Code Review Agent

• Application that analyzes GitHub pull requests. It leverages Celery for asynchronous task processing, Redis as a message broker, and Currently integrated with GROQ with LLM API for code analysis.

---

Features Implemented

• Asynchronous processing of code review tasks with Celery
• Integration with GitHub to fetch pull request details
• GROQ LLM API integration for code review analysis
• Result Caching
• Rate limiting
• Structured logging
• Multi-language support

Project Setup

Prerequisites

• Python: 3.10
• Redis: For task management and caching
• GitHub API Token: Required for accessing private GitHub repositories
• LLM API Key: Required for integrating with the language model (e.g., Groq API)

1. Clone the Repository & Create a Virtual Env

2. Install Dependencies

     pip install -r requirements.txt

2. Set Up Environment Variables

     LLM_API_KEY=llm_api_key
     MODEL_NAME=MODEL_NAME
     REDIS_URL=redis://localhost:6379/0
     BROKER_URL=redis://localhost:6379/0
     BACKEND_URL=redis://localhost:6379/0

3. Run the FastAPI Server

     uvicorn main:app --reload   or   python main.py
     The API will be available at http://localhost:8000.

4. Run Celery Wroker

     celery -A celery_app.celery_app worker --loglevel=info

---

API Documentation

• Endpoints

  • POST /analyze-pr: Initiates a code review task for a specified GitHub pull request.

    • Request Body:
      • repo_url (str): URL of the GitHub repository.
      • pr_number (str): Pull request number.
      • github_token (str, optional): GitHub API token for authentication.
    • Response:
      • Returns a task ID for the initiated review task.

  • GET /status/{task_id}: Checks the status of a code review task.

    • Path Parameter:
      • task_id (str): ID of the task.
    • Response: Status of the task (pending, completed, or failed).

  • GET /results/{task_id}: Retrieves the results of a completed code review task.

---

Design Decisions

• Asynchronous Processing with Celery: Using Celery enables handling long-running tasks (like fetching and analyzing PRs) asynchronously, improving API responsiveness and enabling efficient background processing.
• Redis for Task Management and Caching: Redis serves as both the Celery message broker and a caching layer for task results.
• Structured Logging: JSON-based structured logging is implemented to make logs machine-readable and easier to parse for monitoring or debugging purposes. This helps maintain high observability.
• Rate Limiting with SlowAPI: SlowAPI middleware is used to prevent abuse by limiting the number of requests per minute per client IP. This helps in managing load and preventing excessive use of resources.

---

Future Improvements

• Database for Persistent Storage: Currently, Redis is used for caching task results. Adding a database (e.g., PostgreSQL) could provide persistent storage for task history and user interactions. Authentication and Authorization:

• Authentication (e.g., JWT tokens) to protect certain endpoints, especially if deployed in a production environment where users may need restricted access.

• Webhooks for Real-Time PR Analysis: Implement GitHub webhooks to trigger analysis automatically when a new pull request is created, updated, or merged, providing real-time insights.