preset

var PlainSimilarityScore = factory.PrependStringSanitizerForSimilarityScore(
	transform.LatinExtendedSanitize,
	SimpleSimilarityScore,
)

PlainSimilarityScore computes the similarity score between two input strings but each input string will be sanitized before they are compared to each other.

var SimpleSimilarityScore = editdist.MakeStringSimilarityFunction(
	editdist.MakeOptimalAlignmentDistFunction(editdist.UnitPenalty, editdist.UnitPenalty),
)

SimpleSimilarityScore computes the similarity score between two input strings. Two input strings will be directly compared under optimal alignment distance metric without any pre-processing, and the resulting distance will be re-normalized to a similarity score between 0 and 1 (inclusive).

var ThaiNameSimilarityScore = factory.PrependStringSanitizerForSimilarityScore(
	sanitaryThai.Sanitize,
	factory.MaxFromCandidatesProduct(
		nametitle.GenerateNamesWithoutTitles,
		editdist.MakeStringSimilarityFunction(
			editdist.MakeOptimalAlignmentDistFunction(editdistThai.SubstPenalty, editdistThai.TransPenalty),
		),
	),
)

ThaiNameSimilarityScore computes the similarity score between two input strings with the following functionalities:

1. Each input string will be sanitized via sanitaryThai.Sanitize function (e.g. removing diacritics from latin scripts, removing repeated Thai tonal marks, etc.)
2. Each input string will be used to generate bare names (i.e. attempting to remove English and Thai titles such as Mrs. or dek-chai)
3. For optimal alignment distance metric over string space, the specialized substitution/transposition penalty functions are used instead.