Tests workflow (datachain-ai/datachain)

name: Tests
 
on:
  push:
    branches: [main]
  pull_request:
  workflow_dispatch:
 
env:
  FORCE_COLOR: "1"
 
concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
  cancel-in-progress: true
 
jobs:
  lint:
    timeout-minutes: 30
    runs-on: latchkey-small
    steps:
      - name: Check out the repository
        uses: actions/checkout@v7
        with:
          fetch-depth: 0
          ref: ${{ github.event.pull_request.head.sha || github.ref }}
 
      - name: Set up Python 3.10
        uses: actions/setup-python@v6
        with:
          cache: 'pip'
          python-version: '3.10'
 
      - name: Setup uv
        uses: astral-sh/setup-uv@v7
        with:
          enable-cache: true
          cache-suffix: lint
          cache-dependency-glob: pyproject.toml
 
      - name: Use CPU-only PyTorch on Linux
        run: echo "UV_TORCH_BACKEND=cpu" >> "$GITHUB_ENV"
 
      - name: Install nox
        run: uv pip install nox --system
 
      - name: Cache mypy
        uses: actions/cache@v6
        with:
          path: .mypy_cache
          key: mypy-${{ runner.os }}-${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}
 
      - name: Cache pre-commit hooks
        uses: actions/cache@v6
        with:
          path: ~/.cache/pre-commit
          key: pre-commit-3|${{ env.pythonLocation }}|${{ hashFiles('.pre-commit-config.yaml') }}
 
      - name: Lint code
        run: nox -s lint
 
  datachain:
    timeout-minutes: 40
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest-8-cores]
        pyv: ['3.10', '3.11', '3.12', '3.13', '3.14']
        include:
          - os: macos-latest
            pyv: '3.10'
          - os: macos-latest
            pyv: '3.14'
          - os: windows-latest
            pyv: '3.10'
          - os: windows-latest
            pyv: '3.14'
 
    steps:
      - name: Check out the repository
        uses: actions/checkout@v7
        with:
          fetch-depth: 0
          ref: ${{ github.event.pull_request.head.sha || github.ref }}
 
      - name: Setup PostgreSQL
        if: runner.os != 'Windows'
        uses: ikalnytskyi/action-setup-postgres@c4dda34aae1c821e3a771b68b73b13af3198a7ee # v8
        with:
          username: test
          password: test
          database: test_datachain
          port: 5432
          postgres-version: "17"
        id: postgres
 
      - name: Set PostgreSQL URI
        if: runner.os != 'Windows'
        run: |
          FULL_URI="${{ steps.postgres.outputs.connection-uri }}"
          echo "TEST_POSTGRES_URI=${FULL_URI%/*}" >> "$GITHUB_ENV"
        shell: bash
 
      - name: Set up Python ${{ matrix.pyv }}
        uses: actions/setup-python@v6
        with:
          cache: 'pip'
          python-version: ${{ matrix.pyv }}
 
      - name: Setup uv
        uses: astral-sh/setup-uv@v7
        with:
          enable-cache: true
          cache-suffix: tests-${{ matrix.pyv }}
          cache-dependency-glob: pyproject.toml
 
      - name: Use CPU-only PyTorch on Linux
        if: runner.os == 'Linux'
        # torch 2.11+ on PyPI pulls CUDA deps that aren't fully available on
        # CPU-only runners (missing libnppicc from nvidia-npp).
        run: echo "UV_TORCH_BACKEND=cpu" >> "$GITHUB_ENV"
 
      - name: Install FFmpeg on Windows
        if: runner.os == 'Windows'
        run: choco install ffmpeg --version 7.0.2
 
      - name: Install FFmpeg on macOS
        if: runner.os == 'macOS'
        # Using ffmpeg@7 since torchcodec does not support ffmpeg 8 yet
        # See: https://github.com/pytorch/torchcodec/issues/839
        # On Windows and Linux ffmpeg < 8 is installed by default
        run: |
          brew install ffmpeg@7
          echo 'DYLD_FALLBACK_LIBRARY_PATH=/opt/homebrew/opt/ffmpeg@7/lib' >> "$GITHUB_ENV"
 
