ci.yml

name: Run all checks

on:
  pull_request:
    branches:
      - main
  workflow_dispatch:
    inputs:
      ref:
        description: 'The git ref to build the package for'
        required: false
        default: ''
        type: string
      use_lkg:
        description: 'Whether to use the last known good versions of dependencies'
        required: false
        default: True
        type: boolean
  # nightly
  schedule:
    - cron: '0 0 * * *'

# Only run once per PR, canceling any previous runs
concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
  cancel-in-progress: true

# Precompute the ref if the workflow was triggered by a workflow dispatch rather than copying this logic repeatedly
env:
  ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || null }}
  # we want to use the LKG if that is explicitly requested, or if we're in a PR, but not a nightly run
  # the final `|| ''` is because env vars are always converted to strings and the string 'false' is truthy (!!)
  # (see https://github.com/orgs/community/discussions/25645)
  use_lkg: ${{ (github.event_name == 'workflow_dispatch' && inputs.use_lkg) || github.event_name == 'pull_request' || ''}}

jobs:
  eval: 
    name: Evaluate changes
    runs-on: ubuntu-latest
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        ref: ${{ env.ref }}
        fetch-depth: 2

      # We want to enforce the following rules for PRs:
      # * if all modifications are to README.md
      #     no testing is needed
      # * if there are modifications to docs/* or to any code
      #     then docs need to be built to verify consistency
      # * if there are modifications to notebooks/* or to any code
      #     then notebooks need to be run to verify consistency
      # * for any code changes (or changes to metadata files)
      #     linting and testing should be run
      # For a PR build, HEAD will be the merge commit, and we want to diff against the base branch,
      #  which will be the first parent: HEAD^ 
      # (For non-PR changes, we will always perform all CI tasks)
      # Note that GitHub Actions provides path filters, but they operate at the workflow level, not the job level
    - name: Determine type of code change
      run: |
        if ($env:GITHUB_EVENT_NAME -eq 'pull_request') {
          $editedFiles = git diff HEAD^ --name-only
          $editedFiles # echo edited files to enable easier debugging
          $codeChanges = $false
          $docChanges = $false
          $nbChanges = $false
          $changeType = "none"
          foreach ($file in $editedFiles) {
            switch -Wildcard ($file) {
              "README.md" { Continue }
              ".gitignore" { Continue }
              "econml/_version.py" { Continue }
              "prototypes/*" { Continue }
              "images/*" { Continue }
              "doc/*" { $docChanges = $true; Continue }
              "notebooks/*" { $nbChanges = $true; Continue }
              default { $codeChanges = $true; Continue }
            }
          }
        }
        echo "buildDocs=$(($env:GITHUB_EVENT_NAME -ne 'pull_request') -or ($docChanges -or $codeChanges))" >> $env:GITHUB_OUTPUT
        echo "buildNbs=$(($env:GITHUB_EVENT_NAME -ne 'pull_request') -or ($nbChanges -or $codeChanges))" >> $env:GITHUB_OUTPUT
        echo "testCode=$(($env:GITHUB_EVENT_NAME -ne 'pull_request') -or $codeChanges)" >> $env:GITHUB_OUTPUT
      shell: pwsh
      id: eval
    outputs: 
      buildDocs: ${{ steps.eval.outputs.buildDocs }}
      buildNbs: ${{ steps.eval.outputs.buildNbs }}
      testCode: ${{ steps.eval.outputs.testCode }}

  lint:
    name: Lint code
    needs: [eval]
    if: ${{ needs.eval.outputs.testCode == 'True' }}
    runs-on: ubuntu-latest
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        ref: ${{ env.ref }}
    - name: Setup Python
      uses: actions/setup-python@v5
      with:
        python-version: '3.9'
    - name: Ensure latest pip and setuptools
      run: python -m pip install --upgrade pip && pip install --upgrade setuptools
    - name: Run pycodestyle to check style
      run: 'pip install pycodestyle && pycodestyle econml'

