Merge pull request #27 from pks/master

Tidying (soft) syntax features; loo for C++ extractor; updates for dtrain 

38 files changed, 620 insertions, 759 deletions

diff --git a/decoder/Makefile.am b/decoder/Makefile.am
index f02299e6..8280b22c 100644
--- a/decoder/Makefile.am
+++ b/decoder/Makefile.am
@@ -61,12 +61,10 @@ libcdec_a_SOURCES = \
   ff_ruleshape.h \
   ff_sample_fsa.h \
   ff_soft_syntax.h \
-  ff_soft_syntax2.h \
+  ff_soft_syntax_mindist.h \
   ff_source_path.h \
   ff_source_syntax.h \
   ff_source_syntax2.h \
-  ff_source_syntax2_p.h \
-  ff_source_syntax_p.h \
   ff_spans.h \
   ff_tagger.h \
   ff_wordalign.h \
@@ -127,12 +125,10 @@ libcdec_a_SOURCES = \
   ff_rules.cc \
   ff_ruleshape.cc \
   ff_soft_syntax.cc \
-  ff_soft_syntax2.cc \
+  ff_soft_syntax_mindist.cc \
   ff_source_path.cc \
   ff_source_syntax.cc \
   ff_source_syntax2.cc \
-  ff_source_syntax2_p.cc \
-  ff_source_syntax_p.cc \
   ff_spans.cc \
   ff_tagger.cc \
   ff_wordalign.cc \
diff --git a/decoder/cdec_ff.cc b/decoder/cdec_ff.cc
index 09597e87..d586c1d1 100644
--- a/decoder/cdec_ff.cc
+++ b/decoder/cdec_ff.cc
@@ -15,17 +15,11 @@
 #include "ff_ruleshape.h"
 #include "ff_bleu.h"
 #include "ff_soft_syntax.h"
-#include "ff_soft_syntax2.h"
+#include "ff_soft_syntax_mindist.h"
 #include "ff_source_path.h"
-
-
 #include "ff_parse_match.h"
 #include "ff_source_syntax.h"
-#include "ff_source_syntax_p.h"
 #include "ff_source_syntax2.h"
-#include "ff_source_syntax2_p.h"
-
-
 #include "ff_register.h"
 #include "ff_charset.h"
 #include "ff_wordset.h"
@@ -51,23 +45,12 @@ void register_feature_functions() {
   ff_registry.Register("NgramFeatures", new FFFactory<NgramDetector>());
   ff_registry.Register("RuleContextFeatures", new FFFactory<RuleContextFeatures>());
   ff_registry.Register("RuleIdentityFeatures", new FFFactory<RuleIdentityFeatures>());
-
-
   ff_registry.Register("ParseMatchFeatures", new FFFactory<ParseMatchFeatures>);
-
-  ff_registry.Register("SoftSyntacticFeatures", new FFFactory<SoftSyntacticFeatures>);
-  ff_registry.Register("SoftSyntacticFeatures2", new FFFactory<SoftSyntacticFeatures2>);
-
+  ff_registry.Register("SoftSyntaxFeatures", new FFFactory<SoftSyntaxFeatures>);
+  ff_registry.Register("SoftSyntaxFeaturesMindist", new FFFactory<SoftSyntaxFeaturesMindist>);
   ff_registry.Register("SourceSyntaxFeatures", new FFFactory<SourceSyntaxFeatures>);
-  ff_registry.Register("SourceSyntaxFeatures2", new FFFactory<SourceSyntaxFeatures2>);
-
   ff_registry.Register("SourceSpanSizeFeatures", new FFFactory<SourceSpanSizeFeatures>);
-
-  //ff_registry.Register("PSourceSyntaxFeatures", new FFFactory<PSourceSyntaxFeatures>);
-  //ff_registry.Register("PSourceSpanSizeFeatures", new FFFactory<PSourceSpanSizeFeatures>);
-  //ff_registry.Register("PSourceSyntaxFeatures2", new FFFactory<PSourceSyntaxFeatures2>);
-
-
+  ff_registry.Register("SourceSyntaxFeatures2", new FFFactory<SourceSyntaxFeatures2>);
   ff_registry.Register("CMR2008ReorderingFeatures", new FFFactory<CMR2008ReorderingFeatures>());
   ff_registry.Register("RuleSourceBigramFeatures", new FFFactory<RuleSourceBigramFeatures>());
   ff_registry.Register("RuleTargetBigramFeatures", new FFFactory<RuleTargetBigramFeatures>());
diff --git a/decoder/ff_parse_match.cc b/decoder/ff_parse_match.cc
index ed556b91..58026975 100644
--- a/decoder/ff_parse_match.cc
+++ b/decoder/ff_parse_match.cc
@@ -42,10 +42,8 @@ struct ParseMatchFeaturesImpl {
 
   void InitializeGrids(const string& tree, unsigned src_len) {
     assert(tree.size() > 0);
-    //fids_cat.clear();
     fids_ef.clear();
     src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
     fids_ef.resize(src_len, src_len + 1);
     src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
     ParseTreeString(tree, src_len);
@@ -112,7 +110,7 @@ struct ParseMatchFeaturesImpl {
     int fid_ef = FD::Convert("PM");
     int min_dist; // minimal distance to next syntactic constituent of this rule's LHS
     int summed_min_dists; // minimal distances of LHS and NTs summed up
-    if (TD::Convert(lhs).compare("XX") != 0) 
+    if (TD::Convert(lhs).compare("XX") != 0)
         min_dist= 0;
     // compute the distance to the next syntactical constituent
     else {
@@ -131,7 +129,7 @@ struct ParseMatchFeaturesImpl {
             ok = 1;
             break;
           }
-          // check if removing k words from the rule span will 
+          // check if removing k words from the rule span will
           // lead to a syntactical constituent
           else {
             //cerr << "Hilfe...!" << endl;
@@ -144,7 +142,7 @@ struct ParseMatchFeaturesImpl {
               ok = 1;
               break;
             }
-          }    
+          }
         }
         if (ok) break;
       }
@@ -183,9 +181,9 @@ struct ParseMatchFeaturesImpl {
     return min_dist;
   }
 
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
+  Array2D<WordID> src_tree; // src_tree(i,j) NT = type
   unsigned int src_sent_len;
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
+  mutable Array2D<map<const TRule*, int> > fids_ef; // fires for fully lexicalized
   int scoring_method;
 };
 
@@ -214,5 +212,9 @@ void ParseMatchFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
 }
 
 void ParseMatchFeatures::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
+  ReadFile f = ReadFile(smeta.GetSGMLValue("src_tree"));
+  string tree;
+  f.ReadAll(tree);
+  impl->InitializeGrids(tree, smeta.GetSourceLength());
 }
+
diff --git a/decoder/ff_parse_match.h b/decoder/ff_parse_match.h
index fa73481a..7820b418 100644
--- a/decoder/ff_parse_match.h
+++ b/decoder/ff_parse_match.h
@@ -23,3 +23,4 @@ class ParseMatchFeatures : public FeatureFunction {
 };
 
 #endif
+
diff --git a/decoder/ff_soft_syntax.cc b/decoder/ff_soft_syntax.cc
index 9981fa45..23fe87bd 100644
--- a/decoder/ff_soft_syntax.cc
+++ b/decoder/ff_soft_syntax.cc
@@ -13,16 +13,15 @@
 
 using namespace std;
 
-// Implements the soft syntactic features described in 
+// Implements the soft syntactic features described in
 // Marton and Resnik (2008): "Soft Syntacitc Constraints for Hierarchical Phrase-Based Translation".
 // Source trees must be represented in Penn Treebank format,
 // e.g. (S (NP John) (VP (V left))).
 
-struct SoftSyntacticFeaturesImpl {
-  SoftSyntacticFeaturesImpl(const string& param) {
+struct SoftSyntaxFeaturesImpl {
+  SoftSyntaxFeaturesImpl(const string& param) {
     vector<string> labels = SplitOnWhitespace(param);
-	for (unsigned int i = 0; i < labels.size(); i++) 
-      //cerr << "Labels: " << labels.at(i) << endl;
+  	//for (unsigned int i = 0; i < labels.size(); i++) { cerr << "Labels: " << labels.at(i) << endl; }
     for (unsigned int i = 0; i < labels.size(); i++) {
       string label = labels.at(i);
       pair<string, string> feat_label;
@@ -34,10 +33,8 @@ struct SoftSyntacticFeaturesImpl {
 
   void InitializeGrids(const string& tree, unsigned src_len) {
     assert(tree.size() > 0);
-    //fids_cat.clear();
     fids_ef.clear();
     src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
     fids_ef.resize(src_len, src_len + 1);
     src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
     ParseTreeString(tree, src_len);
@@ -99,7 +96,7 @@ struct SoftSyntacticFeaturesImpl {
     const WordID lhs = src_tree(i,j);
 	string lhs_str = TD::Convert(lhs);
     //cerr << "LHS: " << lhs_str << " from " << i << " to " << j << endl;
-	//cerr << "RULE :"<< rule << endl;
+	  //cerr << "RULE :"<< rule << endl;
     int& fid_ef = fids_ef(i,j)[&rule];
     for (unsigned int i = 0; i < feat_labels.size(); i++) {
       ostringstream os;
@@ -110,10 +107,10 @@ struct SoftSyntacticFeaturesImpl {
       switch(feat_type) {
         case '2':
           if (lhs_str.compare(label) == 0) {
-            os << "SYN:" << label << "_conform";
+            os << "SOFT:" << label << "_conform";
           }
           else {
-            os << "SYN:" << label << "_cross";
+            os << "SOFT:" << label << "_cross";
           }
           fid_ef = FD::Convert(os.str());
           if (fid_ef > 0) {
@@ -122,11 +119,11 @@ struct SoftSyntacticFeaturesImpl {
           }
           break;
         case '_':
-          os << "SYN:" << label;
+          os << "SOFT:" << label;
           fid_ef = FD::Convert(os.str());
           if (lhs_str.compare(label) == 0) {
             if (fid_ef > 0) {
-            //cerr << "Feature: " << os.str() << endl;
+              //cerr << "Feature: " << os.str() << endl;
               feats->set_value(fid_ef, 1.0);
             }
           }
@@ -139,7 +136,7 @@ struct SoftSyntacticFeaturesImpl {
           break;
         case '+':
           if (lhs_str.compare(label) == 0) {
-            os << "SYN:" << label << "_conform";
+            os << "SOFT:" << label << "_conform";
             fid_ef = FD::Convert(os.str());
             if (fid_ef > 0) {
               //cerr << "Feature: " << os.str() << endl;
@@ -147,10 +144,10 @@ struct SoftSyntacticFeaturesImpl {
             }
           }
           break;
-        case '-': 
-          //cerr << "-" << endl;           
+        case '-':
+          //cerr << "-" << endl;
           if (lhs_str.compare(label) != 0) {
-            os << "SYN:" << label << "_cross";
+            os << "SOFT:" << label << "_cross";
             fid_ef = FD::Convert(os.str());
             if (fid_ef > 0) {
               //cerr << "Feature :" << os.str() << endl;
@@ -167,22 +164,22 @@ struct SoftSyntacticFeaturesImpl {
     return lhs;
   }
 
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
+  Array2D<WordID> src_tree; // src_tree(i,j) NT = type
+  mutable Array2D<map<const TRule*, int> > fids_ef; // fires for fully lexicalized
   vector<pair<string, string> > feat_labels;
 };
 
-SoftSyntacticFeatures::SoftSyntacticFeatures(const string& param) :
+SoftSyntaxFeatures::SoftSyntaxFeatures(const string& param) :
     FeatureFunction(sizeof(WordID)) {
-  impl = new SoftSyntacticFeaturesImpl(param);
+  impl = new SoftSyntaxFeaturesImpl(param);
 }
 
-SoftSyntacticFeatures::~SoftSyntacticFeatures() {
+SoftSyntaxFeatures::~SoftSyntaxFeatures() {
   delete impl;
   impl = NULL;
 }
 
-void SoftSyntacticFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
+void SoftSyntaxFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
                                      const Hypergraph::Edge& edge,
                                      const vector<const void*>& ant_contexts,
                                      SparseVector<double>* features,
@@ -196,6 +193,10 @@ void SoftSyntacticFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
      impl->FireFeatures(*edge.rule_, edge.i_, edge.j_, ants, features);
 }
 
-void SoftSyntacticFeatures::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
+void SoftSyntaxFeatures::PrepareForInput(const SentenceMetadata& smeta) {
+  ReadFile f = ReadFile(smeta.GetSGMLValue("src_tree"));
+  string tree;
+  f.ReadAll(tree);
+  impl->InitializeGrids(tree, smeta.GetSourceLength());
 }
+
diff --git a/decoder/ff_soft_syntax.h b/decoder/ff_soft_syntax.h
index 79352f49..e71825d5 100644
--- a/decoder/ff_soft_syntax.h
+++ b/decoder/ff_soft_syntax.h
@@ -1,15 +1,15 @@
-#ifndef _FF_SOFTSYNTAX_H_
-#define _FF_SOFTSYNTAX_H_
+#ifndef _FF_SOFT_SYNTAX_H_
+#define _FF_SOFT_SYNTAX_H_
 
