.pre-commit-config.yml

---
# yaml-language-server: $schema=https://raw.githubusercontent.com/j178/prek/refs/heads/master/prek.schema.json
minimum_prek_version: 0.2.29
default_install_hook_types:
  - pre-commit
default_stages:
  - pre-commit
# define priority: 0 for linters that don't modify sources
# leave fixers with priority: null so they run sequentially
repos:
  # "Built-in Fast Hooks" from prek.
  - repo: builtin
    hooks:
      - id: check-added-large-files
        name: Check for large files
        priority: 0

      - id: check-case-conflict
        name: Check for files that would conflict in case-insensitive filesystems
        priority: 0

      - id: trailing-whitespace
        name: Fix trailing whitespace
        types: [ text ]
        exclude:
          glob:
            - lucene/analysis/common/src/java/org/apache/lucene/analysis/charfilter/HTMLStripCharFilter.java

      - id: end-of-file-fixer
        name: Fix newline at EOF
        exclude:
          glob:
            # generated file
            - versions.lock
            # generated file
            - lucene/analysis/common/src/resources/org/apache/lucene/analysis/snowball/*_stop.txt
            # TODO: do this one later with more free RAM
            - lucene/analysis/common/src/java/org/apache/lucene/analysis/email/ASCIITLD.jflex

      - id: fix-byte-order-marker
        name: Fix UTF-8 byte order marker

      - id: check-json
        name: Check JSON files
        priority: 0
        types: [ json ]

      - id: check-toml
        name: Check TOML files
        priority: 0
        types: [ toml ]

      - id: check-yaml
        name: Check YAML files
        priority: 0
        types: [ yaml ]
        args: [ --allow-multiple-documents ]

      - id: check-xml
        name: Check XML files
        priority: 0
        types: [ xml ]

      - id: mixed-line-ending
        name: Fix line endings
        types: [ text ]

      - id: check-merge-conflict
        name: Check for merge conflicts
        priority: 0

      - id: detect-private-key
        name: Check for private keys
        priority: 0

      - id: check-executables-have-shebangs
        name: Check that (non-binary) executables have a shebang
        priority: 0

  - repo: local
    hooks:
      - id: shellcheck
        name: Check Shell Scripts
        priority: 0
        language: system
        entry: uv
        args: [ 'run', 'shellcheck', '--format', 'gcc' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        types: [ shell ]
        exclude:
          glob:
            # TODO: fix these shell scripts
            - dev-tools/test-patch/*.sh
            - gradle/regenerate/snowball/snowball.sh
            - gradlew
            - lucene/distribution/src/binary-release/bin/luke.sh

      - id: actionlint
        name: Check Github Actions
        priority: 0
        language: system
        entry: uv
        args: [ 'run', 'actionlint' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        # run across all actions if any are touched
        pass_filenames: false
        files:
          glob:
            - .github/**/*.yml

      - id: zizmor
        name: Check Github Actions Security
        priority: 0
        language: system
        entry: uv
        args: [ 'run', 'zizmor', '--pedantic', '--offline', '.' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        # run across all actions if any are touched
        pass_filenames: false
        files:
          glob:
            - .github/**/*.yml

      - id: rumdl
        name: Fix Markdown
        language: system
        entry: uv
        args: [ 'run', 'rumdl', 'check', '--fix' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        types: [ 'markdown']
        require_serial: true

      - id: ruff-check
        name: Fix Python
        language: system
        entry: uv
        args: [ 'run', 'ruff', 'check', '--fix', '--force-exclude' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        types_or: [ python, pyi, jupyter ]
        require_serial: true

      - id: ruff-format
        language: system
        name: Fix Python formatting
        entry: uv
        args: [ 'run', 'ruff', 'format', '--force-exclude' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        types_or: [ python, pyi, jupyter ]
        require_serial: true

      - id: test-ast-grep
        name: Check ast-grep rules
        priority: 0
        language: system
        entry: uv
        args: [ 'run', 'ast-grep', '-c', './gradle/validation/ast-grep/sgconfig.yml', 'test', '--skip-snapshot-tests' ]
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
        # run across all rules if any are touched
        pass_filenames: false
        files:
          glob:
            - gradle/validation/ast-grep/**/*.yml

      - id: ast-grep
        name: Check Sources with ast-grep
        priority: 0
        language: system
        entry: uv
        args: [ 'run', 'ast-grep', '-c', './gradle/validation/ast-grep/sgconfig.yml', 'scan', '--format', 'github' ]
        types_or: [ 'css', 'html', 'java', 'javascript', 'yaml']
        require_serial: true
        env:
          UV_PROJECT: dev-tools
          UV_FROZEN: "1"
  # self-checks for this configuration file
  - repo: meta
    hooks:
      - id: check-hooks-apply
        name: Check prek hooks match at least one file
        priority: 0
        files:
          glob:
            .pre-commit-config.yml

      - id: check-useless-excludes
        name: Check prek excludes match include pattern
        priority: 0
        files:
          glob:
            .pre-commit-config.yml