PR Reviewer Bot

An AI-powered code review assistant that automatically analyzes pull requests and provides feedback on code quality, potential bugs, security issues, and optimization opportunities.

Features

• Automated Code Reviews: Analyzes code changes in pull requests
• Multiple Analysis Types: Detects security vulnerabilities, style issues, bugs, and optimization opportunities
• AI-Powered Insights: Uses AI models to provide intelligent code feedback
• Flexible Output: Generates reports as Markdown files or GitHub PR comments
• Multiple AI Options: Supports commercial APIs (OpenAI, Claude) and self-hosted models (Ollama)

Project Structure

The project is organized as a monorepo with the following packages:

• agents: Code analysis agents that detect different types of issues
• ai-connectors: Connectors for various AI services (OpenAI, Claude, Ollama)
• output: Formatters for delivering review results (file output, GitHub PR comments)
• core: Core functionality and shared utilities
• cli: Command-line interface for running the bot

Installation

# Clone the repository
git clone https://github.com/shurajcodx/PRReviewerBot.git
cd PRReviewerBot

# Install dependencies
make install

# Build the project
make build

Development

Code Style and Linting

The project uses ESLint and Prettier to maintain code quality and consistent style:

# Run linting
make lint

# Fix linting issues
make lint-fix

# Format code
make format

# Check code formatting
make format-check

Setup

Using Commercial AI Services

1. Get API Keys:

   • OpenAI: Sign up at platform.openai.com
   • Claude: Sign up at anthropic.com

2. Configure the Bot: You can set environment variables when running the CLI:

   # For OpenAI
   AI_API_KEY=your-openai-api-key prreviewbot review --ai-model openai [options]
   
   # For Claude (default)
   AI_API_KEY=your-claude-api-key prreviewbot review [options]

Using Self-Hosted Ollama (Free)

1. Install Ollama:

   • Download and install from ollama.ai
   • Start the Ollama server

2. Pull a Model:

   # Pull a general-purpose model
   ollama pull llama3.2

3. Configure the Bot:

   # Use Ollama with a specific model
   prreviewbot review --ai-model ollama --ai-options '{"model":"llama3.2","baseUrl":"http://localhost:11434"}' [options]

GitHub Integration (Optional)

To post review comments directly on GitHub PRs:

1. Create a GitHub Personal Access Token:

   • Go to GitHub Settings → Developer Settings → Personal Access Tokens
   • Create a token with repo scope

2. Add to Environment Variables:

   prreviewbot review --url https://github.com/owner/repo --branch branch-name --token your-github-token

   Or add to your .env file:

   GIT_TOKEN=your-github-token
   

Usage

Basic Usage

To use the PR Reviewer Bot, run the CLI directly:

# Review a branch in a repository
prreviewbot review --url https://github.com/owner/repo --branch feature-branch

# Review with specific AI model
prreviewbot review --url https://github.com/owner/repo --branch feature-branch --ai-model openai --ai-key your-api-key

# Review with Ollama (free, self-hosted)
prreviewbot review --url https://github.com/owner/repo --branch feature-branch --ai-model ollama

# Output results to a specific file
prreviewbot review --url https://github.com/owner/repo --branch feature-branch --output-format file --output-file ./my-review.md

Command Options

The CLI supports the following options:

Options:
  -u, --url <url>               Git repository URL
  -t, --token <token>           Git access token
  -b, --branch <branch>         Branch name to review
  -o, --output-dir <directory>  Output directory for cloned repo (default: "./temp-repo")
  -a, --ai-key <key>            AI API key (or set AI_API_KEY env var)
  -m, --ai-model <model>        AI model to use (claude, openai, ollama) (default: "claude")
  --ai-options <json>           Additional options for the AI model as JSON string
  -f, --output-format <format>  Output format (file, pr-comment) (default: "file")
  --output-file <file>          Output file path (when using file format) (default: "./review-results.md")
  --agents <agents>             Comma-separated list of agents to use (all, security, style, bug, optimization) (default: "all")
  -h, --help                    display help for command

Examples

Example 1: Review a Branch

# Review a branch using Claude
AI_API_KEY=your-claude-key prreviewbot review --url https://github.com/myorg/myrepo --branch feature-branch

This will:

1. Clone the repository
2. Analyze the code in the branch
3. Generate a report in review-results.md

Example 2: Review with Specific Agents

# Review using only security and optimization agents with OpenAI
prreviewbot review --url https://github.com/myorg/myrepo --branch feature-branch --ai-model openai --ai-key your-openai-key --agents security,optimization

This will:

1. Clone the repository
2. Analyze the code using only security and optimization agents
3. Generate a report in review-results.md

Example 3: Review with Ollama (Free, Self-Hosted)

# Review using Ollama with CodeLlama model
prreviewbot review --url https://github.com/myorg/myrepo --branch feature-branch --ai-model ollama --ai-options '{"model":"codellama"}'

This will:

1. Clone the repository
2. Analyze the code using the local Ollama server with CodeLlama
3. Generate a report in review-results.md

Using as a GitHub Action

For GitHub Actions, set up a workflow using environment secrets:

name: Code Review

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
        with:
          fetch-depth: 0
      - name: Set up environment
        run: make install
      - name: Build project
        run: make build

      - name: Run PR Review
        run: prreviewbot review --url https://github.com/${{ github.repository }} --branch ${{ github.head_ref }} --output-format pr-comment
        env:
          GIT_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          AI_API_KEY: ${{ secrets.AI_API_KEY }}
          AI_MODEL: openai

Extending the Bot

Adding a Custom Agent

1. Create a new agent in packages/agents/src/YourCustomAgent.ts
2. Extend the BaseAgent class
3. Implement the analyze method
4. Export your agent in packages/agents/src/index.ts

Adding a Custom AI Connector

1. Create a new connector in packages/ai-connectors/src/connectors/YourConnector.ts
2. Extend the BaseAIConnector class
3. Implement the generateResponse method
4. Update the AIConnectorFactory to support your connector
5. Export your connector in packages/ai-connectors/src/index.ts

Troubleshooting

API Connection Issues

• Authentication Errors: Verify your API keys are correct
• Rate Limiting: Commercial APIs have rate limits; check your usage quotas

Ollama Issues

• Connection Errors: Make sure Ollama is running (ps aux | grep ollama)
• Model Not Found: Ensure you've pulled the model (ollama list)
• Slow Responses: Ollama performance depends on your hardware; consider using a smaller model

GitHub Integration Issues

• Authentication Errors: Verify your GitHub token has the correct permissions
• Rate Limiting: GitHub API has rate limits; consider using a PAT with higher limits

CLI Not Found

• Make sure you've built the project with make build
• Check that the CLI is properly installed and in your PATH
• Try running with the full command: prreviewbot review --help

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.