fuzzy

Documentation

Overview

package fuzzy contains interfaces and code to facilitate fuzzy-matching of name-strings to scientific names collected in gnames database.

Index

• func EditDistance(s1, s2 string) int
• type FuzzyMatcher

Functions

func EditDistance

func EditDistance(s1, s2 string) int

EditDistance calculates edit distance (**ed**) according to Levenshtein algorithm. It also runs additional checks and if they fail, returns -1.

Checks: - result should not exceed maxEditDistance - number of characters divided by ed should be bigger than charsPerED

It assumes that checks have to be applied only to the second string:

EditDistance("Pomatomus", "Pom atomus")

returns -1

EditDistance("Pom atomus", "Pomatomus")

returns 1

It also assumes that number os spaces between words was already normalized to 1 space, and that s1 and s2 always have the same number of words.

Types

type FuzzyMatcher

type FuzzyMatcher interface {
	// Initialize data for the matcher.
	Init()

	// MatchStem takes a stemmed scientific name and max edit distance.
	// The search stops if current edit distance becomes bigger than edit
	// distance. The method returns 0 or more stems that did match the
	// input stem within the edit distance constraint.
	MatchStem(stem string) []string

	// MatchStemExact takes a stem and returns true if the is the exact
	// match of the stem is found.
	MatchStemExact(stem string) bool

	// StemToCanonicals takes a stem and returns back canonicals
	// that correspond to that stem.
	StemToMatchItems(stem string) []mlib.MatchItem
}

FuzzyMatcher describes methods needed for fuzzy matching.