@nozbe/microfuzz
Version:
A tiny, simple, fast fuzzy search library
305 lines (298 loc) • 11.1 kB
JavaScript
import createFuzzySearch, { normalizeText } from '../index'
import { experimentalSmartFuzzyMatch } from '../impl'
describe('createFuzzySearch', () => {
const matches = (text, query, expectedScore, expectedIndices = null) => {
// console.log(`${text} ${query}`)
const results = createFuzzySearch([text])(query)
expect(results.length).toBe(1)
const [result] = results
expect(result.item).toBe(text)
if (expectedScore != null) {
expect(result.score).toBeCloseTo(expectedScore)
}
if (expectedIndices) {
expect(result.matches.length).toBe(1)
expect(result.matches[0]).toEqual(expectedIndices)
}
}
it(`can match by: exact match`, () => {
matches('foo', 'foo', 0, [[0, 2]])
matches('ABC', 'ABC', 0)
matches('ąść', 'ąść', 0)
matches('📚', '📚', 0)
matches('123¡™£§', '123¡™£§', 0)
matches('żabki', 'żabki', 0)
matches('ząbki', 'ząbki', 0)
matches('русский язык', 'русский язык', 0)
matches('汉语', '汉语', 0)
matches('日本語', '日本語', 0)
matches('한국어', '한국어', 0)
matches('ภาษาไทย', 'ภาษาไทย', 0)
matches(' he ', ' he ', 0)
matches(' he\n\t', ' he\n\t', 0) // hard space
})
it(`can match by: full match`, () => {
matches('Foo', 'foo', 0.1)
matches('FOO', 'foo', 0.1)
matches('foo', 'Foo', 0.1)
matches('foo', 'FOO', 0.1)
matches('foo', 'foo ', 0.1)
matches(' foo bar', 'foo bar', 0.1)
// eslint-disable-next-line
// matches('foo bar', 'foo bar', 0.1) // hard space
matches('Żabki', 'żabki', 0.1, [[0, 4]])
matches('Żabki', 'zabki', 0.1)
matches('Ząbki', 'zabki', 0.1)
matches('ZABKI', 'żąbki', 0.1)
matches('Szczegół', 'szczegol', 0.1)
matches('Язык', 'язык', 0.1, [[0, 3]])
// Check for regression - previously highlight would be off by whitespace
matches('foo ', 'foo', 0.1, [[0, 2]])
})
it(`can match by: "Starts with" match`, () => {
// TODO: startsWith with exact diacritics should match more strongly (e.g. ża -> Żabka before Zabawa), but case-insensitively
// (we want `to` to match Tom more than Couto)
matches('Tomasz Kapelak', 'to', 0.5, [[0, 1]])
matches('Żabka - oferta', 'Żab', 0.5, [[0, 2]])
matches('Żabka - oferta', 'Zab', 0.5)
matches('Żabka - oferta', 'zab', 0.5)
matches('Szczegółowe', 'szcz', 0.5, [[0, 3]])
matches('Русский язык', 'рус', 0.5, [[0, 2]])
matches('汉语', '汉', 0.5, [[0, 0]])
matches('日本語', '日', 0.5, [[0, 0]])
// TODO: Fix Hangul highlighting
// matches('한국어', '한', 0.5, [[0, 0]])
matches('ภาษาไทย', 'ภ', 0.5, [[0, 0]])
// Check for regression - previously highlight would be 1 longer
matches('There is no icon', 'the ', 0.5, [[0, 2]])
})
it(`can match by: contains query (at word boundary) exactly`, () => {
matches('[N4] Marketing', 'Market', 0.9, [[5, 10]])
matches('Wypad do Żabki', 'Żabk', 0.9)
matches('русский язык', 'язык', 0.9, [[8, 11]])
})
it(`can match by: contains query (at word boundary)`, () => {
matches('[N4] Marketing', 'market', 1, [[5, 10]])
matches('Wypad do Żabki', 'zabki', 1)
matches('Myjcie ząbki!!', 'zabki', 1, [[7, 11]])
matches('Русский Язык', 'язык', 1, [[8, 11]])
})
it(`can match by: contains query (at any position)`, () => {
matches('Marco Couto', 'To', 2, [[9, 10]])
