gandlf.yml

# This workflow will install Python dependencies, run tests and lint with a single version of Python
# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions

name: GaNDLF TaskRunner

on:
  pull_request:
    branches: [ develop ]
    types: [opened, synchronize, reopened, ready_for_review]

permissions:
  contents: read

env:
  # A workaround for long FQDN names provided by GitHub actions.
  FQDN: "localhost"

jobs:
  build:
    if: github.event.pull_request.draft == false
    runs-on: ubuntu-latest
    timeout-minutes: 15

    steps:
    - uses: actions/checkout@v3
    - name: Set up Python 3.10
      uses: actions/setup-python@v3
      with:
        python-version: "3.10"
    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip
        pip install typer==0.11.1
        pip install torch==2.3.1+cpu torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
        pip install .
    - name: Install GaNDLF
      run: |
        git clone https://github.com/MLCommons/GaNDLF.git ./gandlf
        cd gandlf
        git fetch --tags
        echo "Checkout the latest GaNDLF tag"
        latestTag=$(git describe --tags "$(git rev-list --tags --max-count=1)")
        # Instead of using the latest tag, we are pinning gandlf to an older version as changes are required
        # in torch to work with a later version.
        git checkout 0.1.1
    - name: GaNDLF Task Runner Test
      run: |
        cd gandlf
        pwd
        pip install -e .
        pip uninstall onnx -y
        cat ./GANDLF/version.py
        echo "Download data and Split CSVs into training and validation"
        # python -c "from testing.test_full import test_generic_download_data, test_generic_constructTrainingCSV; test_generic_download_data(); test_generic_constructTrainingCSV()"
        pytest --cov=. --cov-report=xml -k "prepare_data_for_ci"
        head -n 1 testing/data/train_2d_rad_segmentation.csv > /home/runner/work/openfl/openfl/valid.csv
        tail -n +9 testing/data/train_2d_rad_segmentation.csv >> /home/runner/work/openfl/openfl/valid.csv
        head -n 8 testing/data/train_2d_rad_segmentation.csv > /home/runner/work/openfl/openfl/train.csv
        cp testing/config_segmentation.yaml /home/runner/work/openfl/openfl/config_segmentation.yaml
        echo "DEBUG display the config file"
        cat /home/runner/work/openfl/openfl/config_segmentation.yaml
        echo "Initialize OpenFL plan"
        ## from docs
        export WORKSPACE_TEMPLATE=gandlf_seg_test
        export WORKSPACE_PATH=./my_federation
        fx workspace create --prefix ${WORKSPACE_PATH} --template ${WORKSPACE_TEMPLATE}
        cd ${WORKSPACE_PATH}
        mkdir ./data/one
        mkdir ./data/two
        cp /home/runner/work/openfl/openfl/*.csv ./data/one/
        cp /home/runner/work/openfl/openfl/*.csv ./data/two/
        ## from docs
        # fx plan initialize --gandlf_config ../testing/config_segmentation.yaml
        cd /home/runner/work/openfl/openfl
        ls
        file "/home/runner/work/openfl/openfl/config_segmentation.yaml"
        ## for 2d data, only a single change is needed in the gandlf config
        sed -i 's/# n_channels: 3/num_channels: 3/g' "/home/runner/work/openfl/openfl/config_segmentation.yaml"
        ## for 3d data, the following changes are needed in the gandlf config -- commented out for now
        # sed -i 's/dimension: 2/dimension: 3/g' "/home/runner/work/openfl/openfl/config_segmentation.yaml"
        # sed -i 's/0,255/0,1/g' "/home/runner/work/openfl/openfl/config_segmentation.yaml"
        # sed -i 's/128,128/32,32,32/g' "/home/runner/work/openfl/openfl/config_segmentation.yaml"
        python -m tests.github.test_gandlf --template gandlf_seg_test --fed_workspace aggregator --col1 one --col2 two --rounds-to-train 1 --gandlf_config "/home/runner/work/openfl/openfl/config_segmentation.yaml"