class StringSimilarity

# File lib/string_similarity.rb, line 13
def self.bigram_score(string_1, string_2)
  score(string_1, string_2, MIN_NGRAM_LENGTH)
end

# File lib/string_similarity.rb, line 7
def self.ngram_score(string_1, string_2)
  ngram_length = (string_1.length.to_f * 0.1).round
  ngram_length = [MIN_NGRAM_LENGTH, ngram_length].max
  score(string_1, string_2, ngram_length)
end

# File lib/string_similarity.rb, line 46
def self.capitalization_differences(original_phrase, found_phrase)
  (original_phrase.downcase == found_phrase || found_phrase.downcase == original_phrase ||
    original_phrase.upcase == found_phrase || found_phrase.upcase == original_phrase) &&
    found_phrase != original_phrase
end

# File lib/string_similarity.rb, line 52
def self.normalize_score(score, original_phrase, found_phrase)
  capitalization_differences(original_phrase, found_phrase) ? score - 1 : score
end

# File lib/string_similarity.rb, line 39
def self.remove_special_characters(phrase)
  phrase.gsub!(/[[:punct:]<>]+/, "") # remove all punctuation
  phrase.strip! # remove trailing and leading white space
  phrase.gsub!(/\s+/, " ") # replace extended white space with one space
  phrase.downcase
end

# File lib/string_similarity.rb, line 19
def self.score(string_1, string_2, ngram_length)
  string_1 = String.new(string_1)
  string_2 = String.new(string_2)

  cleaned_string_1 = remove_special_characters(string_1)
  cleaned_string_2 = remove_special_characters(string_2)

  string_1_ngrams = cleaned_string_1.each_char.each_cons(ngram_length).to_set
  string_2_ngrams = cleaned_string_2.each_char.each_cons(ngram_length).to_set

  overlap = (string_1_ngrams & string_2_ngrams).size
  total = string_1_ngrams.size + string_2_ngrams.size
  return 0 unless overlap > 0 && total > 0

  sorensen_dice = overlap * 2.0 / total

  score = (sorensen_dice * 100).round
  normalize_score(score, string_1, string_2)
end