[Track] Additional output formats (ORC, Lance)

[mprammer]

Add ORC and Lance as additional columnar outputs alongside the existing Parquet (canonical) and Vortex (optional) artifacts.

Per format

• New convert stage (or generalised convert that switches on format).
• Extend sources.json: per-format convert.<fmt> + convert.<fmt>_skip_reason, mirroring the Vortex pair.
• Update validate_manifest invariants.
• Outputs at outputs/v{n}/<slug>/<fmt>/<slug>.<ext>.
• Regen docs/datasets.md + docs/snapshot.json.

Formats in scope

• ORC — Apache ORC. PyArrow has read-only support; need an external writer (e.g. pyorc, or PyArrow + orc-tools).
• Lance — Lance v2 via pylance.