Underscore mixin for Levenshtein distance of two strings Adapted from [Levenshtein.coffee](https://raw.github.com/phstc/levenshtein/master/src/Levenshtein.coffee) by @phstc (http://coffeescriptcookbook.com/chapters/strings/matching-strings was broken)

levenshteinDistanceMixin

# usage _.levenshteinDistance("PART", "PARTY") or _("PART").levenshteinDistance("PARTY")
_.mixin
  levenshtein : (str1, str2) ->
    return str2.length if str1.length == 0
    distance = []

    for i in [0..str1.length]
      distance[i] = []
      distance[i][0] = i

    for j in [0..str2.length]
      distance[0][j] = j

    for i in [1..str1.length]
      for j in [1..str2.length]
        deletion         = distance[i - 1][j] + 1
        insertion        = distance[i][j - 1] + 1
        substitutionCost = if (str1.charAt(i - 1) == str2.charAt(j - 1)) then 0 else 1
        substitution     = distance[i - 1][j - 1] + substitutionCost
        distance[i][j]   = Math.min(Math.min(deletion, insertion), substitution)

    distance[str1.length][str2.length]