matches('汉语', '语', 2, [[1, 1]])
matches('日本語', '本', 2, [[1, 1]])
// TODO: Fix Hangul highlighting
// matches('한국어', '국', 2, [[1, 1]])
matches('ภาษาไทย', 'า', 2, [[1, 1]])
})
it(`can match by words (in some order)`, () => {
matches('Setting to disable fuzzy search', 'fuzzy setting', 1.9, [
[0, 6],
[19, 23],
])
matches('Setting to disable fuzzy search', 'disable setting fuzzy search', 2.3, [
[0, 6],
[11, 17],
[19, 23],
[25, 30],
])
// matches by words before matching by letters
matches('Cloak Duck, test clock', 'clock test', 1.9, [
[12, 15],
[17, 21],
])
})
it(`can match by: contains letters awww yisss fuzzzzz`, () => {
// score(of)
const s = (score) => 2 + score
// full word
const w = 0.2
// beginning of word
const b = 0.4
// middle of word
const m = 0.8
// middle of word (1 or 2 chars)
const ms = 1.6
// TODO: Matching diacritics should be scored better
matches('Wypad do Żabki', 'wdż', s(b + ms + b), [
[0, 0],
[4, 4],
[9, 9],
])
matches('Wypad do Żabki', 'wdz', s(b + ms + b), [
[0, 0],
[4, 4],
[9, 9],
])
matches('Wypad do Żabki', 'wypażab', s(b + b), [
[0, 3],
[9, 11],
])
matches('Wypad do Żabki', 'wypa żab', s(b + w + b), [
[0, 3],
[5, 5],
[9, 11],
])
matches('Marco Couto', 'mc', s(b + ms))
matches('Marco Couto', 'm c', s(b + b))
matches('Tomasz Kapelak', 'tokp', s(b + b + ms))
// trying to match `Referral` (and not r, e in Marketing) is probably too complex/magic
matches('[Marketing ] Referral Program', 'mrefp', s(ms + ms + ms + ms + b))
matches('[Marketing ] Referral Program', 'm refp', s(ms + w + b + b))
matches('Nozbe.com web site', 'website')
matches('Książka 10 steps EN', '10en', s(w + ms + ms))
matches('Won’t fix', 'wontfix')
matches('[HR] JavaScript Developer', 'jsdev', s(b + ms + b), [
[5, 5],
[9, 9],
[16, 18],
])
matches('MacKay', 'mckay', s(b + m))
matches('Русский Язык', 'уски', s(ms + ms))
matches('Русский Язык', 'руя', s(b + b), [
[0, 1],
[8, 8],
])
matches('汉语的，又称汉文、華文、唐文', '汉语唐', s(b + ms), [
[0, 1],
[12, 12],
])
matches('日本語', '日語', s(b + ms), [
[0, 0],
[2, 2],
])
matches('한국어', '한어', s(b + ms))
// matches('한국어', '한어', 2, [[0, 0], [2, 2]]) // FIXME: Fix highglighting for Hangul
})
const noMatch = (text, query) => {
const results = createFuzzySearch([text])(query)
expect(results.length).toBe(0)
}
it(`can not match everything, okay :(`, () => {
// no stemming
noMatch('recognition', 'recognize')
noMatch('production', 'produce')
noMatch('żołądź', 'żołędzie')
noMatch('take', 'took')
noMatch('produce', 'reproduce')
// no synonyms/alt spellings
noMatch('McKay', 'MacKay')
noMatch('mac', 'macintosh')
noMatch('grey', 'gray')
// no soundex
noMatch('kay', 'kai')
// no substitutions/typo autofix
noMatch('leters', 'letters')
noMatch('letters', 'lettesr')
noMatch('referral', 'referarl')
})
it(`can search by key`, () => {
expect(
createFuzzySearch([{ t: 'foo' }, { t: 'foo2' }, { t: 'bar' }], { key: 't' })('foo'),
).toMatchObject([{ item: { t: 'foo' } }, { item: { t: 'foo2' } }])
})
it(`can search by many keys`, () => {
const u1 = { name: 'foo1', alias: 'fooa1' }
const u2 = { name: 'foo2', alias: 'bar' }
const u3 = { name: 'bar', alias: '3foo' }
