スペル修正プログラム Ruby版 (改)

初はてぶなのに転載...。
いや、スーパーpre記法を試したくて。

これのRuby版。

class SpellChecker
  def initialize(filename=nil)
    @nwords = train(word(File.read(filename))) if filename
  end

  def word(text)
    text.downcase.gsub(/[^a-z]/, ' ').split(' ')
  end

  def train(features)
    model = Hash.new(1)
    features.each{|f| model[f] += 1}
    model
  end

  def edits1(w, a='abcdefghijklmnopqrstuvwxyz'.split(''))
    n = w.size
    [(0...n).map{|i| w[0...i] + w[i+1...n]},                             # deletion
     (0...n-1).map{|i| w[0, i] + w[i+1].chr + w[i].chr + w[i+2, n-i-2]}, # transposition
     a.map{|c| (0...n).map{|i| w[0...i] + c + w[i+1...n]}},              # alteration
     a.map{|c| (0..n).map {|i| w[0...i] + c + w[i...n]}}].flatten.uniq   # insertion
  end

  def known_edits2(word, known_words=@nwords.keys)
    edits1(word).map{|s1| edits1(s1)}.flatten & known_words
  end

  def known(words, known_words=@nwords.keys)
    words & known_words
  end

  def correct(word, nwords=@nwords)
    @nwords = nwords
    candidates = known([word]) | known(edits1(word)) | known_edits2(word) | [word]
    candidates.max{|c1, c2| @nwords[c1] <=> @nwords[c2]}
  end
end

if __FILE__ == $0
  checker = SpellChecker.new 'big.txt'
  if ARGV.empty?
    puts 'Ctrl+c to stop'
    while gets
      puts checker.correct($_)
    end
  else
    puts checker.correct(ARGV[0])
  end
end