 #include "ff.h"
 #include "hg.h"
 
-struct SoftSyntacticFeaturesImpl;
+struct SoftSyntaxFeaturesImpl;
 
-class SoftSyntacticFeatures : public FeatureFunction {
+class SoftSyntaxFeatures : public FeatureFunction {
  public:
-  SoftSyntacticFeatures(const std::string& param);
-  ~SoftSyntacticFeatures();
+  SoftSyntaxFeatures(const std::string& param);
+  ~SoftSyntaxFeatures();
  protected:
   virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
                                      const Hypergraph::Edge& edge,
@@ -19,9 +19,9 @@ class SoftSyntacticFeatures : public FeatureFunction {
                                      void* context) const;
   virtual void PrepareForInput(const SentenceMetadata& smeta);
  private:
-  SoftSyntacticFeaturesImpl* impl;
+  SoftSyntaxFeaturesImpl* impl;
 };
 
 
-
 #endif
+
diff --git a/decoder/ff_soft_syntax2.cc b/decoder/ff_soft_syntax_mindist.cc
index 121bc39b..a23f70f8 100644
--- a/decoder/ff_soft_syntax2.cc
+++ b/decoder/ff_soft_syntax_mindist.cc
@@ -1,4 +1,4 @@
-#include "ff_soft_syntax2.h"
+#include "ff_soft_syntax_mindist.h"
 
 #include <cstdio>
 #include <sstream>
@@ -13,16 +13,18 @@
 
 using namespace std;
 
-// Implements the soft syntactic features described in 
+// Implements the soft syntactic features described in
 // Marton and Resnik (2008): "Soft Syntacitc Constraints for Hierarchical Phrase-Based Translation".
 // Source trees must be represented in Penn Treebank format,
 // e.g. (S (NP John) (VP (V left))).
+//
+// This variant accepts fuzzy matches, choosing the constituent with
+// minimum distance.
 
-struct SoftSyntacticFeatures2Impl {
-  SoftSyntacticFeatures2Impl(const string& param) {
+struct SoftSyntaxFeaturesMindistImpl {
+  SoftSyntaxFeaturesMindistImpl(const string& param) {
     vector<string> labels = SplitOnWhitespace(param);
-	//for (unsigned int i = 0; i < labels.size(); i++) 
-      //cerr << "Labels: " << labels.at(i) << endl;
+	  //for (unsigned int i = 0; i < labels.size(); i++) { cerr << "Labels: " << labels.at(i) << endl; }
     for (unsigned int i = 0; i < labels.size(); i++) {
       string label = labels.at(i);
       pair<string, string> feat_label;
@@ -30,14 +32,12 @@ struct SoftSyntacticFeatures2Impl {
       feat_label.second = label.at(label.size() - 1);
       feat_labels.push_back(feat_label);
     }
-  } 
+  }
 
   void InitializeGrids(const string& tree, unsigned src_len) {
     assert(tree.size() > 0);
-    //fids_cat.clear();
     fids_ef.clear();
     src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
     fids_ef.resize(src_len, src_len + 1);
     src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
     ParseTreeString(tree, src_len);
@@ -99,14 +99,14 @@ struct SoftSyntacticFeatures2Impl {
     const WordID lhs = src_tree(i,j);
 	string lhs_str = TD::Convert(lhs);
     //cerr << "LHS: " << lhs_str << " from " << i << " to " << j << endl;
-	//cerr << "RULE :"<< rule << endl;
+	  //cerr << "RULE :"<< rule << endl;
     int& fid_ef = fids_ef(i,j)[&rule];
     string lhs_to_str = TD::Convert(lhs);
     int min_dist;
     string min_dist_label;
     if (lhs_to_str.compare("XX") != 0) {
       min_dist = 0;
-      min_dist_label = lhs_to_str;      
+      min_dist_label = lhs_to_str;
     }
     else {
       int ok = 0;
@@ -128,7 +128,7 @@ struct SoftSyntacticFeatures2Impl {
               min_dist_label = (TD::Convert(src_tree(l_rem, r_rem)));
               break;
             }
-          }    
+          }
         }
         if (ok) break;
       }
@@ -146,10 +146,10 @@ struct SoftSyntacticFeatures2Impl {
         case '2':
           if (min_dist_label.compare(label) == 0) {
             if (min_dist == 0) {
-              os << "SYN:" << label << "_conform";
+              os << "SOFTM:" << label << "_conform";
             }
             else {
-              os << "SYN:" << label << "_cross";
+              os << "SOFTM:" << label << "_cross";
             }
             fid_ef = FD::Convert(os.str());
             //cerr << "Feature :" << os.str() << endl;
@@ -157,7 +157,7 @@ struct SoftSyntacticFeatures2Impl {
           }
           break;
         case '_':
-          os << "SYN:" << label;
+          os << "SOFTM:" << label;
           fid_ef = FD::Convert(os.str());
           if (min_dist_label.compare(label) == 0) {
             //cerr << "Feature: " << os.str() << endl;
@@ -172,7 +172,7 @@ struct SoftSyntacticFeatures2Impl {
           break;
         case '+':
           if (min_dist_label.compare(label) == 0) {
-            os << "SYN:" << label << "_conform";
+            os << "SOFTM:" << label << "_conform";
             fid_ef = FD::Convert(os.str());
             if (min_dist == 0) {
               //cerr << "Feature: " << os.str() << endl;
@@ -180,10 +180,10 @@ struct SoftSyntacticFeatures2Impl {
             }
           }
           break;
-        case '-': 
-          //cerr << "-" << endl;           
+        case '-':
+          //cerr << "-" << endl;
           if (min_dist_label.compare(label) != 0) {
-            os << "SYN:" << label << "_cross";
+            os << "SOFTM:" << label << "_cross";
             fid_ef = FD::Convert(os.str());
             if (min_dist > 0) {
               //cerr << "Feature :" << os.str() << endl;
@@ -200,22 +200,22 @@ struct SoftSyntacticFeatures2Impl {
     return lhs;
   }
 
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
+  Array2D<WordID> src_tree; // src_tree(i,j) NT = type
+  mutable Array2D<map<const TRule*, int> > fids_ef; // fires for fully lexicalized
   vector<pair<string, string> > feat_labels;
 };
 
-SoftSyntacticFeatures2::SoftSyntacticFeatures2(const string& param) :
+SoftSyntaxFeaturesMindist::SoftSyntaxFeaturesMindist(const string& param) :
     FeatureFunction(sizeof(WordID)) {
-  impl = new SoftSyntacticFeatures2Impl(param);
+  impl = new SoftSyntaxFeaturesMindistImpl(param);
 }
 
-SoftSyntacticFeatures2::~SoftSyntacticFeatures2() {
+SoftSyntaxFeaturesMindist::~SoftSyntaxFeaturesMindist() {
   delete impl;
   impl = NULL;
 }
 
-void SoftSyntacticFeatures2::TraversalFeaturesImpl(const SentenceMetadata& smeta,
+void SoftSyntaxFeaturesMindist::TraversalFeaturesImpl(const SentenceMetadata& smeta,
                                      const Hypergraph::Edge& edge,
                                      const vector<const void*>& ant_contexts,
                                      SparseVector<double>* features,
@@ -229,6 +229,10 @@ void SoftSyntacticFeatures2::TraversalFeaturesImpl(const SentenceMetadata& smeta
      impl->FireFeatures(*edge.rule_, edge.i_, edge.j_, ants, features);
 }
 
-void SoftSyntacticFeatures2::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
+void SoftSyntaxFeaturesMindist::PrepareForInput(const SentenceMetadata& smeta) {
+  ReadFile f = ReadFile(smeta.GetSGMLValue("src_tree"));
+  string tree;
+  f.ReadAll(tree);
+  impl->InitializeGrids(tree, smeta.GetSourceLength());
 }
+
diff --git a/decoder/ff_soft_syntax2.h b/decoder/ff_soft_syntax_mindist.h
index 4de91d86..bf938b38 100644
--- a/decoder/ff_soft_syntax2.h
+++ b/decoder/ff_soft_syntax_mindist.h
@@ -1,15 +1,15 @@
-#ifndef _FF_SOFTSYNTAX2_H_
-#define _FF_SOFTSYNTAX2_H_
+#ifndef _FF_SOFT_SYNTAX_MINDIST_H_
+#define _FF_SOFT_SYNTAX_MINDIST_H_
 
 #include "ff.h"
 #include "hg.h"
 
-struct SoftSyntacticFeatures2Impl;
+struct SoftSyntaxFeaturesMindistImpl;
 
-class SoftSyntacticFeatures2 : public FeatureFunction {
+class SoftSyntaxFeaturesMindist : public FeatureFunction {
  public:
-  SoftSyntacticFeatures2(const std::string& param);
-  ~SoftSyntacticFeatures2();
+  SoftSyntaxFeaturesMindist(const std::string& param);
+  ~SoftSyntaxFeaturesMindist();
  protected:
   virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
                                      const Hypergraph::Edge& edge,
@@ -19,9 +19,9 @@ class SoftSyntacticFeatures2 : public FeatureFunction {
                                      void* context) const;
   virtual void PrepareForInput(const SentenceMetadata& smeta);
  private:
-  SoftSyntacticFeatures2Impl* impl;
+  SoftSyntaxFeaturesMindistImpl* impl;
 };
 
 
-
 #endif
+
diff --git a/decoder/ff_source_syntax.cc b/decoder/ff_source_syntax.cc
index 88f6714c..6b183863 100644
--- a/decoder/ff_source_syntax.cc
+++ b/decoder/ff_source_syntax.cc
@@ -9,7 +9,6 @@
 namespace std { using std::tr1::unordered_set; }
 #endif
 
-#include "hg.h"
 #include "sentence_metadata.h"
 #include "array2d.h"
 #include "filelib.h"
@@ -30,6 +29,17 @@ inline int SpanSizeTransform(unsigned span_size) {
 struct SourceSyntaxFeaturesImpl {
   SourceSyntaxFeaturesImpl() {}
 
+  SourceSyntaxFeaturesImpl(const string& param) {
+    if (!(param.compare("") == 0)) {
+      string triggered_features_fn = param;
+      ReadFile triggered_features(triggered_features_fn);
+      string in;
+      while(getline(*triggered_features, in)) {
+        feature_filter.insert(FD::Convert(in));
+      }
+    }
+  }
+
   void InitializeGrids(const string& tree, unsigned src_len) {
     assert(tree.size() > 0);
     //fids_cat.clear();
@@ -99,7 +109,7 @@ struct SourceSyntaxFeaturesImpl {
     if (fid_ef <= 0) {
       ostringstream os;
       //ostringstream os2;
-      os << "SYN:" << TD::Convert(lhs);
+      os << "SSYN:" << TD::Convert(lhs);
       //os2 << "SYN:" << TD::Convert(lhs) << '_' << SpanSizeTransform(j - i);
       //fid_cat = FD::Convert(os2.str());
       os << ':';
@@ -124,21 +134,28 @@ struct SourceSyntaxFeaturesImpl {
       }
       fid_ef = FD::Convert(os.str());
     }
-    //if (fid_cat > 0)
-    //  feats->set_value(fid_cat, 1.0);
-    if (fid_ef > 0)
-      feats->set_value(fid_ef, 1.0);
+    if (fid_ef > 0) {
+      if (feature_filter.size()>0) {
+        if (feature_filter.find(fid_ef) != feature_filter.end()) {
+          feats->set_value(fid_ef, 1.0);
+        }
+      } else {
+        feats->set_value(fid_ef, 1.0);
+      }
+    }
+    cerr << FD::Convert(fid_ef) << endl;
     return lhs;
   }
 
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  // mutable Array2D<int> fids_cat;   // this tends to overfit baddly
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
+  Array2D<WordID> src_tree; // src_tree(i,j) NT = type
+  // mutable Array2D<int> fids_cat; // this tends to overfit baddly
+  mutable Array2D<map<const TRule*, int> > fids_ef; // fires for fully lexicalized
+  unordered_set<int> feature_filter;
 };
 
 SourceSyntaxFeatures::SourceSyntaxFeatures(const string& param) :
     FeatureFunction(sizeof(WordID)) {
-  impl = new SourceSyntaxFeaturesImpl;
+  impl = new SourceSyntaxFeaturesImpl(param);
 }
 
 SourceSyntaxFeatures::~SourceSyntaxFeatures() {
@@ -161,7 +178,10 @@ void SourceSyntaxFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
 }
 
 void SourceSyntaxFeatures::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
+  ReadFile f = ReadFile(smeta.GetSGMLValue("src_tree"));
+  string tree;
+  f.ReadAll(tree);
+  impl->InitializeGrids(tree, smeta.GetSourceLength());
 }
 
 struct SourceSpanSizeFeaturesImpl {
@@ -236,4 +256,3 @@ void SourceSpanSizeFeatures::PrepareForInput(const SentenceMetadata& smeta) {
   impl->InitializeGrids(smeta.GetSourceLength());
 }
 
