feat: add date-window and validator-version filters to rebuild_missing_validation_reports

[cka-y]

Summary:

Extends rebuild_missing_validation_reports (functions-python/tasks_executor) by adding optional filters so we can re-validate a historical window against the prod validator. Closes #1762.

• filter_downloaded_after / filter_downloaded_before: ISO-date window on downloaded_at.
• filter_validator_version_prefix: target datasets lacking any report matching the prefix (e.g. "8."), not just a non-exact version.
• include_all_datasets: consider all datasets of a feed, not just the latest.

All default to preserve existing behavior. README + unit tests updated.

Expected behavior:

Dry run reports the count of published feeds' datasets in the window missing an 8.* report; non-dry run triggers a validation workflow per candidate against prod and writes the reports. Response params echo the applied filters.

Testing tips:

Fields must be nested under "payload". Dry run first:

{
  "task": "rebuild_missing_validation_reports",
  "payload": {
    "dry_run": true,
    "filter_downloaded_after": "2026-03-01",
    "filter_downloaded_before": "2026-06-01",
    "filter_validator_version_prefix": "8.",
    "include_all_datasets": true,
    "validator_endpoint": "https://gtfs-validator-web-mbzoxaljzq-ue.a.run.app"
  }
}

Check params echo your filters and review total_candidates, then flip dry_run to false to trigger.

• include_all_datasets: true is what covers all datasets in the window (else latest-only).
• Count filters by the feed's operational_status (default published) with an exclusive upper bound, so it won't match a raw downloaded_at BETWEEN query.
• Idempotent: already-triggered and already-8.* datasets are skipped. Try "limit": 10 for a small batch first.
  Notes for testers:
• The count reflects each feed's latest dataset only, operational_status = 'published', and an exclusive upper bound — so it won't match a raw SELECT ... FROM gtfsdataset WHERE downloaded_at BETWEEN ... count.
• Re-running is idempotent: already-triggered datasets are skipped, and any dataset that already has an 8.* report is excluded.
• Try omitting the new fields to confirm the original behavior is unchanged, and try "limit": 10 for a small end-to-end batch first.

Please make sure these boxes are checked before submitting your pull request - thanks!

• Run the unit tests with ./scripts/api-tests.sh to make sure you didn't break anything
• Add or update any needed documentation to the repo
• Format the title like "feat: [new feature short description]". Title must follow the Conventional Commit Specification(https://www.conventionalcommits.org/en/v1.0.0/).
• Linked all relevant issues
• Include screenshot(s) showing how this pull request works and fixes the issue(s)

[davidgamez]

LGTM!