ukui-search/libchinese-segmentation/cppjieba/PosTagger.hpp

85 lines
2.0 KiB
C++
Raw Normal View History

2022-06-18 10:46:14 +08:00
#pragma once
#include "limonp/StringUtil.hpp"
2022-10-21 11:21:35 +08:00
#include "segment-trie/segment-trie.h"
//#include "DictTrie.hpp"
//#include "SegmentTagged.hpp"
2022-06-18 10:46:14 +08:00
namespace cppjieba {
using namespace limonp;
static const char* const POS_M = "m";
static const char* const POS_ENG = "eng";
static const char* const POS_X = "x";
class PosTagger {
public:
PosTagger() {
}
~PosTagger() {
}
bool Tag(const string& src, vector<pair<string, string> >& res, const SegmentTagged& segment) const {
vector<string> CutRes;
segment.CutToStr(src, CutRes);
for (vector<string>::iterator itr = CutRes.begin(); itr != CutRes.end(); ++itr) {
res.push_back(make_pair(*itr, LookupTag(*itr, segment)));
}
return !res.empty();
}
string LookupTag(const string &str, const SegmentTagged& segment) const {
const DictTrie * dict = segment.GetDictTrie();
2022-10-21 11:21:35 +08:00
assert(dict != nullptr);
2022-06-18 10:46:14 +08:00
const auto tmp = dict->Find(str);
2022-10-21 11:21:35 +08:00
if (tmp == nullptr || tmp->GetTag().empty()) {
2022-06-18 10:46:14 +08:00
RuneStrArray runes;
if (!DecodeRunesInString(str, runes)) {
XLOG(ERROR) << "Decode failed.";
return POS_X;
}
return SpecialRule(runes);
} else {
return tmp->GetTag();
}
}
private:
const char* SpecialRule(const RuneStrArray& unicode) const {
size_t m = 0;
size_t eng = 0;
for (size_t i = 0; i < unicode.size() && eng < unicode.size() / 2; i++) {
if (unicode[i].rune < 0x80) {
eng ++;
if ('0' <= unicode[i].rune && unicode[i].rune <= '9') {
m++;
}
}
}
// ascii char is not found
if (eng == 0) {
return POS_X;
}
// all the ascii is number char
if (m == eng) {
return POS_M;
}
// the ascii chars contain english letter
return POS_ENG;
}
}; // class PosTagger
} // namespace cppjieba