  notebooks:
    name: Run notebooks
    needs: [eval]
    if: ${{ needs.eval.outputs.buildNbs == 'True' }}
    runs-on: ubuntu-latest
    strategy:
      matrix:
        kind: [except-customer-scenarios, customer-scenarios]
        include:
        - kind: "except-customer-scenarios"
          extras: "[tf,plt,ray]"
          pattern: "(?!CustomerScenarios)"
          install_graphviz: true
          version: '3.8' # no supported version of tensorflow for 3.9
        - kind: "customer-scenarios"
          extras: "[plt,dowhy]"
          pattern: "CustomerScenarios"
          version: '3.9'
          install_graphviz: false
      fail-fast: false
    env:
      id_string: ${{ matrix.kind }}-${{ matrix.version }}
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        ref: ${{ env.ref }}
    - name: Setup Python
      uses: actions/setup-python@v5
      with:
        python-version: ${{ matrix.version }}
    - name: Install uv
      # check if we're running on windows
      run: ${{ runner.os == 'Windows' && 'irm https://astral.sh/uv/install.ps1 | iex' || 'curl -LsSf https://astral.sh/uv/install.sh | sh' }}
    - name: Install graphviz
      run: sudo apt-get -yq install graphviz
      if: ${{ matrix.install_graphviz }}
    # Add verbose flag to pip installation if in debug mode
    - name: Install econml
      run: uv pip install --system -e .${{ matrix.extras }} ${{ fromJSON('["","-v"]')[runner.debug] }} ${{ env.use_lkg && '-r lkg-notebook.txt' }}
    # Install notebook requirements (if not already done as part of lkg)
    - name: Install notebook requirements
      run: uv pip install --system jupyter jupyter-client nbconvert nbformat seaborn xgboost tqdm
      if: ${{ !env.use_lkg }}
    - name: Save installed packages
      run: pip freeze --exclude-editable > notebooks-${{ env.id_string }}-requirements.txt
    - name: Upload installed packages
      uses: actions/upload-artifact@v4
      with:
        name: requirements-${{ env.id_string }}
        path: notebooks-${{ env.id_string }}-requirements.txt
    - name: Install pytest
      run: uv pip install --system pytest pytest-xdist pytest-cov coverage[toml]
    - name: Run notebook tests
      run: python -m pytest
      id: run_tests
      env:
        PYTEST_ADDOPTS: '-m "notebook"'
        NOTEBOOK_DIR_PATTERN: ${{ matrix.pattern }}
        COVERAGE_PROCESS_START: 'pyproject.toml'
    - name: Make coverage filename unique
      run: mv .coverage .coverage.${{ env.id_string }}
      # Run whether or not the tests passed, but only if they ran at all
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
    - name: Upload coverage report
      uses: actions/upload-artifact@v4
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
      with:
        name: coverage-${{ env.id_string }}
        path: .coverage.${{ env.id_string }}    
    - name: Make test XML filename unique
      run: mv junit/test-results.xml ${{ env.id_string }}-test-results.xml
      # Run whether or not the tests passed, but only if they ran at all
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
    - name: Upload test XML files
      uses: actions/upload-artifact@v4
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
      with:
        name: tests-${{ env.id_string }}
        path: ${{ env.id_string }}-test-results.xml
    - name: Upload notebook outputs
      uses: actions/upload-artifact@v4
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
      with:
        name: notebooks-${{ env.id_string }}
        path: notebooks/output/
  
  tests:
    name: "Run tests"
    needs: [eval]
    if: ${{ needs.eval.outputs.testCode == 'True' }}
    strategy:
      matrix:
        os: [ubuntu-latest, windows-latest, macos-12]
        python-version: ['3.8', '3.9', '3.10', '3.11', '3.12']
        kind: [serial, other, dml, main, treatment, ray]
        exclude:
          # Serial tests fail randomly on mac sometimes, so we don't run them there
          - os: macos-12
            kind: serial
          # Ray tests run out of memory on Windows
          - os: windows-latest
            kind: ray
          # Ray doesn't currently support Python 3.12
          - python-version: '3.12'
            kind: ray
        # Assign the correct package and testing options for each kind of test
        include:
          - kind: serial
            opts: '-m "serial and not ray" -n 1'
            extras: ""
          - kind: other
            opts: '-m "cate_api and not ray" -n auto'
            extras: "[tf,plt]"
          - kind: dml
            opts: '-m "dml and not ray"'
            extras: "[tf,plt]"
          - kind: main
            opts: '-m "not (notebook or automl or dml or serial or cate_api or treatment_featurization or ray)" -n 2'
            extras: "[tf,plt,dowhy]"
          - kind: treatment
            opts: '-m "treatment_featurization and not ray" -n auto'
            extras: "[tf,plt]"
          - kind: ray
            opts: '-m "ray"'
            extras: "[ray]"
      fail-fast: false    
    runs-on: ${{ matrix.os }}
    env:
      id_string: ${{ matrix.os }}-${{ matrix.python-version }}-${{ matrix.kind }}
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        ref: ${{ env.ref }}
    - name: Setup Python
      uses: actions/setup-python@v5
      with:
        python-version: ${{ matrix.python-version }}
    - name: Install uv
      # check if we're running on windows
      run: ${{ runner.os == 'Windows' && 'irm https://astral.sh/uv/install.ps1 | iex' || 'curl -LsSf https://astral.sh/uv/install.sh | sh' }}
    # Add verbose flag to pip installation if in debug mode
    - name: Install econml
      run: uv pip install --system -e .${{ matrix.extras }} ${{ fromJSON('["","-v"]')[runner.debug] }} ${{ env.use_lkg && '-r lkg.txt' }}
    - name: Save installed packages
      run: pip freeze --exclude-editable > tests-${{ env.id_string }}-requirements.txt
    - name: Upload installed packages
      uses: actions/upload-artifact@v4
      with:
        name: requirements-${{ env.id_string }}
        path: tests-${{ env.id_string }}-requirements.txt
    - name: Install pytest
      run: uv pip install --system pytest pytest-xdist pytest-cov coverage[toml]
    - name: Run tests
      run: python -m pytest
      id: run_tests
      env:
        PYTEST_ADDOPTS: ${{ matrix.opts }}
        COVERAGE_PROCESS_START: 'pyproject.toml'
    - name: Make coverage filename unique
      run: mv .coverage .coverage.${{ env.id_string }}
      # Run whether or not the tests passed, but only if they ran at all
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
    - name: Upload coverage report
      uses: actions/upload-artifact@v4
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
      with:
        name: coverage-${{ env.id_string }}
        path: .coverage.${{ env.id_string }}
    - name: Make test XML filename unique
      run: mv junit/test-results.xml ${{ env.id_string }}-test-results.xml
      # Run whether or not the tests passed, but only if they ran at all
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
    - name: Upload test XML files
      uses: actions/upload-artifact@v4
      if: success() || failure() && contains(fromJSON('["success", "failure"]'), steps.run_tests.outcome)
      with:
        name: tests-${{ env.id_string }}
        path: ${{ env.id_string }}-test-results.xml

