1# data/raw — provenance (never committed)23| Item | Provenance | Taken | Integrity |4|---|---|---|---|5| `foodka-snapshot-20261101/foodka-snapshot.db` | Consistent SQLite `.backup` of the production Food-Ka database `M2U64:~/apps/food-ka/data/foodka.db` (WAL mode), opened read-only (`?mode=ro`). Production untouched. | 2026-11-01T07:57:11Z (`snapshot-utc.txt`) | SHA-256 in `foodka-snapshot.sha256`, verified after transfer. Food-Ka code at commit `2c76523` (`foodka-git-head.txt`). |6| `foodka-snapshot-20261101/sources.json` | Registry of the 61 source feeds (57 active, 4 documented as non-connectable). | same | — |7| `foodka-src/` | Read-only copy of the Food-Ka backend (`foodka/`, `run.py`, `tests/`), excluding `.venv`, `.git`, `frontend`, `docs`, `data`. Used only to document the data-generating process (diff engine, normalization, matching). | same | — |89Authorization: the database owner (S.-P. Boucher, Food-Ka author) is the WP37 principal investigator; access is an owner export, not scraping. `robots.txt` of www.food-ka.com disallows `/api/` to crawlers; the public API was queried once (2026-11-01) only to cross-check headline counts and was not used as a data source.10