Repository navigation
Handle unexpected retrieved data with a null-only results schema - #47
Open
mkh-welsch wants to merge 1 commit into
Open
mkh-welsch wants to merge 1 commit into
mkh-welsch wants to merge 1 commit into
Conversation
mkh-welsch
marked this pull request as ready for review
October 3, 2026 07:35
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A retrieval task with
results_schema: {"type": "null"}crashes during evaluation when the actual response contains data: the evaluator passes it to the array normalizer, which rejects the null-only schema. Expected null already bypasses normalization correctly.Preserve nonempty actual data for the existing
retrieved_data_unexpectedassertion so it produces a normal failed evaluation. Keep strict rejection of nonempty expected data and existing empty/missing-data behavior. No response schema, dataset, prompt, or status policy changes are included.Validation on exact commit567fbcb4c039b22ff3e024786b679f0b2f669476:20synthetic parametrized tests passed using Python3.12.13/pydantic2.12.0 and isolated uv.lock-pinned test dependencies. Cases cover valid null, empty/missing data, unexpected values including zero and false, wrong status, invalid expected data, and ordinary array normalization. Existing evaluator environment unchanged.
git diff --checkpassed. No historical benchmark response or evaluation changed or rescored.