-
diff --git a/decoder/ff_source_syntax.h b/decoder/ff_source_syntax.h
index a8c7150a..bdd638c1 100644
--- a/decoder/ff_source_syntax.h
+++ b/decoder/ff_source_syntax.h
@@ -1,7 +1,8 @@
-#ifndef _FF_SOURCE_TOOLS_H_
-#define _FF_SOURCE_TOOLS_H_
+#ifndef _FF_SOURCE_SYNTAX_H_
+#define _FF_SOURCE_SYNTAX_H_
 
 #include "ff.h"
+#include "hg.h"
 
 struct SourceSyntaxFeaturesImpl;
 
@@ -11,7 +12,7 @@ class SourceSyntaxFeatures : public FeatureFunction {
   ~SourceSyntaxFeatures();
  protected:
   virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const HG::Edge& edge,
+                                     const Hypergraph::Edge& edge,
                                      const std::vector<const void*>& ant_contexts,
                                      SparseVector<double>* features,
                                      SparseVector<double>* estimated_features,
@@ -28,7 +29,7 @@ class SourceSpanSizeFeatures : public FeatureFunction {
   ~SourceSpanSizeFeatures();
  protected:
   virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const HG::Edge& edge,
+                                     const Hypergraph::Edge& edge,
                                      const std::vector<const void*>& ant_contexts,
                                      SparseVector<double>* features,
                                      SparseVector<double>* estimated_features,
@@ -39,3 +40,4 @@ class SourceSpanSizeFeatures : public FeatureFunction {
 };
 
 #endif
+
diff --git a/decoder/ff_source_syntax2.cc b/decoder/ff_source_syntax2.cc
index 622c6908..a97e31d8 100644
--- a/decoder/ff_source_syntax2.cc
+++ b/decoder/ff_source_syntax2.cc
@@ -16,7 +16,7 @@ using namespace std;
 
 struct SourceSyntaxFeatures2Impl {
   SourceSyntaxFeatures2Impl(const string& param) {
-    if (!(param.compare("") == 0)) {
+    if (param.compare("") != 0) {
       string triggered_features_fn = param;
       ReadFile triggered_features(triggered_features_fn);
       string in;
@@ -28,10 +28,8 @@ struct SourceSyntaxFeatures2Impl {
 
   void InitializeGrids(const string& tree, unsigned src_len) {
     assert(tree.size() > 0);
-    //fids_cat.clear();
     fids_ef.clear();
     src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
     fids_ef.resize(src_len, src_len + 1);
     src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
     ParseTreeString(tree, src_len);
@@ -39,7 +37,7 @@ struct SourceSyntaxFeatures2Impl {
 
   void ParseTreeString(const string& tree, unsigned src_len) {
     //cerr << "TREE: " << tree << endl;
-    stack<pair<int, WordID> > stk;  // first = i, second = category
+    stack<pair<int, WordID> > stk; // first = i, second = category
     pair<int, WordID> cur_cat; cur_cat.first = -1;
     unsigned i = 0;
     unsigned p = 0;
@@ -91,7 +89,7 @@ struct SourceSyntaxFeatures2Impl {
     const WordID lhs = src_tree(i,j);
     int& fid_ef = fids_ef(i,j)[&rule];
     ostringstream os;
-    os << "SYN:" << TD::Convert(lhs);
+    os << "SSYN2:" << TD::Convert(lhs);
     os << ':';
     unsigned ntc = 0;
     for (unsigned k = 0; k < rule.f_.size(); ++k) {
@@ -99,7 +97,7 @@ struct SourceSyntaxFeatures2Impl {
       if (k > 0 && fj <= 0) os << '_';
       if (fj <= 0) {
         os << '[' << TD::Convert(ants[ntc++]) << ']';
-      } /*else {
+      }/*else {
         os << TD::Convert(fj);
       }*/
     }
@@ -115,16 +113,22 @@ struct SourceSyntaxFeatures2Impl {
     fid_ef = FD::Convert(os.str());
     //cerr << "FEATURE: " << os.str() << endl;
     //cerr << "FID_EF: " << fid_ef << endl;
-    if (feature_filter.find(fid_ef) != feature_filter.end()) {
-      cerr << "SYN-Feature was trigger more than once on training set." << endl;
+    if (feature_filter.size() > 0) {
+      if (feature_filter.find(fid_ef) != feature_filter.end()) {
+        //cerr << "SYN-Feature was trigger more than once on training set." << endl;
+        feats->set_value(fid_ef, 1.0);
+      }
+      //else cerr << "SYN-Feature was triggered less than once on training set." << endli;
+    }
+    else {
       feats->set_value(fid_ef, 1.0);
     }
-    else cerr << "SYN-Feature was triggered less than once on training set." << endl;
+    cerr << FD::Convert(fid_ef) << endl;
     return lhs;
   }
 
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
+  Array2D<WordID> src_tree; // src_tree(i,j) NT = type
+  mutable Array2D<map<const TRule*, int> > fids_ef; // fires for fully lexicalized
   unordered_set<int> feature_filter;
 };
 
@@ -153,5 +157,9 @@ void SourceSyntaxFeatures2::TraversalFeaturesImpl(const SentenceMetadata& smeta,
 }
 
 void SourceSyntaxFeatures2::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
+  ReadFile f = ReadFile(smeta.GetSGMLValue("src_tree"));
+  string tree;
+  f.ReadAll(tree);
+  impl->InitializeGrids(tree, smeta.GetSourceLength());
 }
+
diff --git a/decoder/ff_source_syntax2.h b/decoder/ff_source_syntax2.h
index b6b7dc3d..f606c2bf 100644
--- a/decoder/ff_source_syntax2.h
+++ b/decoder/ff_source_syntax2.h
@@ -1,5 +1,5 @@
-#ifndef _FF_SOURCE_TOOLS2_H_
-#define _FF_SOURCE_TOOLS2_H_
+#ifndef _FF_SOURCE_SYNTAX2_H_
+#define _FF_SOURCE_SYNTAX2_H_
 
 #include "ff.h"
 #include "hg.h"
@@ -23,3 +23,4 @@ class SourceSyntaxFeatures2 : public FeatureFunction {
 };
 
