[PERF] Add include metadata to MetadataReader to avoid unncessary payload size

chroma-core / chroma

the AI-native open-source embedding database

Apache License 2.0

15.45k stars 1.3k forks source link

Description of changes

Summarize the changes made by this PR.

Improvements & Bug fixes
- This PR introduces a "include_metadata" field to the MetadataReader across the local and distributed implementations. This way the filter step can only transmit ids, and only final hydration requests all data. This reduces overall payload sizes.
- I chose this pattern to mirror what is done by the VectorReader with the "include_embeddings" flag
- The correct solution for this is push the query down to one node instead of roundtripping.
New functionality
- None

How are these changes tested? I extended the tests for the metadata segment

[x] Tests pass locally with pytest for python, yarn test for js, cargo test for rust

Are all docstrings for user-facing APIs updated if required? Do we need to make documentation changes in the docs repository?

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?

[ ] 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?

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