「Judy先生の英語科学論文の書き方」付属のCD-ROMのデータをDictionary.app用のXMLデータに変換するスクリプト

judy-to-dictionary.rb

#!/opt/local/bin/ruby1.9
# encoding: utf-8

# 「Judy先生の英語科学論文の書き方」[1]付属のCD-ROMのデータを
# Dictionary.app用のXMLデータに変換するスクリプト。
# Usage :: ruby1.9 judy-to-dictionary.rb > MyDictionary.xml && make && make install
#
# Author :: Masayuki Higashino
#
# References ::
# [1] Judy先生の英語科学論文の書き方
# http://www.amazon.co.jp/dp/4061539523

###### 設定 ################################################

# CD-ROMのルートディレクトリ
JUDY_HOME = '/Users/masayuki/Documents/man/English/Judy/'

############################################################

require 'nokogiri'
require 'rexml/document'
require 'json'

DICT = {}

def format(word)
  key = String.new(word)
  key.gsub!(/  |　　| 　|　 /, '')  # 二重の空白を除去
  key.gsub!('/ ', '/')
  key.gsub!(/\A( |　)*/, '')  # 先頭の空白を削除
  key.gsub!(/( |　)*\z/, '')  # 末尾の空白を削除
  key
end

def en_word_to_key(en_word)
  key = String.new(en_word)
  key.gsub!(/,| |\//, '_')  # カンマや空白やスラッシュはアンダーバーに変換
  key.gsub!(/_{2,}?/, '_')  # 2つ以上の連続したアンダーバーは1つに変換
  key
end

html = open(File.join(JUDY_HOME, 'judy-paper/text/part3.html'), 'r:ascii-8bit'){|i|i.read}
html.encode!(Encoding::UTF_8, Encoding::CP932,
             :undef => :replace,
             :replace => '_',
             :universal_newline => true
             )
html.gsub!(/\n/, ' ') # 変な改行をとりあえず空白にしとく

@event = {}
doc = Nokogiri::HTML(html).xpath('//p').each do |p|
  if p.xpath('font[1]').text =~ /\A∽{5}/
    @event[:type] = :definition
  elsif p.xpath('font[1]').text =~ /\A…{5}/  
    @event[:type] = :description
  elsif p.xpath('font[1]').text =~ /基本例文|基本文例/
    context = p.xpath('font[1]').text.scan(/(基本例文|基本文例).*?：(.+?)＞/).flatten.last
    @event[:context] = context
    @event[:type] = :example_en
  elsif @event[:type] == :definition
    definition = p.children.map{|e|format(e.text)}.reject{|e|e.empty?}
    # '（動詞）' to '動詞'
    definition[2] = definition[2].scan(/（(.+?)）/).flatten.first
    en = definition[0]
    ja = definition[1]
    pos = definition[2]
    key = en_word_to_key(en)
    DICT[key] ||= {}
    DICT[key][:key] ||= key
    DICT[key][:title] ||= en
    DICT[key][:pos] ||= {}
    DICT[key][:pos][pos] ||= {}
    DICT[key][:pos][pos][:en] ||= en
    DICT[key][:pos][pos][:ja] ||= ja
    @event[:key] = key
    @event[:pos] = pos
  elsif @event[:type] == :description
    description = format(p.text)
    description.gsub!(/\A(☆|★)/, '')
    description.gsub!(/\A　|/, '')
    description.gsub!(/　|(， )/, '，')
    key = @event[:key]
    pos = @event[:pos]
    DICT[key][:pos][pos][:description] ||= []
    DICT[key][:pos][pos][:description] << description
  elsif @event[:type] == :example_en
    example = {}
    example_en = format(p.text)
    example[:en] = example_en unless example_en.empty?
    example[:context] = @event[:context] unless @event[:context].nil?
    @event[:example] = example
    @event[:type] = :example_ja
  elsif @event[:type] == :example_ja
    example = @event[:example]
    example_ja = format(p.text)
    example_ja.gsub!(/\A(（|\()/, '')
    example_ja.gsub!(/(）|\))\z/, '')
    example_ja = format(example_ja)
    example[:ja] = example_ja unless example_ja.empty?
    key = @event[:key]
    pos = @event[:pos]
    DICT[key][:pos][pos][:example] ||= []
    DICT[key][:pos][pos][:example] << example
    @event[:example] = example
    @event[:type] = :example_en
  end
end

#puts JSON.pretty_generate(DICT)
#exit
############################################################

CONTEXTS = {}
html = open(File.join(JUDY_HOME, 'judy-paper/text/part2.html'), 'r:ascii-8bit'){|i|i.read}
html.encode!(Encoding::UTF_8, Encoding::CP932,
             :undef => :replace,
             :replace => '_',
             :universal_newline => true)
html.gsub!(/\n/, ' ')  # 変な改行をとりあえず空白にしておく

doc = Nokogiri::HTML(html).xpath('//table').each do |table|
  if table[:width] == '70%'
    table.xpath('tr').each_with_index do |tr, i|
      next if i.zero?  # 1行目は無視
      entry = tr.xpath('td').map{|td|format(td.text)}
      label = entry[1].downcase
      if label.empty?
        unless entry[0].empty?
          @section = entry[0]
        end
      else
        context = ''
        if label =~ /\d\z/  # ラベルの末尾が番号ならサブセクション
          context = [@section, entry[0], entry[2]].compact.join('．')
        else  # そうでなければ単体のセクション
          context = [entry[0], entry[2]].compact.join('．')
        end
        context.gsub!(/ /, '') # 空白除去
        context.gsub!('/FONT>', '定') # タグ入力ミスを修正
        context << '．' unless context =~ /．\z/ # 末尾に句点が無ければ付与
        CONTEXTS[label] = context
      end
    end
    break
  end
end

