summaryrefslogtreecommitdiff
path: root/src/dict.h
diff options
context:
space:
mode:
authorChris Dyer <redpony@gmail.com>2009-12-03 16:33:55 -0500
committerChris Dyer <redpony@gmail.com>2009-12-03 16:33:55 -0500
commit671c21451542e2dd20e45b4033d44d8e8735f87b (patch)
treeb1773b077dd65b826f067a423d26f7942ce4e043 /src/dict.h
initial check in
Diffstat (limited to 'src/dict.h')
-rw-r--r--src/dict.h40
1 files changed, 40 insertions, 0 deletions
diff --git a/src/dict.h b/src/dict.h
new file mode 100644
index 00000000..bae9debe
--- /dev/null
+++ b/src/dict.h
@@ -0,0 +1,40 @@
+#ifndef DICT_H_
+#define DICT_H_
+
+#include <cassert>
+#include <cstring>
+#include <tr1/unordered_map>
+#include <string>
+#include <vector>
+
+#include <boost/functional/hash.hpp>
+
+#include "wordid.h"
+
+class Dict {
+ typedef std::tr1::unordered_map<std::string, WordID, boost::hash<std::string> > Map;
+ public:
+ Dict() : b0_("<bad0>") { words_.reserve(1000); }
+ inline int max() const { return words_.size(); }
+ inline WordID Convert(const std::string& word) {
+ Map::iterator i = d_.find(word);
+ if (i == d_.end()) {
+ words_.push_back(word);
+ d_[word] = words_.size();
+ return words_.size();
+ } else {
+ return i->second;
+ }
+ }
+ inline const std::string& Convert(const WordID& id) const {
+ if (id == 0) return b0_;
+ assert(id <= words_.size());
+ return words_[id-1];
+ }
+ private:
+ const std::string b0_;
+ std::vector<std::string> words_;
+ Map d_;
+};
+
+#endif