action.yml

name: 'gpuci - GPU Kernel Testing'
description: 'Test CUDA kernels across multiple GPUs via SSH'
author: 'RightNow AI'
branding:
  icon: 'cpu'
  color: 'green'

inputs:
  kernel:
    description: 'Path to CUDA kernel file(s). Supports glob patterns.'
    required: true
  config:
    description: 'Path to gpuci.yml config file'
    required: false
    default: 'gpuci.yml'
  target:
    description: 'Filter to specific target(s) by name or GPU type'
    required: false
  runs:
    description: 'Number of benchmark iterations'
    required: false
    default: '10'
  warmup:
    description: 'Number of warmup iterations'
    required: false
    default: '3'
  ssh-key:
    description: 'SSH private key for connecting to GPU machines'
    required: false
  brev-token:
    description: 'Brev.dev API token (if using Brev provider)'
    required: false
  post-comment:
    description: 'Post results as PR comment (true/false)'
    required: false
    default: 'true'
  fail-on-error:
    description: 'Fail the workflow if any GPU test fails'
    required: false
    default: 'true'

outputs:
  results:
    description: 'JSON results from gpuci'
    value: ${{ steps.run-gpuci.outputs.results }}
  passed:
    description: 'Whether all tests passed (true/false)'
    value: ${{ steps.run-gpuci.outputs.passed }}

runs:
  using: 'composite'
  steps:
    - name: Set up Python
      uses: actions/setup-python@v5
      with:
        python-version: '3.11'

    - name: Install gpuci
      shell: bash
      run: |
        pip install gpuci
        gpuci --version

    - name: Setup SSH key
      if: inputs.ssh-key != ''
      shell: bash
      run: |
        mkdir -p ~/.ssh
        echo "${{ inputs.ssh-key }}" > ~/.ssh/gpuci_key
        chmod 600 ~/.ssh/gpuci_key
        # Add to ssh-agent
        eval "$(ssh-agent -s)"
        ssh-add ~/.ssh/gpuci_key

    - name: Setup Brev CLI
      if: inputs.brev-token != ''
      shell: bash
      run: |
        curl -fsSL https://raw.githubusercontent.com/brevdev/brev-cli/main/bin/install-latest.sh | bash
        echo "${{ inputs.brev-token }}" | brev login --token

    - name: Run gpuci tests
      id: run-gpuci
      shell: bash
      run: |
        set +e  # Don't exit on error, we handle it

        # Build command
        CMD="gpuci test ${{ inputs.kernel }}"
        CMD="$CMD --config ${{ inputs.config }}"
        CMD="$CMD --runs ${{ inputs.runs }}"
        CMD="$CMD --warmup ${{ inputs.warmup }}"

        if [ -n "${{ inputs.target }}" ]; then
          CMD="$CMD --target ${{ inputs.target }}"
        fi

        # Run and capture output
        OUTPUT=$(eval $CMD 2>&1)
        EXIT_CODE=$?

        echo "$OUTPUT"

        # Save output for comment
        echo "results<<EOF" >> $GITHUB_OUTPUT
        echo "$OUTPUT" >> $GITHUB_OUTPUT
        echo "EOF" >> $GITHUB_OUTPUT

        if [ $EXIT_CODE -eq 0 ]; then
          echo "passed=true" >> $GITHUB_OUTPUT
        else
          echo "passed=false" >> $GITHUB_OUTPUT
        fi

        # Fail if configured to do so
        if [ "${{ inputs.fail-on-error }}" = "true" ] && [ $EXIT_CODE -ne 0 ]; then
          exit $EXIT_CODE
        fi

    - name: Post PR comment
      if: inputs.post-comment == 'true' && github.event_name == 'pull_request'
      uses: actions/github-script@v7
      with:
        script: |
          const results = `${{ steps.run-gpuci.outputs.results }}`;
          const passed = '${{ steps.run-gpuci.outputs.passed }}' === 'true';

          const status = passed ? '✅ All GPU tests passed' : '❌ Some GPU tests failed';

          const body = `## GPU Kernel Test Results

          ${status}

          <details>
          <summary>Full Results</summary>

          \`\`\`
          ${results}
          \`\`\`

          </details>

          ---
          *Tested with [gpuci](https://github.com/rightnow-ai/gpuci)*`;

          // Find existing comment
          const { data: comments } = await github.rest.issues.listComments({
            owner: context.repo.owner,
            repo: context.repo.repo,
            issue_number: context.issue.number,
          });

          const botComment = comments.find(c =>
            c.user.type === 'Bot' && c.body.includes('GPU Kernel Test Results')
          );

          if (botComment) {
            // Update existing comment
            await github.rest.issues.updateComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              comment_id: botComment.id,
              body: body
            });
          } else {
            // Create new comment
            await github.rest.issues.createComment({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: context.issue.number,
              body: body
            });
          }

    - name: Cleanup SSH key
      if: always() && inputs.ssh-key != ''
      shell: bash
      run: |
        rm -f ~/.ssh/gpuci_key