.pre-commit-config.yaml

# MLSysBook Pre-commit Configuration
# =============================================================================
# Combined configuration for both Book and TinyTorch projects
#
# Install: pip install pre-commit && pre-commit install
# Run all: pre-commit run --all-files
# Run specific hook: pre-commit run <hook-id> --all-files
#
# STRUCTURE:
#   1. GLOBAL HOOKS - Apply to entire repo
#   2. BOOK HOOKS - Quarto content validation (book/ directory)
#   3. TINYTORCH HOOKS - CLI and module validation (tinytorch/ directory)
# =============================================================================

repos:
  # ===========================================================================
  # SECTION 1: GLOBAL HOOKS (apply to entire repo)
  # ===========================================================================

  - repo: https://github.com/pre-commit/pre-commit-hooks
    rev: v5.0.0
    hooks:
      - id: trailing-whitespace
        name: "Global: Trim trailing whitespace"
        # Only run on markdown files to avoid breaking templates, extensions, etc.
        files: "\\.(md|qmd)$"
        exclude: "^(_site/|_book/|node_modules/)"
      - id: end-of-file-fixer
        name: "Global: Fix end of file newlines"
        # Only run on markdown files to avoid breaking templates, extensions, etc.
        files: "\\.(md|qmd)$"
        exclude: "^(_site/|_book/|node_modules/)"
      - id: check-json
        name: "Global: Validate JSON syntax"
      - id: check-yaml
        name: "Global: Validate YAML syntax"
        exclude: "^(_site/)"

  - repo: https://github.com/codespell-project/codespell
    rev: v2.3.0
    hooks:
      - id: codespell
        name: "Global: Check for common misspellings"
        args: ["--skip", "*.json,*.bib,*.js,*.tex,_site,_book,node_modules,.venv,htmlcov", "--ignore-words", ".codespell-ignore-words.txt"]
        exclude: "^(_site/|_book/|htmlcov/|.*\\.js$)"

  # ===========================================================================
  # SECTION 2: BOOK HOOKS (quarto content validation)
  # Files: book/quarto/contents/**/*.qmd
  # ===========================================================================

  # --- Book: Auto-formatters ---
  - repo: https://github.com/executablebooks/mdformat
    rev: 0.7.9
    hooks:
      - id: mdformat
        name: "Book: Format quarto markdown"
        additional_dependencies: [mdformat-frontmatter]
        files: ^book/quarto/contents/.*\.qmd$
        pass_filenames: true

  - repo: https://github.com/FlamingTempura/bibtex-tidy
    rev: v1.14.0
    hooks:
      - id: bibtex-tidy
        name: "Book: Tidy bibtex files"
        args: [
          '--align=space',
          '--curly',
          '--sort=key',
          '--sort-fields',
          '--duplicates=key',
          '--remove-empty-fields',
          '--space=2',
          '--trailing-commas',
          '--escape',
          '--wrap=100',
          '--blank-lines'
        ]
        files: ^book/quarto/contents/.*\.bib$

  # --- Book: Content validators ---
  - repo: local
    hooks:
      - id: book-collapse-blank-lines
        name: "Book: Collapse extra blank lines"
        entry: python book/tools/scripts/content/format_blank_lines.py
        language: python
        pass_filenames: true
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-format-python
        name: "Book: Format Python code blocks (Black, 70 chars)"
        entry: python book/tools/scripts/content/format_python_in_qmd.py
        language: python
        additional_dependencies: [black>=23.0.0]
        pass_filenames: true
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-check-list-formatting
        name: "Book: Fix markdown list formatting"
        entry: python book/tools/scripts/utilities/check_list_formatting.py --fix
        language: python
        pass_filenames: true
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-validate-json
        name: "Book: Validate JSON files"
        entry: python book/tools/scripts/utilities/validate_json.py
        language: python
        files: ^book/.*\.json$
        pass_filenames: true

