Extract words and explanations from epub version of Merriam Webster's Vocabulary Builder. See http://blog.yxwang.me/notes/languages/vocabulary-builder.html for more details.

extract.rb

require 'nokogiri'

unit = ARGV[0].to_i
started = false
titles = []
1.upto(3000) do |i|
  filename = "html/Merriam-Webster_s_Vocabulary_Bu_split_#{'%03d' % i}.html"
  doc = Nokogiri::HTML(open(filename))
  title = doc.css('div span.bold').first
  next if title == nil
  word = title.content.strip
  if word == "Unit #{unit}"
    started = true
    titles = doc.css('div p.calibre36 a').map(&:content)
    titles = titles[0...titles.find_index { |e| e[/^Quiz/] }].map(&:strip)
  end
  break if word == "Unit #{unit + 1}"
  next unless started
  next if word[/^([ABC]\.)|(Quiz)/]

  begin
    if word.start_with?('Unit ')
      word = doc.css('div p.calibre12 span').first.content.strip
      explanation = doc.css('div p.calibre12').first.content.strip
    elsif word == titles[-1]
      word = doc.css('div p.calibre34 span.bold')[1].content.strip
      explanation = doc.css('div p.calibre34 br').first.previous.content.strip
    elsif titles.include?(word)
      explanation = doc.css('div p').first.content.strip
    else
      explanation = doc.css('div p br').first.previous.content.strip
    end
    puts "#{word}:::#{explanation}"
  rescue => e
    puts word
    puts e
    break
  end
end

See http://blog.yxwang.me/notes/languages/vocabulary-builder.html for more details.