python-package.yml

name: geosPythonPackages CI
on: pull_request

# Cancels in-progress workflows for a PR when updated
concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true


jobs:
  # Checks if PR title follows conventional semantics
  semantic_pull_request:
    permissions:
      pull-requests: write  # for amannn/action-semantic-pull-request to analyze PRs and
      statuses: write  # for amannn/action-semantic-pull-request to mark status of analyzed PR
      contents: read
    runs-on: ubuntu-latest

    steps:
      - name: Check if the PR name has conventional semantics
        if: github.event_name == 'pull_request'
        uses: amannn/action-semantic-pull-request@v5.5.3
        id: lint_pr_title
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        with:
          wip: true
          # Configure that a scope doesn't need to be provided.
          requireScope: false

      - name: Skip the check on main branch
        if:  github.ref_name == 'main'
        run: |
          echo "This is not a Pull-Request, skipping"

  build:
    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      max-parallel: 3
      matrix:
        python-version: ["3.10", "3.11", "3.12"]
        package-name:
          - geos-ats
          - geos-utils
          - geos-geomechanics
          - geos-mesh
          - geos-posp
          - geos-timehistory
          - geos-trame
          - geos-xml-tools
          - geos-xml-viewer
          - hdf5-wrapper
          - pygeos-tools
        include:
          - package-name: geos-geomechanics
            dependencies: "geos-utils"
          - package-name: geos-mesh
            dependencies: "geos-utils geos-geomechanics"
          - package-name: geos-posp
            dependencies: "geos-utils geos-mesh geos-geomechanics"
          - package-name: pygeos-tools
            dependencies: "geos-utils geos-mesh"
          - package-name: geos-timehistory
            dependencies: "hdf5-wrapper"
    steps:
    - uses: actions/checkout@v4
    - uses: mpi4py/setup-mpi@v1
    - name: Set up Python ${{ matrix.python-version }}
      uses: actions/setup-python@v5
      with:
        python-version: ${{ matrix.python-version }}
        cache: 'pip'
    - name: Install package
      # working-directory: ./${{ matrix.package-name }}
      run: |
        python -m pip install --upgrade pip
        python -m pip install pytest yapf toml

        DEPS="${{ matrix.dependencies || '' }}"

        if [ -n "$DEPS" ]; then
        echo "Installing additional dependencies: $DEPS"
        for dep in $DEPS; do
          python -m pip install ./$dep
        done
        fi

        echo "Installing main package..."
        python -m pip install ./${{ matrix.package-name }}/[test]

    - name: Lint with yapf
      # working-directory: ./${{ matrix.package-name }}
      run: |
        yapf -r --diff ./${{ matrix.package-name }} --style .style.yapf
    - name: Test with pytest
      #working-directory: ./${{ matrix.package-name }}
      run:
      # python -m pytest ./${{ matrix.package-name }} --doctest-modules --junitxml=junit/test-results.xml --cov-report=xml --cov-report=html |
      # wrap pytest to avoid error when no tests in the package
        sh -c 'python -m pytest ./${{ matrix.package-name }}; ret=$?; [ $ret = 5 ] && exit 0 || exit $ret'

  # Step 3: Check if GEOS integration is required based on changed files
  check_geos_integration_required:
    name: Check GEOS Integration Required
    runs-on: ubuntu-latest
    needs: [semantic_pull_request, build]
    if: github.event_name == 'pull_request'
    outputs:
      geos_integration_required: ${{ steps.check_changes.outputs.required }}
      skip_reason: ${{ steps.check_changes.outputs.skip_reason }}
    steps:
    - name: Checkout code
      uses: actions/checkout@v4
      with:
        fetch-depth: 0  # Fetch all history to compare with base branch

    - name: Check if GEOS integration is required
      id: check_changes
      run: |
        echo "Analyzing changed files to determine if GEOS integration test is required..."

        # Get list of changed files
        git fetch origin ${{ github.base_ref }}
        CHANGED_FILES=$(git diff --name-only origin/${{ github.base_ref }}...HEAD)

        echo "Changed files:"
        echo "$CHANGED_FILES"
        echo ""

        # Define packages that are integrated into GEOS (from GEOS/scripts/setupPythonEnvironment.bash)
        GEOS_INTEGRATED_PACKAGES=(
          "geos-utils"
          "geos-mesh"
          "geos-xml-tools"
          "hdf5-wrapper"
          "pygeos-tools"
          "geos-ats"
        )

        # Define patterns that DON'T require GEOS integration testing
        SKIP_PATTERNS=(
          "^docs/"
          "^\.github/workflows/doc-test\.yml$"
          "^\.github/workflows/typing-check\.yml$"
          "^README\.md$"
          "^\.readthedocs\.yml$"
          "^\.gitignore$"
          "^\.gitattributes$"
          "^\.style\.yapf$"
          "^\.ruff\.toml$"
          "^\.mypy\.ini$"
          # Packages not used in GEOS
          "^geos-geomechanics/"
          "^geos-posp/"
          "^geos-pv/"
          "^geos-timehistory/"
          "^geos-trame/"
          "^geos-xml-viewer/"
        )

