aws_runner.yml

name: aws_runner
on:
  issues:
    types: [ assigned, labeled ] 

jobs:
  start-runner:
    name: Start EC2 runner
    if: ${{ startsWith(github.event.issue.title, 'New Model:') && github.event.label.name != 'failed' }}
    runs-on: ubuntu-latest
    outputs:
      label: ${{ steps.start-ec2-runner.outputs.label }}
      ec2-instance-id: ${{ steps.start-ec2-runner.outputs.ec2-instance-id }}
    steps:
      - name: Configure AWS credentials
        uses: aws-actions/configure-aws-credentials@v1
        with:
          aws-access-key-id: ${{ secrets.AWS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_KEY_SECRET }}
          aws-region: ${{ vars.AWS_REGION }}
      - name: Start EC2 runner
        id: start-ec2-runner
        uses: machulav/ec2-github-runner@v2
        with:
          mode: start
          github-token: ${{ secrets.GH_TOKEN }}
          ec2-image-id: ${{ vars.AWS_IMAGE_ID }}
          ec2-instance-type: ${{ vars.AWS_INSTANCE_TYPE }}
          subnet-id: ${{ vars.AWS_SUBNET }}
          security-group-id: ${{ vars.AWS_SECURITY_GROUP }}

  create_new_branch:
    needs: [start-runner]
    runs-on: ${{ needs.start-runner.outputs.label }} # run the job on the newly created runner
    steps:
      - name: Cleanup disk space for large docker images
        run: |
          sudo rm -rf /usr/share/dotnet
          sudo rm -rf /opt/ghc
          sudo rm -rf "/usr/local/share/boost"
          sudo rm -rf /opt/hostedtoolcache
          sudo rm -rf "$AGENT_TOOLSDIRECTORY"
            
      - name: Free Disk Space (Ubuntu)
        uses: jlumbroso/free-disk-space@main
        with:
          # this might remove tools that are actually needed,
          # if set to "true" but frees about 6 GB
          tool-cache: false
          
          # all of these default to true, but feel free to set to
          # "false" if necessary for your workflow
          android: true
          dotnet: true
          haskell: true
          large-packages: true
          docker-images: false

      - name: Install singularity
        run: |
          wget https://github.com/sylabs/singularity/releases/download/v4.0.0/singularity-ce_4.0.0-focal_amd64.deb

          sudo apt install ./singularity-ce_4.0.0-focal_amd64.deb -y
          singularity version

      - name: Install Datalad
        id: datalad-testing
        run: |
          apt install unzip datalad python3-pip subversion tree -y
          python3 -m pip install datalad-installer datalad-osf
          python3 -m pip install pyyaml oyaml linkml
          
          datalad-installer --sudo ok git-annex -m datalad/git-annex:release
          sudo git config --global filter.annex.process "git-annex filter-process"

      - name: Configure git username and email
        run: | 
          git config --system user.name "trained_models"
          git config --system user.email "trained_models"

      # This will automatically create a new branch from this issue, using custom config at /.github/issue-branch.yml 🟢
      - name: Create Issue Branch
        id: branch
        uses: robvanderleek/create-issue-branch@main
        env:
          GITHUB_TOKEN: ${{ secrets.GH_TOKEN }}
          
      - name: Clone repo and checkout branch
        uses: actions/checkout@v4
        with:
          ref: ${{ steps.branch.outputs.branchName }}
          fetch-depth: 0
          
      - name: Parse github issue body
        uses: stefanbuck/github-issue-parser@v3
        id: issue-parser
        with:
          template-path: .github/ISSUE_TEMPLATE/addModel.yml # optional but recommended
        env:
          HOME: "/home/ubuntu"

      - name: Get the Path
        id: get_path
        run: echo "path=${{ steps.issue-parser.outputs.issueparser_path }}" >> $GITHUB_OUTPUT
      
      - name: Get the Weights
        id: get_weights
        run: echo "weights=${{ steps.issue-parser.outputs.issueparser_weights }}" >> $GITHUB_OUTPUT
      
      - name: Get docker folder
        id: get_docker
        run: echo "docker=${{ steps.issue-parser.outputs.issueparser_docker }}" >> $GITHUB_OUTPUT

      - name: Get Python Scripts
        id: python_scripts
        run: |
          echo "pythons<<EOF" >> $GITHUB_ENV
          echo "${{ steps.issue-parser.outputs.issueparser_python-scripts }}" >> $GITHUB_ENV
          echo "EOF" >> $GITHUB_ENV

      - name: Get Model Info
        id: get_model_info
        run: |
          echo "model_info<<EOF" >> $GITHUB_ENV
          echo "${{ steps.issue-parser.outputs.issueparser_model_info }}" >> $GITHUB_ENV
          echo "EOF" >> $GITHUB_ENV

      - name: Get Test Command
        id: get_test_command
        run: |
          echo "test_command=${{ steps.issue-parser.outputs.issueparser_testCommand }}" >> $GITHUB_OUTPUT

