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

prettify Trie.hpp ing

Browse Source
This commit is contained in:
wyy 2014-03-16 21:00:51 +08:00
parent 762495f5f4
commit 89c955c1d6
2 changed files with 18 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
ChangeLog.md
View File

@ -2,6 +2,7 @@
1. 修改了设计上的问题，删除了`TrieManager`这个类，以避免造成一些可能的隐患。
2. 增加`stop_words.utf8`词典，并修改`KeywordExtractor`的初始化函数用以使用此词典。
3. 优化了Trie树相关部分代码结构。
## v2.3.2

49
src/Trie.hpp
View File

@ -104,18 +104,17 @@ namespace CppJieba
public:
const TrieNodeInfo* find(Unicode::const_iterator begin, Unicode::const_iterator end)const
{
TrieNodeMap::const_iterator citer;
TrieNode* p = _root;
for(Unicode::const_iterator it = begin; it != end; it++)
{
uint16_t chUni = *it;
if(p->hmap.find(chUni) == p-> hmap.end())
citer = p->hmap.find(chUni);
if(p-> hmap.end() == citer)
{
return NULL;
}
else
{
p = p->hmap[chUni];
}
p = citer->second;
}
if(p->isLeaf)
{
@ -126,26 +125,19 @@ namespace CppJieba
bool find(Unicode::const_iterator begin, Unicode::const_iterator end, vector<pair<size_t, const TrieNodeInfo*> >& res) const
{
TrieNodeMap::const_iterator citer;
TrieNode* p = _root;
for (Unicode::const_iterator itr = begin; itr != end; itr++)
{
if(p->hmap.find(*itr) == p-> hmap.end())
citer = p->hmap.find(*itr);
if(p->hmap.end() == citer)
{
break;
}
p = p->hmap[*itr];
p = citer->second;
if(p->isLeaf)
{
size_t pos = p->nodeInfoPos;
if(pos < _nodeInfos.size())
{
res.push_back(make_pair(itr-begin, &_nodeInfos[pos]));
}
else
{
LogFatal("node's nodeInfoPos is out of _nodeInfos's range");
return false;
}
res.push_back(make_pair(itr-begin, &_nodeInfos[p->nodeInfoPos]));
}
}
return !res.empty();
@ -154,25 +146,18 @@ namespace CppJieba
bool find(Unicode::const_iterator begin, Unicode::const_iterator end, size_t offset, DagType & res) const
{
TrieNode* p = _root;
TrieNodeMap::const_iterator citer;
for (Unicode::const_iterator itr = begin; itr != end; itr++)
{
if(p->hmap.find(*itr) == p-> hmap.end())
citer = p->hmap.find(*itr);
if(p->hmap.end() == citer)
{
break;
}
p = p->hmap[*itr];
p = citer->second;
if(p->isLeaf)
{
size_t pos = p->nodeInfoPos;
if(pos < _nodeInfos.size())
{
res[itr-begin + offset] = &_nodeInfos[pos];
}
else
{
LogFatal("node's nodeInfoPos is out of _nodeInfos's range");
return false;
}
res[itr - begin + offset] = &_nodeInfos[p->nodeInfoPos];
}
}
return !res.empty();
@ -184,11 +169,11 @@ namespace CppJieba
private:
void _insert(const TrieNodeInfo& nodeInfo, size_t nodeInfoPos)
{
const Unicode& uintVec = nodeInfo.word;
const Unicode& unico = nodeInfo.word;
TrieNode* p = _root;
for(size_t i = 0; i < uintVec.size(); i++)
for(size_t i = 0; i < unico.size(); i++)
{
uint16_t cu = uintVec[i];
uint16_t cu = unico[i];
assert(p);
if(!isIn(p->hmap, cu))
{