增加更详细的错误日志，在初始化过程中合理使用LogFatal。

ChangeLog.md

@@ -1,10 +1,11 @@
 # CppJieba ChangeLog
 
-## v2.4.5 (untagged)
+## v2.5.0 (untagged)
 
 1. 使得 QuerySegment 支持自定义词典（可选参数）。
 2. 使得 KeywordExtractor 支持自定义词典（可选参数）。
 3. 修改 Code Style ，参照 google code style 。 
+4. 增加更详细的错误日志，在初始化过程中合理使用LogFatal。
 
 ## v2.4.4
 

src/DictTrie.hpp

@@ -39,7 +39,7 @@ class DictTrie {
     }
   }
 
-  bool init(const string& dictPath, const string& userDictPath = "") {
+  void init(const string& dictPath, const string& userDictPath = "") {
     if(trie_ != NULL) {
       LogFatal("trie already initted");
     }
@@ -54,7 +54,6 @@ class DictTrie {
     shrink_(nodeInfos_);
     trie_ = createTrie_(nodeInfos_);
     assert(trie_);
-    return true;
   }
 
   const DictUnit* find(Unicode::const_iterator begin, Unicode::const_iterator end) const {

src/HMMSegment.hpp

@@ -6,7 +6,6 @@
 #include <memory.h>
 #include <cassert>
 #include "Limonp/StringUtil.hpp"
-#include "Limonp/Logger.hpp"
 #include "TransCode.hpp"
 #include "ISegment.hpp"
 #include "SegmentBase.hpp"
@@ -26,11 +25,11 @@ class HMMSegment: public SegmentBase {
  public:
   HMMSegment() {}
   explicit HMMSegment(const string& filePath) {
-    LIMONP_CHECK(init(filePath));
+    init(filePath);
   }
   virtual ~HMMSegment() {}
  public:
-  bool init(const string& filePath) {
+  void init(const string& filePath) {
     memset(startProb_, 0, sizeof(startProb_));
     memset(transProb_, 0, sizeof(transProb_));
     statMap_[0] = 'B';
@@ -41,9 +40,8 @@ class HMMSegment: public SegmentBase {
     emitProbVec_.push_back(&emitProbE_);
     emitProbVec_.push_back(&emitProbM_);
     emitProbVec_.push_back(&emitProbS_);
-    LIMONP_CHECK(loadModel_(filePath.c_str()));
+    loadModel_(filePath.c_str());
     LogInfo("HMMSegment init(%s) ok.", filePath.c_str());
-    return true;
   }
  public:
   using SegmentBase::cut;
@@ -212,19 +210,21 @@ class HMMSegment: public SegmentBase {
 
     return true;
   }
-  bool loadModel_(const char* const filePath) {
+  void loadModel_(const char* const filePath) {
     ifstream ifile(filePath);
+    if(!ifile.is_open()) {
+      LogFatal("open %s failed.", filePath);
+    }
     string line;
     vector<string> tmp;
     vector<string> tmp2;
     //load startProb_
     if(!getLine_(ifile, line)) {
-      return false;
+      LogFatal("load startProb_");
     }
     split(line, tmp, " ");
     if(tmp.size() != STATUS_SUM) {
-      LogError("start_p illegal");
+      LogFatal("start_p illegal");
-      return false;
     }
     for(size_t j = 0; j< tmp.size(); j++) {
       startProb_[j] = atof(tmp[j].c_str());
@@ -233,12 +233,11 @@ class HMMSegment: public SegmentBase {
     //load transProb_
     for(size_t i = 0; i < STATUS_SUM; i++) {
       if(!getLine_(ifile, line)) {
-        return false;
+        LogFatal("load transProb_ failed.");
       }
       split(line, tmp, " ");
       if(tmp.size() != STATUS_SUM) {
-        LogError("trans_p illegal");
+        LogFatal("trans_p illegal");
-        return false;
       }
       for(size_t j =0; j < STATUS_SUM; j++) {
         transProb_[i][j] = atof(tmp[j].c_str());
@@ -247,25 +246,23 @@ class HMMSegment: public SegmentBase {
 
     //load emitProbB_
     if(!getLine_(ifile, line) || !loadEmitProb_(line, emitProbB_)) {
-      return false;
+      LogFatal("load emitProbB_ failed.");
     }
 
     //load emitProbE_
     if(!getLine_(ifile, line) || !loadEmitProb_(line, emitProbE_)) {
-      return false;
+      LogFatal("load emitProbE_ failed.");
     }
 
     //load emitProbM_
     if(!getLine_(ifile, line) || !loadEmitProb_(line, emitProbM_)) {
-      return false;
+      LogFatal("load emitProbM_ failed.");
     }
 
