Integrate Lexi's LR Score into tuning

lm/read_arpa.cc

@@ -11,6 +11,10 @@
 #include <string.h>
 #include <stdint.h>
 
+#ifdef WIN32
+#include <float.h>
+#endif
+
 namespace lm {
 
 // 1 for '\t', '\n', and ' '.  This is stricter than isspace.  
@@ -95,8 +99,13 @@ void ReadBackoff(util::FilePiece &in, float &backoff) {
       backoff = in.ReadFloat();
       if (backoff == ngram::kExtensionBackoff) backoff = ngram::kNoExtensionBackoff;
       {
+#ifdef WIN32
+		int float_class = _fpclass(backoff);
+        UTIL_THROW_IF(float_class == _FPCLASS_SNAN || float_class == _FPCLASS_QNAN || float_class == _FPCLASS_NINF || float_class == _FPCLASS_PINF, FormatLoadException, "Bad backoff " << backoff);
+#else
         int float_class = fpclassify(backoff);
         UTIL_THROW_IF(float_class == FP_NAN || float_class == FP_INFINITE, FormatLoadException, "Bad backoff " << backoff);
+#endif
       }
       UTIL_THROW_IF(in.get() != '\n', FormatLoadException, "Expected newline after backoff");
       break;

mert/Data.cpp

@@ -140,7 +140,7 @@ void Data::loadNBest(const string &file)
     throw runtime_error("Unable to open: " + file);
 
   ScoreStats scoreentry;
-  string line, sentence_index, sentence, feature_str;
+  string line, sentence_index, sentence, feature_str, alignment;
 
   while (getline(inp, line, '\n')) {
     if (line.empty()) continue;
@@ -151,7 +151,21 @@ void Data::loadNBest(const string &file)
     getNextPound(line, sentence, "|||");       // second field
     getNextPound(line, feature_str, "|||");    // third field
 
+    if (line.length() > 0) {
+      string temp;
+      getNextPound(line, temp, "|||"); //fourth field sentence score
+      if (line.length() > 0) {
+        getNextPound(line, alignment, "|||"); //fourth field only there if alignment scorer
+      }
+    }
+    //TODO check alignment exists if scorers need it
+
+    if (m_scorer->useAlignment()) {
+      sentence += "|||";
+      sentence += alignment;
+    }
     m_scorer->prepareStats(sentence_index, sentence, scoreentry);
+    
     m_score_data->add(scoreentry, sentence_index);
 
     // examine first line for name of features

mert/InterpolatedScorer.cpp

@@ -66,6 +66,17 @@ InterpolatedScorer::InterpolatedScorer(const string& name, const string& config)
   cerr <<endl;
 }
 
+bool InterpolatedScorer::useAlignment() const {
+  //cout << "InterpolatedScorer::useAlignment" << endl;
+  for (vector<Scorer*>::const_iterator itsc =  m_scorers.begin(); itsc < m_scorers.end(); itsc++) {
+    if ((*itsc)->useAlignment()) {
+      //cout <<"InterpolatedScorer::useAlignment Returning true"<<endl;
+      return true;
+    }
+  }
+  return false;
+};
+
 void InterpolatedScorer::setScoreData(ScoreData* data)
 {
   size_t last = 0;
@@ -156,11 +167,23 @@ void InterpolatedScorer::setReferenceFiles(const vector<string>& referenceFiles)
 void InterpolatedScorer::prepareStats(size_t sid, const string& text, ScoreStats& entry)
 {
   stringstream buff;
+  string align = text;
+  string sentence = "";
+  size_t alignmentData = text.find("|||");
+  //Get sentence and alignment parts
+  if(alignmentData != string::npos) {
+    getNextPound(align,sentence, "|||");
+  }
+
   int i = 0;
-  for (ScopedVector<Scorer>::iterator itsc = m_scorers.begin();
-       itsc != m_scorers.end(); ++itsc) {
+  for (ScopedVector<Scorer>::iterator itsc = m_scorers.begin(); itsc != m_scorers.end(); ++itsc) {
     ScoreStats tempEntry;
-    (*itsc)->prepareStats(sid, text, tempEntry);
+    if ((*itsc)->useAlignment()) {
+      (*itsc)->prepareStats(sid, text, tempEntry);
+    }
+    else {
+      (*itsc)->prepareStats(sid, sentence, tempEntry);
+    }
     if (i > 0) buff <<  " ";
     buff << tempEntry;
     i++;

mert/InterpolatedScorer.h

@@ -46,6 +46,8 @@ public:
 
   virtual void setFilter(const std::string& filterCommand);
 
+  bool useAlignment() const;
+
 protected:
   ScopedVector<Scorer> m_scorers;
 

mert/Scorer.h

@@ -103,6 +103,15 @@ class Scorer
     m_score_data = data;
   }
 
+  /**
+   * The scorer returns if it uses the reference alignment data
+   * for permutation distance scores
+   **/
+  virtual bool useAlignment() const {
+    //cout << "Scorer::useAlignment returning false " << endl;
+    return false;
+  };
+
   /**
    * Set the factors, which should be used for this metric
    */

mert/ScorerFactory.cpp

@@ -46,8 +46,8 @@ Scorer* ScorerFactory::getScorer(const string& type, const string& config) {
     return new SemposScorer(config);
   } else if (type == "MERGE") {
     return new MergeScorer(config);
-  } else if (type == "LRSCORE") {
-    return new PermutationScorer(config);
+  } else if ((type == "HAMMING") || (type == "KENDALL")) {
+    return (PermutationScorer*) new PermutationScorer(type, config);
   } else {
     if (type.find(',') != string::npos) {
       return new InterpolatedScorer(type, config);