const u4 = { name: 'bar', alias: 'bar' }
expect(
createFuzzySearch([u1, u2, u3, u4], {
getText: (item) => [item.name, item.alias],
})('foo'),
).toMatchObject([
{
item: u1,
matches: [[[0, 2]], [[0, 2]]],
},
{ item: u2, matches: [[[0, 2]], null] },
{ item: u3, matches: [null, [[1, 3]]] },
])
})
it(`sorts searches by score`, () => {
expect(
createFuzzySearch([
'[Marketing] Żabka etc.',
'Zabawny Katar',
'Żal Betoniarka',
'Ząbka',
'Żabka',
'Żabowe Karabiny',
'Żabka - oferta współpracy',
'żabka',
'[Marketing] żabka',
])('żabka'),
).toMatchObject([
{ item: 'żabka' /* score: 0 */ },
{ item: 'Ząbka' /* score: 0.1 */ },
{ item: 'Żabka' /* score: 0.1 */ },
{ item: 'Żabka - oferta współpracy' /* score: 0.5 */ },
{ item: '[Marketing] żabka' /* score: 0.9 */ },
{ item: '[Marketing] Żabka etc.' /* score: 1 */ },
{ item: 'Zabawny Katar' /* score: 2 */ },
{ item: 'Żabowe Karabiny' /* score: 2 */ },
{ item: 'Żal Betoniarka' /* score: 3 */ },
])
})
it(`sorts searches by score for many keys`, () => {
expect(
createFuzzySearch(
[
{ name: 'Matt', alias: 'Matthias Obst-Mölinger' },
{ name: 'Marco Couto', alias: null },
{ name: 'Tomasz Kapelak', alias: 'Tom' },
{ name: 'tommy' },
{ name: 'Jacob Tom Belinger', alias: 'Jake' },
],
{
getText: (item) => [item.name, item.alias],
},
)('tom'),
).toMatchObject([
{ item: { name: 'Tomasz Kapelak', alias: 'Tom' } /* score: 0.1 */ },
{ item: { name: 'tommy' } /* score: 0.5 */ },
{ item: { name: 'Jacob Tom Belinger', alias: 'Jake' } /* score: 1 */ },
{ item: { name: 'Matt', alias: 'Matthias Obst-Mölinger' } /* score: 3 */ },
])
})
it(`returns empty array for empty query`, () => {
expect(createFuzzySearch(['a', 'b', 'c', 'd'])('')).toEqual([])
})
const matchesNew = (text, query, ...expectedIndices) => {
const result = experimentalSmartFuzzyMatch(normalizeText(text), normalizeText(query))
expect(result).not.toBe(null)
if (expectedIndices.length) {
const [, indices] = result
expect(indices).toEqual(expectedIndices)
}
}
it(`experimentalSmartFuzzyMatch`, () => {
matchesNew('Wypad do Żabki', 'wdż', [0, 0], [6, 6], [9, 9])
matchesNew('Wypad do Żabki', 'wdz', [0, 0], [6, 6], [9, 9])
matchesNew('Wypad do Żabki', 'wypażab', [0, 3], [9, 11])
matchesNew('Wypad do Żabki', 'wypa żab', [0, 3], [8, 11])
matchesNew(
'Marco Couto',
'mc',
[0, 0],
[3, 3], // NOTE: Ideally would be (6,6), but remaining query is len 1, so [3] matches
)
matchesNew('Marco Couto', 'm c', [0, 0], [5, 6])
matchesNew('Tomasz Kapelak', 'tokp', [0, 1], [7, 7], [9, 9])
matchesNew('[Marketing ] Referral Program', 'mrefp', [1, 1], [13, 15], [22, 22])
matchesNew('[Marketing ] Referral Program', 'm refp', [1, 1], [12, 15], [22, 22])
matchesNew('Nozbe.com web site', 'website', [10, 12], [14, 17])
matchesNew('Książka 10 steps EN', '10en')
// FIXME:
// matchesNew('Won’t fix', 'wontfix')
// matchesNew('[HR] JavaScript Developer', 'jsdev')
matchesNew('MacKay', 'mckay')
matchesNew('Русский Язык', 'усс')
matchesNew('Русский Язык', 'руя')
matchesNew('汉语的，又称汉文、華文、唐文', '汉语唐')
matchesNew('日本語', '日語')
matchesNew('한국어', '한어')
// new cases
matchesNew('GH - Growth Hacking', 'growha', [0, 0], [6, 8], [12, 13])
})
})