fix: sweep cleanup — deps, docs, tool count, brain_expand, AIDEV-NOTEs

[EtanHey]

Summary

• pyproject.toml: Remove unused tree-sitter from core deps; add faiss-cpu to [brain]; add new [style] extra with chromadb; [kg] extra now documented in README
• README.md: Tool count 7→8 everywhere (badge, tagline, mermaid, comparison table); test badge 698→715; add brain_expand to MCP Tools table; fix clustering name (HDBSCAN→Leiden); add [kg]/[style] to Optional Extras; add Groq backend to Configuration table
• CLAUDE.md: Remove non-existent --http flag from brainlayer serve; expand MCP tools listing to all 8 tools
• docs/mcp-tools.md: Update tool count 7→8; add full brain_expand documentation
• CHANGELOG.md: Add [Unreleased] section documenting all recent changes
• brain_graph.py: Add AIDEV-NOTE explaining the correlated subquery → literal optimization (PR Fix brain_search excluding brain_store chunks #62 fix)
• kg/init.py: Add AIDEV-NOTE marking as canonical KG spec shared with 6PM Convex backend
• greptile.json: Commit previously untracked Greptile config file

Test plan

• pyproject.toml parses cleanly: pip install -e ".[brain,style,kg,ast,dev]" (or uv sync)
• README renders correctly — check badge URLs, mermaid diagram, tables
• docs/mcp-tools.md has brain_expand section with correct parameter table
• CHANGELOG.md has [Unreleased] section at top
• No functionality changed — only docs, deps, and uncommitted files

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Added a chunk expansion tool.
  • Introduced Groq enrichment backend with new environment variables.
  • Added optional extras for ChromaDB vector store and FAISS-based ANN.

• Bug Fixes

  • Improved performance for large datasets by eliminating an N+1 query pattern.
  • Corrected CLI documentation by removing an invalid flag.

• Documentation

  • Added an “Unreleased” changelog section.
  • Updated tool count, test count, and clustering description to Leiden + UMAP.
  • Documented new environment variables and optional installation extras.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 800bc01 and 4895ab7.

📒 Files selected for processing (8)

• CHANGELOG.md
• CLAUDE.md
• README.md
• docs/mcp-tools.md
• greptile.json
• pyproject.toml
• src/brainlayer/kg/__init__.py
• src/brainlayer/pipeline/brain_graph.py

🔥 Files not summarized due to errors (4)

• docs/mcp-tools.md: Error: Server error: no LLM provider could handle the message
• CLAUDE.md: Error: Server error: no LLM provider could handle the message
• greptile.json: Error: Server error: no LLM provider could handle the message
• src/brainlayer/kg/init.py: Error: Server error: no LLM provider could handle the message

---

📝 Walkthrough

Walkthrough

This PR documents and implements feature additions and performance optimizations: introducing a new brain_expand MCP tool, Groq enrichment backend support, ChromaDB and FAISS vector store options, and an N+1 query performance fix in the brain graph module.

Changes

Cohort / File(s)	Summary
Documentation Updates CHANGELOG.md, README.md	Updated tool counts (7→8), test counts (698→715), clustering algorithm (HDBSCAN→Leiden), added Groq environment variables (GROQ_API_KEY, BRAINLAYER_GROQ_URL, BRAINLAYER_GROQ_MODEL), and documented new optional extras ([kg], [style]).
Dependency Configuration pyproject.toml	Removed tree-sitter>=0.21.0 from core dependencies, added faiss-cpu>=1.7.0 to [brain] extras, and created new [style] extras group with chromadb>=0.4.0.
Performance Optimization src/brainlayer/pipeline/brain_graph.py	Replaced correlated subquery in load_sessions() with hardcoded literal 'claude_code' to eliminate N+1 subqueries on large databases; includes explanatory comment on future multi-source CTE approach.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• fix: README tagline — 7 tools + KG #42: Modifies README tool-count documentation (PR sets to "7 tools + KG", this PR updates to 8 total)
• docs: Phase 9b — README rewrite, CONTRIBUTING update #19: Updates README.md and pyproject.toml with tool counts, optional extras, and project metadata changes

Poem

🐰 Eight tools now bound in our expanding brain,
Groq's whispers quicken the data chain,
FAISS hops swift where N+1 once crawled,
ChromaDB styles the vectors called,
Performance springs from optimization's art! 🌟

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/sweep-cleanup

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[greptile-apps]

Your free trial has ended. If you'd like to continue receiving code reviews, you can add a payment method here.