ci.yml

name: compose-ci

on: [push, pull_request]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          submodules: true

      - uses: actions/setup-python@v5
        with:
          python-version: "3.11"

      - name: Install test dependencies
        run: |
          pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu
          pip install pytest xarray netCDF4 numpy matplotlib scikit-learn pillow

      # Create a dummy checkpoint so setup_device_and_model hits the except
      # branch and downloads DINO weights from Facebook's hub instead.
      - name: Create dummy checkpoint for DINO hub fallback
        run: touch checkpoint.pth

      - name: Run tests
        run: pytest tests/ -v

  pipeline:
    runs-on: ubuntu-latest

    steps:
      #--------------------------------------------------------------
      # 1. Checkout (include submodules such as raw_consumer/pyEcholab)
      #--------------------------------------------------------------
      - name: Checkout repository (with submodules)
        uses: actions/checkout@v4
        with:
          submodules: recursive
          fetch-depth: 0

      #--------------------------------------------------------------
      # 2. (Optional) pull a sample .raw file so the pipeline has input
      #--------------------------------------------------------------
      - name: Download test data
        run: |
          mkdir -p data/input
          aws s3 cp --no-sign-request \
            "s3://noaa-wcsd-pds/data/raw/Bell_M._Shimada/SH2306/EK80/Hake-D20230811-T165727.raw" \
            data/input/
          touch inference/checkpoint.pth

      #--------------------------------------------------------------
      # 3. Build images & start RAW → PRE → INFER stack detached
      #--------------------------------------------------------------
      - name: Start Docker pipeline (detached)
        run: |
          docker compose -f docker-compose.yml up --build -d \
            raw preprocessing infer

      #--------------------------------------------------------------
      # 4. Poll for inference output (max 5 min)
      #--------------------------------------------------------------
      - name: Wait for inference output (max 5 min)
        run: |
          end=$((SECONDS+300))
          echo "Polling for PNGs in data/inference ..."
          while [ $SECONDS -lt $end ]; do
            cnt=$(find data/inference -type f -name '*.png' | wc -l)
            if [ "$cnt" -gt 0 ]; then
              echo "✅  Found $cnt PNG(s) – pipeline succeeded."
              exit 0
            fi
            sleep 10
          done
          echo "❌  No PNGs produced in 5 minutes."
          docker compose -f docker-compose.yml logs --tail 50
          exit 1

      #--------------------------------------------------------------
      # 5. Upload all generated PNGs as artifacts
      #--------------------------------------------------------------
      - name: Upload preprocessing PNGs
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: preprocessing-pngs
          path: data/preprocessing/**/*.jpg
          retention-days: 7

      - name: Upload inference PNGs
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: inference-pngs
          path: data/inference/**/*.png
          retention-days: 7

      #--------------------------------------------------------------
      # 6. Tear down containers & volumes (always runs)
      #--------------------------------------------------------------
      - name: Teardown
        if: always()
        run: docker compose -f docker-compose.yml down -v