Arnaud
b243cc3b8b
Ingredient-alias matching (pantry, can-cook, auto-deduct, shopping-list generation) now covers common bilingual EN/FR ingredients across all 8 grocery categories, up from ~10 staples. Also: matching now ignores accents as well as case (NFD-normalize + strip combining marks) in both the alias resolver (lib/ingredient-match.ts) and the older name-fallback matcher (pantry-shopping-match.ts) — "café"/"Café"/"cafe" and "Épinard"/"epinard" all recognize as the same ingredient without needing every accent variant manually listed as an alias. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
54 lines
2.4 KiB
TypeScript
54 lines
2.4 KiB
TypeScript
import { db, ingredients } from "@epicure/db";
|
|
|
|
export type IngredientAliasIndex = Map<string, string>;
|
|
|
|
// Case- and accent-insensitive so "café"/"cafe" and "Épinard"/"epinard"
|
|
// compare equal — NFD splits an accented letter into base letter +
|
|
// combining mark, then the combining marks (U+0300-U+036F) are stripped.
|
|
function normalize(name: string): string {
|
|
return name.trim().toLowerCase().normalize("NFD").replace(/[̀-ͯ]/g, "");
|
|
}
|
|
|
|
/**
|
|
* Loads every canonical ingredient's name + aliases into a flat
|
|
* normalized-string -> canonical-ingredient-id map, once per request. Used
|
|
* to recognize that "sel", "sel fin", and "table salt" are all the same
|
|
* ingredient, without requiring every recipe/pantry row to carry a stored
|
|
* ingredientId (they don't — this resolves purely from the free-text name
|
|
* at comparison time).
|
|
*/
|
|
export async function loadIngredientAliasIndex(): Promise<IngredientAliasIndex> {
|
|
const rows = await db.select({ id: ingredients.id, name: ingredients.name, aliases: ingredients.aliases }).from(ingredients);
|
|
const index: IngredientAliasIndex = new Map();
|
|
for (const row of rows) {
|
|
index.set(normalize(row.name), row.id);
|
|
for (const alias of row.aliases) {
|
|
index.set(normalize(alias), row.id);
|
|
}
|
|
}
|
|
return index;
|
|
}
|
|
|
|
/** Canonical ingredient id if `rawName` matches a known name/alias exactly
|
|
* (case/accent/whitespace-insensitive); otherwise the normalized rawName
|
|
* itself, so unmatched items still compare equal to other unmatched items
|
|
* with the same text (today's behavior, unchanged for anything not seeded). */
|
|
export function resolveIngredientKey(rawName: string, index: IngredientAliasIndex): string {
|
|
const normalized = normalize(rawName);
|
|
return index.get(normalized) ?? normalized;
|
|
}
|
|
|
|
/** Single-name lookup (pantry add/edit). Loads the same small table as
|
|
* loadIngredientAliasIndex and compares in JS rather than in SQL — accent
|
|
* stripping via NFD has no simple SQL equivalent without the `unaccent`
|
|
* extension, which isn't guaranteed to be installed. The ingredients table
|
|
* is small (tens to low hundreds of rows), so this is cheap. Returns null
|
|
* when there's no canonical match, meaning the item stays a plain freeform
|
|
* pantry entry. */
|
|
export async function findIngredientIdByName(rawName: string): Promise<string | null> {
|
|
const normalized = normalize(rawName);
|
|
if (!normalized) return null;
|
|
const index = await loadIngredientAliasIndex();
|
|
return index.get(normalized) ?? null;
|
|
}
|