third_party/hunspell/google/bdict_writer.h

   1 // Copyright (c) 2011 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 #ifndef THIRD_PARTY_HUNSPELL_GOOGLE_BDICT_WRITER_H_
   6 #define THIRD_PARTY_HUNSPELL_GOOGLE_BDICT_WRITER_H_
   7
   8 #include <string>
   9 #include <vector>
  10
  11 #include "base/macros.h"
  12
  13 namespace hunspell {
  14
  15 class DicNode;
  16
  17 // Class for creating a binary dictionary file from data read from Hunspell
  18 // spellchecker files.
  19 class BDictWriter {
  20  public:
  21   typedef std::vector< std::pair<std::string, std::vector<int> > > WordList;
  22
  23   BDictWriter();
  24   ~BDictWriter();
  25
  26   // Affix setters.
  27   void SetComment(const std::string& comment) {
  28     comment_ = comment;
  29   }
  30   void SetAffixRules(const std::vector<std::string>& rules) {
  31     affix_rules_ = rules;
  32   }
  33   void SetAffixGroups(const std::vector<std::string>& groups) {
  34     affix_groups_ = groups;
  35   }
  36   void SetReplacements(
  37       const std::vector< std::pair<std::string, std::string> >& replacements) {
  38     replacements_ = replacements;
  39   }
  40   void SetOtherCommands(const std::vector<std::string>& commands) {
  41     other_commands_ = commands;
  42   }
  43
  44   // The words must be sorted already.
  45   void SetWords(const WordList& words);
  46
  47   // Returns the serialized data for the entire file. You must call all the
  48   // setters above before this.
  49   std::string GetBDict() const;
  50
  51  private:
  52   // Converts the affix members to a string.
  53   void SerializeAff(std::string* output) const;
  54
  55   // Affix members.
  56   std::string comment_;
  57   std::vector<std::string> affix_rules_;
  58   std::vector<std::string> affix_groups_;
  59   std::vector< std::pair<std::string, std::string> > replacements_;
  60   std::vector<std::string> other_commands_;
  61
  62   // Root of the generated trie. Filled by SetWords.
  63   DicNode* trie_root_;
  64
  65   DISALLOW_COPY_AND_ASSIGN(BDictWriter);
  66 };
  67
  68 }  // namespace hunspell
  69
  70 #endif  // THIRD_PARTY_HUNSPELL_GOOGLE_BDICT_WRITER_H_