.coderabbit.yaml

language: "en-US"

reviews:
  profile: "chill"
  high_level_summary: true
  collapse_walkthrough: true
  poem: false
  review_status: true
  assess_linked_issues: true
  related_prs: true
  suggested_labels: true
  suggested_reviewers: true

  auto_title_instructions: |
    Format: "<category>: <title>". Category must be one of: feat, fix,
    docs, style, refactor, perf, test, build, ci, chore, revert.
    Title should be concise (<= 80 chars).

  path_filters:
    # Lock and config files
    - "!**/*.lock"
    - "!**/package-lock.json"
    - "!**/pnpm-lock.yaml"
    - "!**/yarn.lock"
    - "!uv.lock"

    # Build artifacts and caches
    - "!**/node_modules/**"
    - "!**/__pycache__/**"
    - "!**/.pytest_cache/**"
    - "!**/.mypy_cache/**"
    - "!**/.ruff_cache/**"
    - "!**/dist/**"
    - "!**/build/**"
    - "!docs/_build/**"
    - "!docs/build/**"

    # Generated/bundled frontend assets
    - "!massgen/frontend/web/static/assets/**"
    - "!**/*.min.js"
    - "!**/*.min.css"
    - "!**/*.map"

    # Binary/media files
    - "!**/*.png"
    - "!**/*.jpg"
    - "!**/*.jpeg"
    - "!**/*.gif"
    - "!**/*.ico"
    - "!**/*.svg"
    - "!**/*.woff"
    - "!**/*.woff2"
    - "!**/*.ttf"
    - "!**/*.eot"

    # MassGen runtime logs and workspaces (agent-generated content)
    - "!.massgen/**"
    - "!massgen_logs/**"

    # Website static images
    - "!website/images/**"

  path_instructions:
    # Core Python code
    - path: "massgen/**/*.py"
      instructions: |
        Review Python code for:
        - PEP 8 compliance and Google-style docstrings
        - Proper async/await patterns
        - Type hints where appropriate
        - No security vulnerabilities (command injection, etc.)
        - Consistency with existing code patterns

    # Backend implementations - check capabilities registry
    - path: "massgen/backend/**/*.py"
      instructions: |
        Backend code review - CHECK FOR THESE UPDATES:
        1. If adding new model support → ensure massgen/backend/capabilities.py is updated
        2. If adding new capabilities → update supported_capabilities in capabilities.py
        3. Verify inheritance from base classes is correct
        4. Check streaming buffer usage where applicable
        5. Ensure error handling follows existing patterns

    # Token manager - check pricing updates
    - path: "massgen/token_manager/**/*.py"
      instructions: |
        Token manager changes - VERIFY:
        1. New models should check LiteLLM database first before adding to PROVIDER_PRICING
        2. Pricing is in per-1K tokens format (not per-token)
        3. Provider names use correct casing: "OpenAI", "Anthropic", "Google", "xAI"
        4. Context windows and max output tokens are accurate

    # Config validator - ensure schema is updated
    - path: "massgen/config_validator.py"
      instructions: |
        Config validator changes - CHECK:
        1. New parameters added here must also be added to:
           - massgen/backend/base.py → get_base_excluded_config_params()
           - massgen/api_params_handler/_api_params_handler_base.py → get_base_excluded_config_params()
        2. Ensure validation error messages are helpful

    # Tests
    - path: "massgen/tests/**/*.py"
      instructions: |
        Verify tests use pytest best practices:
        - @pytest.mark.expensive for API-calling tests
        - @pytest.mark.docker for Docker-dependent tests
        - @pytest.mark.asyncio for async tests
        - Clear test descriptions and proper fixtures

    # YAML configs - validate structure
    - path: "massgen/configs/**/*.yaml"
      instructions: |
        Validate YAML configs follow MassGen conventions:
        1. Property placement: cwd at BACKEND-level, context_paths at ORCHESTRATOR-level
        2. Prefer cost-effective models (gpt-5-nano, gpt-5-mini, gemini-2.5-flash)
        3. All agents should have identical system_message for multi-agent setups
        4. Never reference legacy paths or massgen v1
        5. Include "What happens" comments explaining execution flow