 #endif
+
diff --git a/decoder/ff_source_syntax2_p.cc b/decoder/ff_source_syntax2_p.cc
deleted file mode 100644
index 6a2ae742..00000000
--- a/decoder/ff_source_syntax2_p.cc
+++ /dev/null
@@ -1,170 +0,0 @@
-#include "ff_source_syntax2_p.h"
-
-#include <sstream>
-#include <stack>
-#include <string>
-#ifndef HAVE_OLD_CPP
-# include <unordered_set>
-#else
-# include <tr1/unordered_set>
-namespace std { using std::tr1::unordered_set; }
-#endif
-
-#include "sentence_metadata.h"
-#include "array2d.h"
-#include "filelib.h"
-
-using namespace std;
-
-// implements the source side syntax features described in Blunsom et al. (EMNLP 2008)
-// source trees must be represented in Penn Treebank format, e.g.
-//     (S (NP John) (VP (V left)))
-
-struct PSourceSyntaxFeatures2Impl {
-  PSourceSyntaxFeatures2Impl(const string& param) {
-    if (param.compare("") != 0) {
-      string triggered_features_fn = param;
-      ReadFile triggered_features(triggered_features_fn);
-      string in;
-      while(getline(*triggered_features, in)) {
-        feature_filter.insert(FD::Convert(in));
-      }
-    }
-    /*cerr << "find(\"One\") == " << boolalpha << (table.find("One") != table.end()) << endl; 
-    cerr << "find(\"Three\") == " << boolalpha << (table.find("Three") != table.end()) << endl;*/
-  }
-
-  void InitializeGrids(const string& tree, unsigned src_len) {
-    assert(tree.size() > 0);
-    //fids_cat.clear();
-    fids_ef.clear();
-    src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
-    fids_ef.resize(src_len, src_len + 1);
-    src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
-    ParseTreeString(tree, src_len);
-  }
-
-  void ParseTreeString(const string& tree, unsigned src_len) {
-    //cerr << "TREE: " << tree << endl;
-    stack<pair<int, WordID> > stk;  // first = i, second = category
-    pair<int, WordID> cur_cat; cur_cat.first = -1;
-    unsigned i = 0;
-    unsigned p = 0;
-    while(p < tree.size()) {
-      const char cur = tree[p];
-      if (cur == '(') {
-        stk.push(cur_cat);
-        ++p;
-        unsigned k = p + 1;
-        while (k < tree.size() && tree[k] != ' ') { ++k; }
-        cur_cat.first = i;
-        cur_cat.second = TD::Convert(tree.substr(p, k - p));
-        // cerr << "NT: '" << tree.substr(p, k-p) << "' (i=" << i << ")\n";
-        p = k + 1;
-      } else if (cur == ')') {
-        unsigned k = p;
-        while (k < tree.size() && tree[k] == ')') { ++k; }
-        const unsigned num_closes = k - p;
-        for (unsigned ci = 0; ci < num_closes; ++ci) {
-          src_tree(cur_cat.first, i) = cur_cat.second;
-          cur_cat = stk.top();
-          stk.pop();
-        }
-        p = k;
-        while (p < tree.size() && (tree[p] == ' ' || tree[p] == '\t')) { ++p; }
-      } else if (cur == ' ' || cur == '\t') {
-        cerr << "Unexpected whitespace in: " << tree << endl;
-        abort();
-      } else { // terminal symbol
-        unsigned k = p + 1;
-        do {
-          while (k < tree.size() && tree[k] != ')' && tree[k] != ' ') { ++k; }
-          // cerr << "TERM: '" << tree.substr(p, k-p) << "' (i=" << i << ")\n";
-          ++i;
-          assert(i <= src_len);
-          while (k < tree.size() && tree[k] == ' ') { ++k; }
-          p = k;
-        } while (p < tree.size() && tree[p] != ')');
-      }
-    //cerr << "i=" << i << "  src_len=" << src_len << endl;
-    }
-    //cerr << "i=" << i << "  src_len=" << src_len << endl;
-    assert(i == src_len);  // make sure tree specified in src_tree is
-                           // the same length as the source sentence
-  }
-
-  WordID FireFeatures(const TRule& rule, const int i, const int j, const WordID* ants, SparseVector<double>* feats) {
-    //cerr << "fire features: " << rule.AsString() << " for " << i << "," << j << endl;
-    const WordID lhs = src_tree(i,j);
-    int& fid_ef = fids_ef(i,j)[&rule];
-    ostringstream os;
-    os << "SYN:" << TD::Convert(lhs);
-    os << ':';
-    unsigned ntc = 0;
-    for (unsigned k = 0; k < rule.f_.size(); ++k) {
-      int fj = rule.f_[k];
-      if (k > 0 && fj <= 0) os << '_';
-      if (fj <= 0) {
-        os << '[' << TD::Convert(ants[ntc++]) << ']';
-      } /*else {
-        os << TD::Convert(fj);
-      }*/
-    }
-    os << ':';
-    for (unsigned k = 0; k < rule.e_.size(); ++k) {
-      const int ei = rule.e_[k];
-      if (k > 0) os << '_';
-      if (ei <= 0)
-        os << '[' << (1-ei) << ']';
-      else
-        os << TD::Convert(ei);
-    }
-    fid_ef = FD::Convert(os.str());
-    //cerr << "FEATURE: " << os.str() << endl;
-    //cerr << "FID_EF: " << fid_ef << endl;
-    if (feature_filter.size() > 0) {
-      if (feature_filter.find(fid_ef) != feature_filter.end()) {
-        //cerr << "SYN-Feature was trigger more than once on training set." << endl;
-        feats->set_value(fid_ef, 1.0);
-      }  
-      //else cerr << "SYN-Feature was triggered less than once on training set." << endli;
-    }
-    else {
-      feats->set_value(fid_ef, 1.0);
-    }
-    return lhs;
-  }
-
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
-  unordered_set<int> feature_filter;
-};
-
-PSourceSyntaxFeatures2::PSourceSyntaxFeatures2(const string& param) :
-    FeatureFunction(sizeof(WordID)) {
-  impl = new PSourceSyntaxFeatures2Impl(param);
-}
-
-PSourceSyntaxFeatures2::~PSourceSyntaxFeatures2() {
-  delete impl;
-  impl = NULL;
-}
-
-void PSourceSyntaxFeatures2::TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const {
-  WordID ants[8];
-  for (unsigned i = 0; i < ant_contexts.size(); ++i)
-    ants[i] = *static_cast<const WordID*>(ant_contexts[i]);
-
-  *static_cast<WordID*>(context) =
-     impl->FireFeatures(*edge.rule_, edge.i_, edge.j_, ants, features);
-}
-
-void PSourceSyntaxFeatures2::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
-}
diff --git a/decoder/ff_source_syntax2_p.h b/decoder/ff_source_syntax2_p.h
deleted file mode 100644
index d56ecab0..00000000
--- a/decoder/ff_source_syntax2_p.h
+++ /dev/null
@@ -1,25 +0,0 @@
-#ifndef _FF_SOURCE_TOOLS2_H_
-#define _FF_SOURCE_TOOLS2_H_
-
-#include "ff.h"
-#include "hg.h"
-
-struct PSourceSyntaxFeatures2Impl;
-
-class PSourceSyntaxFeatures2 : public FeatureFunction {
- public:
-  PSourceSyntaxFeatures2(const std::string& param);
-  ~PSourceSyntaxFeatures2();
- protected:
-  virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const std::vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const;
-  virtual void PrepareForInput(const SentenceMetadata& smeta);
- private:
-  PSourceSyntaxFeatures2Impl* impl;
-};
-
-#endif
diff --git a/decoder/ff_source_syntax_p.cc b/decoder/ff_source_syntax_p.cc
deleted file mode 100644
index c094de59..00000000
--- a/decoder/ff_source_syntax_p.cc
+++ /dev/null
@@ -1,250 +0,0 @@
-#include "ff_source_syntax_p.h"
-
-#include <sstream>
-#include <stack>
-#ifndef HAVE_OLD_CPP
-# include <unordered_set>
-#else
-# include <tr1/unordered_set>
-namespace std { using std::tr1::unordered_map; using std::tr1::unordered_set; }
-#endif
-
-#include "sentence_metadata.h"
-#include "array2d.h"
-#include "filelib.h"
-
-using namespace std;
-
-// implements the source side syntax features described in Blunsom et al. (EMNLP 2008)
-// source trees must be represented in Penn Treebank format, e.g.
-//     (S (NP John) (VP (V left)))
-
-// log transform to make long spans cluster together
-// but preserve differences
-inline int SpanSizeTransform(unsigned span_size) {
-  if (!span_size) return 0;
-  return static_cast<int>(log(span_size+1) / log(1.39)) - 1;
-}
-
-struct PSourceSyntaxFeaturesImpl {
-  PSourceSyntaxFeaturesImpl() {}
-
-  PSourceSyntaxFeaturesImpl(const string& param) {
-    if (!(param.compare("") == 0)) {
-      string triggered_features_fn = param;
-      ReadFile triggered_features(triggered_features_fn);
-      string in;
-      while(getline(*triggered_features, in)) {
-        feature_filter.insert(FD::Convert(in));
-      }
-    }
-  }
-
-  void InitializeGrids(const string& tree, unsigned src_len) {
-    assert(tree.size() > 0);
-    //fids_cat.clear();
-    fids_ef.clear();
-    src_tree.clear();
-    //fids_cat.resize(src_len, src_len + 1);
-    fids_ef.resize(src_len, src_len + 1);
-    src_tree.resize(src_len, src_len + 1, TD::Convert("XX"));
-    ParseTreeString(tree, src_len);
-  }
-
-  void ParseTreeString(const string& tree, unsigned src_len) {
-    stack<pair<int, WordID> > stk;  // first = i, second = category
-    pair<int, WordID> cur_cat; cur_cat.first = -1;
-    unsigned i = 0;
-    unsigned p = 0;
-    while(p < tree.size()) {
-      const char cur = tree[p];
-      if (cur == '(') {
-        stk.push(cur_cat);
-        ++p;
-        unsigned k = p + 1;
-        while (k < tree.size() && tree[k] != ' ') { ++k; }
-        cur_cat.first = i;
-        cur_cat.second = TD::Convert(tree.substr(p, k - p));
-        // cerr << "NT: '" << tree.substr(p, k-p) << "' (i=" << i << ")\n";
-        p = k + 1;
-      } else if (cur == ')') {
-        unsigned k = p;
-        while (k < tree.size() && tree[k] == ')') { ++k; }
-        const unsigned num_closes = k - p;
-        for (unsigned ci = 0; ci < num_closes; ++ci) {
-          // cur_cat.second spans from cur_cat.first to i
-          // cerr << TD::Convert(cur_cat.second) << " from " << cur_cat.first << " to " << i << endl;
-          // NOTE: unary rule chains end up being labeled with the top-most category
-          src_tree(cur_cat.first, i) = cur_cat.second;
-          cur_cat = stk.top();
-          stk.pop();
-        }
-        p = k;
-        while (p < tree.size() && (tree[p] == ' ' || tree[p] == '\t')) { ++p; }
-      } else if (cur == ' ' || cur == '\t') {
-        cerr << "Unexpected whitespace in: " << tree << endl;
-        abort();
-      } else { // terminal symbol
-        unsigned k = p + 1;
-        do {
-          while (k < tree.size() && tree[k] != ')' && tree[k] != ' ') { ++k; }
-          // cerr << "TERM: '" << tree.substr(p, k-p) << "' (i=" << i << ")\n";
-          ++i;
-          assert(i <= src_len);
-          while (k < tree.size() && tree[k] == ' ') { ++k; }
-          p = k;
-        } while (p < tree.size() && tree[p] != ')');
-      }
-    }
-    // cerr << "i=" << i << "  src_len=" << src_len << endl;
-    assert(i == src_len);  // make sure tree specified in src_tree is
-                           // the same length as the source sentence
-  }
-
-  WordID FireFeatures(const TRule& rule, const int i, const int j, const WordID* ants, SparseVector<double>* feats) {
-    //cerr << "fire features: " << rule.AsString() << " for " << i << "," << j << endl;
-    const WordID lhs = src_tree(i,j);
-    //int& fid_cat = fids_cat(i,j);
-    int& fid_ef = fids_ef(i,j)[&rule];
-    if (fid_ef <= 0) {
-      ostringstream os;
-      //ostringstream os2;
-      os << "SYN:" << TD::Convert(lhs);
-      //os2 << "SYN:" << TD::Convert(lhs) << '_' << SpanSizeTransform(j - i);
-      //fid_cat = FD::Convert(os2.str());
-      os << ':';
-      unsigned ntc = 0;
-      for (unsigned k = 0; k < rule.f_.size(); ++k) {
-        if (k > 0) os << '_';
-        int fj = rule.f_[k];
-        if (fj <= 0) {
-          os << '[' << TD::Convert(ants[ntc++]) << ']';
-        } else {
-          os << TD::Convert(fj);
-        }
-      }
-      os << ':';
-      for (unsigned k = 0; k < rule.e_.size(); ++k) {
-        const int ei = rule.e_[k];
-        if (k > 0) os << '_';
-        if (ei <= 0)
-          os << '[' << (1-ei) << ']';
-        else
-          os << TD::Convert(ei);
-      }
-      fid_ef = FD::Convert(os.str());
-    }
-    //if (fid_cat > 0)
-    //  feats->set_value(fid_cat, 1.0);
-    if (fid_ef > 0 && (feature_filter.find(fid_ef) != feature_filter.end()))
-      feats->set_value(fid_ef, 1.0);
-    return lhs;
-  }
-
-  Array2D<WordID> src_tree;  // src_tree(i,j) NT = type
-  // mutable Array2D<int> fids_cat;   // this tends to overfit baddly
-  mutable Array2D<map<const TRule*, int> > fids_ef;    // fires for fully lexicalized
-  unordered_set<int> feature_filter;
-};
-
-PSourceSyntaxFeatures::PSourceSyntaxFeatures(const string& param) :
-    FeatureFunction(sizeof(WordID)) {
-  impl = new PSourceSyntaxFeaturesImpl(param);
-}
-
-PSourceSyntaxFeatures::~PSourceSyntaxFeatures() {
-  delete impl;
-  impl = NULL;
-}
-
-void PSourceSyntaxFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const {
-  WordID ants[8];
-  for (unsigned i = 0; i < ant_contexts.size(); ++i)
-    ants[i] = *static_cast<const WordID*>(ant_contexts[i]);
-
-  *static_cast<WordID*>(context) =
-     impl->FireFeatures(*edge.rule_, edge.i_, edge.j_, ants, features);
-}
-
-void PSourceSyntaxFeatures::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSGMLValue("src_tree"), smeta.GetSourceLength());
-}
-
-struct PSourceSpanSizeFeaturesImpl {
-  PSourceSpanSizeFeaturesImpl() {}
-
-  void InitializeGrids(unsigned src_len) {
-    fids.clear();
-    fids.resize(src_len, src_len + 1);
-  }
-
-  int FireFeatures(const TRule& rule, const int i, const int j, const WordID* ants, SparseVector<double>* feats) {
-    if (rule.Arity() > 0) {
-      int& fid = fids(i,j)[&rule];
-      if (fid <= 0) {
-        ostringstream os;
-        os << "SSS:";
-        unsigned ntc = 0;
-        for (unsigned k = 0; k < rule.f_.size(); ++k) {
-          if (k > 0) os << '_';
-          int fj = rule.f_[k];
-          if (fj <= 0) {
-            os << '[' << TD::Convert(-fj) << ants[ntc++] << ']';
-          } else {
-            os << TD::Convert(fj);
-          }
-        }
-        os << ':';
-        for (unsigned k = 0; k < rule.e_.size(); ++k) {
-          const int ei = rule.e_[k];
-          if (k > 0) os << '_';
-          if (ei <= 0)
-            os << '[' << (1-ei) << ']';
-          else
-            os << TD::Convert(ei);
-        }
-        fid = FD::Convert(os.str());
-      }
-      if (fid > 0)
-        feats->set_value(fid, 1.0);
-    }
-    return SpanSizeTransform(j - i);
-  }
-
-  mutable Array2D<map<const TRule*, int> > fids;
-};
-
-PSourceSpanSizeFeatures::PSourceSpanSizeFeatures(const string& param) :
-    FeatureFunction(sizeof(char)) {
-  impl = new PSourceSpanSizeFeaturesImpl;
-}
-
-PSourceSpanSizeFeatures::~PSourceSpanSizeFeatures() {
-  delete impl;
-  impl = NULL;
-}
-
-void PSourceSpanSizeFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const {
-  int ants[8];
-  for (unsigned i = 0; i < ant_contexts.size(); ++i)
-    ants[i] = *static_cast<const char*>(ant_contexts[i]);
-
-  *static_cast<char*>(context) =
-     impl->FireFeatures(*edge.rule_, edge.i_, edge.j_, ants, features);
-}
-
-void PSourceSpanSizeFeatures::PrepareForInput(const SentenceMetadata& smeta) {
-  impl->InitializeGrids(smeta.GetSourceLength());
-}
-
-
diff --git a/decoder/ff_source_syntax_p.h b/decoder/ff_source_syntax_p.h
deleted file mode 100644
index 2dd9094a..00000000
--- a/decoder/ff_source_syntax_p.h
+++ /dev/null
@@ -1,42 +0,0 @@
-#ifndef _FF_SOURCE_TOOLS_H_
-#define _FF_SOURCE_TOOLS_H_
-
-#include "ff.h"
-#include "hg.h"
-
-struct PSourceSyntaxFeaturesImpl;
-
-class PSourceSyntaxFeatures : public FeatureFunction {
- public:
-  PSourceSyntaxFeatures(const std::string& param);
-  ~PSourceSyntaxFeatures();
- protected:
-  virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const std::vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const;
-  virtual void PrepareForInput(const SentenceMetadata& smeta);
- private:
-  PSourceSyntaxFeaturesImpl* impl;
-};
-
-struct PSourceSpanSizeFeaturesImpl;
-class PSourceSpanSizeFeatures : public FeatureFunction {
- public:
-  PSourceSpanSizeFeatures(const std::string& param);
-  ~PSourceSpanSizeFeatures();
- protected:
-  virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
-                                     const Hypergraph::Edge& edge,
-                                     const std::vector<const void*>& ant_contexts,
-                                     SparseVector<double>* features,
-                                     SparseVector<double>* estimated_features,
-                                     void* context) const;
-  virtual void PrepareForInput(const SentenceMetadata& smeta);
- private:
-  PSourceSpanSizeFeaturesImpl* impl;
-};
-
-#endif
diff --git a/extractor/grammar_extractor.cc b/extractor/grammar_extractor.cc
index 8050ce7b..1fbdee5b 100644
--- a/extractor/grammar_extractor.cc
+++ b/extractor/grammar_extractor.cc
@@ -3,11 +3,13 @@
 #include <iterator>
 #include <sstream>
 #include <vector>
+#include <unordered_set>
 
 #include "grammar.h"
 #include "rule.h"
 #include "rule_factory.h"
 #include "vocabulary.h"
+#include "data_array.h"
 
 using namespace std;
 
@@ -32,10 +34,10 @@ GrammarExtractor::GrammarExtractor(
     vocabulary(vocabulary),
     rule_factory(rule_factory) {}
 
-Grammar GrammarExtractor::GetGrammar(const string& sentence) {
+Grammar GrammarExtractor::GetGrammar(const string& sentence, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array) {
   vector<string> words = TokenizeSentence(sentence);
   vector<int> word_ids = AnnotateWords(words);
-  return rule_factory->GetGrammar(word_ids);
+  return rule_factory->GetGrammar(word_ids, blacklisted_sentence_ids, source_data_array);
 }
 
 vector<string> GrammarExtractor::TokenizeSentence(const string& sentence) {
diff --git a/extractor/grammar_extractor.h b/extractor/grammar_extractor.h
index b36ceeb9..6c0aafbf 100644
--- a/extractor/grammar_extractor.h
+++ b/extractor/grammar_extractor.h
@@ -4,6 +4,7 @@
 #include <memory>
 #include <string>
 #include <vector>
+#include <unordered_set>
 
 using namespace std;
 
