gpt-4o-mini-2024-07-18· OpenAI LMArena Human Preference 140K
Give me feedback on this github action. What can be improved? [code]
- User
Give me feedback on this github action. What can be improved?
name: Build and Push Docker Image on: workflow_dispatch: inputs: vllm_tag: description: 'vLLM version to use. Must be the name of a tag in https://github.com/vllm-project/vllm' required: true default: 'v0.8.2' image_tag: description: 'Tag for the resulting Docker image. Normally, the vLLM tag will be used, but this can be overridden. For releases, an override is not allowed. Please note that for the arm64 version, a suffix "-arm" will be appended to the tag.' required: false default: '' push_image: description: 'Push Docker image to registry' type: boolean required: false default: false is_release: description: 'Perform a release. I.e., push to the "intrafind" section instead of the "dev" section of the registry. Requires that this action is triggered on a Git tag.' type: boolean required: false default: false jobs: build-and-push: runs-on: ubuntu-24.04 permissions: contents: write env: BUILD_TAG: ${{ github.run_id }} steps: - name: Display parameters run: | echo "Current workflow commit: ${{ github.sha }}" echo "vLLM tag: ${{ github.event.inputs.vllm_tag }}" echo "Image tag: ${{ github.event.inputs.image_tag }}" echo "Push image: ${{ github.event.inputs.push_image }}" echo "Is release: ${{ github.event.inputs.is_release }}" - name: Releases must have a Git tag if: ${{ github.event.inputs.is_release == 'true' && !startsWith(github.ref, 'refs/tags/') }} run: | echo "Error: When release is requested, the commit must have a Git tag." exit 1 - name: Releases must not override the Docker tag if: ${{ github.event.inputs.is_release == 'true' && github.event.inputs.image_tag != '' }} run: | echo "Error: When release is requested, the image tag must not be overridden, but it has been set to '${{ github.event.inputs.image_tag }}'." exit 1 - name: Determine destination and tag for resulting Docker image run: | # tag if [ -z "${{ github.event.inputs.image_tag }}" ]; then echo "IMAGE_TAG=${{ github.event.inputs.vllm_tag }}" >> $GITHUB_ENV else echo "IMAGE_TAG=${{ github.event.inputs.image_tag }}" >> $GITHUB_ENV fi # destination REPO=${{ github.event.inputs.is_release == 'true' && 'intrafind' || 'dev' }} DESTINATION=${{ vars.IF_REGISTRY }}/${REPO}/vllm-cpu echo "DESTINATION=$DESTINATION" echo "DESTINATION=$DESTINATION" >> $GITHUB_ENV - name: Determine path of Dockerfiles based on vLLM version run: | # Set Dockerfile path according to vLLM version layout VER=${{ github.event.inputs.vllm_tag }} if [ "$(printf '%s\n' "v0.8.2" "$VER" | sort -V | head -n1)" != "$VER" ]; then echo "DOCKERFILE_PATH=vllm/docker/Dockerfile" >> $GITHUB_ENV else echo "DOCKERFILE_PATH=vllm/Dockerfile" >> $GITHUB_ENV fi - name: Log in to Docker Registry for pushing the image uses: docker/login-action@v3 if: ${{ github.event.inputs.push_image == 'true' }} with: username: ${{ secrets.IF_REGISTRY_USER }} password: ${{ secrets.IF_REGISTRY_PASSWORD }} registry: ${{ vars.IF_REGISTRY }} - name: Set up Docker Buildx uses: docker/setup-buildx-action@v3 - name: Checkout vLLM uses: actions/checkout@v4 with: repository: vllm-project/vllm ref: ${{ github.event.inputs.vllm_tag }} path: vllm - name: Modify CMake files to fix arm64 build run: | # Replace the condition that checks for MacOS with a constant TRUE # Otherwise the build will fail with "unrecognized command-line option ‘-mavx2’" because it tries to make use of the AVX2 instruction set # # Replace: # if (MACOSX_FOUND AND CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64") # set(APPLE_SILICON_FOUND TRUE) # with: # if (TRUE AND (CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64")) # set(APPLE_SILICON_FOUND TRUE) # sed -i 's/if (MACOSX_FOUND AND CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64")/if (TRUE AND (CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64"))/g' vllm/cmake/cpu_extension.cmake # Show the changed part to make it possible to validate it cat vllm/cmake/cpu_extension.cmake -n | grep "set(APPLE_SILICON_FOUND TRUE)" -C 10 - name: Build and Push Docker Image for amd64 uses: docker/build-push-action@v6 with: context: vllm file: ${{ env.DOCKERFILE_PATH }}.cpu platforms: linux/amd64 push: ${{ github.event.inputs.push_image == 'true' }} tags: | ${{ env.DESTINATION }}:${{ env.IMAGE_TAG }} labels: | org.opencontainers.image.title=${{ github.repository }} org.opencontainers.image.description=vLLM