From 4313e335b534102f18c3d6308b60caa1ffdcfc98 Mon Sep 17 00:00:00 2001
From: hieuhoang1972 <hieuhoang1972@1f5c12ca-751b-0410-a591-d2e778427230>
Date: Fri, 16 Sep 2011 17:13:34 +0000
Subject: [PATCH] print out span widths of non-terms. Extra argument
 --OutputNTLengths

git-svn-id: https://mosesdecoder.svn.sourceforge.net/svnroot/mosesdecoder/trunk@4230 1f5c12ca-751b-0410-a591-d2e778427230
---
 .../training/phrase-extract/consolidate.cpp   | 10 +++++++++-
 scripts/training/phrase-extract/score.cpp     | 19 +++++++++++--------
 2 files changed, 20 insertions(+), 9 deletions(-)

diff --git a/scripts/training/phrase-extract/consolidate.cpp b/scripts/training/phrase-extract/consolidate.cpp
index 8d31a1d27..cb4b96659 100644
--- a/scripts/training/phrase-extract/consolidate.cpp
+++ b/scripts/training/phrase-extract/consolidate.cpp
@@ -40,6 +40,7 @@ bool lowCountFlag = false;
 bool goodTuringFlag = false;
 bool kneserNeyFlag = false;
 bool logProbFlag = false;
+bool outputNTLengths = false;
 inline float maybeLogProb( float a ) { return logProbFlag ? log(a) : a; }
 
 char line[LINE_MAX_LENGTH];
@@ -54,7 +55,7 @@ int main(int argc, char* argv[])
        << "consolidating direct and indirect rule tables\n";
 
   if (argc < 4) {
-    cerr << "syntax: consolidate phrase-table.direct phrase-table.indirect phrase-table.consolidated [--Hierarchical] [--OnlyDirect]\n";
+    cerr << "syntax: consolidate phrase-table.direct phrase-table.indirect phrase-table.consolidated [--Hierarchical] [--OnlyDirect] [--OutputNTLengths] \n";
     exit(1);
   }
   char* &fileNameDirect = argv[1];
@@ -94,6 +95,8 @@ int main(int argc, char* argv[])
     } else if (strcmp(argv[i],"--LogProb") == 0) {
       logProbFlag = true;
       cerr << "using log-probabilities\n";
+    } else if (strcmp(argv[i],"--OutputNTLengths") == 0) {
+      outputNTLengths = true;
     } else {
       cerr << "ERROR: unknown option " << argv[i] << endl;
       exit(1);
@@ -271,6 +274,11 @@ void processFiles( char* fileNameDirect, char* fileNameIndirect, char* fileNameC
     // counts, for debugging
     fileConsolidated << "||| " << countE << " " << countF; // << " " << countEF;
 
+    if (outputNTLengths)
+    {
+      fileConsolidated << " ||| " << itemDirect[5];
+    }
+    
     fileConsolidated << endl;
   }
   fileDirect.Close();
diff --git a/scripts/training/phrase-extract/score.cpp b/scripts/training/phrase-extract/score.cpp
index 00f6dfcf6..5cf71bf91 100644
--- a/scripts/training/phrase-extract/score.cpp
+++ b/scripts/training/phrase-extract/score.cpp
@@ -529,15 +529,18 @@ void outputPhrasePair( vector< PhraseAlignment* > &phrasePair, float totalCount,
   // nt lengths  
   if (outputNTLengths)
   {
-    map<size_t, map<size_t, float> > sourceProb, targetProb;
-    // 1st sourcePos, 2nd = length, 3rd = prob
-
-    calcNTLengthProb(phrasePair, sourceProb, targetProb);
-    
     phraseTableFile << " ||| ";
-    outputNTLengthProbs(phraseTableFile, sourceProb, "S");
-    outputNTLengthProbs(phraseTableFile, targetProb, "T");
-    
+
+    if (!inverseFlag)
+    {
+      map<size_t, map<size_t, float> > sourceProb, targetProb;
+      // 1st sourcePos, 2nd = length, 3rd = prob
+
+      calcNTLengthProb(phrasePair, sourceProb, targetProb);
+      
+      outputNTLengthProbs(phraseTableFile, sourceProb, "S");
+      outputNTLengthProbs(phraseTableFile, targetProb, "T");
+    }    
   }
   
   phraseTableFile << endl;