode::const_iterator it = begin; it != end; it++)
//{
// segContext.push_back(SegmentChar(*it));
//}
for(Unicode::const_iterator it = begin; it != end; it++)
//for(uint i = 0; i < segContext.size(); i++)
{
//unicode.clear();
//for(uint j = i; j < segContext.size(); j++)
//{
// unicode.push_back(segContext[j].uniCh);
//}
segContext.push_back(SegmentChar(*it));
SegmentChar& back = segContext.back();
int i = it - begin;
vp.clear();
if(_trie.find(unicode, vp))
if(_trie.find(it, end, vp))
{
for(uint j = 0; j < vp.size(); j++)
{
uint nextp = vp[j].first + i;
segContext[i].dag[nextp] = vp[j].second;
back.dag[nextp] = vp[j].second;
}
}
if(segContext[i].dag.end() == segContext[i].dag.find(i))
if(back.dag.end() == back.dag.find(i))
{
segContext[i].dag[i] = NULL;
back.dag[i] = NULL;
}
}
return true;

16
src/Trie.hpp
View File

@ -244,14 +244,14 @@ namespace CppJieba
return !res.empty();
}
bool find(const Unicode& unico, vector<pair<uint, const TrieNodeInfo*> >& res)const
{
if (!unico.empty())
{
return find(unico.begin(), unico.end(), res);
}
return false;
}
//bool find(const Unicode& unico, vector<pair<uint, const TrieNodeInfo*> >& res)const
//{
// if (!unico.empty())
// {
// return find(unico.begin(), unico.end(), res);
// }
// return false;
//}
public:
double getMinLogFreq()const{return _minLogFreq;};