Merge pull request #36 from pks/master

RuleWordAlignmentFeatures, minor dtrain changes

3 files changed, 37 insertions, 0 deletions

diff --git a/decoder/cdec_ff.cc b/decoder/cdec_ff.cc
index d586c1d1..b2541722 100644
--- a/decoder/cdec_ff.cc
+++ b/decoder/cdec_ff.cc
@@ -45,6 +45,7 @@ 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("RuleWordAlignmentFeatures", new FFFactory<RuleWordAlignmentFeatures>());
   ff_registry.Register("ParseMatchFeatures", new FFFactory<ParseMatchFeatures>);
   ff_registry.Register("SoftSyntaxFeatures", new FFFactory<SoftSyntaxFeatures>);
   ff_registry.Register("SoftSyntaxFeaturesMindist", new FFFactory<SoftSyntaxFeaturesMindist>);
diff --git a/decoder/ff_rules.cc b/decoder/ff_rules.cc
index 410e083c..7bccf084 100644
--- a/decoder/ff_rules.cc
+++ b/decoder/ff_rules.cc
@@ -12,6 +12,7 @@
 #include "verbose.h"
 #include "tdict.h"
 #include "hg.h"
+#include "trule.h"
 
 using namespace std;
 
@@ -68,6 +69,28 @@ void RuleIdentityFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
   features->add_value(it->second, 1);
 }
 
+RuleWordAlignmentFeatures::RuleWordAlignmentFeatures(const std::string& param) {
+}
+
+void RuleWordAlignmentFeatures::PrepareForInput(const SentenceMetadata& smeta) {
+}
+
+void RuleWordAlignmentFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
+                                         const Hypergraph::Edge& edge,
+                                         const vector<const void*>& ant_contexts,
+                                         SparseVector<double>* features,
+                                         SparseVector<double>* estimated_features,
+                                         void* context) const {
+  const TRule& rule = *edge.rule_;
+  ostringstream os;
+  vector<AlignmentPoint> als = rule.als(); 
+  std::vector<AlignmentPoint>::const_iterator xx = als.begin();
+  for (; xx != als.end(); ++xx) {
+    os << "WA:" <<  TD::Convert(rule.f_[xx->s_]) << ":" << TD::Convert(rule.e_[xx->t_]);
+  }
+  features->add_value(FD::Convert(Escape(os.str())), 1);
+}
+
 RuleSourceBigramFeatures::RuleSourceBigramFeatures(const std::string& param) {
 }
 
diff --git a/decoder/ff_rules.h b/decoder/ff_rules.h
index f210dc65..324d7a39 100644
--- a/decoder/ff_rules.h
+++ b/decoder/ff_rules.h
@@ -24,6 +24,19 @@ class RuleIdentityFeatures : public FeatureFunction {
   mutable std::map<const TRule*, int> rule2_fid_;
 };
 
+class RuleWordAlignmentFeatures : public FeatureFunction {
+ public:
+  RuleWordAlignmentFeatures(const std::string& param);
+ protected:
+  virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta,
+                                     const HG::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);
+};
+
 class RuleSourceBigramFeatures : public FeatureFunction {
  public:
   RuleSourceBigramFeatures(const std::string& param);