Repository navigation
fix(asap-tools): reject two planner families for one recommended query - #785
Merged
Merged
Conversation
Experiment names carry no sketch family, so a selected query with both a CMS and a KLL row in recommendations.csv (sketch-bench emits both for a query id with keys and values forms, e.g. rt_p99_by_msname) wrote one config over the other. generate now raises for such a query. Co-Authored-By: Claude Opus 5.5 <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
recommended_sketch_configs.py generatenames experiments<dataset>_<query>_<range>_recommended, with no sketch family. If a selected query had both a CMS and a KLL row inrecommendations.csv, the second config silently overwrote the first.(query_id, range, kind), and some query ids (e.g.rt_p99_by_msname) have keys and values forms. The current CSV has no such rows.generatenow raises for a selected query with more than one planner-family row. Collisions on queries that were not selected are ignored.load_querieskeeps only the first form of a duplicated id, so both experiments would run the same PromQL.Not addressed here:
results_summary.csvstill has the p99 runs' old value-error numbers. Regenerating it needs the run outputs from the CloudLab node.Test plan
pytest tests/test_recommended_sketch_configs.py(19 passed), with tests for the collision and for an unselected colliding query🤖 Generated with Claude Code