[SPARK-57124][INFRA] Fix update_build_status.yml to map non-standard workflow run statuses

[dongjoon-hyun]

What changes were proposed in this pull request?

This PR fixes the update_build_status.yml so that it coerces non-standard workflow run statuses (requested, waiting, pending) to a value accepted by the PATCH /repos/{owner}/{repo}/check-runs/{check_run_id} endpoint.

The else branch now maps run.data.status to either in_progress (when it is exactly in_progress) or queued (for any other non-completed value), instead of passing the raw value through.

Why are the changes needed?

The scheduled job has been failing intermittently with HTTP 422.

• https://github.com/apache/spark/actions/workflows/update_build_status.yml?query=is%3Afailure

For example, run 26552065653 failed with:

Run 78216352873: set status (pending)
RequestError [HttpError]: Invalid request.
  No subschema in "anyOf" matched.
  pending is not a member of ["queued", "in_progress", "completed"].

GET /repos/{owner}/{repo}/actions/runs/{run_id} can return any of completed, in_progress, queued, requested, waiting, or pending, while PATCH /check-runs/{id} only accepts queued, in_progress, or completed. Forwarding the raw value caused the workflow to crash whenever a fork's run was in pending (or requested/waiting) state, and the corresponding PR check-runs were not synchronized for that cycle.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Manual review of the workflow script. The change is limited to mapping a string value before passing it to the GitHub REST API; behavior for the existing completed and in_progress paths is unchanged.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: Claude Opus 4.7

[dongjoon-hyun]

cc @HyukjinKwon, @viirya , @peter-toth

[viirya]

Small, safe, well-explained infra fix with the right blast radius. +1.

[dongjoon-hyun]

Thank you, @viirya .

[dongjoon-hyun]

Merged to master to fix the infra flakiness.