llama-stack/.github/workflows/publish-to-test-pypi.yml

name: Publish Python 🐍 distribution 📦 to TestPyPI

on:
  workflow_dispatch:  # Keep manual trigger
    inputs:
      version:
        description: 'Version number (e.g. 0.0.63.dev20250111)'
        required: true
        type: string
  schedule:
    - cron: "0 0 * * *"  # Run every day at midnight

jobs:
  trigger-client-and-models-build:
    name: Trigger llama-stack-client and llama-models build
    runs-on: ubuntu-latest
    outputs:
      version: ${{ steps.version.outputs.version }}
      client_run_id: ${{ steps.trigger-client.outputs.workflow_id }}
      model_run_id: ${{ steps.trigger-models.outputs.workflow_id }}
    steps:
    - uses: actions/checkout@v4
      with:
        persist-credentials: false
    - name: Get date
      id: date
      run: echo "date=$(date +'%Y%m%d')" >> $GITHUB_OUTPUT
    - name: Compute version based on dispatch event
      id: version
      run: |
        # Read base version from pyproject.toml
        version=$(sed -n 's/.*version="\([^"]*\)".*/\1/p' setup.py)
        if [ "${{ github.event_name }}" = "schedule" ]; then
          echo "version=${version}.dev${{ steps.date.outputs.date }}" >> $GITHUB_OUTPUT
        elif [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
          echo "version=${{ inputs.version }}" >> $GITHUB_OUTPUT
        else
          echo "version=${version}.dev$(shuf -i 10000000-99999999 -n 1)" >> $GITHUB_OUTPUT
        fi
    - name: Trigger llama-stack-client workflow
      id: trigger-client
      run: |
        response=$(curl -X POST https://api.github.com/repos/meta-llama/llama-stack-client-python/dispatches \
        -H 'Accept: application/vnd.github.everest-preview+json' \
        -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
        --data "{\"event_type\": \"build-client-package\", \"client_payload\": {\"source\": \"llama-stack-nightly\", \"version\": \"${{ steps.version.outputs.version }}\"}}" \
        -w "\n%{http_code}")

        http_code=$(echo "$response" | tail -n1)
        if [ "$http_code" != "204" ]; then
          echo "Failed to trigger client workflow"
          exit 1
        fi

        # Get the run ID of the triggered workflow
        sleep 5  # Wait for workflow to be created
        run_id=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                 "https://api.github.com/repos/meta-llama/llama-stack-client-python/actions/runs?event=repository_dispatch" \
                 | jq '.workflow_runs[0].id')
        echo "workflow_id=$run_id" >> $GITHUB_OUTPUT

    - name: Trigger llama-models workflow
      id: trigger-models
      run: |
        response=$(curl -X POST https://api.github.com/repos/meta-llama/llama-models/dispatches \
        -H 'Accept: application/vnd.github.everest-preview+json' \
        -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
        --data "{\"event_type\": \"build-models-package\", \"client_payload\": {\"source\": \"llama-stack-nightly\", \"version\": \"${{ steps.version.outputs.version }}\"}}" \
        -w "\n%{http_code}")

        http_code=$(echo "$response" | tail -n1)
        if [ "$http_code" != "204" ]; then
          echo "Failed to trigger models workflow"
          exit 1
        fi

        # Get the run ID of the triggered workflow
        sleep 5  # Wait for workflow to be created
        run_id=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                 "https://api.github.com/repos/meta-llama/llama-models/actions/runs?event=repository_dispatch" \
                 | jq '.workflow_runs[0].id')
        echo "workflow_id=$run_id" >> $GITHUB_OUTPUT

  wait-for-workflows:
    name: Wait for triggered workflows
    needs: trigger-client-and-models-build
    runs-on: ubuntu-latest
    steps:
    - name: Wait for client workflow
      run: |
        while true; do
          status=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                   "https://api.github.com/repos/meta-llama/llama-stack-client-python/actions/runs/${{ needs.trigger-client-and-models-build.outputs.client_run_id }}" \
                   | jq -r '.status')
          conclusion=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                      "https://api.github.com/repos/meta-llama/llama-stack-client-python/actions/runs/${{ needs.trigger-client-and-models-build.outputs.client_run_id }}" \
                      | jq -r '.conclusion')

          echo "llama-stack-client-python workflow status: $status, conclusion: $conclusion"

          if [ "$status" = "completed" ]; then
            if [ "$conclusion" != "success" ]; then
              echo "llama-stack-client-python workflow failed"
              exit 1
            fi
            break
          fi

