ngram.rb

require 'pp'
$KCODE = 'UTF8'

module Ngram
  def run(str)
    analyze(cut(str,3))
  end

  def cut(str, n)
    list = str.split(//)
    hash = { }
    hash.default = 0
    0.upto(list.size - n) do |counter|
      token = list[counter..counter+n-1].join('')
      next if token =~ /\n|\r/
      hash[token] += 1
    end
    hash
  end

  def analyze(hash)
    hash.sort_by{ |key, value| value}.reverse
  end
end

include Ngram
str = ARGF.read
pp Ngram.run(str) #.select{ |key, value| value > 1}