fix(search): place searches find a name typed without its accents (#3075)

"Obidos" did not find "Óbidos", nor "Sao Bras" "São Brás". The search of the
places list, a collection, the saved-place picker, the import from a trip
and the location picker now compare without case and Latin accents, with
letters such as ß, ø or ł folded to their plain spelling. Marks outside Latin
script stay, since there they change the word.

client/src/utils/placeSearch.test.ts

@@ -1,6 +1,6 @@
 // FE-UTIL-PLACESEARCH-001 to FE-UTIL-PLACESEARCH-004
 import { describe, it, expect } from 'vitest'
-import { placeMatchesSearch } from './placeSearch'
+import { foldForSearch, placeMatchesSearch, textMatchesSearch } from './placeSearch'
 
 const place = {
   name: 'Café Central',
@@ -32,3 +32,25 @@ describe('placeMatchesSearch', () => {
     expect(placeMatchesSearch(place, 'pizza')).toBe(false)
   })
 })
+
+describe('search without accents (#3075)', () => {
+  it('FE-UTIL-PLACESEARCH-005: a query without accents finds the accented name, and the other way round', () => {
+    for (const [query, name] of [['Obidos', 'Óbidos'], ['Evora', 'Évora'], ['Sao Bras', 'São Brás'], ['Loule', 'Loulé']]) {
+      expect(placeMatchesSearch({ name, address: null, description: null, notes: null }, query)).toBe(true)
+      expect(placeMatchesSearch({ name: query, address: null, description: null, notes: null }, name)).toBe(true)
+    }
+  })
+
+  it('FE-UTIL-PLACESEARCH-006: letters with the mark built in fold to their plain spelling', () => {
+    expect(foldForSearch('Straße')).toBe('strasse')
+    expect(foldForSearch('Łódź')).toBe('lodz')
+    expect(foldForSearch('Ærø')).toBe('aero')
+    expect(textMatchesSearch(['Kraków'], 'krakow')).toBe(true)
+  })
+
+  it('FE-UTIL-PLACESEARCH-007: marks that change a word outside Latin script stay, so が is not か', () => {
+    expect(foldForSearch('が')).toBe('が')
+    expect(textMatchesSearch(['かき'], 'がき')).toBe(false)
+    expect(foldForSearch('한국')).toBe('한국')
+  })
+})

client/src/utils/placeSearch.ts

@@ -10,9 +10,33 @@ type Searchable = Pick<Place, 'name' | 'address' | 'description' | 'notes'>
  * can never disagree about what a search finds.
  */
 export function placeMatchesSearch(place: Searchable, query: string): boolean {
-  const q = query.trim().toLowerCase()
-  if (!q) return true
-  return [place.name, place.address, place.description, place.notes].some(
-    field => !!field && field.toLowerCase().includes(q),
-  )
+  return textMatchesSearch([place.name, place.address, place.description, place.notes], query)
+}
+
+/** The combining accents of Latin script (U+0300 to U+036F), the marks NFD splits off a letter. */
+const LATIN_MARKS = new RegExp(`[${String.fromCharCode(0x300)}-${String.fromCharCode(0x36f)}]`, 'g')
+
+/** Letters that carry their mark in the letter itself, so taking the marks off leaves them as they are. */
+const FOLDED_LETTERS: Record<string, string> = { ß: 'ss', ø: 'o', æ: 'ae', œ: 'oe', ł: 'l', đ: 'd', ð: 'd', þ: 'th', ı: 'i' }
+
+/**
+ * Text as a search compares it: lower case and without accents, so "Obidos" finds
+ * "Óbidos", "Sao Bras" finds "São Brás" and the other way round (#3075). Only
+ * Latin marks go: a kana's voicing mark or a Hangul syllable stays whole, since
+ * there it changes the word.
+ */
+export function foldForSearch(text: string): string {
+  return text
+    .toLocaleLowerCase()
+    .normalize('NFD')
+    .replace(LATIN_MARKS, '')
+    .replace(/[ßøæœłđðþı]/g, ch => FOLDED_LETTERS[ch] ?? ch)
+    .normalize('NFC')
+}
+
+/** Whether any of the fields holds the query, accents and case aside. An empty query matches everything. */
+export function textMatchesSearch(fields: readonly (string | null | undefined)[], query: string): boolean {
+  const q = foldForSearch(query.trim())
+  if (!q) return true
+  return fields.some(field => !!field && foldForSearch(field).includes(q))
 }