clean up tagger a bit

2 files changed, 15 insertions, 3 deletions

diff --git a/decoder/ff_tagger.cc b/decoder/ff_tagger.cc
index 019315a2..fd9210fa 100644
--- a/decoder/ff_tagger.cc
+++ b/decoder/ff_tagger.cc
@@ -8,6 +8,17 @@
 
 using namespace std;
 
+namespace {
+  string Escape(const string& x) {
+    string y = x;
+    for (int i = 0; i < y.size(); ++i) {
+      if (y[i] == '=') y[i]='_';
+      if (y[i] == ';') y[i]='_';
+    }
+    return y;
+  }
+}
+
 Tagger_BigramIndicator::Tagger_BigramIndicator(const std::string& param) :
   FeatureFunction(sizeof(WordID)) {
    no_uni_ = (LowercaseString(param) == "no_uni");
@@ -28,7 +39,7 @@ void Tagger_BigramIndicator::FireFeature(const WordID& left,
       os << '_';
       if (right < 0) { os << "EOS"; } else { os << TD::Convert(right); }
     }
-    fid = FD::Convert(os.str());
+    fid = FD::Convert(Escape(os.str()));
   }
   features->set_value(fid, 1.0);
 }
@@ -90,7 +101,7 @@ void LexicalPairIndicator::FireFeature(WordID src,
   if (!fid) {
     ostringstream os;
     os << name_ << ':' << TD::Convert(src) << ':' << TD::Convert(trg);
-    fid = FD::Convert(os.str());
+    fid = FD::Convert(Escape(os.str()));
   }
   features->set_value(fid, 1.0);
 }
@@ -127,7 +138,7 @@ void OutputIndicator::FireFeature(WordID trg,
     if (escape.count(trg)) trg = escape[trg];
     ostringstream os;
     os << "T:" << TD::Convert(trg);
-    fid = FD::Convert(os.str());
+    fid = FD::Convert(Escape(os.str()));
   }
   features->set_value(fid, 1.0);
 }
diff --git a/decoder/tagger.cc b/decoder/tagger.cc
index 54890e85..63e855c8 100644
--- a/decoder/tagger.cc
+++ b/decoder/tagger.cc
@@ -54,6 +54,7 @@ struct TaggerImpl {
       const int new_node_id = forest->AddNode(kXCAT)->id_;
       for (int k = 0; k < tagset_.size(); ++k) {
         TRulePtr rule(TRule::CreateLexicalRule(src, tagset_[k]));
+        rule->lhs_ = kXCAT;
         Hypergraph::Edge* edge = forest->AddEdge(rule, Hypergraph::TailNodeVector());
         edge->i_ = i;
         edge->j_ = i+1;