stash/vendor/github.com/agnivade/levenshtein
SmallCoccinelle 45f700d6ea
Support Go 1.18: Upgrade gqlgen to v0.17.2 (#2443)
* Upgrade gqlgen to v0.17.2

This enables builds on Go 1.18. github.com/vektah/gqlparser is upgraded
to the newest version too.

Getting this to work is a bit of a hazzle. I had to first remove
vendoring from the repository, perform the upgrade and then re-introduce
the vendor directory. I think gqlgens analysis went wrong for some
reason on the upgrade. It would seem a clean-room installation fixed it.

* Bump project to 1.18

* Update all packages, address gqlgenc breaking changes

* Let `go mod tidy` handle the go.mod file

* Upgrade linter to 1.45.2

* Introduce v1.45.2 of the linter

The linter now correctly warns on `strings.Title` because it isn't
unicode-aware. Fix this by using the suggested fix from x/text/cases
to produce unicode-aware strings.

The mapping isn't entirely 1-1 as this new approach has a larger iface:
it spans all of unicode rather than just ASCII. It coincides for ASCII
however, so things should be largely the same.

* Ready ourselves for errchkjson and contextcheck.

* Revert dockerfile golang version changes for now

Co-authored-by: Kermie <kermie@isinthe.house>
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
2022-04-02 18:08:14 +11:00
..
.gitignore Updated dependencies 2019-02-09 16:56:50 -08:00
.travis.yml Support Go 1.18: Upgrade gqlgen to v0.17.2 (#2443) 2022-04-02 18:08:14 +11:00
License.txt Updated dependencies 2019-02-09 16:56:50 -08:00
Makefile Stash box client interface (#751) 2020-09-17 19:57:18 +10:00
README.md Support Go 1.18: Upgrade gqlgen to v0.17.2 (#2443) 2022-04-02 18:08:14 +11:00
levenshtein.go Support Go 1.18: Upgrade gqlgen to v0.17.2 (#2443) 2022-04-02 18:08:14 +11:00

README.md

levenshtein Build Status Go Report Card PkgGoDev

Go package to calculate the Levenshtein Distance

The library is fully capable of working with non-ascii strings. But the strings are not normalized. That is left as a user-dependant use case. Please normalize the strings before passing it to the library if you have such a requirement.

Limitation

As a performance optimization, the library can handle strings only up to 65536 characters (runes). If you need to handle strings larger than that, please pin to version 1.0.3.

Install

go get github.com/agnivade/levenshtein

Example

package main

import (
	"fmt"
	"github.com/agnivade/levenshtein"
)

func main() {
	s1 := "kitten"
	s2 := "sitting"
	distance := levenshtein.ComputeDistance(s1, s2)
	fmt.Printf("The distance between %s and %s is %d.\n", s1, s2, distance)
	// Output:
	// The distance between kitten and sitting is 3.
}

Benchmarks

name              time/op
Simple/ASCII-4     330ns ± 2%
Simple/French-4    617ns ± 2%
Simple/Nordic-4   1.16µs ± 4%
Simple/Tibetan-4  1.05µs ± 1%

name              alloc/op
Simple/ASCII-4     96.0B ± 0%
Simple/French-4     128B ± 0%
Simple/Nordic-4     192B ± 0%
Simple/Tibetan-4    144B ± 0%

name              allocs/op
Simple/ASCII-4      1.00 ± 0%
Simple/French-4     1.00 ± 0%
Simple/Nordic-4     1.00 ± 0%
Simple/Tibetan-4    1.00 ± 0%

Comparisons with other libraries

name                     time/op
Leven/ASCII/agniva-4      353ns ± 1%
Leven/ASCII/arbovm-4      485ns ± 1%
Leven/ASCII/dgryski-4     395ns ± 0%
Leven/French/agniva-4     648ns ± 1%
Leven/French/arbovm-4     791ns ± 0%
Leven/French/dgryski-4    682ns ± 0%
Leven/Nordic/agniva-4    1.28µs ± 1%
Leven/Nordic/arbovm-4    1.52µs ± 1%
Leven/Nordic/dgryski-4   1.32µs ± 1%
Leven/Tibetan/agniva-4   1.12µs ± 1%
Leven/Tibetan/arbovm-4   1.31µs ± 0%
Leven/Tibetan/dgryski-4  1.16µs ± 0%