        # Check if label is present (overrides automatic detection)
        HAS_LABEL=$(echo '${{ toJson(github.event.pull_request.labels.*.name) }}' | grep -q "test-geos-integration" && echo "true" || echo "false")

        if [[ "$HAS_LABEL" == "true" ]]; then
          echo "✓ Label 'test-geos-integration' found - GEOS integration test will run"
          echo "required=true" >> "$GITHUB_OUTPUT"
          echo "skip_reason=none" >> "$GITHUB_OUTPUT"
          exit 0
        fi

        # Check if any changed file affects GEOS-integrated packages
        REQUIRES_GEOS_TEST=false
        AFFECTED_PACKAGES=""

        for file in $CHANGED_FILES; do
          # Check if file matches any skip pattern
          SHOULD_SKIP=false
          for pattern in "${SKIP_PATTERNS[@]}"; do
            if echo "$file" | grep -qE "$pattern"; then
              SHOULD_SKIP=true
              break
            fi
          done

          if [[ "$SHOULD_SKIP" == "false" ]]; then
            # Check if file is in a GEOS-integrated package
            for package in "${GEOS_INTEGRATED_PACKAGES[@]}"; do
              if echo "$file" | grep -qE "^${package}/"; then
                REQUIRES_GEOS_TEST=true
                if [[ ! "$AFFECTED_PACKAGES" =~ "$package" ]]; then
                  AFFECTED_PACKAGES="$AFFECTED_PACKAGES $package"
                fi
              fi
            done

            # Check for CI workflow changes that affect GEOS integration
            if echo "$file" | grep -qE "^\.github/workflows/(python-package\.yml|test_geos_integration\.yml)$"; then
              REQUIRES_GEOS_TEST=true
              AFFECTED_PACKAGES="$AFFECTED_PACKAGES [CI-workflows]"
            fi

            # Check for root-level scripts that might affect integration
            if echo "$file" | grep -qE "^install_packages\.sh$"; then
              REQUIRES_GEOS_TEST=true
              AFFECTED_PACKAGES="$AFFECTED_PACKAGES [install-scripts]"
            fi
          fi
        done

        if [[ "$REQUIRES_GEOS_TEST" == "true" ]]; then
          echo "✓ GEOS integration test REQUIRED"
          echo "  Affected packages/components:$AFFECTED_PACKAGES"
          echo "  These packages are integrated into GEOS and require testing"
          echo "required=true" >> "$GITHUB_OUTPUT"
          echo "skip_reason=none" >> "$GITHUB_OUTPUT"
        else
          echo "⊘ GEOS integration test NOT required"
          echo "  All changes are in documentation, non-integrated packages, or config files"
          echo "  To force GEOS integration testing, add the 'test-geos-integration' label"
          echo "required=false" >> "$GITHUB_OUTPUT"
          echo "skip_reason=no-geos-integrated-changes" >> "$GITHUB_OUTPUT"
        fi

  # Step 4: Run GEOS integration tests (only if required or label present)
  geos_integration_test:
    name: GEOS Integration Test
    needs: [check_geos_integration_required]
    if: needs.check_geos_integration_required.outputs.geos_integration_required == 'true'
    uses: ./.github/workflows/test_geos_integration.yml

  # Final validation - Summarize CI results
  final_validation:
    name: Final CI Validation
    runs-on: ubuntu-latest
    needs: [check_geos_integration_required, geos_integration_test]
    if: always() && github.event_name == 'pull_request'
    steps:
    - name: Validate CI completion
      run: |
        echo "Final CI Validation"
        echo "==================="

        GEOS_REQUIRED="${{ needs.check_geos_integration_required.outputs.geos_integration_required }}"
        SKIP_REASON="${{ needs.check_geos_integration_required.outputs.skip_reason }}"
        GEOS_RESULT="${{ needs.geos_integration_test.result }}"

        if [[ "$GEOS_REQUIRED" == "true" ]]; then
          echo "GEOS integration test was required and triggered"
          if [[ "$GEOS_RESULT" == "success" ]]; then
            echo "✓ GEOS integration test PASSED"
            echo "✓ All CI requirements satisfied - PR can be merged"
          else
            echo "✗ GEOS integration test FAILED or was skipped"
            echo "✗ CI FAILED - PR cannot be merged until GEOS integration passes"
            exit 1
          fi
        else
          echo "GEOS integration test was NOT required"
          echo "Reason: $SKIP_REASON"
          echo ""
          echo "Changed files do not affect GEOS-integrated packages:"
          echo "  - geos-utils, geos-mesh, geos-xml-tools"
          echo "  - hdf5-wrapper, pygeos-tools, geos-ats"
          echo ""
          echo "If you want to run GEOS integration tests anyway,"
          echo "add the 'test-geos-integration' label to this PR"
          echo ""
          echo "✓ CI requirements satisfied - PR can be merged"
        fi