[ENH] Materialize before bf knn

[HammadB]

Description of changes

This makes the bf knn operator use materialized logs. This makes the operator valid for cases like an invalid add etc, since those are filtered out of the log.

• We update the output type of bf knn to not depend on logs, this is suboptimal and highlights how the materialized log abstraction is lossy. We should rethink this long term, and perhaps present less of a log interface and more of a query-able layer. The lifetimes also make this clunky, as we can't easily ship materialized log records around between operators. For now, it follows the pattern of creating the materializer and independently materializing logs for its operation. This is both clunky and inefficient, but is functionally correct.
• For now, the bf operator returns the columns needed by downstream, which has minimal impact on the orchestration and doesn't add any new data cloning, since before we cloned off the log for returning.

Tasks This is a WIP (BUT Is now ready for review minus tests, will add those shortly)

• [x] Materialize before we send to log, ideally in separate operator
• [x] Check inputs
• [x] Refactor consumers
• [x] Rewrite tests to use Mat Log

[github-actions[bot]]

Reviewer Checklist

Please leverage this checklist to ensure your code review is thorough before approving

Testing, Bugs, Errors, Logs, Documentation

• [ ] Can you think of any use case in which the code does not behave as intended? Have they been tested?
• [ ] Can you think of any inputs or external events that could break the code? Is user input validated and safe? Have they been tested?
• [ ] If appropriate, are there adequate property based tests?
• [ ] If appropriate, are there adequate unit tests?
• [ ] Should any logging, debugging, tracing information be added or removed?
• [ ] Are error messages user-friendly?
• [ ] Have all documentation changes needed been made?
• [ ] Have all non-obvious changes been commented?

  System Compatibility

• [ ] Are there any potential impacts on other parts of the system or backward compatibility?
• [ ] Does this change intersect with any items on our roadmap, and if so, is there a plan for fitting them together?

  Quality

• [ ] Is this code of a unexpectedly high quality (Readability, Modularity, Intuitiveness)

[HammadB]

This PR is not ready for review. I wouldn’t waste your time - it’s a draft. I’ll mark it when it’s ready.

Hammad Bashir EECS | Cal

On Sat, Jun 15, 2024 at 7:16 PM Sanket Kedia @.***> wrote:

@.**** commented on this pull request.

In rust/worker/src/execution/operators/brute_force_knn.rs https://github.com/chroma-core/chroma/pull/2347#discussion_r1641595435:

         if !input.allowed_ids.is_empty()

&& !input .allowed_ids_brute_force

• .contains(&log_record.record.id)
• .contains(&log_record.merged_user_id())

There is a merged_user_id_ref() that does not clone() that can be used here probably?

— Reply to this email directly, view it on GitHub https://github.com/chroma-core/chroma/pull/2347#pullrequestreview-2121070927, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABKW32LACTRZ5NVKZVGMXVTZHTYPVAVCNFSM6AAAAABJLQ2PM2VHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDCMRRGA3TAOJSG4 . You are receiving this because you authored the thread.Message ID: @.***>

[HammadB]

This is now ready for prelim review. I have my own cleanup and test cleanup to do but an eager eye is welcome to look.