9 files changed, 123 insertions, 8 deletions

diff --git a/mert/BleuScorer.cpp b/mert/BleuScorer.cpp
index cae390a18..58450cffb 100644
--- a/mert/BleuScorer.cpp
+++ b/mert/BleuScorer.cpp
@@ -10,6 +10,7 @@
 #include "Ngram.h"
 #include "Reference.h"
 #include "Util.h"
+#include "Vocabulary.h"
 
 namespace {
 
@@ -64,7 +65,7 @@ void BleuScorer::setReferenceFiles(const vector<string>& referenceFiles)
 {
   // Make sure reference data is clear
   m_references.reset();
-  ClearVocabulary();
+  mert::VocabularyFactory::GetVocabulary()->clear();
 
   //load reference data
   for (size_t i = 0; i < referenceFiles.size(); ++i) {
diff --git a/mert/Jamfile b/mert/Jamfile
index ab105db43..874ff8014 100644
--- a/mert/Jamfile
+++ b/mert/Jamfile
@@ -37,6 +37,7 @@ TER/tools.cpp
 TerScorer.cpp
 CderScorer.cpp
 MergeScorer.cpp
+Vocabulary.cpp
 ../util//kenutil m ..//z ;
 
 exe mert : mert.cpp mert_lib ../moses/src//ThreadPool ;
@@ -54,6 +55,7 @@ unit-test data_test : DataTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test ngram_test : NgramTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test optimizer_factory_test : OptimizerFactoryTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test reference_test : ReferenceTest.cpp mert_lib ..//boost_unit_test_framework ;
+unit-test singleton_test : SingletonTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test timer_test : TimerTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test util_test : UtilTest.cpp mert_lib ..//boost_unit_test_framework ;
 unit-test vocabulary_test : VocabularyTest.cpp mert_lib ..//boost_unit_test_framework ;
diff --git a/mert/Scorer.cpp b/mert/Scorer.cpp
index 6142a88dd..3aaae54c0 100644
--- a/mert/Scorer.cpp
+++ b/mert/Scorer.cpp
@@ -3,6 +3,7 @@
 #include <limits>
 #include "Vocabulary.h"
 #include "Util.h"
+#include "Singleton.h"
 
 namespace {
 
@@ -36,14 +37,14 @@ inline float score_average(const statscores_t& scores, size_t start, size_t end)
 
 Scorer::Scorer(const string& name, const string& config)
     : m_name(name),
-      m_vocab(new mert::Vocabulary),
+      m_vocab(mert::VocabularyFactory::GetVocabulary()),
       m_score_data(0),
       m_enable_preserve_case(true) {
   InitConfig(config);
 }
 
 Scorer::~Scorer() {
-  delete m_vocab;
+  Singleton<mert::Vocabulary>::Delete();
 }
 
 void Scorer::InitConfig(const string& config) {
@@ -81,8 +82,6 @@ void Scorer::TokenizeAndEncode(const string& line, vector<int>& encoded) {
   }
 }
 
-void Scorer::ClearVocabulary() { m_vocab->clear(); }
-
 /**
  * Set the factors, which should be used for this metric
  */
diff --git a/mert/Scorer.h b/mert/Scorer.h
index f57293cb8..e9f0e1fe4 100644
--- a/mert/Scorer.h
+++ b/mert/Scorer.h
@@ -144,8 +144,6 @@ class Scorer
    * Note: We assume that all tokens are separated by whitespaces.
    */
   void TokenizeAndEncode(const string& line, vector<int>& encoded);
-
-  void ClearVocabulary();
 };
 
