fix: don't modify candidates #773
Merged
Conversation
This file contains 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
Since a recent change, we use finalCands in our BM25 scoring, however finalCands is modified in fillChunkMatches which led to suprising scores. Test plan: updated unit test
jtibshirani
reviewed
Apr 30, 2024
| // keyword-score:1.63 (sum-tf: 6.00, length-ratio: 2.00) | ||
| wantScore: 1.63, | ||
| // keyword-score:1.69 (sum-tf: 7.00, length-ratio: 2.00) | ||
| wantScore: 1.69, |
There was a problem hiding this comment.
Wondering why I didn't catch this 🤦
There was a problem hiding this comment.
Should we add more info to the debug message? We could print the term frequency map instead of sum-tf. This wouldn't have helped for this test case, because we are just searching for one term, but for a more complex query it might be easier to debug.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
While working on ranking, I noticed that sum-tf is wrong if we have filename and content matches.
We use
finalCandsin our BM25 scoring, however,finalCandsis modified infillChunkMatchesandfillMatcheswhich can lead to surprising scores.Test plan:
updated unit test