@@ -44,7 +45,7 @@ class GrammarExtractor {
 
   // Converts the sentence to a vector of word ids and uses the RuleFactory to
   // extract the SCFG rules which may be used to decode the sentence.
-  Grammar GetGrammar(const string& sentence);
+  Grammar GetGrammar(const string& sentence, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array);
 
  private:
   // Splits the sentence in a vector of words.
diff --git a/extractor/grammar_extractor_test.cc b/extractor/grammar_extractor_test.cc
index 823bb8b4..f32a9599 100644
--- a/extractor/grammar_extractor_test.cc
+++ b/extractor/grammar_extractor_test.cc
@@ -39,12 +39,15 @@ TEST(GrammarExtractorTest, TestAnnotatingWords) {
   vector<Rule> rules;
   vector<string> feature_names;
   Grammar grammar(rules, feature_names);
-  EXPECT_CALL(*factory, GetGrammar(word_ids))
+  unordered_set<int> blacklisted_sentence_ids;
+  shared_ptr<DataArray> source_data_array;
+  EXPECT_CALL(*factory, GetGrammar(word_ids, blacklisted_sentence_ids, source_data_array))
       .WillOnce(Return(grammar));
 
   GrammarExtractor extractor(vocabulary, factory);
   string sentence = "Anna has many many apples .";
-  extractor.GetGrammar(sentence);
+
+  extractor.GetGrammar(sentence, blacklisted_sentence_ids, source_data_array);
 }
 
 } // namespace
diff --git a/extractor/mocks/mock_rule_factory.h b/extractor/mocks/mock_rule_factory.h
index 7389b396..86a084b5 100644
--- a/extractor/mocks/mock_rule_factory.h
+++ b/extractor/mocks/mock_rule_factory.h
@@ -7,7 +7,7 @@ namespace extractor {
 
 class MockHieroCachingRuleFactory : public HieroCachingRuleFactory {
  public:
-  MOCK_METHOD1(GetGrammar, Grammar(const vector<int>& word_ids));
+  MOCK_METHOD3(GetGrammar, Grammar(const vector<int>& word_ids, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array));
 };
 
 } // namespace extractor
diff --git a/extractor/rule_factory.cc b/extractor/rule_factory.cc
index 8c30fb9e..e52019ae 100644
--- a/extractor/rule_factory.cc
+++ b/extractor/rule_factory.cc
@@ -17,6 +17,7 @@
 #include "suffix_array.h"
 #include "time_util.h"
 #include "vocabulary.h"
+#include "data_array.h"
 
 using namespace std;
 using namespace chrono;
@@ -100,7 +101,7 @@ HieroCachingRuleFactory::HieroCachingRuleFactory() {}
 
 HieroCachingRuleFactory::~HieroCachingRuleFactory() {}
 
-Grammar HieroCachingRuleFactory::GetGrammar(const vector<int>& word_ids) {
+Grammar HieroCachingRuleFactory::GetGrammar(const vector<int>& word_ids, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array) {
   Clock::time_point start_time = Clock::now();
   double total_extract_time = 0;
   double total_intersect_time = 0;
@@ -192,7 +193,7 @@ Grammar HieroCachingRuleFactory::GetGrammar(const vector<int>& word_ids) {
       Clock::time_point extract_start = Clock::now();
       if (!state.starts_with_x) {
         // Extract rules for the sampled set of occurrences.
-        PhraseLocation sample = sampler->Sample(next_node->matchings);
+        PhraseLocation sample = sampler->Sample(next_node->matchings, blacklisted_sentence_ids, source_data_array);
         vector<Rule> new_rules =
             rule_extractor->ExtractRules(next_phrase, sample);
         rules.insert(rules.end(), new_rules.begin(), new_rules.end());
diff --git a/extractor/rule_factory.h b/extractor/rule_factory.h
index 52e8712a..c7332720 100644
--- a/extractor/rule_factory.h
+++ b/extractor/rule_factory.h
@@ -3,6 +3,7 @@
 
 #include <memory>
 #include <vector>
+#include <unordered_set>
 
 #include "matchings_trie.h"
 
@@ -71,7 +72,7 @@ class HieroCachingRuleFactory {
 
   // Constructs SCFG rules for a given sentence.
   // (See class description for more details.)
-  virtual Grammar GetGrammar(const vector<int>& word_ids);
+  virtual Grammar GetGrammar(const vector<int>& word_ids, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array);
 
  protected:
   HieroCachingRuleFactory();
diff --git a/extractor/rule_factory_test.cc b/extractor/rule_factory_test.cc
index 08af3dcd..f26cc567 100644
--- a/extractor/rule_factory_test.cc
+++ b/extractor/rule_factory_test.cc
@@ -76,7 +76,9 @@ TEST_F(RuleFactoryTest, TestGetGrammarDifferentWords) {
       .WillRepeatedly(Return(PhraseLocation(0, 1)));
 
   vector<int> word_ids = {2, 3, 4};
-  Grammar grammar = factory->GetGrammar(word_ids);
+  unordered_set<int> blacklisted_sentence_ids;
+  shared_ptr<DataArray> source_data_array;
+  Grammar grammar = factory->GetGrammar(word_ids, blacklisted_sentence_ids, source_data_array);
   EXPECT_EQ(feature_names, grammar.GetFeatureNames());
   EXPECT_EQ(7, grammar.GetRules().size());
 }
@@ -94,7 +96,9 @@ TEST_F(RuleFactoryTest, TestGetGrammarRepeatingWords) {
       .WillRepeatedly(Return(PhraseLocation(0, 1)));
 
   vector<int> word_ids = {2, 3, 4, 2, 3};
-  Grammar grammar = factory->GetGrammar(word_ids);
+  unordered_set<int> blacklisted_sentence_ids;
+  shared_ptr<DataArray> source_data_array;
+  Grammar grammar = factory->GetGrammar(word_ids, blacklisted_sentence_ids, source_data_array);
   EXPECT_EQ(feature_names, grammar.GetFeatureNames());
   EXPECT_EQ(28, grammar.GetRules().size());
 }
diff --git a/extractor/run_extractor.cc b/extractor/run_extractor.cc
index 8a9ca89d..6eb55073 100644
--- a/extractor/run_extractor.cc
+++ b/extractor/run_extractor.cc
@@ -75,7 +75,9 @@ int main(int argc, char** argv) {
     ("max_samples", po::value<int>()->default_value(300),
         "Maximum number of samples")
     ("tight_phrases", po::value<bool>()->default_value(true),
-        "False if phrases may be loose (better, but slower)");
+        "False if phrases may be loose (better, but slower)")
+    ("leave_one_out", po::value<bool>()->zero_tokens(),
+        "do leave-one-out estimation of grammars (e.g. for extracting grammars for the training set");
 
   po::variables_map vm;
   po::store(po::parse_command_line(argc, argv, desc), vm);
@@ -96,6 +98,11 @@ int main(int argc, char** argv) {
     return 1;
   }
 
+  bool leave_one_out = false;
+  if (vm.count("leave_one_out")) {
+    leave_one_out = true;
+  }
+
   int num_threads = vm["threads"].as<int>();
   cerr << "Grammar extraction will use " << num_threads << " threads." << endl;
 
@@ -223,7 +230,9 @@ int main(int argc, char** argv) {
     }
     suffixes[i] = suffix;
 
-    Grammar grammar = extractor.GetGrammar(sentences[i]);
+    unordered_set<int> blacklisted_sentence_ids;
+    if (leave_one_out) blacklisted_sentence_ids.insert(i);
+    Grammar grammar = extractor.GetGrammar(sentences[i], blacklisted_sentence_ids, source_data_array);
     ofstream output(GetGrammarFilePath(grammar_path, i).c_str());
     output << grammar;
   }
diff --git a/extractor/sample_source.txt b/extractor/sample_source.txt
new file mode 100644
index 00000000..971baf6d
--- /dev/null
+++ b/extractor/sample_source.txt
@@ -0,0 +1,2 @@
+ana are mere .
+ana bea mult lapte .
diff --git a/extractor/sampler.cc b/extractor/sampler.cc
index d81956b5..d332dd90 100644
--- a/extractor/sampler.cc
+++ b/extractor/sampler.cc
@@ -12,7 +12,7 @@ Sampler::Sampler() {}
 
 Sampler::~Sampler() {}
 
-PhraseLocation Sampler::Sample(const PhraseLocation& location) const {
+PhraseLocation Sampler::Sample(const PhraseLocation& location, unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array) const {
   vector<int> sample;
   int num_subpatterns;
   if (location.matchings == NULL) {
@@ -20,8 +20,37 @@ PhraseLocation Sampler::Sample(const PhraseLocation& location) const {
     num_subpatterns = 1;
     int low = location.sa_low, high = location.sa_high;
     double step = max(1.0, (double) (high - low) / max_samples);
-    for (double i = low; i < high && sample.size() < max_samples; i += step) {
-      sample.push_back(suffix_array->GetSuffix(Round(i)));
+    double i = low, last = i;
+    bool found;
+    while (sample.size() < max_samples && i < high) {
+      int x = suffix_array->GetSuffix(Round(i));
+      int id = source_data_array->GetSentenceId(x);
+      if (find(blacklisted_sentence_ids.begin(), blacklisted_sentence_ids.end(), id) != blacklisted_sentence_ids.end()) {
+        found = false;
+        double backoff_step = 1;
+        while (true) {
+          if ((double)backoff_step >= step) break;
+          double j = i - backoff_step;
+          x = suffix_array->GetSuffix(Round(j));
+          id = source_data_array->GetSentenceId(x);
+          if (x >= 0 && j > last && find(blacklisted_sentence_ids.begin(), blacklisted_sentence_ids.end(), id) == blacklisted_sentence_ids.end()) {
+            found = true; last = i; break;
+          }
+          double k = i + backoff_step;
+          x = suffix_array->GetSuffix(Round(k));
+          id = source_data_array->GetSentenceId(x);
+          if (k < min(i+step, (double)high) && find(blacklisted_sentence_ids.begin(), blacklisted_sentence_ids.end(), id) == blacklisted_sentence_ids.end()) {
+            found = true; last = k; break;
+          }
+          if (j <= last && k >= high) break;
+          backoff_step++;
+        }
+      } else {
+        found = true;
+        last = i;
+      }
+      if (found) sample.push_back(x);
+      i += step;
     }
   } else {
     // Sample vector of occurrences.
diff --git a/extractor/sampler.h b/extractor/sampler.h
index be4aa1bb..30e747fd 100644
--- a/extractor/sampler.h
+++ b/extractor/sampler.h
@@ -2,6 +2,9 @@
 #define _SAMPLER_H_
 
 #include <memory>
+#include <unordered_set>
+
+#include "data_array.h"
 
 using namespace std;
 
@@ -20,7 +23,7 @@ class Sampler {
   virtual ~Sampler();
 
   // Samples uniformly at most max_samples phrase occurrences.
-  virtual PhraseLocation Sample(const PhraseLocation& location) const;
+  virtual PhraseLocation Sample(const PhraseLocation& location, const unordered_set<int> blacklisted_sentence_ids, const shared_ptr<DataArray> source_data_array) const;
 
  protected:
   Sampler();
diff --git a/extractor/sampler_test.cc b/extractor/sampler_test.cc
index e9abebfa..965567ba 100644
--- a/extractor/sampler_test.cc
+++ b/extractor/sampler_test.cc
@@ -3,6 +3,7 @@
 #include <memory>
 
 #include "mocks/mock_suffix_array.h"
+#include "mocks/mock_data_array.h"
 #include "phrase_location.h"
 #include "sampler.h"
 
@@ -15,6 +16,8 @@ namespace {
 class SamplerTest : public Test {
  protected:
   virtual void SetUp() {
+    source_data_array = make_shared<MockDataArray>();
+    EXPECT_CALL(*source_data_array, GetSentenceId(_)).WillRepeatedly(Return(9999));
     suffix_array = make_shared<MockSuffixArray>();
     for (int i = 0; i < 10; ++i) {
       EXPECT_CALL(*suffix_array, GetSuffix(i)).WillRepeatedly(Return(i));
@@ -23,51 +26,54 @@ class SamplerTest : public Test {
 
   shared_ptr<MockSuffixArray> suffix_array;
   shared_ptr<Sampler> sampler;
+  shared_ptr<MockDataArray> source_data_array;
 };
 
 TEST_F(SamplerTest, TestSuffixArrayRange) {
   PhraseLocation location(0, 10);
+  unordered_set<int> blacklist;
 
   sampler = make_shared<Sampler>(suffix_array, 1);
   vector<int> expected_locations = {0};
-  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 2);
   expected_locations = {0, 5};
-  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 3);
   expected_locations = {0, 3, 7};
-  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 4);
   expected_locations = {0, 3, 5, 8};
-  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 100);
   expected_locations = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
-  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
 }
 
