From 5775a40beedfddd5ae3d52b236c8698ecd6c38fb Mon Sep 17 00:00:00 2001
From: t-k- <clock126@126.com>
Date: Wed, 6 Jul 2016 02:44:56 -0600
Subject: [PATCH] Add LookupTag function for single token tag lookup.

---
 include/cppjieba/Jieba.hpp      |  3 +++
 include/cppjieba/MixSegment.hpp |  4 ++++
 include/cppjieba/PosTagger.hpp  | 19 ++++++++++++-------
 test/demo.cpp                   |  7 +++++++
 4 files changed, 26 insertions(+), 7 deletions(-)

diff --git a/include/cppjieba/Jieba.hpp b/include/cppjieba/Jieba.hpp
index 76f12b6..9c87ce9 100644
--- a/include/cppjieba/Jieba.hpp
+++ b/include/cppjieba/Jieba.hpp
@@ -62,6 +62,9 @@ class Jieba {
   void Tag(const string& sentence, vector<pair<string, string> >& words) const {
     mix_seg_.Tag(sentence, words);
   }
+  string LookupTag(const string &str) const {
+    return mix_seg_.LookupTag(str);
+  }
   bool InsertUserWord(const string& word, const string& tag = UNKNOWN_TAG) {
     return dict_trie_.InsertUserWord(word, tag);
   }
diff --git a/include/cppjieba/MixSegment.hpp b/include/cppjieba/MixSegment.hpp
index 3e18b73..8fd24e9 100644
--- a/include/cppjieba/MixSegment.hpp
+++ b/include/cppjieba/MixSegment.hpp
@@ -93,6 +93,10 @@ class MixSegment: public SegmentTagged {
     return tagger_.Tag(src, res, *this);
   }
 
+  string LookupTag(const string &str) const {
+    return tagger_.LookupTag(str, *this);
+  }
+
  private:
   MPSegment mpSeg_;
   HMMSegment hmmSeg_;
diff --git a/include/cppjieba/PosTagger.hpp b/include/cppjieba/PosTagger.hpp
index 7113297..78853d5 100644
--- a/include/cppjieba/PosTagger.hpp
+++ b/include/cppjieba/PosTagger.hpp
@@ -23,24 +23,29 @@ class PosTagger {
     vector<string> CutRes;
     segment.Cut(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 DictUnit *tmp = NULL;
     RuneStrArray runes;
     const DictTrie * dict = segment.GetDictTrie();
     assert(dict != NULL);
-    for (vector<string>::iterator itr = CutRes.begin(); itr != CutRes.end(); ++itr) {
-      if (!DecodeRunesInString(*itr, runes)) {
+      if (!DecodeRunesInString(str, runes)) {
         XLOG(ERROR) << "Decode failed.";
-        return false;
+        return POS_X;
       }
       tmp = dict->Find(runes.begin(), runes.end());
       if (tmp == NULL || tmp->tag.empty()) {
-        res.push_back(make_pair(*itr, SpecialRule(runes)));
+        return SpecialRule(runes);
       } else {
-        res.push_back(make_pair(*itr, tmp->tag));
+        return tmp->tag;
       }
-    }
-    return !res.empty();
   }
+
  private:
   const char* SpecialRule(const RuneStrArray& unicode) const {
     size_t m = 0;
diff --git a/test/demo.cpp b/test/demo.cpp
index dddd7fc..bae5ef3 100644
--- a/test/demo.cpp
+++ b/test/demo.cpp
@@ -51,6 +51,13 @@ int main(int argc, char** argv) {
   jieba.CutForSearch(s, jiebawords, true);
   cout << jiebawords << endl;
 
+  cout << "[demo] Lookup Tag for Single Token" << endl;
+  vector<pair<string, string> > LookupTagres = {{"拖拉机", ""}, {"CEO", ""}, {".",""}};
+  LookupTagres[0].second = jieba.LookupTag(LookupTagres[0].first);
+  LookupTagres[1].second = jieba.LookupTag(LookupTagres[1].first);
+  LookupTagres[2].second = jieba.LookupTag(LookupTagres[2].first);
+  cout << LookupTagres << endl;;
+
   cout << "[demo] Tagging" << endl;
   vector<pair<string, string> > tagres;
   s = "我是拖拉机学院手扶拖拉机专业的。不用多久，我就会升职加薪，当上CEO，走上人生巅峰。";