fix(agents-api): Add limit to kNN search. Update search_docs_by_embedding.py

[creatorrr]

Important

Adds a limit to kNN search in search_docs_by_embedding to enhance diversity based on mmr_strength.

• Behavior:
  • Adds a limit to kNN search in search_docs_by_embedding to retrieve k*(3 if mmr_strength else 1) candidates for diversity.
  • Adjusts the limit in both the search_query and normal_interim_query sections.
• Misc:
  • Updates comments to reflect the new limit logic.

^{This description was created by for edc471d. It will automatically update as commits are pushed.}

[sweep-ai]

Hey @creatorrr, here are examples of how you can ask me to improve this pull request:

@Sweep Fix the CI errors.

@Sweep Add unit tests for `search_docs_by_embedding` that verify:
1. The kNN search limit is correctly set to `k*3` when `mmr_strength` is non-zero
2. The kNN search limit is correctly set to `k` when `mmr_strength` is zero
3. The diversity of results improves when using the increased limit with `mmr_strength`

📖 For more information on how to use Sweep, please read our documentation.

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to edc471d in 12 seconds

More details

• Looked at 22 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. agents-api/agents_api/models/docs/search_docs_by_embedding.py:213

• Draft comment:
  The docstring is outdated. It should reflect the new logic for the limit parameter, which now depends on mmr_strength.
• Reason this comment was not posted:
  Confidence changes required: 50%
  The PR introduces a change to the limit in the kNN search query to account for MMR strength. This change is applied in multiple places, but the logic is consistent and correct. However, the docstring is outdated and does not reflect the new logic for the limit parameter.

Workflow ID: wflow_phC67VO9fQgUx0Qk

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.