#puts JSON.pretty_generate(CONTEXTS)
#exit
############################################################

html = open(File.join(JUDY_HOME, 'judy-paper/text/ex/abc-all.html'), 'r:ascii-8bit'){|i|i.read}
html.encode!(Encoding::UTF_8, Encoding::CP932,
             :undef => :replace,
             :replace => '_',
             :universal_newline => true)
html.gsub!(/\n/, ' ')  # 変な改行をとりあえず空白にしておく
doc = Nokogiri::HTML(html).xpath('//tr').each do |p|
  entry   = p.xpath('td/font').map{|e|e.text}
  en      = format(entry[0])
  ja      = format(entry[1])
  pos     = format(entry[2])
  label   = format(entry[3])
  example = format(entry[4])
  key     = en_word_to_key(en)
  DICT[key] ||= {}
  DICT[key][:key] ||= key
  DICT[key][:title] ||= en
  DICT[key][:pos] ||= {}
  DICT[key][:pos][pos] ||= {}
  DICT[key][:pos][pos][:en] ||= en
  DICT[key][:pos][pos][:ja] ||= ja
  DICT[key][:pos][pos][:description] ||= []
  DICT[key][:pos][pos][:example] ||= []
  map = {}
  map[:label] = label
  map[:context] = CONTEXTS[label] if CONTEXTS.include?(label)
  map[:en] = example
  DICT[key][:pos][pos][:example] << map
end

#puts JSON.pretty_generate(DICT)
#exit
############################################################

xml = REXML::Document.new
xml.context[:attribute_quote] = :quote
xml << REXML::XMLDecl.new('1.0', 'UTF-8')
dic = xml.add_element('d:dictionary',{
                        'xmlns' =>'http://www.w3.org/1999/xhtml',
                        'xmlns:d' => 'http://www.apple.com/DTDs/DictionaryService-1.0.rng'
                      })

##### コンテキスト

CONTEXTS.each do |key, value|
  id = key.upcase
  description = value
  entry = dic.add_element('d:entry', {
                            'id' => id,
                            'd:title' => description
                          })
  entry.add_element('d:index', {
                      'd:value'=>id,
                      'd:title'=>id
                    })
  entry.add_element('h1').add_text(id)
  entry.add_element('p').add_text(description)
end

##### 辞書

DICT.each do |key, value|
  ### :)
  # value[:key]
  # value[:title]
  # value[:pos].each do |pos, pos_value|
  #   pos_value[:en]
  #   pos_value[:ja]
  #   pos_value[:description].each do |pos_description|
  #   end
  #   pos_value[:example].each do |pos_example|
  #     pos_example[:label]
  #     pos_example[:context]
  #     pos_example[:en]
  #     pos_example[:ja]
  #   end
  # end
  ###
  entry = dic.add_element('d:entry', {
                            'id' => value[:key],
                            'd:title' => value[:title]
                          })
  # インデックスを付ける
  entry.add_element('d:index', {
                      'd:value' => value[:title],
                      'd:title' => value[:title]
                    })
  # 単語分割してインデックスを付ける
  value[:title].split(/[ ,.-;:]/).reject{|e|e.empty?}.map{|e|e.downcase}.each do |word|
    entry.add_element('d:index', {
                        'd:value'=>word,
                        'd:title'=>value[:title]
                      })
  end
  # タイトル
  entry.add_element('h1').add_text(value[:title])
  # 品詞別
  value[:pos].each do |pos, pos_value|
    # 品詞名
    entry.add_element('span', {'class'=>'encl'}).add_text(pos)

    ul = entry.add_element('ul')
    # 意味
    if pos_value.include?(:description) && !pos_value[:description].empty?
      ul.add_element('li').add_text([pos_value[:ja], '▼', pos_value[:description]].flatten.join(''))
    else
      ul.add_element('li').add_text(pos_value[:ja])
    end
    # 用例
    pos_value[:example].each do |pos_example|
      li = ul.add_element('li')
      li.add_element('span').add_text(pos_example[:en])
      if pos_example.include?(:ja)
        li.add_element('br')
        li.add_element('span').add_text(pos_example[:ja])
      end
      if pos_example.include?(:label)
        li.add_element('br')
        li.add_element('span', {'class'=>'context'}).add_text(pos_example[:label].upcase)
      elsif pos_example.include?(:context)
        li.add_element('br')
        li.add_element('span', {'class'=>'context'}).add_text(pos_example[:context])
      end
    end
  end
end

xml.write($stdout, 4)
puts
exit