      - name: Get Sample Data
        id: get_data
        run: echo "sample_data=${{ steps.issue-parser.outputs.issueparser_sample-data }}" >> $GITHUB_OUTPUT
      
      - name: Get Config
        id: get_model_config
        run: |
          echo "model_config=${{ steps.issue-parser.outputs.issueparser_modelConfig }}" >> $GITHUB_OUTPUT

      # Get svn url for exporting docker 🟢
      - name: Clone docker folder
        run: |
          url="${{ steps.get_docker.outputs.docker }}"
          
          svn_url=$(echo "$url" | sed -E 's|/tree/[^/]+|/trunk|; s|/blob/[^/]+|/trunk|')
          svn export --force $svn_url ./${{ steps.get_path.outputs.path }}/docker
      
      # Get svn url for exporting 🟢
      - name: Generate Python SVN URLs
        id: generate_urls
        run: |
          echo "pythons2<<EOF" >> $GITHUB_ENV
          python3 ./.github/workflows/getPythonScripts.py >> $GITHUB_ENV
          echo "EOF" >> $GITHUB_ENV
        env:
            pythons: ${{ env.pythons }}
      
      # Export the urls/clone the scripts 🟢
      - name: Clone python scripts
        run: |
          cd ./${{ steps.get_path.outputs.path }}
          svn_urls="${{ env.pythons2 }}"
          for svn_url in $svn_urls; do
            svn export --force $svn_url
          done
      
      # Get svn urls for exporting card and spec urls 🟢
      - name: Generate Model Info SVN URLs
        id: generate_model_info_urls
        run: |
          echo "model_info2<<EOF" >> $GITHUB_ENV
          python3 ./.github/workflows/getModelInfo.py >> $GITHUB_ENV
          echo "EOF" >> $GITHUB_ENV
        env:
            model_info: ${{ env.model_info }}

      # Export the urls/clone the model info 🟢
      - name: Clone model info
        run: |
          model_info="${{ env.model_info2 }}"
          for svn_url in $model_info; do
            svn export --force $svn_url ./${{ steps.get_path.outputs.path }}
          done

      # Clone config files
      - name: Config file clone
        # Only run if needs.create_new_branch.outputs.CONFIG is not empty
        if: ${{ steps.get_model_config.outputs.model_config != '' }}
        run: |
          mkdir ./${{ steps.get_path.outputs.path }}/config/
          url="${{ steps.get_model_config.outputs.model_config }}"
              
          svn_url=$(echo "$url" | sed -E 's|/tree/[^/]+|/trunk|; s|/blob/[^/]+|/trunk|')
    
          svn export --force $svn_url ./${{ steps.get_path.outputs.path }}/config
    
      - name: Validate model card and spec files
        run: |
            echo "## Model Card and Spec Validation :white_check_mark:" >> $GITHUB_STEP_SUMMARY
            echo "Model Card and spec yaml files are being validated here with LinkML Schema" >> $GITHUB_STEP_SUMMARY
            python3 ./.github/workflows/create_model_card_and_spec.py
        env:
            model_path: ${{ steps.get_path.outputs.path }}
        
      - name: Update yaml file
        run: |
          model_name=$(echo "${{ steps.get_path.outputs.path }}" | awk -F '/' '{print $(NF-1)}')
          python3 ./.github/workflows/update_yaml_info.py ${{ steps.get_path.outputs.path }}/docker $model_name
      
      - name: Set docker image name
        id: set_image_name
        run: |
          full_path="${{ steps.get_path.outputs.path }}"

          org_name=$(echo "$full_path" | awk -F '/' '{print $1}')
          model_name=$(echo "$full_path" | awk -F '/' '{print $2}')
          model_version=$(echo "$full_path" | awk -F '/' '{print $3}')
              
          echo "image_name=neuronets/$model_name" >> $GITHUB_OUTPUT
          
          echo "org_name=$org_name" >> $GITHUB_OUTPUT
          echo "model_name=$model_name" >> $GITHUB_OUTPUT

          echo "model_version=$model_version" >> $GITHUB_OUTPUT

      - name: Add weights
        env:
          OSF_TOKEN: ${{ secrets.OSF_TOKEN }}
        run: |
            datalad download-url -o --nosave "${{ steps.get_weights.outputs.weights }}" -O ./${{ steps.get_path.outputs.path }}/weights/
            datalad download-url -o --nosave "${{ steps.get_data.outputs.sample_data }}" -O ./${{ steps.get_path.outputs.path }}/example-data/

            # unzip files in place and delete after
            find . -name '*.zip' -not -path "./.git/*" -execdir unzip '{}' ';' -delete
      
      - name: print directory structure
        run: tree ${{ steps.set_image_name.outputs.org_name }}

      - name: Build Docker image
        run: |
          docker build -t ${{ steps.set_image_name.outputs.model_name }} ./${{ steps.get_path.outputs.path }}/docker

