Skip to main content

Ingredient seed/import scaffolding

This folder is the starting point for importing “raw materials” datasets (fermentables, hops, yeast, salts, acids, water profiles).

Why this exists​

Recipes and the water calculator need a canonical ingredient database. We want to seed it from sources with clear licensing (see docs/modules/verticals/brewery/RAW-MATERIALS-SEEDABLE-SOURCES.md).

Data strategy (high level)​

Canonical tables (long-term goal)​

  • fermentable
  • hop
  • yeast
  • salt
  • acid
  • water_profile

Crosswalk/provenance (required)​

For every imported record, preserve provenance:

  • source_name, source_url, source_license, retrieved_at
  • source_key (original identifier/name)
  • raw_payload (JSON blob) in a staging table
  • confidence_score and conflict_notes

Crosswalk idea:

  • ingredient_source_map(ingredient_id, source_name, source_key, confidence, notes)
  1. BeerProto dataset (MIT) as the base seed.
  2. Optionally enrich with other sources only after licensing is clear.
  3. Salts/acids should be curated and maintained by us (small set, high correctness).

Next implementation steps​

  1. Add Prisma models for canonical ingredient tables + staging/provenance tables.
  2. Implement the CLI entrypoint: npm run seed:import (see src/cli/seed-import.ts).
  3. Add source-specific parsers under src/seed/sources/*.