      - id: book-check-unreferenced-labels
        name: "Book: Check for unreferenced labels"
        entry: python book/tools/scripts/content/check_unreferenced_labels.py ./book/quarto/contents/core
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-check-duplicate-labels
        name: "Book: Check for duplicate labels"
        entry: python book/tools/scripts/content/check_duplicate_labels.py
        args: ['-d', 'book/quarto/contents/', '--figures', '--tables', '--listings', '--quiet', '--strict']
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-validate-citations
        name: "Book: Validate citation references"
        entry: python book/tools/scripts/content/validate_citations.py --quiet
        language: python
        pass_filenames: true
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-validate-footnotes
        name: "Book: Validate footnote references"
        entry: python book/tools/scripts/content/footnote_cleanup.py -d book/quarto/contents/ --validate
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-check-forbidden-footnotes
        name: "Book: Check for footnotes in tables/captions"
        entry: python book/tools/scripts/content/check_forbidden_footnotes.py -d book/quarto/contents/
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-check-purpose-unnumbered
        name: "Book: Ensure Purpose sections are unnumbered"
        entry: bash -c 'grep -n "^## Purpose" "$@" | grep -v "\.unnumbered" && exit 1 || exit 0' --
        language: system
        pass_filenames: true
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-check-table-formatting
        name: "Book: Check table formatting"
        entry: python book/tools/scripts/content/format_tables.py --check -d book/quarto/contents/
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-validate-part-keys
        name: "Book: Validate part keys"
        entry: python book/tools/scripts/utilities/validate_part_keys.py
        language: python
        additional_dependencies: [pyyaml]
        pass_filenames: false
        files: ^book/.*\.qmd$

  # --- Book: Image validators ---
  - repo: local
    hooks:
      - id: book-validate-images
        name: "Book: Validate image files"
        entry: python book/tools/scripts/images/manage_images.py
        language: python
        additional_dependencies: [pillow, rich]
        pass_filenames: true
        files: ^book/quarto/contents/.*\.(png|jpg|jpeg|gif)$

      - id: book-validate-external-images
        name: "Book: Check for external images"
        entry: python book/tools/scripts/images/manage_external_images.py --validate book/quarto/contents/
        language: python
        additional_dependencies: [requests]
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-validate-image-references
        name: "Book: Check image references exist"
        entry: python book/tools/scripts/images/validate_image_references.py -d book/quarto/contents/ --quiet
        language: python
        pass_filenames: false
        files: ^book/quarto/contents/.*\.qmd$

      - id: book-prevent-svg
        name: "Book: Prevent manual SVG files"
        entry: bash -c 'echo "❌ Manual SVG files are not allowed. Convert to PNG."; exit 1'
        language: system
        files: ^book/quarto/contents/.*\.svg$
        exclude: ".*_files/mediabag/.*\\.svg$"

  # --- Book: System checks ---
  - repo: local
    hooks:
      - id: book-cleanup-artifacts
        name: "Book: Auto-cleanup build artifacts"
        entry: python book/tools/scripts/maintenance/cleanup_build_artifacts.py
        language: python
        additional_dependencies: [rich]
        pass_filenames: false
        stages: [pre-commit]

  # ===========================================================================
  # SECTION 3: TINYTORCH HOOKS (CLI and module validation)
  # Files: tinytorch/**/*.md, tinytorch/**/*.py
  # ===========================================================================

  # --- TinyTorch: Auto-formatters ---
  - repo: https://github.com/FlamingTempura/bibtex-tidy
    rev: v1.14.0
    hooks:
      - id: bibtex-tidy
        name: "TinyTorch: Tidy paper bibtex"
        args: [
          '--align=space',
          '--curly',
          '--sort=key',
          '--sort-fields',
          '--duplicates=key',
          '--remove-empty-fields',
          '--space=2',
          '--trailing-commas',
          '--escape',
          '--wrap=100',
          '--blank-lines'
        ]
        files: ^tinytorch/paper/.*\.bib$

  # --- TinyTorch: Content validators ---
  - repo: local
    hooks:
      - id: tinytorch-validate-cli-docs
        name: "TinyTorch: Validate CLI docs"
        description: "Ensure tito commands in docs match actual CLI"
        entry: python3 tinytorch/tools/dev/validate_cli_docs.py
        language: system
        files: ^tinytorch/.*\.md$
        pass_filenames: false

  # ===========================================================================
  # DISABLED/OPTIONAL HOOKS
  # ===========================================================================

  # - repo: https://github.com/igorshubovych/markdownlint-cli
  #   rev: v0.45.0
  #   hooks:
  #     - id: markdownlint
  #       name: "Book: Lint quarto markdown"
  #       files: ^book/quarto/contents/.*\.qmd$
  #       args: ["--quiet", "-c", "config/linting/.mdlintconfig.yml"]

  # - id: book-yamllint
  #   name: "Book: Validate YAML files"
  #   entry: yamllint
  #   language: system
  #   args: [--config-file=.yamllint]
  #   files: ^book/.*\.(yml|yaml)$