      - name: Install FFmpeg on Ubuntu
        if: runner.os == 'Linux'
        run: |
          sudo apt update
          sudo apt install -y ffmpeg
 
      - name: Install nox
        run: uv pip install nox --system
 
      - name: Skip flaky azure, gs remotes on macOS
        if: runner.os == 'macOS'
        run: echo 'DISABLE_REMOTES_ARG=--disable-remotes=azure,gs' >> "$GITHUB_ENV"
 
      - name: Skip all remotes on Windows
        if: runner.os == 'Windows'
        run: echo 'DISABLE_REMOTES_ARG=--disable-remotes=azure,gs' >> $env:GITHUB_ENV
 
      - name: Run tests
        run: nox -s tests-${{ matrix.pyv }} -- -m "not e2e and not examples" $DISABLE_REMOTES_ARG
        shell: bash
 
      - name: Run E2E tests
        run: nox -s e2e-${{ matrix.pyv }}
        shell: bash
 
      - name: Upload coverage report
        uses: codecov/codecov-action@v7
        with:
          token: ${{ secrets.CODECOV_TOKEN }}
          files: coverage.xml
          flags: datachain
 
      - name: Build package
        run: nox -s build
 
      - name: Build docs
        run: nox -s docs
 
  examples:
    runs-on: ${{ matrix.os }}
    timeout-minutes: 60
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest, windows-latest]
        pyv: ['3.11', '3.14']
        group: ['get_started', 'computer_vision', 'multimodal']
        exclude:
          - {os: ubuntu-latest, pyv: '3.11', group: 'multimodal'}
          - {os: ubuntu-latest, pyv: '3.14', group: 'multimodal'}
        include:
          # HF runs against actual API - thus run it only once
          - {os: ubuntu-latest, pyv: "3.14", group: llm_and_nlp}
          - {os: ubuntu-latest-4-cores, pyv: "3.11", group: multimodal}
          - {os: ubuntu-latest-4-cores, pyv: "3.14", group: multimodal}
 
    steps:
      - uses: actions/checkout@v7
        with:
          ref: ${{ github.event.pull_request.head.sha || github.ref }}
 
      - name: Set up Python ${{ matrix.pyv }}
        uses: actions/setup-python@v6
        with:
          cache: 'pip'
          python-version: ${{ matrix.pyv }}
 
      - name: Setup uv
        uses: astral-sh/setup-uv@v7
        with:
          enable-cache: true
          cache-suffix: examples-${{ matrix.pyv }}
          cache-dependency-glob: pyproject.toml
 
      - name: Use CPU-only PyTorch on Linux
        if: runner.os == 'Linux'
        run: echo "UV_TORCH_BACKEND=cpu" >> "$GITHUB_ENV"
 
      - name: Install nox
        run: uv pip install nox --system
 
      - name: Install FFmpeg on Windows
        if: runner.os == 'Windows'
        run: choco install ffmpeg
 
      - name: Install FFmpeg on macOS
        if: runner.os == 'macOS'
        run: |
          brew install ffmpeg
          echo 'DYLD_FALLBACK_LIBRARY_PATH=/opt/homebrew/lib' >> "$GITHUB_ENV"
 
      - name: Install FFmpeg on Ubuntu
        if: runner.os == 'Linux'
        run: |
          sudo apt update
          sudo apt install -y ffmpeg
 
      - name: Run examples
        env:
          # Pass HF_TOKEN to all example groups: several groups download public
          # models/datasets from the HF Hub anonymously (transformers, open_clip),
          # which now hits HTTP 429 rate limits. Authenticating raises the limit.
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
          ANTHROPIC_API_KEY: ${{ matrix.group == 'llm_and_nlp' && secrets.ANTHROPIC_API_KEY || '' }}
          OPENAI_API_KEY: ${{ matrix.group == 'multimodal' && secrets.OPENAI_API_KEY || '' }}
        run: nox -s examples -p ${{ matrix.pyv }} -- -m "${{ matrix.group }}"
 
  check:
    timeout-minutes: 30
    if: always()
    needs: [lint, datachain, examples]
    runs-on: latchkey-small
    steps:
      - uses: re-actors/alls-green@release/v1
        with:
          allowed-failures: examples
          jobs: ${{ toJSON(needs) }}