 TEST_F(SamplerTest, TestSubstringsSample) {
   vector<int> locations = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
+  unordered_set<int> blacklist;
   PhraseLocation location(locations, 2);
 
   sampler = make_shared<Sampler>(suffix_array, 1);
   vector<int> expected_locations = {0, 1};
-  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 2);
   expected_locations = {0, 1, 6, 7};
-  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 3);
   expected_locations = {0, 1, 4, 5, 6, 7};
-  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location, blacklist, source_data_array));
 
   sampler = make_shared<Sampler>(suffix_array, 7);
   expected_locations = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
-  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location));
+  EXPECT_EQ(PhraseLocation(expected_locations, 2), sampler->Sample(location, blacklist, source_data_array));
 }
 
 } // namespace
diff --git a/extractor/sampler_test_blacklist.cc b/extractor/sampler_test_blacklist.cc
new file mode 100644
index 00000000..3305b990
--- /dev/null
+++ b/extractor/sampler_test_blacklist.cc
@@ -0,0 +1,102 @@
+#include <gtest/gtest.h>
+
+#include <memory>
+
+#include "mocks/mock_suffix_array.h"
+#include "mocks/mock_data_array.h"
+#include "phrase_location.h"
+#include "sampler.h"
+
+using namespace std;
+using namespace ::testing;
+
+namespace extractor {
+namespace {
+
+class SamplerTestBlacklist : public Test {
+ protected:
+  virtual void SetUp() {
+    source_data_array = make_shared<MockDataArray>();
+    for (int i = 0; i < 10; ++i) {
+      EXPECT_CALL(*source_data_array, GetSentenceId(i)).WillRepeatedly(Return(i));
+    }
+    for (int i = -10; i < 0; ++i) {
+      EXPECT_CALL(*source_data_array, GetSentenceId(i)).WillRepeatedly(Return(0));
+    }
+    suffix_array = make_shared<MockSuffixArray>();
+    for (int i = -10; i < 10; ++i) {
+      EXPECT_CALL(*suffix_array, GetSuffix(i)).WillRepeatedly(Return(i));
+    }
+  }
+
+  shared_ptr<MockSuffixArray> suffix_array;
+  shared_ptr<Sampler> sampler;
+  shared_ptr<MockDataArray> source_data_array;
+};
+
+TEST_F(SamplerTestBlacklist, TestSuffixArrayRange) {
+  PhraseLocation location(0, 10);
+  unordered_set<int> blacklist;
+  vector<int> expected_locations;
+   
+  blacklist.insert(0);
+  sampler = make_shared<Sampler>(suffix_array, 1);
+  expected_locations = {1};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+  
+  for (int i = 0; i < 9; i++) {
+    blacklist.insert(i);
+  }
+  sampler = make_shared<Sampler>(suffix_array, 1);
+  expected_locations = {9};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+
+  blacklist.insert(0);
+  blacklist.insert(5);
+  sampler = make_shared<Sampler>(suffix_array, 2);
+  expected_locations = {1, 4};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+
+  blacklist.insert(0);
+  blacklist.insert(1);
+  blacklist.insert(2);
+  blacklist.insert(3);
+  sampler = make_shared<Sampler>(suffix_array, 2);
+  expected_locations = {4, 5};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+
+  blacklist.insert(0);
+  blacklist.insert(3);
+  blacklist.insert(7);
+  sampler = make_shared<Sampler>(suffix_array, 3);
+  expected_locations = {1, 2, 6};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+
+  blacklist.insert(0);
+  blacklist.insert(3);
+  blacklist.insert(5);
+  blacklist.insert(8);
+  sampler = make_shared<Sampler>(suffix_array, 4);
+  expected_locations = {1, 2, 4, 7};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+  
+  blacklist.insert(0);
+  sampler = make_shared<Sampler>(suffix_array, 100);
+  expected_locations = {1, 2, 3, 4, 5, 6, 7, 8, 9};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+  blacklist.clear();
+ 
+  blacklist.insert(9);
+  sampler = make_shared<Sampler>(suffix_array, 100);
+  expected_locations = {0, 1, 2, 3, 4, 5, 6, 7, 8};
+  EXPECT_EQ(PhraseLocation(expected_locations, 1), sampler->Sample(location, blacklist, source_data_array));
+}
+
+} // namespace
+} // namespace extractor
diff --git a/training/dtrain/Makefile.am b/training/dtrain/Makefile.am
index 844c790d..ecb6c128 100644
--- a/training/dtrain/Makefile.am
+++ b/training/dtrain/Makefile.am
@@ -1,7 +1,7 @@
 bin_PROGRAMS = dtrain
 
 dtrain_SOURCES = dtrain.cc score.cc dtrain.h kbestget.h ksampler.h pairsampling.h score.h
-dtrain_LDADD   = ../../decoder/libcdec.a ../../klm/search/libksearch.a ../../mteval/libmteval.a ../../utils/libutils.a ../../klm/lm/libklm.a ../../klm/util/libklm_util.a ../../klm/util/double-conversion/libklm_util_double.a
+dtrain_LDADD   = ../../decoder/libcdec.a ../../klm/search/libksearch.a ../../mteval/libmteval.a ../../utils/libutils.a ../../klm/lm/libklm.a ../../klm/util/libklm_util.a ../../klm/util/double-conversion/libklm_util_double.a -lboost_regex
 
 AM_CPPFLAGS = -W -Wall -Wno-sign-compare -I$(top_srcdir)/utils -I$(top_srcdir)/decoder -I$(top_srcdir)/mteval
 
diff --git a/training/dtrain/README.md b/training/dtrain/README.md
index 2bae6b48..aa1ab3e7 100644
--- a/training/dtrain/README.md
+++ b/training/dtrain/README.md
@@ -1,10 +1,15 @@
 This is a simple (and parallelizable) tuning method for cdec
-which is able to train the weights of very many (sparse) features.
-It was used here:
-  "Joint Feature Selection in Distributed Stochastic
-   Learning for Large-Scale Discriminative Training in
-   SMT"
-(Simianer, Riezler, Dyer; ACL 2012)
+which is able to train the weights of very many (sparse) features
+on the training set.
+
+It was used in these papers:
+>  "Joint Feature Selection in Distributed Stochastic
+>   Learning for Large-Scale Discriminative Training in
+>   SMT" (Simianer, Riezler, Dyer; ACL 2012)
+>
+>  "Multi-Task Learning for Improved Discriminative
+>   Training in SMT" (Simianer, Riezler; WMT 2013)
+>
 