 /**
diff --git a/mert/Singleton.h b/mert/Singleton.h
new file mode 100644
index 000000000..9fef3e639
--- /dev/null
+++ b/mert/Singleton.h
@@ -0,0 +1,33 @@
+#ifndef MERT_SINGLETON_H_
+#define MERT_SINGLETON_H_
+
+#include <cstdlib>
+
+// thread *un*safe singleton.
+// TODO: replace this with thread-safe singleton.
+template <typename T>
+class Singleton {
+ public:
+  static T* GetInstance() {
+    if (m_instance == NULL) {
+      m_instance = new T;
+    }
+    return m_instance;
+  }
+
+  static void Delete() {
+    if (m_instance) {
+      delete m_instance;
+      m_instance = NULL;
+    }
+  }
+
+ private:
+  Singleton();
+  static T* m_instance;
+};
+
+template <typename T>
+T* Singleton<T>::m_instance = NULL;
+
+#endif  // MERT_SINGLETON_H_
diff --git a/mert/SingletonTest.cpp b/mert/SingletonTest.cpp
new file mode 100644
index 000000000..2c44bdc1f
--- /dev/null
+++ b/mert/SingletonTest.cpp
@@ -0,0 +1,27 @@
+#include "Singleton.h"
+
+#define BOOST_TEST_MODULE MertSingleton
+#include <boost/test/unit_test.hpp>
+
+namespace {
+
+static int g_count = 0;
+
+class Instance {
+ public:
+  Instance() { ++g_count; }
+  ~Instance() {}
+};
+
+} // namespace
+
+BOOST_AUTO_TEST_CASE(singleton_basic) {
+  Instance* instance1 = Singleton<Instance>::GetInstance();
+  Instance* instance2 = Singleton<Instance>::GetInstance();
+  Instance* instance3 = Singleton<Instance>::GetInstance();
+  BOOST_REQUIRE(instance1 == instance2);
+  BOOST_REQUIRE(instance2 == instance3);
+  BOOST_CHECK_EQUAL(1, g_count);
+
+  Singleton<Instance>::Delete();
+}
diff --git a/mert/Vocabulary.cpp b/mert/Vocabulary.cpp
new file mode 100644
index 000000000..40b04bf99
--- /dev/null
+++ b/mert/Vocabulary.cpp
@@ -0,0 +1,21 @@
+#include "Vocabulary.h"
+#include "Singleton.h"
+
+namespace mert {
+namespace {
+Vocabulary* g_vocab = NULL;
+} // namespace
+
+Vocabulary* VocabularyFactory::GetVocabulary() {
+  if (g_vocab == NULL) {
+    return Singleton<Vocabulary>::GetInstance();
+  } else {
+    return g_vocab;
+  }
+}
+
+void VocabularyFactory::SetVocabulary(Vocabulary* vocab) {
+  g_vocab = vocab;
+}
+
+} // namespace mert
diff --git a/mert/Vocabulary.h b/mert/Vocabulary.h
index f9493823d..12c8c1727 100644
--- a/mert/Vocabulary.h
+++ b/mert/Vocabulary.h
@@ -64,6 +64,16 @@ class Vocabulary {
   std::map<std::string, int> m_vocab;
 };
 
+class VocabularyFactory {
+ public:
+  static Vocabulary* GetVocabulary();
+  static void SetVocabulary(Vocabulary* vocab);
+
+ private:
+  VocabularyFactory() {}
+  virtual ~VocabularyFactory() {}
+};
+
 } // namespace mert
 
 #endif  // MERT_VOCABULARY_H_
diff --git a/mert/VocabularyTest.cpp b/mert/VocabularyTest.cpp
index 2439301a5..0e67ba62a 100644
--- a/mert/VocabularyTest.cpp
+++ b/mert/VocabularyTest.cpp
@@ -3,8 +3,19 @@
 #define BOOST_TEST_MODULE MertVocabulary
 #include <boost/test/unit_test.hpp>
 
+#include "Singleton.h"
+
+namespace mert {
+namespace {
+
+void TearDown() {
+  Singleton<Vocabulary>::Delete();
+}
+
+} // namespace
+
 BOOST_AUTO_TEST_CASE(vocab_basic) {
-  mert::Vocabulary vocab;
+  Vocabulary vocab;
   BOOST_REQUIRE(vocab.empty());
   vocab.clear();
 
@@ -26,3 +37,16 @@ BOOST_AUTO_TEST_CASE(vocab_basic) {
   BOOST_CHECK(!vocab.Lookup("hello", &v));
   BOOST_CHECK(!vocab.Lookup("world", &v));
 }
+
+BOOST_AUTO_TEST_CASE(vocab_factory_test) {
+  Vocabulary* vocab1 = VocabularyFactory::GetVocabulary();
+  Vocabulary* vocab2 = VocabularyFactory::GetVocabulary();
+  Vocabulary* vocab3 = VocabularyFactory::GetVocabulary();
+
+  BOOST_REQUIRE(vocab1 != NULL);
+  BOOST_CHECK(vocab1 == vocab2);
+  BOOST_CHECK(vocab2 == vocab3);
+
+  TearDown();
+}
+} // namespace mert