     //load emitProbS_
     if(!getLine_(ifile, line) || !loadEmitProb_(line, emitProbS_)) {
-      return false;
+      LogFatal("load emitProbS_ failed.");
     }
-
-    return true;
   }
   bool getLine_(ifstream& ifile, string& line) {
     while(getline(ifile, line)) {

src/KeywordExtractor.hpp

@@ -20,7 +20,7 @@ class KeywordExtractor {
   void init(const string& dictPath, const string& hmmFilePath, const string& idfPath, const string& stopWordPath, const string& userDict = "") {
     loadIdfDict_(idfPath);
     loadStopWordDict_(stopWordPath);
-    LIMONP_CHECK(segment_.init(dictPath, hmmFilePath, userDict));
+    segment_.init(dictPath, hmmFilePath, userDict);
   };
 
   bool extract(const string& str, vector<string>& keywords, size_t topN) const {

src/MPSegment.hpp

@@ -16,14 +16,13 @@ class MPSegment: public SegmentBase {
  public:
   MPSegment() {};
   MPSegment(const string& dictPath, const string& userDictPath = "") {
-    LIMONP_CHECK(init(dictPath, userDictPath));
+    init(dictPath, userDictPath);
   };
   virtual ~MPSegment() {};
 
-  bool init(const string& dictPath, const string& userDictPath = "") {
+  void init(const string& dictPath, const string& userDictPath = "") {
-    LIMONP_CHECK(dictTrie_.init(dictPath, userDictPath));
+    dictTrie_.init(dictPath, userDictPath);
     LogInfo("MPSegment init(%s) ok", dictPath.c_str());
-    return true;
   }
   bool isUserDictSingleChineseWord(const Unicode::value_type & value) const {
     return dictTrie_.isUserDictSingleChineseWord(value);

src/MixSegment.hpp

@@ -12,15 +12,14 @@ class MixSegment: public SegmentBase {
   MixSegment() {
   }
   MixSegment(const string& mpSegDict, const string& hmmSegDict, const string& userDict = "") {
-    LIMONP_CHECK(init(mpSegDict, hmmSegDict, userDict));
+    init(mpSegDict, hmmSegDict, userDict);
   }
   virtual ~MixSegment() {
   }
-  bool init(const string& mpSegDict, const string& hmmSegDict, const string& userDict = "") {
+  void init(const string& mpSegDict, const string& hmmSegDict, const string& userDict = "") {
-    LIMONP_CHECK(mpSeg_.init(mpSegDict, userDict));
+    mpSeg_.init(mpSegDict, userDict);
-    LIMONP_CHECK(hmmSeg_.init(hmmSegDict));
+    hmmSeg_.init(hmmSegDict);
     LogInfo("MixSegment init(%s, %s)", mpSegDict.c_str(), hmmSegDict.c_str());
-    return true;
   }
   using SegmentBase::cut;
   virtual bool cut(Unicode::const_iterator begin, Unicode::const_iterator end, vector<Unicode>& res) const {

src/PosTagger.hpp

@@ -30,7 +30,7 @@ class PosTagger {
     const string& hmmFilePath,
     const string& userDictPath = ""
   ) {
-    LIMONP_CHECK(segment_.init(dictPath, hmmFilePath, userDictPath));
+    segment_.init(dictPath, hmmFilePath, userDictPath);
     dictTrie_ = segment_.getDictTrie();
     LIMONP_CHECK(dictTrie_);
   };

src/QuerySegment.hpp

@@ -21,12 +21,11 @@ class QuerySegment: public SegmentBase {
     init(dict, model, maxWordLen, userDict);
   };
   virtual ~QuerySegment() {};
-  bool init(const string& dict, const string& model, size_t maxWordLen, const string& userDict = "") {
+  void init(const string& dict, const string& model, size_t maxWordLen, const string& userDict = "") {
-    LIMONP_CHECK(mixSeg_.init(dict, model, userDict));
+    mixSeg_.init(dict, model, userDict);
-    LIMONP_CHECK(fullSeg_.init(mixSeg_.getDictTrie()));
+    fullSeg_.init(mixSeg_.getDictTrie());
     assert(maxWordLen);
     maxWordLen_ = maxWordLen;
-    return true;
   }
   using SegmentBase::cut;
   bool cut(Unicode::const_iterator begin, Unicode::const_iterator end, vector<Unicode>& res) const {

test/unittest/TTrie.cpp

@@ -18,7 +18,7 @@ TEST(DictTrieTest, Test1) {
 
   string s1, s2;
   DictTrie trie;
-  ASSERT_TRUE(trie.init(DICT_FILE));
+  trie.init(DICT_FILE);
   ASSERT_LT(trie.getMinWeight() + 15.6479, 0.001);
   string word("来到");
   Unicode uni;