Update dependency wouter to v3.4.0 #293
Workflow file for this run
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Review Pull Request with llama.cpp | |
| on: | |
| pull_request: | |
| types: [opened, synchronize, reopened] | |
| branches: ["main"] | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }} | |
| cancel-in-progress: true | |
| jobs: | |
| llama-cpp: | |
| if: ${{ !contains(github.event.pull_request.labels.*.name, 'skip-ai-review') }} | |
| continue-on-error: true | |
| runs-on: ubuntu-latest | |
| name: llama.cpp | |
| permissions: | |
| pull-requests: write | |
| contents: read | |
| timeout-minutes: 120 | |
| env: | |
| HF_MODEL_NAME: Qwen2.5.1-Coder-7B-Instruct-GGUF | |
| steps: | |
| - name: Checkout Repository | |
| uses: actions/checkout@v4 | |
| - name: Create temporary directory | |
| run: mkdir -p /tmp/llama_review | |
| - name: Process PR description | |
| id: process_pr | |
| run: | | |
| PR_BODY_ESCAPED=$(cat << 'EOF' | |
| ${{ github.event.pull_request.body }} | |
| EOF | |
| ) | |
| PROCESSED_BODY=$(echo "$PR_BODY_ESCAPED" | sed -E 's/\[(.*?)\]\(.*?\)/\1/g') | |
| echo "$PROCESSED_BODY" > /tmp/llama_review/processed_body.txt | |
| - name: Fetch branches and output the diff in this step | |
| run: | | |
| git fetch origin main:main | |
| git fetch origin pull/${{ github.event.pull_request.number }}/head:pr-branch | |
| git diff main..pr-branch > /tmp/llama_review/diff.txt | |
| - name: Write prompt to file | |
| id: build_prompt | |
| run: | | |
| PR_TITLE=$(echo "${{ github.event.pull_request.title }}" | sed 's/[()]/\\&/g') | |
| DIFF_CONTENT=$(cat /tmp/llama_review/diff.txt) | |
| PROCESSED_BODY=$(cat /tmp/llama_review/processed_body.txt) | |
| echo "<|im_start|>system | |
| You are an experienced developer reviewing a Pull Request. You focus only on what matters and provide concise, actionable feedback. | |
| Review Context: | |
| Repository Name: \"${{ github.event.repository.name }}\" | |
| Repository Description: \"${{ github.event.repository.description }}\" | |
| Branch: \"${{ github.event.pull_request.head.ref }}\" | |
| PR Title: \"$PR_TITLE\" | |
| Guidelines: | |
| 1. Only comment on issues that: | |
| - Could cause bugs or security issues | |
| - Significantly impact performance | |
| - Make the code harder to maintain | |
| - Violate critical best practices | |
| 2. For each issue: | |
| - Point to the specific line/file | |
| - Explain why it's a problem | |
| - Suggest a concrete fix | |
| 3. Praise exceptional solutions briefly, only if truly innovative | |
| 4. Skip commenting on: | |
| - Minor style issues | |
| - Obvious changes | |
| - Working code that could be marginally improved | |
| - Things that are just personal preference | |
| Remember: | |
| Less is more. If the code is good and working, just say so, with a short message.<|im_end|> | |
| <|im_start|>user | |
| This is the description of the pull request: | |
| \`\`\`markdown | |
| $PROCESSED_BODY | |
| \`\`\` | |
| And here is the diff of the changes, for you to review: | |
| \`\`\`diff | |
| $DIFF_CONTENT | |
| \`\`\` | |
| <|im_end|> | |
| <|im_start|>assistant | |
| ### Overall Summary | |
| " > /tmp/llama_review/prompt.txt | |
| - name: Show Prompt | |
| run: cat /tmp/llama_review/prompt.txt | |
| - name: Set up Homebrew | |
| uses: Homebrew/actions/setup-homebrew@master | |
| - name: Install and cache Homebrew tools | |
| uses: tecolicom/actions-use-homebrew-tools@v1 | |
| with: | |
| tools: llama.cpp | |
| - name: Cache the LLM | |
| id: cache_llama_cpp | |
| uses: actions/cache@v4 | |
| with: | |
| path: ~/.cache/llama.cpp/ | |
| key: llama-cpp-${{ env.HF_MODEL_NAME }} | |
| - name: Download and cache the LLM | |
| if: steps.cache_llama_cpp.outputs.cache-hit != 'true' | |
| run: | | |
| mkdir -p ~/.cache/llama.cpp/ | |
| curl -L -o ~/.cache/llama.cpp/model.gguf https://huggingface.co/bartowski/${{ env.HF_MODEL_NAME }}/resolve/main/Qwen2.5.1-Coder-7B-Instruct-IQ4_XS.gguf | |
| - name: Run llama.cpp | |
| run: | | |
| llama-server \ | |
| --model ~/.cache/llama.cpp/model.gguf \ | |
| --ctx-size 32768 \ | |
| --threads -1 \ | |
| --predict -1 \ | |
| --temp 0.5 \ | |
| --top-p 0.9 \ | |
| --min-p 0.1 \ | |
| --top-k 0 \ | |
| --cache-type-k q8_0 \ | |
| --cache-type-v q8_0 \ | |
| --flash-attn \ | |
| --no-warmup \ | |
| --port 11434 & | |
| - name: cURL llama-server to get the completion and timings | |
| run: | | |
| DATA=$(jq -n --arg prompt "$(cat /tmp/llama_review/prompt.txt)" '{"prompt": $prompt}') | |
| echo -e '### Review\n\n' > /tmp/llama_review/response.txt | |
| # Save the full response to a temporary file | |
| curl \ | |
| --silent \ | |
| --request POST \ | |
| --url http://localhost:11434/completion \ | |
| --header "Content-Type: application/json" \ | |
| --data "$DATA" > /tmp/llama_review/full_response.json | |
| # Extract and append content to response.txt | |
| jq -r '.content' /tmp/llama_review/full_response.json >> /tmp/llama_review/response.txt | |
| # Pretty print the timings information | |
| echo "=== Performance Metrics ===" | |
| jq -r '.timings | to_entries | .[] | "\(.key): \(.value)"' /tmp/llama_review/full_response.json | |
| - name: Show Response | |
| run: cat /tmp/llama_review/response.txt | |
| - name: Find Comment | |
| uses: peter-evans/find-comment@v3 | |
| id: find_comment | |
| with: | |
| issue-number: ${{ github.event.pull_request.number }} | |
| comment-author: "github-actions[bot]" | |
| body-includes: "### Review" | |
| - name: Post or Update PR Review | |
| uses: peter-evans/create-or-update-comment@v4 | |
| with: | |
| comment-id: ${{ steps.find_comment.outputs.comment-id }} | |
| issue-number: ${{ github.event.pull_request.number }} | |
| body-path: /tmp/llama_review/response.txt | |
| edit-mode: replace |