Standard metadata: Frictionless datapackage.json + dataset-level citations

[dlebauer]

Goal

• Provide machine-readable metadata for the datasets using a standard (Frictionless Data), not bespoke prose.

Tasks

• Create inst/metadata/datapackage.json describing:
  • each dataset (resource) shipped in the package (traitsview + support tables)
  • schema fields (name + type) derived from the exported data objects (do not hand-write where possible)
  • licensing for data (ODC-By / CC-BY as appropriate, separate from code license)
  • sources and attribution fields
• Add minimal code in data-raw/make-data.R (or helper script) to:
  • generate/refresh datapackage.json from the current exported objects
  • validate JSON (optional: via frictionless tooling if lightweight)
• Add a short README snippet pointing users to inst/metadata/datapackage.json.

Acceptance criteria

• inst/metadata/datapackage.json exists and enumerates all shipped datasets.
• Schema entries match the current shipped objects (no obvious drift).
• A rebuild step exists that regenerates/updates the metadata alongside data rebuild.

Non-goals

• No separate schema spec markdown files.