From 38a5bee71f6b49515cd105a9467ff602ff9dee64 Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@cs.cmu.edu>
Date: Tue, 13 Sep 2011 13:25:46 +0100
Subject: optional support for doing perfect hashing of feature strings to save
 lots of memory

---
 utils/fdict.h | 36 ++++++++++++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)

(limited to 'utils/fdict.h')

diff --git a/utils/fdict.h b/utils/fdict.h
index f9673023..771e8b91 100644
--- a/utils/fdict.h
+++ b/utils/fdict.h
@@ -1,23 +1,56 @@
 #ifndef _FDICT_H_
 #define _FDICT_H_
 
+#include "config.h"
+
+#include <iostream>
 #include <string>
 #include <vector>
 #include "dict.h"
 
+#ifdef HAVE_CMPH
+#include "perfect_hash.h"
+#include "string_to.h"
+#endif
+
 struct FD {
   // once the FD is frozen, new features not already in the
   // dictionary will return 0
   static void Freeze() {
     frozen_ = true;
   }
+  static bool UsingPerfectHashFunction() {
+#ifdef HAVE_CMPH
+    return hash_;
+#else
+    return false;
+#endif
+  }
+  static void EnableHash(const std::string& cmph_file) {
+#ifdef HAVE_CMPH
+    hash_ = new PerfectHashFunction(cmph_file);
+#endif
+  }
   static inline int NumFeats() {
+#ifdef HAVE_CMPH
+    if (hash_) return hash_->number_of_keys();
+#endif
     return dict_.max() + 1;
   }
   static inline WordID Convert(const std::string& s) {
+#ifdef HAVE_CMPH
+    if (hash_) return (*hash_)(s);
+#endif
     return dict_.Convert(s, frozen_);
   }
   static inline const std::string& Convert(const WordID& w) {
+#ifdef HAVE_CMPH
+    if (hash_) {
+      static std::string tls;
+      tls = to_string(w);
+      return tls;
+    }
+#endif
     return dict_.Convert(w);
   }
   static std::string Convert(WordID const *i,WordID const* e);
@@ -29,6 +62,9 @@ struct FD {
   static Dict dict_;
  private:
   static bool frozen_;
+#ifdef HAVE_CMPH
+  static PerfectHashFunction* hash_;
+#endif
 };
 
 #endif
-- 
cgit v1.2.3