Class: Okura::Serializer::FormatInfo

Inherits:
Object
  • Object
show all
Defined in:
lib/okura/serializer.rb

Overview

辞書ファイルのコンパイル形式を表現し、コンパイルとロードの制御を担当する

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initializeFormatInfo

Returns a new instance of FormatInfo.



9
10
11
12
13
14
15
16
# File 'lib/okura/serializer.rb', line 9

def initialize
  @word_dic=:DoubleArray
  @unk_dic=:Marshal
  @features=:Marshal
  @char_types=:Marshal
  @matrix=:Marshal
  @encoding='EUC-JP'
end

Instance Attribute Details

#char_typesObject

Returns the value of attribute char_types.



20
21
22
# File 'lib/okura/serializer.rb', line 20

def char_types
  @char_types
end

#encodingObject

Returns the value of attribute encoding.



22
23
24
# File 'lib/okura/serializer.rb', line 22

def encoding
  @encoding
end

#featuresObject

Returns the value of attribute features.



19
20
21
# File 'lib/okura/serializer.rb', line 19

def features
  @features
end

#matrixObject

Returns the value of attribute matrix.



21
22
23
# File 'lib/okura/serializer.rb', line 21

def matrix
  @matrix
end

#unk_dicObject

Returns the value of attribute unk_dic.



18
19
20
# File 'lib/okura/serializer.rb', line 18

def unk_dic
  @unk_dic
end

#word_dicObject

Returns the value of attribute word_dic.



17
18
19
# File 'lib/okura/serializer.rb', line 17

def word_dic
  @word_dic
end

Class Method Details

.create_tagger(bin_dir) ⇒ Object

指定されたディレクトリにあるコンパイル済み辞書をロードし、Taggerを作成する



65
66
67
68
# File 'lib/okura/serializer.rb', line 65

def self.create_tagger bin_dir
  format_info=File.open(File.join(bin_dir,'format-info')){|f| self.load f }
  format_info.create_tagger bin_dir
end

.load(io) ⇒ Object

シリアライズされたFormatInfoオブジェクトを復元する



100
101
102
103
104
105
106
107
108
109
# File 'lib/okura/serializer.rb', line 100

def self.load io
  data=YAML.load(io)
  fi=FormatInfo.new
  fi.word_dic=data[:word_dic]
  fi.unk_dic=data[:unk_dic]
  fi.features=data[:features]
  fi.char_types=data[:char_types]
  fi.matrix=data[:matrix]
  fi
end

Instance Method Details

#compile(io) ⇒ Object

このFormatInfoオブジェクトをシリアライズする



90
91
92
93
94
95
96
97
98
# File 'lib/okura/serializer.rb', line 90

def compile io
  YAML.dump({
    word_dic: word_dic,
    unk_dic: unk_dic,
    features: features,
    char_types: char_types,
    matrix: matrix
  },io)
end

#compile_dict(src_dir, bin_dir) ⇒ Object

指定されたディレクトリにあるソースをコンパイルする



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
# File 'lib/okura/serializer.rb', line 25

def compile_dict src_dir,bin_dir
  open_dest(bin_dir,'format-info'){|dest| self.compile dest}
  features_l=open_src(src_dir,'left-id.def'){|src|
    open_dest(bin_dir,'left-id.bin'){|dest|
      serializer_for('Features',features).compile(src,dest)
    }
  }

  features_r=open_src(src_dir,'right-id.def'){|src|
    open_dest(bin_dir,'right-id.bin'){|dest|
      serializer_for('Features',features).compile(src,dest)
    }
  }

  word_src_files=
    Dir.chdir(src_dir){ Dir.glob('*.csv') }.
    map{|file|File.join(src_dir,file)}
  open_dest(bin_dir,'word_dic.bin'){|dest|
    serializer_for('WordDic',word_dic).compile(features_l,features_r,word_src_files,encoding,dest)
  }

  char_types=open_src(src_dir,'char.def'){|src|
    open_dest(bin_dir,'char_types.bin'){|dest|
      serializer_for('CharTypes',@char_types).compile(src,dest)
    }
  }

  open_src(src_dir,'unk.def'){|src|
    open_dest(bin_dir,'unk_dic.bin'){|dest|
      serializer_for('UnkDic',unk_dic).compile(char_types,features_l,features_r,src,dest)
    }
  }

  open_src(src_dir,'matrix.def'){|src|
    open_dest(bin_dir,'matrix.bin'){|dest|
      serializer_for('Matrix',matrix).compile(src,dest)
    }
  }
end

#create_tagger(bin_dir) ⇒ Object



69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
# File 'lib/okura/serializer.rb', line 69

def create_tagger bin_dir
  features_l=open_bin(bin_dir,'left-id.bin'){|bin|
    serializer_for('Features',features).load(bin)
  }
  features_r=open_bin(bin_dir,'right-id.bin'){|bin|
    serializer_for('Features',features).load(bin)
  }
  wd=open_bin(bin_dir,'word_dic.bin'){|f|
    serializer_for('WordDic',word_dic).load(f)
  }
  ud=open_bin(bin_dir,'unk_dic.bin'){|f|
    serializer_for('UnkDic',unk_dic).load(f)
  }
  mat=open_bin(bin_dir,'matrix.bin'){|f|
    serializer_for('Matrix',matrix).load(f)
  }
  dic=Okura::Dic.new wd,ud
  tagger=Okura::Tagger.new dic,mat
  tagger
end