          sleep 10
        done

    - name: Wait for models workflow
      run: |
        while true; do
          status=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                   "https://api.github.com/repos/meta-llama/llama-models/actions/runs/${{ needs.trigger-client-and-models-build.outputs.model_run_id }}" \
                   | jq -r '.status')
          conclusion=$(curl -s -H "authorization: Bearer ${{ secrets.PAT_TOKEN }}" \
                      "https://api.github.com/repos/meta-llama/llama-models/actions/runs/${{ needs.trigger-client-and-models-build.outputs.model_run_id }}" \
                      | jq -r '.conclusion')

          echo "llama-models workflow status: $status, conclusion: $conclusion"

          if [ "$status" = "completed" ]; then
            if [ "$conclusion" != "success" ]; then
              echo "llama-models workflow failed"
              exit 1
            fi
            break
          fi

          sleep 10
        done

  build:
    name: Build distribution 📦
    needs:
      - wait-for-workflows
      - trigger-client-and-models-build
    runs-on: ubuntu-latest

    steps:
    - uses: actions/checkout@v4
      with:
        persist-credentials: false
    - name: Get date
      id: date
      run: echo "date=$(date +'%Y%m%d')" >> $GITHUB_OUTPUT
    - name: Update version for nightly
      run: |
        sed -i 's/version="\([^"]*\)"/version="${{ needs.trigger-client-and-models-build.outputs.version }}"/' setup.py
        sed -i 's/llama-stack-client>=\([^"]*\)/llama-stack-client==${{ needs.trigger-client-and-models-build.outputs.version }}/' requirements.txt
        sed -i 's/llama-models>=\([^"]*\)/llama-models==${{ needs.trigger-client-and-models-build.outputs.version }}/' requirements.txt
    - name: Set up Python
      uses: actions/setup-python@v5
      with:
        python-version: "3.11"
    - name: Install pypa/build
      run: >-
        python3 -m
        pip install
        build
        --user
    - name: Build a binary wheel and a source tarball
      run: python3 -m build
    - name: Store the distribution packages
      uses: actions/upload-artifact@v4
      with:
        name: python-package-distributions
        path: dist/

  publish-to-testpypi:
    name: Publish Python 🐍 distribution 📦 to TestPyPI
    needs:
    - build
    runs-on: ubuntu-latest

    environment:
      name: testrelease
      url: https://test.pypi.org/p/llama-stack

    permissions:
      id-token: write  # IMPORTANT: mandatory for trusted publishing

    steps:
    - name: Download all the dists
      uses: actions/download-artifact@v4
      with:
        name: python-package-distributions
        path: dist/
    - name: Publish distribution 📦 to TestPyPI
      uses: pypa/gh-action-pypi-publish@release/v1
      with:
        repository-url: https://test.pypi.org/legacy/

  test-published-package:
    name: Test published package
    needs:
      - publish-to-testpypi
      - trigger-client-and-models-build
    runs-on: ubuntu-latest
    env:
      TOGETHER_API_KEY: ${{ secrets.TOGETHER_API_KEY }}
      TAVILY_SEARCH_API_KEY: ${{ secrets.TAVILY_SEARCH_API_KEY }}
    steps:
    - uses: actions/checkout@v4
      with:
        persist-credentials: false
    - name: Install the package
      run: |
        max_attempts=6
        attempt=1
        while [ $attempt -le $max_attempts ]; do
          echo "Attempt $attempt of $max_attempts to install package..."
          if pip install --no-cache --index-url https://pypi.org/simple/ --extra-index-url https://test.pypi.org/simple/ llama-stack==${{ needs.trigger-client-and-models-build.outputs.version }}; then
            echo "Package installed successfully"
            break
          fi
          if [ $attempt -ge $max_attempts ]; then
            echo "Failed to install package after $max_attempts attempts"
            exit 1
          fi
          attempt=$((attempt + 1))
          sleep 10
        done
    - name: Test the package versions
      run: |
        pip list | grep llama_
    - name: Test CLI commands
      run: |
        llama model list
        llama stack build --list-templates
        llama model prompt-format -m Llama3.2-11B-Vision-Instruct
        llama stack list-apis
        llama stack list-providers inference
        llama stack list-providers telemetry
    - name: Test Notebook
      run: |
        pip install pytest nbval
        llama stack build --template together --image-type venv
        pytest -v -s --nbval-lax ./docs/getting_started.ipynb
        pytest -v -s --nbval-lax ./docs/notebooks/Llama_Stack_Benchmark_Evals.ipynb

    # TODO: add trigger for integration test workflow & docker builds