detect_typo.rb

require 'levenshtein'

content = ARGF.read.split(/[^\w]/).grep(/^\w+$/)
h = {}
content.each do |c|
  h[c] ||= 0
  h[c] += 1
end

words = h.keys
alone = h.select{|k, v| v == 1}.keys

alone.each do |word|
  next if word.size < 4
  like = h.keys.find{|k| word != k && Levenshtein.normalized_distance(word, k) < 0.2}
  next unless like
  puts "Detect Typo!: #{word}"
  puts "Did you mean #{like} ?"
end