Eval runs

Preview a failed-cases rerun

What POST /projects/{projectId}/eval-runs/{runId}/rerun with scope: "failed_cases" would execute, before anything spends. The platform selects every case with a trial that is not completed and passed; cancelled and skipped trials do not count. reasons and excluded count trials; selectedCaseCount counts cases. rerunnable is false while the run is still in flight (including grading) and when nothing qualifies.

get/projects/{projectId}/eval-runs/{runId}/rerun-preview

Path parameters

projectIdstring required

ID of the hosted project that contains the server.

runIdstring required

Eval run ID, as returned by POST /eval-runs.

Response

The rerun selection.

runIdstring required
suiteIdstring required
scope'failed_cases' required
sourceStatusstring required
sourceTerminalboolean required

false while the source run is still executing or grading.

totalCaseCountinteger required
selectedCaseCountinteger required
selectedCaseIdsstring[] required
reasonsobject required

Qualifying trials by the most specific reason: failed, evaluator_error, timed_out, execution_failed, setup_failed, pending.

excludedobject required

Trials that did not qualify: passed, cancelled, skipped.

rerunnableboolean required

The source run finished and at least one case qualifies.

Changes

Changed in 1 of the 133 revisions of this API.1