new data source

intake + examination + activation validate + preview + run structural checks. Examine to see what it'd do, then Approve to stage it in config (ADR-005 B1).

▸ candidate data source kinds introspected from the CONNECTORS registry

domain the business area — picks which checks run (products / orders in the demo) name this source any label you choose — it's just an identifier

where is the data? (connector) csv / jsonl = a file on disk · sql = a database query · sample = generated demo data which system is this from? a SHORT tag for the source system (e.g. pim, shop) — not a file path

file location / connection (JSON) file path on disk, e.g. {"path": "/tmp/products.csv"} (optional "delimiter": ",") file path on disk, e.g. {"path": "/tmp/products.jsonl"} a database URL + SQL query, e.g. {"url": "sqlite:///db.sqlite", "query": "SELECT * FROM products"} generated demo data — n rows + random seed field mapping (JSON) identity_field = which column is the unique ID; source_to_canonical = your column names → field names

source · provenance required

kind author

citation — why this source / where it comes from