Merge branch 'master' of https://github.com/redpony/cdec

1 files changed, 43 insertions, 5 deletions

diff --git a/decoder/ff_rules.cc b/decoder/ff_rules.cc
index 0aafb0ba..6716d3da 100644
--- a/decoder/ff_rules.cc
+++ b/decoder/ff_rules.cc
@@ -68,15 +68,15 @@ void RuleIdentityFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
   features->add_value(it->second, 1);
 }
 
-RuleNgramFeatures::RuleNgramFeatures(const std::string& param) {
+RuleSourceBigramFeatures::RuleSourceBigramFeatures(const std::string& param) {
 }
 
-void RuleNgramFeatures::PrepareForInput(const SentenceMetadata& smeta) {
+void RuleSourceBigramFeatures::PrepareForInput(const SentenceMetadata& smeta) {
 //  std::map<const TRule*, SparseVector<double> >
   rule2_feats_.clear();
 }
 
-void RuleNgramFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
+void RuleSourceBigramFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
                                          const Hypergraph::Edge& edge,
                                          const vector<const void*>& ant_contexts,
                                          SparseVector<double>* features,
@@ -94,14 +94,52 @@ void RuleNgramFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
       assert(w > 0);
       const string& cur = TD::Convert(w);
       ostringstream os;
-      os << "RB:" << prev << '_' << cur;
+      os << "RBS:" << prev << '_' << cur;
       const int fid = FD::Convert(Escape(os.str()));
       if (fid <= 0) return;
       f.add_value(fid, 1.0);
       prev = cur;
     }
     ostringstream os;
-    os << "RB:" << prev << '_' << "</r>";
+    os << "RBS:" << prev << '_' << "</r>";
+    f.set_value(FD::Convert(Escape(os.str())), 1.0);
+  }
+  (*features) += it->second;
+}
+
+RuleTargetBigramFeatures::RuleTargetBigramFeatures(const std::string& param) {
+}
+
+void RuleTargetBigramFeatures::PrepareForInput(const SentenceMetadata& smeta) {
+  rule2_feats_.clear();
+}
+
+void RuleTargetBigramFeatures::TraversalFeaturesImpl(const SentenceMetadata& smeta,
+                                         const Hypergraph::Edge& edge,
+                                         const vector<const void*>& ant_contexts,
+                                         SparseVector<double>* features,
+                                         SparseVector<double>* estimated_features,
+                                         void* context) const {
+  map<const TRule*, SparseVector<double> >::iterator it = rule2_feats_.find(edge.rule_.get());
+  if (it == rule2_feats_.end()) {
+    const TRule& rule = *edge.rule_;
+    it = rule2_feats_.insert(make_pair(&rule, SparseVector<double>())).first;
+    SparseVector<double>& f = it->second;
+    string prev = "<r>";
+    for (int i = 0; i < rule.e_.size(); ++i) {
+      WordID w = rule.e_[i];
+      if (w < 0) w = -w;
+      if (w == 0) return;
+      const string& cur = TD::Convert(w);
+      ostringstream os;
+      os << "RBT:" << prev << '_' << cur;
+      const int fid = FD::Convert(Escape(os.str()));
+      if (fid <= 0) return;
+      f.add_value(fid, 1.0);
+      prev = cur;
+    }
+    ostringstream os;
+    os << "RBT:" << prev << '_' << "</r>";
     f.set_value(FD::Convert(Escape(os.str())), 1.0);
   }
   (*features) += it->second;