 
 Building
@@ -17,20 +22,9 @@ To build only parts needed for dtrain do
   cd training/dtrain/; make
 ```
 
-Ideas
------
- * get approx_bleu to work?
- * implement minibatches (Minibatch and Parallelization for Online Large Margin Structured Learning)
- * learning rate 1/T?
- * use an oracle? mira-like (model vs. BLEU), feature repr. of reference!? 
- * implement lc_bleu properly
- * merge kbest lists of previous epochs (as MERT does)
- * ``walk entire regularization path''
- * rerank after each update?
-
 Running
 -------
-See directories under test/ .
+See directories under examples/ .
 
 Legal
 -----
diff --git a/training/dtrain/dtrain.cc b/training/dtrain/dtrain.cc
index 0ee2f124..0a27a068 100644
--- a/training/dtrain/dtrain.cc
+++ b/training/dtrain/dtrain.cc
@@ -12,8 +12,9 @@ dtrain_init(int argc, char** argv, po::variables_map* cfg)
 {
   po::options_description ini("Configuration File Options");
   ini.add_options()
-    ("input",             po::value<string>()->default_value("-"),                                             "input file (src)")
+    ("input",             po::value<string>(),                                                                 "input file (src)")
     ("refs,r",            po::value<string>(),                                                                       "references")
+    ("bitext,b",          po::value<string>(),                                                            "bitext: 'src ||| tgt'")
     ("output",            po::value<string>()->default_value("-"),                          "output weights file, '-' for STDOUT")
     ("input_weights",     po::value<string>(),                                "input weights file (e.g. from previous iteration)")
     ("decoder_config",    po::value<string>(),                                                      "configuration file for cdec")
@@ -40,6 +41,10 @@ dtrain_init(int argc, char** argv, po::variables_map* cfg)
     ("scale_bleu_diff",   po::value<bool>()->zero_tokens(),                      "learning rate <- bleu diff of a misranked pair")
     ("loss_margin",       po::value<weight_t>()->default_value(0.),  "update if no error in pref pair but model scores this near")
     ("max_pairs",         po::value<unsigned>()->default_value(std::numeric_limits<unsigned>::max()), "max. # of pairs per Sent.")
+    ("pclr",              po::value<string>()->default_value("no"),         "use a (simple|adagrad) per-coordinate learning rate")
+    ("batch",             po::value<bool>()->zero_tokens(),                                               "do batch optimization")
+    ("repeat",            po::value<unsigned>()->default_value(1),          "repeat optimization over kbest list this number of times")
+    //("test-k-best",       po::value<bool>()->zero_tokens(),                       "check if optimization works (use repeat >= 2)")
     ("noup",              po::value<bool>()->zero_tokens(),                                               "do not update weights");
   po::options_description cl("Command Line Options");
   cl.add_options()
@@ -72,13 +77,17 @@ dtrain_init(int argc, char** argv, po::variables_map* cfg)
     cerr << "Wrong 'pair_sampling' param: '" << (*cfg)["pair_sampling"].as<string>() << "'." << endl;
     return false;
   }
-  if(cfg->count("hi_lo") && (*cfg)["pair_sampling"].as<string>() != "XYX") {
+  if (cfg->count("hi_lo") && (*cfg)["pair_sampling"].as<string>() != "XYX") {
     cerr << "Warning: hi_lo only works with pair_sampling XYX." << endl;
   }
-  if((*cfg)["hi_lo"].as<float>() > 0.5 || (*cfg)["hi_lo"].as<float>() < 0.01) {
+  if ((*cfg)["hi_lo"].as<float>() > 0.5 || (*cfg)["hi_lo"].as<float>() < 0.01) {
     cerr << "hi_lo must lie in [0.01, 0.5]" << endl;
     return false;
   }
+  if ((cfg->count("input")>0 || cfg->count("refs")>0) && cfg->count("bitext")>0) {
+    cerr << "Provide 'input' and 'refs' or 'bitext', not both." << endl;
+    return false;
+  }
   if ((*cfg)["pair_threshold"].as<score_t>() < 0) {
     cerr << "The threshold must be >= 0!" << endl;
     return false;
@@ -120,10 +129,16 @@ main(int argc, char** argv)
   const float hi_lo = cfg["hi_lo"].as<float>();
   const score_t approx_bleu_d = cfg["approx_bleu_d"].as<score_t>();
   const unsigned max_pairs = cfg["max_pairs"].as<unsigned>();
+  int repeat = cfg["repeat"].as<unsigned>();
+  //bool test_k_best = false;
+  //if (cfg.count("test-k-best")) test_k_best = true;
   weight_t loss_margin = cfg["loss_margin"].as<weight_t>();
+  bool batch = false;
+  if (cfg.count("batch")) batch = true;
   if (loss_margin > 9998.) loss_margin = std::numeric_limits<float>::max();
   bool scale_bleu_diff = false;
   if (cfg.count("scale_bleu_diff")) scale_bleu_diff = true;
+  const string pclr = cfg["pclr"].as<string>();
   bool average = false;
   if (select_weights == "avg")
     average = true;
@@ -131,7 +146,6 @@ main(int argc, char** argv)
   if (cfg.count("print_weights"))
     boost::split(print_weights, cfg["print_weights"].as<string>(), boost::is_any_of(" "));
 
-
   // setup decoder
   register_feature_functions();
   SetSilent(true);
@@ -178,17 +192,16 @@ main(int argc, char** argv)
   observer->SetScorer(scorer);
 
   // init weights
-  vector<weight_t>& dense_weights = decoder.CurrentWeightVector();
+  vector<weight_t>& decoder_weights = decoder.CurrentWeightVector();
   SparseVector<weight_t> lambdas, cumulative_penalties, w_average;
-  if (cfg.count("input_weights")) Weights::InitFromFile(cfg["input_weights"].as<string>(), &dense_weights);
-  Weights::InitSparseVector(dense_weights, &lambdas);
+  if (cfg.count("input_weights")) Weights::InitFromFile(cfg["input_weights"].as<string>(), &decoder_weights);
+  Weights::InitSparseVector(decoder_weights, &lambdas);
 
   // meta params for perceptron, SVM
   weight_t eta = cfg["learning_rate"].as<weight_t>();
   weight_t gamma = cfg["gamma"].as<weight_t>();
 
   // faster perceptron: consider only misranked pairs, see
-  // DO NOT ENABLE  WITH SVM (gamma > 0) OR loss_margin!
   bool faster_perceptron = false;
   if (gamma==0 && loss_margin==0) faster_perceptron = true;
 
@@ -208,13 +221,24 @@ main(int argc, char** argv)
   // output
   string output_fn = cfg["output"].as<string>();
   // input
-  string input_fn = cfg["input"].as<string>();
+  bool read_bitext = false;
+  string input_fn;
+  if (cfg.count("bitext")) {
+    read_bitext = true;
+    input_fn = cfg["bitext"].as<string>();
+  } else {
+    input_fn = cfg["input"].as<string>();
+  }
   ReadFile input(input_fn);
   // buffer input for t > 0
   vector<string> src_str_buf;          // source strings (decoder takes only strings)
   vector<vector<WordID> > ref_ids_buf; // references as WordID vecs
-  string refs_fn = cfg["refs"].as<string>();
-  ReadFile refs(refs_fn);
+  ReadFile refs;
+  string refs_fn;
+  if (!read_bitext) {
+    refs_fn = cfg["refs"].as<string>();
+    refs.Init(refs_fn);
+  }
 
   unsigned in_sz = std::numeric_limits<unsigned>::max(); // input index, input size
   vector<pair<score_t, score_t> > all_scores;
@@ -229,6 +253,7 @@ main(int argc, char** argv)
     cerr << setw(25) << "k " << k << endl;
     cerr << setw(25) << "N " << N << endl;
     cerr << setw(25) << "T " << T << endl;
+    cerr << setw(25) << "batch " << batch << endl;
     cerr << setw(26) << "scorer '" << scorer_str << "'" << endl;
     if (scorer_str == "approx_bleu")
       cerr << setw(25) << "approx. B discount " << approx_bleu_d << endl;
@@ -249,10 +274,14 @@ main(int argc, char** argv)
       cerr << setw(25) << "l1 reg " << l1_reg << " '" << cfg["l1_reg"].as<string>() << "'" << endl;
     if (rescale)
       cerr << setw(25) << "rescale " << rescale << endl;
+    cerr << setw(25) << "pclr " << pclr << endl;
     cerr << setw(25) << "max pairs " << max_pairs << endl;
+    cerr << setw(25) << "repeat " << repeat << endl;
+    //cerr << setw(25) << "test k-best " << test_k_best << endl;
     cerr << setw(25) << "cdec cfg " << "'" << cfg["decoder_config"].as<string>() << "'" << endl;
     cerr << setw(25) << "input " << "'" << input_fn << "'" << endl;
-    cerr << setw(25) << "refs " << "'" << refs_fn << "'" << endl;
+    if (!read_bitext)
+      cerr << setw(25) << "refs " << "'" << refs_fn << "'" << endl;
     cerr << setw(25) << "output " << "'" << output_fn << "'" << endl;
     if (cfg.count("input_weights"))
       cerr << setw(25) << "weights in " << "'" << cfg["input_weights"].as<string>() << "'" << endl;
@@ -261,6 +290,11 @@ main(int argc, char** argv)
     if (!verbose) cerr << "(a dot represents " << DTRAIN_DOTS << " inputs)" << endl;
   }
 
+  // pclr
+  SparseVector<weight_t> learning_rates;
+  // batch
+  SparseVector<weight_t> batch_updates;
+  score_t batch_loss;
 
   for (unsigned t = 0; t < T; t++) // T epochs
   {
@@ -269,16 +303,24 @@ main(int argc, char** argv)
   time(&start);
   score_t score_sum = 0.;
   score_t model_sum(0);
-  unsigned ii = 0, rank_errors = 0, margin_violations = 0, npairs = 0, f_count = 0, list_sz = 0;
+  unsigned ii = 0, rank_errors = 0, margin_violations = 0, npairs = 0, f_count = 0, list_sz = 0, kbest_loss_improve = 0;
+  batch_loss = 0.;
   if (!quiet) cerr << "Iteration #" << t+1 << " of " << T << "." << endl;
 
   while(true)
   {
 
     string in;
+    string ref;
     bool next = false, stop = false; // next iteration or premature stop
     if (t == 0) {
       if(!getline(*input, in)) next = true;
+      if(read_bitext) {
+        vector<string> strs;
+        boost::algorithm::split_regex(strs, in, boost::regex(" \\|\\|\\| "));
+        in = strs[0];
+        ref = strs[1];
+      }
     } else {
       if (ii == in_sz) next = true; // stop if we reach the end of our input
     }
@@ -310,15 +352,16 @@ main(int argc, char** argv)
     if (next || stop) break;
 
     // weights
-    lambdas.init_vector(&dense_weights);
+    lambdas.init_vector(&decoder_weights);
 
     // getting input
     vector<WordID> ref_ids; // reference as vector<WordID>
     if (t == 0) {
-      string r_;
-      getline(*refs, r_);
+      if (!read_bitext) {
+        getline(*refs, ref);
+      }
       vector<string> ref_tok;
-      boost::split(ref_tok, r_, boost::is_any_of(" "));
+      boost::split(ref_tok, ref, boost::is_any_of(" "));
       register_and_convert(ref_tok, ref_ids);
       ref_ids_buf.push_back(ref_ids);
       src_str_buf.push_back(in);
@@ -348,8 +391,10 @@ main(int argc, char** argv)
       }
     }
 
-    score_sum += (*samples)[0].score; // stats for 1best
-    model_sum += (*samples)[0].model;
+    if (repeat == 1) {
+      score_sum += (*samples)[0].score; // stats for 1best
+      model_sum += (*samples)[0].model;
+    }
 
     f_count += observer->get_f_count();
     list_sz += observer->get_sz();
@@ -364,30 +409,74 @@ main(int argc, char** argv)
         partXYX(samples, pairs, pair_threshold, max_pairs, faster_perceptron, hi_lo);
       if (pair_sampling == "PRO")
         PROsampling(samples, pairs, pair_threshold, max_pairs);
-      npairs += pairs.size();
+      int cur_npairs = pairs.size();
+      npairs += cur_npairs;
+
+      score_t kbest_loss_first, kbest_loss_last = 0.0;
 
-      SparseVector<weight_t> lambdas_copy;
+      for (vector<pair<ScoredHyp,ScoredHyp> >::iterator it = pairs.begin();
+           it != pairs.end(); it++) {
+        score_t model_diff = it->first.model - it->second.model;
+        kbest_loss_first += max(0.0, -1.0 * model_diff);
+      }
+
+      for (int ki=0; ki < repeat; ki++) {
+
+      score_t kbest_loss = 0.0; // test-k-best
+      SparseVector<weight_t> lambdas_copy; // for l1 regularization
+      SparseVector<weight_t> sum_up; // for pclr
       if (l1naive||l1clip||l1cumul) lambdas_copy = lambdas;
 
       for (vector<pair<ScoredHyp,ScoredHyp> >::iterator it = pairs.begin();
            it != pairs.end(); it++) {
-        bool rank_error;
+        score_t model_diff = it->first.model - it->second.model;
+        if (repeat > 1) {
+          model_diff = lambdas.dot(it->first.f) - lambdas.dot(it->second.f);
+          kbest_loss += max(0.0, -1.0 * model_diff);
+        }
+        bool rank_error = false;
         score_t margin;
         if (faster_perceptron) { // we only have considering misranked pairs
           rank_error = true; // pair sampling already did this for us
           margin = std::numeric_limits<float>::max();
         } else {
-          rank_error = it->first.model <= it->second.model;
-          margin = fabs(it->first.model - it->second.model);
+          rank_error = model_diff<=0.0;
+          margin = fabs(model_diff);
           if (!rank_error && margin < loss_margin) margin_violations++;
         }
-        if (rank_error) rank_errors++;
+        if (rank_error && ki==1) rank_errors++;
         if (scale_bleu_diff) eta = it->first.score - it->second.score;
         if (rank_error || margin < loss_margin) {
           SparseVector<weight_t> diff_vec = it->first.f - it->second.f;
-          lambdas.plus_eq_v_times_s(diff_vec, eta);
-          if (gamma)
-            lambdas.plus_eq_v_times_s(lambdas, -2*gamma*eta*(1./npairs));
+          if (batch) {
+            batch_loss += max(0., -1.0*model_diff);
+            batch_updates += diff_vec;
+            continue;
+          }
+          if (pclr != "no") {
+            sum_up += diff_vec;
+          } else {
+            lambdas.plus_eq_v_times_s(diff_vec, eta);
+            if (gamma) lambdas.plus_eq_v_times_s(lambdas, -2*gamma*eta*(1./cur_npairs));
+          }
+        }
+      }
+
+      // per-coordinate learning rate
+      if (pclr != "no") {
+        SparseVector<weight_t>::iterator it = sum_up.begin();
+        for (; it != sum_up.end(); ++it) {
+          if (pclr == "simple") {
+           lambdas[it->first] += it->second / max(1.0, learning_rates[it->first]);
+           learning_rates[it->first]++;
+          } else if (pclr == "adagrad") {
+            if (learning_rates[it->first] == 0) {
+             lambdas[it->first] +=  it->second * eta;
+            } else {
+             lambdas[it->first] +=  it->second * eta * learning_rates[it->first];
+            }
+            learning_rates[it->first] += pow(it->second, 2.0);
+          }
         }
       }
 
@@ -395,14 +484,16 @@ main(int argc, char** argv)
       // please note that this regularizations happen
       // after a _sentence_ -- not after each example/pair!
       if (l1naive) {
-        FastSparseVector<weight_t>::iterator it = lambdas.begin();
+        SparseVector<weight_t>::iterator it = lambdas.begin();
         for (; it != lambdas.end(); ++it) {
