Seeding the Database

Ingredients are a preset catalog owned by the food-api service (the cookest_food database). It is the single source of truth: the pantry, recipes, and the AI can only reference catalog ingredients — they never create free-text entries. The app-api database keeps a same-id mirror of any ingredient it touches.

This page covers the ways to populate that catalog.

1. Base seed (automatic)

On startup, if the ingredients table is empty, food-api loads a small curated set of common ingredients (~140) bundled in the binary at crates/food-api/seed/ingredients.csv. This gives a fresh install a usable catalog out of the box. It only runs when the table is empty, so it never overwrites an existing catalog.

2. Import an ingredient list from a folder

For larger or custom lists, drop CSV files into a folder the food-api process can read and import them from the admin dashboard.

With Docker Compose the host folder cookest-backend/data/imports/ is mounted into the container at /data/imports:

# docker-compose.yml (food-api service)
volumes:
  - ./data/imports:/data/imports

CSV format

A header row is required. Columns (extra columns are ignored; numeric cells are optional):

name,category,calories,protein_g,carbs_g,fat_g
Greek yogurt,dairy,59,10,3.6,0.4
Quinoa,grain,120,4.4,21,1.9
Kale,vegetable,49,4.3,9,0.9

Macros are per 100 g. Rows are upserted by name, so re-running an import updates existing ingredients instead of duplicating them. A ready-to-edit template lives at cookest-backend/data/imports/example-ingredients.csv.

From the admin dashboard

1. Put your .csv files in cookest-backend/data/imports/ on the host.
2. Open the admin dashboard → Ingredients → Import.
3. Keep the folder as /data/imports (or point it at another server folder), click Scan.
4. Click Import next to the file. The result (rows imported) is shown and the list refreshes.

From the API

The dashboard calls these admin endpoints (JWT + admin required):

GET  /admin/ingredients/import/scan?folder=/data/imports
POST /admin/ingredients/import/execute      { "folder": "/data/imports", "filename": "my-list.csv" }

3. Add ingredients individually

Admin dashboard → Ingredients → Add Ingredient lets you create or edit a single ingredient (name, category, optional per-100g macros). Deleting an ingredient that is used by a recipe is blocked (returns 409).

4. Bulk import from Open Food Facts

food-api also understands Open Food Facts exports for large catalogs with barcodes and nutrition. Place a tab-separated OFF file whose name contains products or openfoodfacts (e.g. the repo's en.openfoodfacts.org.products.csv.gz) in the import folder and import it from the admin Database page. Up to 100,000 rows are imported per run.

5. Resetting the app-db mirror (development)

The app database (cookest_app) mirrors catalog ingredients using the same ids as the master. If it still contains legacy free-text rows from before the preset-catalog change, those can collide with mirror inserts on the unique name constraint. To clear them in development:

psql "$APP_DATABASE_URL" -f cookest-backend/scripts/reset_app_ingredient_mirror.sql
# e.g. APP_DATABASE_URL=postgresql://postgres:postgres@localhost:5433/cookest_app

This truncates the app-db ingredients table and its dependents (pantry, recipe links, shopping lists); they repopulate automatically from the master catalog as users reference ingredients. The master catalog in cookest_food is not affected.