{"product":"string-similarity","free":true,"input":{"a":"kitten","b":"sitting"},"output":{"ok":true,"result":{"valid":true,"reason":null,"unit":"code_point","ignore_case":false,"a_length":6,"b_length":7,"levenshtein":{"distance":3,"similarity":0.5714285714285714,"matches":4,"substitutions":2,"deletions":0,"insertions":1,"edit_script":[{"op":"substitute","a_index":0,"b_index":0,"a":"k","b":"s"},{"op":"substitute","a_index":4,"b_index":4,"a":"e","b":"i"},{"op":"insert","a_index":6,"b_index":6,"a":null,"b":"g"}]},"osa":{"distance":3,"similarity":0.5714285714285714},"damerau":{"distance":3,"similarity":0.5714285714285714},"hamming":{"distance":null,"similarity":null},"jaro":{"similarity":0.746031746031746,"matches":4,"transpositions":0},"jaro_winkler":{"similarity":0.746031746031746,"prefix_length":0,"boost_applied":false},"lcs":{"length":4,"distance":5,"similarity":0.6153846153846154},"formulas":{"levenshtein_similarity":"1 - distance / max(len_a, len_b); 1 when both are empty","osa_similarity":"1 - distance / max(len_a, len_b); 1 when both are empty","damerau_similarity":"1 - distance / max(len_a, len_b); 1 when both are empty","hamming_similarity":"1 - distance / len; 1 when both are empty; null when the lengths differ","jaro":"(m / len_a + m / len_b + (m - t) / m) / 3 with m matches within floor(max(len_a, len_b) / 2) - 1 positions (scanning a against b, first free match) and t = floor(out-of-order matches / 2); 1 when both are empty, 0 when m = 0","jaro_winkler":"jaro + l * 0.1 * (1 - jaro) when jaro > 0.7, else jaro; l = common prefix length, at most 4","lcs_distance":"len_a + len_b - 2 * lcs_length (edit distance with insertions and deletions only)","lcs_similarity":"2 * lcs_length / (len_a + len_b); 1 when both are empty"},"standard":"Levenshtein (1966) edit distance; optimal string alignment (restricted Damerau-Levenshtein, no substring edited twice); unrestricted Damerau-Levenshtein (Lowrance-Wagner 1975); Hamming (1950); Jaro (1989) and Jaro-Winkler (1990, p = 0.1, prefix up to 4, boost only when Jaro > 0.7, as in strcmp95); longest common subsequence by dynamic programming. Units are Unicode code points or extended grapheme clusters (Unicode UAX #29 via Intl.Segmenter, so grapheme boundaries follow the runtime's Unicode/ICU version); ignore_case is lowercasing with String.prototype.toLowerCase, not full Unicode case folding (ß and SS stay different); no Unicode normalization."},"warnings":["Hamming distance is defined only for sequences of equal length; the lengths differ (6 and 7 units)"]},"paid_url":"https://forge.fizzl.eu/p/string-similarity"}