global features

2 files changed, 61 insertions, 0 deletions

diff --git a/rst_parser/global_ff.cc b/rst_parser/global_ff.cc
new file mode 100644
index 00000000..bb715f18
--- /dev/null
+++ b/rst_parser/global_ff.cc
@@ -0,0 +1,43 @@
+#include "global_ff.h"
+
+#include <iostream>
+
+#include "tdict.h"
+
+using namespace std;
+
+struct GFFImpl {
+  void PrepareForInput(const TaggedSentence& sentence) {
+  }
+  void Features(const TaggedSentence& sentence,
+                const EdgeSubset& tree,
+                SparseVector<double>* feats) const {
+    const vector<WordID>& words = sentence.words;
+    const vector<WordID>& tags = sentence.pos;
+    const vector<pair<short,short> >& hms = tree.h_m_pairs;
+    assert(words.size() == tags.size());
+    vector<int> mods(words.size());
+    for (int i = 0; i < hms.size(); ++i) {
+      mods[hms[i].first]++;        // first = head, second = modifier
+    }
+    for (int i = 0; i < mods.size(); ++i) {
+      ostringstream os;
+      os << "NM:" << TD::Convert(tags[i]) << "_" << mods[i];
+      feats->add_value(FD::Convert(os.str()), 1.0);
+    }
+  }
+};
+
+GlobalFeatureFunctions::GlobalFeatureFunctions() {}
+GlobalFeatureFunctions::~GlobalFeatureFunctions() { delete pimpl; }
+
+void GlobalFeatureFunctions::PrepareForInput(const TaggedSentence& sentence) {
+  pimpl->PrepareForInput(sentence);
+}
+
+void GlobalFeatureFunctions::Features(const TaggedSentence& sentence,
+                                      const EdgeSubset& tree,
+                                      SparseVector<double>* feats) const {
+  pimpl->Features(sentence, tree, feats);
+}
+
diff --git a/rst_parser/global_ff.h b/rst_parser/global_ff.h
new file mode 100644
index 00000000..d71d0fa1
--- /dev/null
+++ b/rst_parser/global_ff.h
@@ -0,0 +1,18 @@
+#ifndef _GLOBAL_FF_H_
+#define _GLOBAL_FF_H_
+
+#include "arc_factored.h"
+
+struct GFFImpl;
+struct GlobalFeatureFunctions {
+  GlobalFeatureFunctions();
+  ~GlobalFeatureFunctions();
+  void PrepareForInput(const TaggedSentence& sentence);
+  void Features(const TaggedSentence& sentence,
+                const EdgeSubset& tree,
+                SparseVector<double>* feats) const;
+ private:
+  GFFImpl* pimpl;
+};
+
+#endif