HTTP API
EvalCaseDiff
One case’s outcome on both sides of an EvalRunDiff.
Shape: object
| Property | Required | Type | Description | Rules |
|---|---|---|---|---|
caseId |
yes | string (uuid) |
The identifier of the case being compared. | — |
kind |
yes | EvalCaseDiffKind |
Which bucket this case falls into. | — |
baselinePassed |
no | boolean |
Whether the case passed on the baseline side. null when the case is EvalCaseDiffKind.Added. |
— |
candidatePassed |
no | boolean |
Whether the case passed on the candidate side. null when the case is EvalCaseDiffKind.Removed. |
— |
baselineRunId |
no | string (uuid) |
The identifier of the agent run the baseline result came from, so a regression can be traced to the exact conversation. | — |
candidateRunId |
no | string (uuid) |
The identifier of the agent run the candidate result came from. | — |
baselineFailureReason |
no | string |
The baseline side’s failure reason, when it failed. | — |
candidateFailureReason |
no | string |
The candidate side’s failure reason, when it failed. | — |