Text 0.3.2 levenshtein normalizeddistance

Lodestar.Text 0.3.2. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

Levenshtein.NormalizedDistance

Scales the distance into [0, 1] by dividing it by the length of the longer input.

public static double NormalizedDistance(ReadOnlySpan<char> a, ReadOnlySpan<char> b, TextElement element = TextElement.Utf16Unit)

Parameters — a and b are the two strings to compare; element says what counts as one character, and it decides both the distance and the lengths it is divided by.

Returns — double in [0, 1]: 0 when the two are equal, 1 when nothing can be reused. Two empty inputs give 0 rather than a division by zero.

Example — three edits over the seven characters of the longer input.

using Lodestar.Text.Distances;

double d = Levenshtein.NormalizedDistance("kitten", "sitting");   // => 0.4285…

Remarks — this is the number to threshold on and the number to compare across pairs; the raw Distance is neither. It matches Levenshtein.normalized_distance in rapidfuzz exactly.

The trap is the same divisor mismatch that catches people between measures: max(len(a), len(b)) here, len(a) + len(b) in Indel.NormalizedDistance. A cut-off of 0.3 is a much stricter demand here than there.

Applies to — net10.0, netstandard2.0.

See also — Levenshtein.Distance, Levenshtein.NormalizedSimilarity, Indel.NormalizedDistance, the Python equivalence table.