Docker Image for CPUs org.opencontainers.image.vendor=IntraFind Software AG org.opencontainers.image.source=${{ github.server_url }}/${{ github.repository }} org.opencontainers.image.revision=${{ github.sha }} org.opencontainers.image.version=${{ github.ref }} com.intrafind.vllmcpu.vllm.tag=${{ github.event.inputs.vllm_tag }} - name: Build and Push Docker Image for arm64 uses: docker/build-push-action@v6 with: context: vllm file: ${{ env.DOCKERFILE_PATH }}.arm platforms: linux/arm64 push: ${{ github.event.inputs.push_image == 'true' }} tags: | ${{ env.DESTINATION }}:${{ env.IMAGE_TAG }}-arm build-args: | CMAKE_VERBOSE_MAKEFILE=ON labels: | org.opencontainers.image.title=${{ github.repository }} org.opencontainers.image.description=vLLM Docker Image for CPUs org.opencontainers.image.vendor=IntraFind Software AG org.opencontainers.image.source=${{ github.server_url }}/${{ github.repository }} org.opencontainers.image.revision=${{ github.sha }} org.opencontainers.image.version=${{ github.ref }} com.intrafind.vllmcpu.vllm.tag=${{ github.event.inputs.vllm_tag }} - Assistant
Your GitHub Action for building and pushing Docker images is well-structured and covers a lot of important aspects. However, there are several areas where you could improve clarity, maintainability, and functionality. Here are some suggestions:
1. Input Validation
- Type Checking: You are using string comparisons for boolean inputs (
is_releaseandpush_image). Consider usingif: ${{ github.event.inputs.is_release }}directly, as GitHub Actions will treat it as a boolean. - Input Constraints: You could add more validation for the
vllm_taginput to ensure it matches the expected format or exists in the repository.
2. Error Handling
- Consistent Error Messages: Instead of using
echofollowed byexit 1, consider usingfailorerrorcommands to provide clearer feedback. This can help in debugging. - Use of
set -e: Consider addingset -eat the beginning of your shell scripts to ensure that the script exits immediately if any command fails.
3. Environment Variables
- Use of
GITHUB_ENV: You are usingecho "VAR=value" >> $GITHUB_ENVmultiple times. You could consolidate these into a single step to reduce redundancy and improve readability. - Variable Naming: Ensure that environment variable names are consistent and descriptive. For example,
DESTINATIONcould be renamed toDOCKER_IMAGE_DESTINATIONfor clarity.
4. Docker Build Steps
- Conditional Builds: If the
push_imageinput is false, you might want to skip the build steps entirely. This can save time and resources. - Use of BuildKit: Ensure that Docker BuildKit is enabled for better performance and caching. You can set
DOCKER_BUILDKIT=1in the environment.
5. Documentation and Comments
- Inline Comments: While you have some comments, consider adding more detailed comments explaining the purpose of each step, especially for complex logic.
- README Update: Ensure that the README or documentation for this action is updated to reflect any changes in inputs or behavior.
6. Versioning
- Action Versions: You are using specific versions for actions (e.g.,
docker/login-action@v3). Consider using the latest stable version or a specific commit SHA for better stability. - Docker Image Tags: Ensure that the tags used for Docker images are versioned appropriately to avoid confusion in the future.
7. Testing and Validation
- Test Cases: If possible, create a separate job to run tests on the built images to ensure they work as expected.
- Linting: Consider adding a step to lint your Dockerfiles to catch potential issues early.
8. Performance Optimization
- Caching: If your builds take a long time, consider using caching strategies for dependencies or layers to speed up the build process.
9. Security Considerations
- Secrets Management: Ensure that sensitive information (like Docker credentials) is handled securely and not exposed in logs.
- Docker Image Scanning: Consider adding a step to scan the built images for vulnerabilities.
10. Final Touches
- Consistent Formatting: Ensure consistent formatting throughout the YAML file for better readability.
- Use of
jobs.<job_id>.steps[*].if: Instead of usingifconditions in therunsteps, consider using theifattribute directly on the steps to make the logic clearer.
By implementing these suggestions, you can enhance the robustness, maintainability, and clarity of your GitHub Action.
- Type Checking: You are using string comparisons for boolean inputs (
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.