Class: Tataki::Converter::SkkJisyo

Inherits:
Base
  • Object
show all
Defined in:
lib/tataki/converters/skk_jisyo.rb

Constant Summary collapse

DEFAULT_CONFIG_PATH =
"../../../../data/skk-jisyo.yml"
DEFAULT_JISYO_SUFFIXES =
%w[M]

Instance Method Summary collapse

Constructor Details

#initialize(jisyo_types = DEFAULT_JISYO_SUFFIXES) ⇒ SkkJisyo

Returns a new instance of SkkJisyo.



12
13
14
15
16
17
18
19
20
21
22
23
# File 'lib/tataki/converters/skk_jisyo.rb', line 12

def initialize(jisyo_types = DEFAULT_JISYO_SUFFIXES)
  @jisyo_paths = jisyo_types.map{|suffix| Skk::Jisyo.path(suffix) }
  @table_cache_path = table_cache_path(jisyo_types.join("_"))

  config_file = File.expand_path(DEFAULT_CONFIG_PATH, __FILE__)
  config_data = YAML.load_file(config_file)
  @roman_data = config_data["roman_table"]
  @ignore_kana = config_data["ignore_kana"]
  tables = setup_jisyo
  @match_table = tables[0].freeze
  @okurigana_table = tables[1].freeze
end

Instance Method Details

#add_jisyo(match_table, okurigana_table, jisyo_path) ⇒ Object



41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
# File 'lib/tataki/converters/skk_jisyo.rb', line 41

def add_jisyo(match_table, okurigana_table, jisyo_path)
  File.open(jisyo_path, "rb:euc-jp") do |jisyo_file|
    jisyo_file.each_line do |line|
      next if line.empty? || line[0] == ";" || line.include?("#")
      kana, kanji_part = line.encode("utf-8").split(" ")
      next unless kana && kanji_part
      kana.gsub!(/[^ぁ-んa-z]/, "")
      next if kana.empty? || !(kana =~ /^[ぁ-ん]+[a-z]?/) || @ignore_kana.include?(kana)
      kanji_part.gsub!(/^\/|;.+|\/$/, "")

      table = kana =~ /^(.+)([a-z])$/ ? okurigana_table : match_table
      kanji_part.split("/").each do |kanji|
        kanji_prefix = kanji[0]
        table_entry = table[kanji_prefix]
        table[kanji_prefix] = table_entry = [] unless table_entry
        table_entry.push($2 ? [kanji, $1, $2] : [kanji, kana])
        table_entry.sort_by!{|entry| - (entry[0].size) }
      end
    end
  end
end

#jisyo_pathObject



63
64
65
# File 'lib/tataki/converters/skk_jisyo.rb', line 63

def jisyo_path
  File.expand_path("../../../../data/jisyo", __FILE__)
end

#jisyo_timestamp(path) ⇒ Object



71
72
73
# File 'lib/tataki/converters/skk_jisyo.rb', line 71

def jisyo_timestamp(path)
  Time.parse(File.read("#{path}.timestamp"))
end

#setup_jisyoObject



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
# File 'lib/tataki/converters/skk_jisyo.rb', line 25

def setup_jisyo
  if File.exist?(@table_cache_path)
    tables = Marshal.load(File.read(@table_cache_path))
  else
    match_table = {}
    okurigana_table = {}
    @jisyo_paths.each do |jisyo_path|
      add_jisyo(match_table, okurigana_table, jisyo_path)
    end
    tables = [match_table, okurigana_table]
    File.binwrite(@table_cache_path, Marshal.dump(tables))
    File.write("#{@table_cache_path}.timestamp", Time.now.to_s)
  end
  tables
end

#table_cache_path(name) ⇒ Object



67
68
69
# File 'lib/tataki/converters/skk_jisyo.rb', line 67

def table_cache_path(name)
  File.join(jisyo_path, "SKK-JISYO.#{name}.table.cache")
end

#to_kana(sentence) ⇒ Object



75
76
77
# File 'lib/tataki/converters/skk_jisyo.rb', line 75

def to_kana(sentence)
  _to_kana(sentence, "")
end