      - name: Convert docker image to singularity image
        run: |
          singularity build ../${{ steps.set_image_name.outputs.model_name }}.sif docker-daemon://${{ steps.set_image_name.outputs.model_name }}:latest

      - name: Collect Workflow Telemetry
        uses: runforesight/workflow-telemetry-action@v1
        with:
          job_summary: true
          proc_trace_sys_enable: true
          proc_trace_table_show: true
  
      - name: Run test command in Singularity
        id: singularity_test 
        run: |
          # Parent directory as a bind path in a env variable
          singularity exec --nv ../${{ steps.set_image_name.outputs.model_name }}.sif ${{ steps.get_test_command.outputs.test_command }}

      - name: Run test command in Docker
        id: docker_test
        run: |
          docker run --gpus all -v $(pwd):/output \
          -w /output \
          ${{ steps.set_image_name.outputs.model_name }} "${{ steps.get_test_command.outputs.test_command }}"
          
      - name: Push Docker image
        if: steps.docker_test.outcome == 'success' && steps.singularity_test.outcome == 'success'
        run: |
          docker login -u ${{ secrets.DOCKER_USERNAME }} -p ${{ secrets.DOCKER_PASSWORD }}
          docker tag ${{ steps.set_image_name.outputs.model_name }} ${{ steps.set_image_name.outputs.image_name }}:${{ steps.set_image_name.outputs.model_version }}
          docker push ${{ steps.set_image_name.outputs.image_name }}:${{ steps.set_image_name.outputs.model_version }}

      - name: Commit changes
        if: steps.docker_test.outcome == 'success' && steps.singularity_test.outcome == 'success'
        run: |
          datalad save -m "SUCCESS: Adding Model" ${{ steps.set_image_name.outputs.org_name }}
          datalad push -f all --to osf-annex1106-storage
          datalad push --to origin
        env:
          OSF_TOKEN: ${{ secrets.OSF_TOKEN }}

  stop-runner:
    name: Stop EC2 runner
    needs: [start-runner, create_new_branch] # required to wait when the main job is done
    runs-on: ubuntu-latest
    if: ${{ always() }}
    # required to stop the runner even if the error happened in the previous jobs
    steps:
      - name: Configure AWS credentials
        uses: aws-actions/configure-aws-credentials@v1
        with:
            aws-access-key-id: ${{ secrets.AWS_KEY_ID }}
            aws-secret-access-key: ${{ secrets.AWS_KEY_SECRET }}
            aws-region: ${{ vars.AWS_REGION }}
      - name: Stop EC2 runner
        uses: machulav/ec2-github-runner@v2
        with:
          mode: stop
          github-token: ${{ secrets.GH_TOKEN }}
          label: ${{ needs.start-runner.outputs.label }}
          ec2-instance-id: ${{ needs.start-runner.outputs.ec2-instance-id }}

  failed:
    runs-on: ubuntu-latest
    needs: [start-runner, create_new_branch]
    if: ${{ failure() }}
    steps:
      - name: Set labels
        uses: actions-cool/issues-helper@v3
        with:
          actions: 'set-labels'
          token: ${{ secrets.GITHUB_TOKEN }}
          issue-number: ${{ github.event.issue.number }}
          labels: 'failed'

      - name: Create comment
        uses: actions-cool/issues-helper@v3
        with:
          actions: 'create-comment'
          token: ${{ secrets.GITHUB_TOKEN }}
          issue-number: ${{ github.event.issue.number }}
          body: |
              [This is an automated message] Hello @${{ github.event.issue.user.login }},
              🔴 The folders/scripts you provided did not pass our tests. Please review the Action, and modify this issue's urls accordingly. 
              When ready, simply append "Ready XX" in the issue title (where XX is a number incrementing with 01 each time a fix has been applied).
          emoji: '+1,eyes'
        
  success:
    needs: [start-runner, create_new_branch, stop-runner]
    runs-on: ubuntu-latest
    steps:
      - name: Set labels
        uses: actions-cool/issues-helper@v3
        with:
          actions: 'set-labels'
          token: ${{ secrets.GITHUB_TOKEN }}
          issue-number: ${{ github.event.issue.number }}
          labels: 'success'
      
      - name: Create comment
        uses: actions-cool/issues-helper@v3
        with:
          actions: 'create-comment'
          token: ${{ secrets.GITHUB_TOKEN }}
          issue-number: ${{ github.event.issue.number }}
          body: |
              [This is an automated message] Hello @${{ github.event.issue.user.login }}. The workflow linked to adding your model finished successfully! Please double check that this issue's tag is "success."  
              🟢 A Draft PR should be linked to this issue. Now that your model passed the checks, feel free to change the status of the PR to "ready for review."
              ⭐ Thank you for adding a model to Nobrainer-Zoo!
          emoji: '+1,hooray,rocket'