From 851e389dffdd6996ea32d70defb8906de80b9edc Mon Sep 17 00:00:00 2001
From: Chris Dyer <redpony@gmail.com>
Date: Mon, 14 Dec 2009 20:35:11 -0500
Subject: few small fixes of alignment tools, add new orthographic similarity
 feature for word aligner, final naming of directories, libraries in cdec

---
 decoder/grammar.h | 83 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
 create mode 100644 decoder/grammar.h

(limited to 'decoder/grammar.h')

diff --git a/decoder/grammar.h b/decoder/grammar.h
new file mode 100644
index 00000000..3471e3f1
--- /dev/null
+++ b/decoder/grammar.h
@@ -0,0 +1,83 @@
+#ifndef GRAMMAR_H_
+#define GRAMMAR_H_
+
+#include <vector>
+#include <map>
+#include <set>
+#include <boost/shared_ptr.hpp>
+
+#include "lattice.h"
+#include "trule.h"
+
+struct RuleBin {
+  virtual ~RuleBin();
+  virtual int GetNumRules() const = 0;
+  virtual TRulePtr GetIthRule(int i) const = 0;
+  virtual int Arity() const = 0;
+};
+
+struct GrammarIter {
+  virtual ~GrammarIter();
+  virtual const RuleBin* GetRules() const = 0;
+  virtual const GrammarIter* Extend(int symbol) const = 0;
+};
+
+struct Grammar {
+  typedef std::map<WordID, std::vector<TRulePtr> > Cat2Rules;
+  static const std::vector<TRulePtr> NO_RULES;
+
+  virtual ~Grammar();
+  virtual const GrammarIter* GetRoot() const = 0;
+  virtual bool HasRuleForSpan(int i, int j, int distance) const;
+
+  // cat is the category to be rewritten
+  inline const std::vector<TRulePtr>& GetAllUnaryRules() const {
+    return unaries_;
+  }
+
+  // get all the unary rules that rewrite category cat
+  inline const std::vector<TRulePtr>& GetUnaryRulesForRHS(const WordID& cat) const {
+    Cat2Rules::const_iterator found = rhs2unaries_.find(cat);
+    if (found == rhs2unaries_.end())
+      return NO_RULES;
+    else
+      return found->second;
+  }
+
+ protected:
+  Cat2Rules rhs2unaries_;     // these must be filled in by subclasses!
+  std::vector<TRulePtr> unaries_;
+};
+
+typedef boost::shared_ptr<Grammar> GrammarPtr;
+
+class TGImpl;
+struct TextGrammar : public Grammar {
+  TextGrammar();
+  TextGrammar(const std::string& file);
+  void SetMaxSpan(int m) { max_span_ = m; }
+  virtual const GrammarIter* GetRoot() const;
+  void AddRule(const TRulePtr& rule);
+  void ReadFromFile(const std::string& filename);
+  virtual bool HasRuleForSpan(int i, int j, int distance) const;
+  const std::vector<TRulePtr>& GetUnaryRules(const WordID& cat) const;
+ private:
+  int max_span_;
+  boost::shared_ptr<TGImpl> pimpl_;
+};
+
+struct GlueGrammar : public TextGrammar {
+  // read glue grammar from file
+  explicit GlueGrammar(const std::string& file);
+  GlueGrammar(const std::string& goal_nt, const std::string& default_nt);  // "S", "X"
+  virtual bool HasRuleForSpan(int i, int j, int distance) const;
+};
+
+struct PassThroughGrammar : public TextGrammar {
+  PassThroughGrammar(const Lattice& input, const std::string& cat);
+  virtual bool HasRuleForSpan(int i, int j, int distance) const;
+ private:
+  std::vector<std::set<int> > has_rule_;  // index by [i][j]
+};
+
+#endif
-- 
cgit v1.2.3