Source Types Cookbook

This page shows how to describe different kinds of enterprise data systems using DSP.

For every system, the pattern is the same:

1. Define a Source (source_id, connection)
2. Define one or more Datasets (dataset_id, kind, physical_name)
3. Define Fields (name, type, nullable, mapping)
4. Optionally add Semantics (domain, entity_type, PII, join hints)
5. Optionally add Access metadata (SQL/search/stream/file)

DSP uses one JSON shape for all of them – the difference is mostly in:

• connection.type
• connection.driver
• dataset.kind
• dataset.access details

You can either hand-write the JSON or use the DSP SDK profiles where available.

(Full cookbook content omitted here for brevity in this comment, but in your actual project you can paste the full Source Types Cookbook as previously generated.)