fix(eval): fix search_tool correctness always scoring 0%#1675
Open
Tomkess wants to merge 2 commits into
Open
Conversation
_args_match checked emit_widget (renamed to user_requested_search in the tool schema) and limit (optional, server-side default). Both mismatched on every real model call, so tool_correctness was always False regardless of whether the model used the right keywords and object types. Fix: evaluate only keywords (case-insensitive) and object_types — the two fields that actually determine whether the search was semantically correct.
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthrough
ChangesSearch Tool Argument Matching
Estimated code review effort: 1 (Trivial) | ~5 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches📝 Generate docstrings
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@packages/gooddata-eval/src/gooddata_eval/core/evaluators/search_tool.py`:
- Around line 12-16: The _args_match comparison is not defensive enough and can
crash on malformed tool arguments from parsed_arguments(). Update _args_match to
validate and normalize actual_args["keywords"] and actual_args["object_types"]
before lowercasing or sorting, treating any non-string or unexpected entry as a
mismatch instead of raising. Keep the existing matching behavior for valid
inputs, but ensure bad JSON like mixed types or numeric keywords returns False
rather than aborting evaluation.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 1e3e175f-7f97-436d-9802-bd0a292bd4bb
📒 Files selected for processing (1)
packages/gooddata-eval/src/gooddata_eval/core/evaluators/search_tool.py
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #1675 +/- ##
=======================================
Coverage 79.21% 79.21%
=======================================
Files 232 232
Lines 15809 15809
=======================================
Hits 12523 12523
Misses 3286 3286 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
parsed_arguments() returns raw model-emitted JSON, so a bad tool call
like {"keywords":[1]} or mixed-type object_types would raise on
.lower()/sorted() and abort the whole eval run instead of scoring
tool_correctness=False. Add _normalize_str_list to drop non-string
entries defensively; valid comparisons (incl. case-insensitive keyword
match) are unchanged.
Addresses CodeRabbit review on PR #1675.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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.
Summary
The
tool_correctnessmetric insearch_toolevaluations was alwaysFalse(0%), making the dashboard metric useless.Root cause:
_args_matchchecked two fields that never matched real model calls:emit_widget— this parameter was renamed touser_requested_searchin the tool schema; the model never sendsemit_widget, soactual_args.get("emit_widget")was alwaysNonevs the expectedfalselimit— declared as optional (int | None) in the schema; the model omits it and relies on the server default, while fixtures hardcode10Fix: Only check
keywords(case-insensitive) andobject_types— the two fields that determine whether the search was semantically correct.limitand the widget display flag are not part of search correctness.Test Plan
uv run pytest tests/ -x -qinpackages/gooddata-evaltool_correctnessnow reflects actual model behaviour (models that call with the right keywords and object types should score True)Risk
Low — single function change in the evaluator, no effect on eval runner or production code.
Summary by CodeRabbit