Extract shared knowledge base search and evaluation logic to submodule

[brikin01]

No description provided.

[Copilot]

Pull request overview

Extracts the knowledge-base search utilities (previously in mcp-local/utils/search_utils.py) and retrieval evaluation helpers into a new standalone arm_kb_search Python package, and updates mcp-local/server.py, the Dockerfile, and the evaluation script to consume it.

Changes:

• New arm_kb_search package with search, loaders, evaluation, config modules and a pyproject.toml.
• mcp-local/server.py and Dockerfile updated to import from arm_kb_search; DISTANCE_THRESHOLD/K_RESULTS removed from mcp-local/utils/config.py.
• embedding-generation/evaluate_retrieval.py replaced by evaluate_local_retrieval.py, which delegates to arm_kb_search.evaluation; shell script and README updated accordingly.

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
pyproject.toml	Declares the new arm-kb-search package and its dependencies.
arm_kb_search/init.py	Public API re-exports for the new submodule.
arm_kb_search/config.py	Hosts moved DISTANCE_THRESHOLD/K_RESULTS constants.
arm_kb_search/loaders.py	Index/metadata loader helpers moved out of search.py.
arm_kb_search/search.py	Removes loader helpers and json/os imports.
arm_kb_search/evaluation.py	New shared retrieval evaluation helpers with dataclasses.
mcp-local/server.py	Switches imports to arm_kb_search.
mcp-local/utils/config.py	Removes constants that moved to arm_kb_search.config.
mcp-local/Dockerfile	Copies the arm_kb_search package into the image.
embedding-generation/evaluate_local_retrieval.py	New evaluation entrypoint built on arm_kb_search.
embedding-generation/evaluate_retrieval.py	Removed (replaced by new script).
embedding-generation/run-question-eval.sh	Invokes the renamed evaluation script.
embedding-generation/README.md	Updates the script name mentioned in docs.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[brikin01]

@JoeStech Addressed the comments and also updated the evaluation function to output both the summary and actually include the individual cases so that we can do more detailed breakdowns later.