Add and maintain a changelog for the Zenodo public data dump

[bencap]

Context

We published v5 of the Zenodo bulk-data archive with a substantial set of changes (annotation CSVs, mapped-variant JSON, VA-Spec NDJSON, namespaced columns) but no changelog, so data consumers had no way to see what changed between versions. We want a changelog to be a standing part of every release going forward.

Done (in working tree — needs PRs)

mavedb-api:

• Add src/mavedb/scripts/resources/CHANGELOG.md (canonical, Keep-a-Changelog style) with the v5 entry.
• export_public_data.py should bundles CHANGELOG.md into the zip and its docstring should point at the changelog + runbook.
• Add CHANGELOG.md to the archive-structure diagram in resources/README.md.

deployment:

• Add docs/zenodo-release.md release runbook (write changelog first → generate → verify → upload as a new Zenodo version → mirror notes → update docs), with a checklist.

Outstanding

• Backfill the already-published v5 Zenodo record: paste the v5 CHANGELOG.md entry into the record's version notes/description (v5 shipped without them).
• Follow the runbook for v6 onward.