  coverage-report:
    name: "Coverage report"
    needs: [tests, notebooks]
    if: success() || failure()
    runs-on: ubuntu-latest
    steps:
    - name: Checkout repository
      uses: actions/checkout@v4
      with:
        ref: ${{ env.ref }}
    - name: Get coverage reports
      uses: actions/download-artifact@v4
      with:
        pattern: coverage-*
        path: coverage
        merge-multiple: true
    - name: Setup Python
      uses: actions/setup-python@v5
      with:
        python-version: '3.8'
    - name: Install coverage
      run: pip install coverage[toml]
    - name: Combine coverage reports
      run: coverage combine coverage/
    - name: Generate coverage report
      run: coverage report -m --format=markdown > $GITHUB_STEP_SUMMARY
    - name: Generate coverage html --fail-under=86
      run: coverage html
    - name: Upload coverage report
      uses: actions/upload-artifact@v4
      with:
        name: coverage-report
        path: htmlcov
      
  merge-artifacts:
    name: "Merge artifacts"
    needs: [coverage-report, tests, notebooks]
    if: success() || failure()
    strategy:
      matrix:
        artifact: [requirements, tests, coverage, notebooks]
    runs-on: ubuntu-latest
    steps:
    - name: "Merge artifacts"
      uses: actions/upload-artifact/merge@v4
      with:
        name: ${{ matrix.artifact }}
        pattern: "${{ matrix.artifact }}-*"
        delete-merged: true
      
  build:
    name: Build package
    needs: [eval]
    if: ${{ needs.eval.outputs.testCode == 'True' }}
    uses: ./.github/workflows/publish-package.yml
    with:
      publish: false
      environment: test
      # don't have access to env context here for some reason
      ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || null }}
      # can't use env context here so need to duplicate expression, but these are true boolean values so don't need extra string logic
      use_lkg: ${{ (github.event_name == 'workflow_dispatch' && inputs.use_lkg) || github.event_name == 'pull_request' }}

  docs:
    name: Build documentation
    needs: [eval]
    if: ${{ needs.eval.outputs.buildDocs == 'True' }}
    uses: ./.github/workflows/publish-documentation.yml
    with:
      publish: false
      environment: test
      # don't have access to env context here for some reason
      ref: ${{ github.event_name == 'workflow_dispatch' && inputs.ref || null }}
      # can't use env context here so need to duplicate expression, but these are true boolean values so don't need extra string logic
      use_lkg: ${{ (github.event_name == 'workflow_dispatch' && inputs.use_lkg) || github.event_name == 'pull_request' }}
      
  verify:
    name: Verify CI checks
    needs: [lint, notebooks, tests, build, docs]
    if: always()
    runs-on: ubuntu-latest
    steps:
    - name: At least one check failed or was cancelled
      run: exit 1
      if: ${{ contains(needs.*.result, 'failure') || contains(needs.*.result, 'cancelled') }}
    - name: All checks passed
      run: exit 0
      if: ${{ !(contains(needs.*.result, 'failure') || contains(needs.*.result, 'cancelled')) }}