           if (!lambdas_copy.get(it->first) || lambdas_copy.get(it->first)!=it->second) {
+              it->second *= max(0.0000001, eta/(eta+learning_rates[it->first])); // FIXME
+              learning_rates[it->first]++;
             it->second -= sign(it->second) * l1_reg;
           }
         }
       } else if (l1clip) {
-        FastSparseVector<weight_t>::iterator it = lambdas.begin();
+        SparseVector<weight_t>::iterator it = lambdas.begin();
         for (; it != lambdas.end(); ++it) {
           if (!lambdas_copy.get(it->first) || lambdas_copy.get(it->first)!=it->second) {
             if (it->second != 0) {
@@ -417,7 +508,7 @@ main(int argc, char** argv)
         }
       } else if (l1cumul) {
         weight_t acc_penalty = (ii+1) * l1_reg; // ii is the index of the current input
-        FastSparseVector<weight_t>::iterator it = lambdas.begin();
+        SparseVector<weight_t>::iterator it = lambdas.begin();
         for (; it != lambdas.end(); ++it) {
           if (!lambdas_copy.get(it->first) || lambdas_copy.get(it->first)!=it->second) {
             if (it->second != 0) {
@@ -435,7 +526,28 @@ main(int argc, char** argv)
         }
       }
 
-    }
+      if (ki==repeat-1) { // done
+        kbest_loss_last = kbest_loss;
+        if (repeat > 1) {
+          score_t best_score = -1.;
+          score_t best_model = -std::numeric_limits<score_t>::max();
+          unsigned best_idx;
+          for (unsigned i=0; i < samples->size(); i++) {
+            score_t s = lambdas.dot((*samples)[i].f);
+            if (s > best_model) {
+              best_idx = i;
+              best_model = s;
+            }
+          }
+          score_sum += (*samples)[best_idx].score;
+          model_sum += best_model;
+        }
+      }
+    } // repeat
+
+    if ((kbest_loss_first - kbest_loss_last) >= 0) kbest_loss_improve++;
+
+    } // noup
 
     if (rescale) lambdas /= lambdas.l2norm();
 
@@ -443,14 +555,19 @@ main(int argc, char** argv)
 
   } // input loop
 
-  if (average) w_average += lambdas;
+  if (t == 0) in_sz = ii; // remember size of input (# lines)
 
-  if (scorer_str == "approx_bleu" || scorer_str == "lc_bleu") scorer->Reset();
 
-  if (t == 0) {
-    in_sz = ii; // remember size of input (# lines)
+  if (batch) {
+    lambdas.plus_eq_v_times_s(batch_updates, eta);
+    if (gamma) lambdas.plus_eq_v_times_s(lambdas, -2*gamma*eta*(1./npairs));
+    batch_updates.clear();
   }
 
+  if (average) w_average += lambdas;
+
+  if (scorer_str == "approx_bleu" || scorer_str == "lc_bleu") scorer->Reset();
+
   // print some stats
   score_t score_avg = score_sum/(score_t)in_sz;
   score_t model_avg = model_sum/(score_t)in_sz;
@@ -477,13 +594,15 @@ main(int argc, char** argv)
     cerr << _np << " 1best avg model score: " << model_avg;
     cerr << _p << " (" << model_diff << ")" << endl;
     cerr << "           avg # pairs: ";
-    cerr << _np << npairs/(float)in_sz;
+    cerr << _np << npairs/(float)in_sz << endl;
+    cerr << "        avg # rank err: ";
+    cerr << rank_errors/(float)in_sz;
     if (faster_perceptron) cerr << " (meaningless)";
     cerr << endl;
-    cerr << "        avg # rank err: ";
-    cerr << rank_errors/(float)in_sz << endl;
     cerr << "     avg # margin viol: ";
     cerr << margin_violations/(float)in_sz << endl;
+    if (batch) cerr << "            batch loss: " << batch_loss << endl;
+    cerr << "       k-best loss imp: " << ((float)kbest_loss_improve/in_sz)*100 << "%" << endl;
     cerr << "    non0 feature count: " <<  nonz << endl;
     cerr << "           avg list sz: " << list_sz/(float)in_sz << endl;
     cerr << "           avg f count: " << f_count/(float)list_sz << endl;
@@ -510,9 +629,9 @@ main(int argc, char** argv)
 
   // write weights to file
   if (select_weights == "best" || keep) {
-    lambdas.init_vector(&dense_weights);
+    lambdas.init_vector(&decoder_weights);
     string w_fn = "weights." + boost::lexical_cast<string>(t) + ".gz";
-    Weights::WriteToFile(w_fn, dense_weights, true);
+    Weights::WriteToFile(w_fn, decoder_weights, true);
   }
 
   } // outer loop
diff --git a/training/dtrain/dtrain.h b/training/dtrain/dtrain.h
index 3981fb39..ccb5ad4d 100644
--- a/training/dtrain/dtrain.h
+++ b/training/dtrain/dtrain.h
@@ -9,6 +9,8 @@
 #include <string.h>
 
 #include <boost/algorithm/string.hpp>
+#include <boost/regex.hpp>
+#include <boost/algorithm/string/regex.hpp>
 #include <boost/program_options.hpp>
 
 #include "decoder.h"
diff --git a/training/dtrain/examples/standard/dtrain.ini b/training/dtrain/examples/standard/dtrain.ini
index 23e94285..fc83f08e 100644
--- a/training/dtrain/examples/standard/dtrain.ini
+++ b/training/dtrain/examples/standard/dtrain.ini
@@ -1,5 +1,6 @@
-input=./nc-wmt11.de.gz
-refs=./nc-wmt11.en.gz
+#input=./nc-wmt11.de.gz
+#refs=./nc-wmt11.en.gz
+bitext=./nc-wmt11.gz
 output=-                  # a weights file (add .gz for gzip compression) or STDOUT '-'
 select_weights=VOID       # output average (over epochs) weight vector
 decoder_config=./cdec.ini # config for cdec
@@ -10,11 +11,11 @@ print_weights=Glue WordPenalty LanguageModel LanguageModel_OOV PhraseModel_0 Phr
 stop_after=10 # stop epoch after 10 inputs
 
 # interesting stuff
-epochs=2                 # run over input 2 times
+epochs=3                 # run over input 3 times
 k=100                    # use 100best lists
 N=4                      # optimize (approx) BLEU4
 scorer=fixed_stupid_bleu # use 'stupid' BLEU+1
-learning_rate=1.0        # learning rate, don't care if gamma=0 (perceptron)
+learning_rate=0.1        # learning rate, don't care if gamma=0 (perceptron) and loss_margin=0 (not margin perceptron)
 gamma=0                  # use SVM reg
 sample_from=kbest        # use kbest lists (as opposed to forest)
 filter=uniq              # only unique entries in kbest (surface form)
@@ -22,3 +23,5 @@ pair_sampling=XYX        #
 hi_lo=0.1                # 10 vs 80 vs 10 and 80 vs 10 here
 pair_threshold=0         # minimum distance in BLEU (here: > 0)
 loss_margin=0            # update if correctly ranked, but within this margin
+repeat=1                 # repeat training on a kbest list 1 times 
+#batch=true              # batch tuning, update after accumulating over all sentences and all kbest lists
diff --git a/training/dtrain/examples/standard/expected-output b/training/dtrain/examples/standard/expected-output
index 21f91244..75f47337 100644
--- a/training/dtrain/examples/standard/expected-output
+++ b/training/dtrain/examples/standard/expected-output
@@ -4,17 +4,18 @@ Reading ./nc-wmt11.en.srilm.gz
 ----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100
 ****************************************************************************************************
   Example feature: Shape_S00000_T00000
-Seeding random number sequence to 970626287
+Seeding random number sequence to 3751911392
 
 dtrain
 Parameters:
                        k 100
                        N 4
-                       T 2
+                       T 3
+                   batch 0
                   scorer 'fixed_stupid_bleu'
              sample from 'kbest'
                   filter 'uniq'
-           learning rate 1
+           learning rate 0.1
                    gamma 0
              loss margin 0
        faster perceptron 1
@@ -23,69 +24,99 @@ Parameters:
           pair threshold 0
           select weights 'VOID'
                   l1 reg 0 'none'
+                    pclr no
                max pairs 4294967295
+                  repeat 1
                 cdec cfg './cdec.ini'
-                   input './nc-wmt11.de.gz'
-                    refs './nc-wmt11.en.gz'
+                   input './nc-wmt11.gz'
                   output '-'
               stop_after 10
 (a dot represents 10 inputs)
-Iteration #1 of 2.
+Iteration #1 of 3.
  . 10
 Stopping after 10 input sentences.
 WEIGHTS
-              Glue = -614
-       WordPenalty = +1256.8
-     LanguageModel = +5610.5
- LanguageModel_OOV = -1449
-     PhraseModel_0 = -2107
-     PhraseModel_1 = -4666.1
-     PhraseModel_2 = -2713.5
-     PhraseModel_3 = +4204.3
-     PhraseModel_4 = -1435.8
-     PhraseModel_5 = +916
-     PhraseModel_6 = +190
-       PassThrough = -2527
+              Glue = -110
+       WordPenalty = -8.2082
+     LanguageModel = -319.91
+ LanguageModel_OOV = -19.2
+     PhraseModel_0 = +312.82
+     PhraseModel_1 = -161.02
+     PhraseModel_2 = -433.65
+     PhraseModel_3 = +291.03
+     PhraseModel_4 = +252.32
+     PhraseModel_5 = +50.6
+     PhraseModel_6 = +146.7
+       PassThrough = -38.7
         ---
-       1best avg score: 0.17874 (+0.17874)
- 1best avg model score: 88399 (+88399)
-           avg # pairs: 798.2 (meaningless)
-        avg # rank err: 798.2
+       1best avg score: 0.16966 (+0.16966)
+ 1best avg model score: 29874 (+29874)
+           avg # pairs: 906.3
+        avg # rank err: 0 (meaningless)
      avg # margin viol: 0
-    non0 feature count: 887
+       k-best loss imp: 100%
+    non0 feature count: 832
            avg list sz: 91.3
-           avg f count: 126.85
-(time 0.33 min, 2 s/S)
+           avg f count: 139.77
+(time 0.35 min, 2.1 s/S)
 
-Iteration #2 of 2.
+Iteration #2 of 3.
  . 10
 WEIGHTS
-              Glue = -1025
-       WordPenalty = +1751.5
-     LanguageModel = +10059
- LanguageModel_OOV = -4490
-     PhraseModel_0 = -2640.7
-     PhraseModel_1 = -3757.4
-     PhraseModel_2 = -1133.1
-     PhraseModel_3 = +1837.3
-     PhraseModel_4 = -3534.3
-     PhraseModel_5 = +2308
-     PhraseModel_6 = +1677
-       PassThrough = -6222
+              Glue = -122.1
+       WordPenalty = +83.689
+     LanguageModel = +233.23
+ LanguageModel_OOV = -145.1
+     PhraseModel_0 = +150.72
+     PhraseModel_1 = -272.84
+     PhraseModel_2 = -418.36
+     PhraseModel_3 = +181.63
+     PhraseModel_4 = -289.47
+     PhraseModel_5 = +140.3
+     PhraseModel_6 = +3.5
+       PassThrough = -109.7
         ---
-       1best avg score: 0.30764 (+0.12891)
- 1best avg model score: -2.5042e+05 (-3.3882e+05)
-           avg # pairs: 725.9 (meaningless)
-        avg # rank err: 725.9
+       1best avg score: 0.17399 (+0.004325)
+ 1best avg model score: 4936.9 (-24937)
+           avg # pairs: 662.4
+        avg # rank err: 0 (meaningless)
      avg # margin viol: 0
-    non0 feature count: 1499
+       k-best loss imp: 100%
+    non0 feature count: 1240
            avg list sz: 91.3
-           avg f count: 114.34
-(time 0.32 min, 1.9 s/S)
+           avg f count: 125.11
+(time 0.27 min, 1.6 s/S)
+
+Iteration #3 of 3.
+ . 10
+WEIGHTS
+              Glue = -157.4
+       WordPenalty = -1.7372
+     LanguageModel = +686.18
+ LanguageModel_OOV = -399.7
+     PhraseModel_0 = -39.876
+     PhraseModel_1 = -341.96
+     PhraseModel_2 = -318.67
+     PhraseModel_3 = +105.08
+     PhraseModel_4 = -290.27
+     PhraseModel_5 = -48.6
+     PhraseModel_6 = -43.6
+       PassThrough = -298.5
+        ---
+       1best avg score: 0.30742 (+0.13343)
+ 1best avg model score: -15393 (-20329)
+           avg # pairs: 623.8
+        avg # rank err: 0 (meaningless)
+     avg # margin viol: 0
+       k-best loss imp: 100%
+    non0 feature count: 1776
+           avg list sz: 91.3
+           avg f count: 118.58
+(time 0.28 min, 1.7 s/S)
 
 Writing weights file to '-' ...
 done
 
 ---
-Best iteration: 2 [SCORE 'fixed_stupid_bleu'=0.30764].
-This took 0.65 min.
+Best iteration: 3 [SCORE 'fixed_stupid_bleu'=0.30742].
+This took 0.9 min.
diff --git a/training/dtrain/examples/standard/nc-wmt11.gz b/training/dtrain/examples/standard/nc-wmt11.gz
new file mode 100644
index 00000000..c39c5aef
--- /dev/null
+++ b/training/dtrain/examples/standard/nc-wmt11.gz
diff --git a/training/dtrain/parallelize.rb b/training/dtrain/parallelize.rb
index 285f3c9b..60ca9422 100755
--- a/training/dtrain/parallelize.rb
+++ b/training/dtrain/parallelize.rb
@@ -21,6 +21,8 @@ opts = Trollop::options do
   opt :qsub, "use qsub", :type => :bool, :default => false
   opt :dtrain_binary, "path to dtrain binary", :type => :string
   opt :extra_qsub, "extra qsub args", :type => :string, :default => ""
+  opt :per_shard_decoder_configs, "give special decoder config per shard", :type => :string, :short => '-o'
+  opt :first_input_weights, "input weights for first iter", :type => :string, :default => '', :short => '-w'
 end
 usage if not opts[:config]&&opts[:shards]&&opts[:input]&&opts[:references]
 
@@ -41,9 +43,11 @@ epochs     = opts[:epochs]
 rand       = opts[:randomize]
 reshard    = opts[:reshard]
 predefined_shards = false
+per_shard_decoder_configs = false
 if opts[:shards] == 0
   predefined_shards = true
   num_shards = 0
+  per_shard_decoder_configs = true if opts[:per_shard_decoder_configs]
 else
   num_shards = opts[:shards]
 end
@@ -51,6 +55,7 @@ input = opts[:input]
 refs  = opts[:references]
 use_qsub       = opts[:qsub]
 shards_at_once = opts[:processes_at_once]
+first_input_weights  = opts[:first_input_weights]
 
 `mkdir work`
 
@@ -101,6 +106,9 @@ refs_files = []
 if predefined_shards
   input_files = File.new(input).readlines.map {|i| i.strip }
   refs_files = File.new(refs).readlines.map {|i| i.strip }
+  if per_shard_decoder_configs
+    decoder_configs = File.new(opts[:per_shard_decoder_configs]).readlines.map {|i| i.strip}
+  end
   num_shards = input_files.size
 else
   input_files, refs_files = make_shards input, refs, num_shards, 0, rand
@@ -126,10 +134,18 @@ end
       else
         local_end = "2>work/out.#{shard}.#{epoch}"
       end
+      if per_shard_decoder_configs
+        cdec_cfg = "--decoder_config #{decoder_configs[shard]}"
+      else
+        cdec_cfg = ""
+      end
+      if first_input_weights!='' && epoch == 0
+        input_weights = "--input_weights #{first_input_weights}"
+      end
       pids << Kernel.fork {
-        `#{qsub_str_start}#{dtrain_bin} -c #{ini}\
+        `#{qsub_str_start}#{dtrain_bin} -c #{ini} #{cdec_cfg} #{input_weights}\
           --input #{input_files[shard]}\
-          --refs #{refs_files[shard]} #{input_weights}\
+          --refs #{refs_files[shard]}\
           --output work/weights.#{shard}.#{epoch}#{qsub_str_end} #{local_end}`
       }
       weights_files << "work/weights.#{shard}.#{epoch}"
diff --git a/utils/filelib.h b/utils/filelib.h
index b9ea3940..4fa69760 100644
--- a/utils/filelib.h
+++ b/utils/filelib.h
@@ -75,7 +75,10 @@ class ReadFile : public BaseFile<std::istream> {
       }
     }
   }
-
+  void ReadAll(std::string& s) {
+    getline(*stream(), s, (char) EOF);
+    if (s.size() > 0) s.resize(s.size()-1);
+  }
 };
 
 class WriteFile : public BaseFile<std::ostream> {