Change the Encoder class to Vocabulary.

- Introduce the namespace to avoid naming collisions. The class name
  is used in KenLM.
- Add the unit test.

1 files changed, 0 insertions, 55 deletions

diff --git a/mert/Ngram.h b/mert/Ngram.h
index d2c5f3932..4b64c327b 100644
--- a/mert/Ngram.h
+++ b/mert/Ngram.h
@@ -5,61 +5,6 @@
 #include <map>
 #include <string>
 
-/**
- * A map to manage vocaburaries.
- */
-class Encoder {
- public:
-  typedef std::map<std::string, int>::iterator iterator;
-  typedef std::map<std::string, int>::const_iterator const_iterator;
-
-  Encoder() {}
-  virtual ~Encoder() {}
-
-  /** Returns the assiged id for given "token". */
-  int Encode(const std::string& token) {
-    iterator it = m_vocab.find(token);
-    int encoded_token;
-    if (it == m_vocab.end()) {
-      // Add an new entry to the vocaburary.
-      encoded_token = static_cast<int>(m_vocab.size());
-      m_vocab[token] = encoded_token;
-    } else {
-      encoded_token = it->second;
-    }
-    return encoded_token;
-  }
-
-  /**
-   * Return true iff the specified "str" is found in the container.
-   */
-  bool Lookup(const std::string&str , int* v) const {
-    const_iterator it = m_vocab.find(str);
-    if (it == m_vocab.end()) return false;
-    *v = it->second;
-    return true;
-  }
-
-  void clear() { m_vocab.clear(); }
-
-  bool empty() const { return m_vocab.empty(); }
-
-  size_t size() const { return m_vocab.size(); }
-
-  iterator find(const std::string& str) { return m_vocab.find(str); }
-  const_iterator find(const std::string& str) const { return m_vocab.find(str); }
-
-  int& operator[](const std::string& str) { return m_vocab[str]; }
-
-  iterator begin() { return m_vocab.begin(); }
-  const_iterator begin() const { return m_vocab.begin(); }
-  iterator end() { return m_vocab.end(); }
-  const_iterator end() const { return m_vocab.end(); }
-
- private:
-  std::map<std::string, int> m_vocab;
-};
-
 /** A simple STL-std::map based n-gram counts. Basically, we provide
  * typical accessors and mutaors, but we intentionally does not allow
  * erasing elements.