[score boosting] fix proxy segment leakage #6170

coszio · 2025-03-14T15:51:00Z

The signature for rescore_with_formula included a deleted bitslice only because the proxy segment would introduce it to handle the write and wrapped results.

In this PR, we remove the abstraction leak, and handle the logic to prefer write results only within the proxy segment.

coszio · 2025-03-14T15:53:08Z

lib/collection/src/collection_manager/holders/proxy_segment.rs


-        // Just join both results, they will be deduplicated and top-k'd later
-        write_results.append(&mut wrapped_results);


Here we also fix the fact that we were returning the wrong Vec. We were returning the empty one 💀

coderabbitai · 2025-03-14T15:54:20Z

📝 Walkthrough

Walkthrough

This pull request removes the _wrapped_deleted parameter from several implementations of the rescore_with_formula method across the codebase. The changes affect multiple modules, including the proxy segment, segment entry, and formula rescoring components. In the ProxySegment implementation, the method now always calls the wrapped segment’s rescore functionality without considering deleted points explicitly, and the merging process between wrapped and write segments is refined. Additionally, the call sites, such as in the segments searcher, have been updated to remove the now unnecessary argument, and the trait definition in the segment entry has been simplified. Overall, the changes streamline the rescoring process by eliminating the optional filtering logic based on deleted points.

Possibly related PRs

Add BitSliceExt::get_bit #6162: The changes in the main PR, which involve the removal of the _wrapped_deleted parameter from the rescore_with_formula method and the restructuring of how deleted points are handled, are related to the modifications in the retrieved PR that introduce the get_bit method for BitSlice, as both involve simplifying the handling of deletion logic in the context of scoring operations.

Suggested reviewers

timvisee

Tip

⚡🧪 Multi-step agentic review comment chat (experimental)

We're introducing multi-step agentic chat in review comments. This experimental feature enhances review discussions with the CodeRabbit agentic chat by enabling advanced interactions, including the ability to create pull requests directly from comments.
- To enable this feature, set early_access to true under in the settings.

📜 Recent review details

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 8505591 and 57afa04.

📒 Files selected for processing (1)

lib/collection/src/collection_manager/holders/proxy_segment.rs (1 hunks)

⏰ Context from checks skipped due to timeout of 90000ms (13)

GitHub Check: Basic TLS/HTTPS tests
GitHub Check: test-snapshot-operations-s3-minio
GitHub Check: test-shard-snapshot-api-s3-minio
GitHub Check: test-low-resources
GitHub Check: test-consistency
GitHub Check: test (macos-latest)
GitHub Check: test
GitHub Check: test (windows-latest)
GitHub Check: test-consensus
GitHub Check: test (ubuntu-latest)
GitHub Check: test-consensus-compose
GitHub Check: test
GitHub Check: test

🔇 Additional comments (1)

lib/collection/src/collection_manager/holders/proxy_segment.rs (1)

453-487: Good refactoring of the rescoring logic!

The implementation now correctly handles the merging of results from the wrapped and write segments. The approach of conditionally filtering wrapped results based on the deleted points state is clean and straightforward. This change effectively removes the abstraction leak by eliminating the _wrapped_deleted parameter and instead using the proxy segment's knowledge of deleted points directly.

The fix is especially important for correctly returning the merged results (as noted in your previous comment) - now we're properly returning write_results which contains the combined results from both segments.

✨ Finishing Touches

📝 Generate Docstrings

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

Review comments: Directly reply to a review comment made by CodeRabbit. Example:
- I pushed a fix in commit <commit_id>, please review it.
- Generate unit testing code for this file.
- Open a follow-up GitHub issue for this discussion.
Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
- @coderabbitai generate unit testing code for this file.
- @coderabbitai modularize this function.
PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
- @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
- @coderabbitai read src/utils.ts and generate unit testing code.
- @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.
- @coderabbitai help me debug CodeRabbit configuration file.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (Invoked using PR comments)

@coderabbitai pause to pause the reviews on a PR.
@coderabbitai resume to resume the paused reviews.
@coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
@coderabbitai full review to do a full review from scratch and review all the files again.
@coderabbitai summary to regenerate the summary of the PR.
@coderabbitai generate docstrings to generate docstrings for this PR.
@coderabbitai resolve resolve all the CodeRabbit review comments.
@coderabbitai configuration to show the current CodeRabbit configuration for the repository.
@coderabbitai help to get help.