    # Sphinx documentation
    - path: "docs/source/**/*.rst"
      instructions: |
        RST documentation review:
        1. Ensure proper RST syntax
        2. Check cross-references are valid
        3. Verify code examples are runnable and match current implementation
        4. For new features, ensure docs/source/index.rst "Recent Releases" is updated

        Documentation quality:
        - Keep explanations concise - avoid bloated prose
        - Remove redundant or duplicate information
        - Ensure file paths and parameter names match actual code
        - Flag any examples that may be outdated or inconsistent with implementation

    # User guides - check completeness and accuracy
    - path: "docs/source/user_guide/**/*.rst"
      instructions: |
        User guide documentation review:

        Completeness:
        - Clear explanations with working code examples
        - Expected output where applicable
        - Links to related configuration examples

        Consistency with implementation:
        - Parameter names must match actual code
        - File paths must be accurate
        - Behavior descriptions must match what the code does
        - Flag any discrepancies between docs and implementation

        Usability:
        - Include runnable commands users can try immediately
        - Provide architecture diagrams for complex features
        - Show expected output so users know what to expect

        Conciseness:
        - Avoid over-documentation and bloated explanations
        - One clear explanation is better than multiple redundant ones
        - Remove filler text and unnecessary verbosity

    # Case studies
    - path: "docs/source/examples/case_studies/**/*.md"
      instructions: |
        Case study review:
        1. Follow the template in case-study-template.md
        2. Include real usage examples
        3. Show configuration used
        4. Document outcomes and results

    # README updates (only required for release PRs)
    - path: "README.md"
      instructions: |
        README changes are typically only required for release PRs (branches named dev/v0.1.X, e.g., dev/v0.1.33).
        For release PRs, verify:
        1. Recent Achievements section is updated for new features
        2. Case study links are added if applicable
        3. Configuration examples are accurate
        Note: README_PYPI.md is auto-synced via pre-commit hook

    # CHANGELOG (only required for release PRs)
    - path: "CHANGELOG.md"
      instructions: |
        CHANGELOG updates are typically only required for release PRs (branches named dev/v0.1.X, e.g., dev/v0.1.33).
        For release PRs, verify:
        1. Follow Keep a Changelog format
        2. Categories: Added, Changed, Fixed, Documentations/Configurations/Resources, Technical Details
        3. Include file paths for all changes
        4. Reference issue/PR numbers

    # Skills
    - path: "massgen/skills/**/*.md"
      instructions: |
        Skill documentation review:
        - Verify SKILL.md frontmatter is complete (name, description)
        - Check workflow steps are clear and actionable
        - Ensure reference files are accurate

  auto_review:
    enabled: true
    auto_incremental_review: true
    drafts: false
    ignore_title_keywords:
      - "WIP"
      - "DO NOT MERGE"
      - "[skip ci]"
      - "[skip review]"

  pre_merge_checks:
    custom_checks:
      - name: "Documentation Updated"
        mode: "warning"
        instructions: |
          Verify appropriate documentation is included for the changes:

          User-facing changes:
          - New features → docs/source/user_guide/ RST updates
          - New YAML params → docs/source/reference/yaml_schema.rst
          - New models → massgen/backend/capabilities.py (and token_manager.py if pricing needed)
          - Breaking changes → migration guide in docs

          Internal/architectural changes:
          - Complex features → design doc in docs/dev_notes/
          - New/changed functions → Google-style docstrings
          - New config options → example YAML in massgen/configs/

          Note: README.md and CHANGELOG.md are only updated in release PRs (dev/v0.1.X branches).

      - name: "Capabilities Registry Check"
        mode: "warning"
        instructions: |
          If backend or model changes are detected:
          1. Check massgen/backend/capabilities.py is updated with new models/capabilities
          2. Check massgen/token_manager/token_manager.py if pricing needed
          3. Run: uv run pytest massgen/tests/test_backend_capabilities.py -v

      - name: "Config Parameter Sync"
        mode: "warning"
        instructions: |
          If new YAML parameters are added, verify BOTH files are updated:
          1. massgen/backend/base.py → get_base_excluded_config_params()
          2. massgen/api_params_handler/_api_params_handler_base.py → get_base_excluded_config_params()

knowledge_base:
  code_guidelines:
    enabled: true
    filePatterns:
      - "CLAUDE.md"
      - "docs/dev_notes/release_checklist.md"
      - "docs/source/development/writing_configs.rst"

chat:
  auto_reply: true

early_access: false