Other keywords and placeholders

Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
Add @coderabbitai anywhere in the PR title to generate the title automatically.

CodeRabbit Configuration File (`.coderabbit.yaml`)

You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
Please see the configuration documentation for more information.
If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

Visit our Documentation for detailed information on how to use CodeRabbit.
Join our Discord Community to get help, request features, and share feedback.
Follow us on X/Twitter for updates and announcements.

generall · 2025-03-17T09:55:01Z

lib/collection/src/collection_manager/holders/proxy_segment.rs

+        } else {
+            // Create a hashmap of write results by point ID for faster lookup
+            // We expect write results to be shorter than wrapped results
+            let mut write_results_map: AHashMap<_, _> = write_results


I would prefer to avoid allocating HashMaps when possible. This is potentially an expensive operation.

Duh, this is much better, thanks

* remove deleted bitslice from rescore_with_formula signature * dont allocate unnecessary ahash --------- Co-authored-by: generall <andrey@vasnetsov.com>

remove deleted bitslice from rescore_with_formula signature

8505591

coszio requested review from timvisee and generall March 14, 2025 15:51

coszio added this to the score boosting milestone Mar 14, 2025

coszio commented Mar 14, 2025

View reviewed changes

github-actions bot mentioned this pull request Mar 14, 2025

Flaky test multivector_filtrable_hnsw_test::test_multi_filterable_hnsw::case_5_recommend_eq #5902

Closed

generall reviewed Mar 17, 2025

View reviewed changes

dont allocate unnecessary ahash

57afa04

generall approved these changes Mar 17, 2025

View reviewed changes

coszio merged commit a2dec6f into dev Mar 17, 2025
17 checks passed

coszio deleted the unleak-deleted-from-proxy-segment branch March 17, 2025 12:27

coderabbitai bot mentioned this pull request Mar 19, 2025

Measure Payload Index IO Writes #6137

Merged

timvisee pushed a commit that referenced this pull request Mar 21, 2025

[score boosting] fix proxy segment leakage (#6170)

d289a92

* remove deleted bitslice from rescore_with_formula signature * dont allocate unnecessary ahash --------- Co-authored-by: generall <andrey@vasnetsov.com>

coderabbitai bot mentioned this pull request Mar 31, 2025

[query] Optimize fetching payloads/vectors at local shard #6279

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[score boosting] fix proxy segment leakage #6170

[score boosting] fix proxy segment leakage #6170

Uh oh!

coszio commented Mar 14, 2025

Uh oh!

coszio Mar 14, 2025

Uh oh!

coderabbitai bot commented Mar 14, 2025 •

edited

Loading

Walkthrough

Possibly related PRs

Suggested reviewers

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

CodeRabbit Configuration File (`.coderabbit.yaml`)

Documentation and Community

Uh oh!

generall Mar 17, 2025

Uh oh!

coszio Mar 17, 2025

Uh oh!

Uh oh!

Uh oh!


		// Just join both results, they will be deduplicated and top-k'd later
		write_results.append(&mut wrapped_results);

[score boosting] fix proxy segment leakage #6170

[score boosting] fix proxy segment leakage #6170

Uh oh!

Conversation

coszio commented Mar 14, 2025

Uh oh!

coszio Mar 14, 2025

Choose a reason for hiding this comment

Uh oh!

coderabbitai bot commented Mar 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Possibly related PRs

Suggested reviewers

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

CodeRabbit Configuration File (.coderabbit.yaml)

Documentation and Community

Uh oh!

generall Mar 17, 2025

Choose a reason for hiding this comment

Uh oh!

coszio Mar 17, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

coderabbitai bot commented Mar 14, 2025 •

edited

Loading

CodeRabbit Configuration File (`.coderabbit.yaml`)