REVCD (Review Code)

A command-line utility to list and review git commit changes, built with Node.js and TypeScript. This tool uses OpenAI-compatible APIs to provide instant feedback on your code changes, thus helping you improve your code quality. The generated feedback is displayed in the terminal and saved to a file named REVIEW.md.

REVCD supports any OpenAI-compatible API endpoint, including OpenAI, Ollama, LM Studio, and other compatible providers. You can configure your preferred model and API endpoint through environment variables.

Features

• Commit Message Generation: Automatically suggests descriptive commit messages based on your code changes
• AI-Powered Code Reviews: Get instant feedback on your code changes using any OpenAI-compatible API. Reviews include:
  • Mermaid sequence diagram if the code is deemed complex enough
  • Code quality assessment
  • Performance optimization suggestions
  • Security vulnerability detection
  • Best practices recommendations
  • Lists code smells
  • Improved code alternatives
• Selective Review: Choose to review specific files, line ranges, or only changed parts (hunks)
• Staged & Unstaged Changes: Review both staged and unstaged changes in your git repository
• Language Support: Built-in support for TypeScript, JavaScript, Python, Go, and configurable for other languages
• Simple CLI Interface: Easy-to-use command-line interface with intuitive options
• Flexible API Support: Works with OpenAI, Ollama, LM Studio, and any OpenAI-compatible endpoint
• Customizable: Configure AI parameters, API endpoint, and code review settings through environment variables

Prerequisites

• Node.js (v16 or higher)
• An OpenAI-compatible API endpoint (e.g., OpenAI, Ollama, LM Studio, or other compatible services)

Setting up an LLM Provider

Option 1: OpenAI

• Sign up at OpenAI
• Get your API key from the dashboard
• Set LLM_API_BASE_URL=https://api.openai.com/v1 and LLM_API_KEY=your-api-key

Option 2: Ollama (Local)

Using Docker:

docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama
docker exec -it ollama ollama run codellama:7b

Using Homebrew:

brew install ollama
brew services start ollama
ollama run codellama:7b

Option 3: LM Studio (Local)

• Download from LM Studio
• Load a model and start the local server
• Set LLM_API_BASE_URL=http://localhost:1234/v1

Installation

From npm (recommended)

You can install the package directly from npm:

# Install globally
npm install -g revcd

# Or with yarn
yarn global add revcd

From Source

After cloning the repository, you can install the CLI globally on your system:

# Install dependencies
yarn install

# Build the project
yarn build

# Link the package globally
yarn link

Usage

Once installed, you can use the revcd command from anywhere in your terminal.

Available Commands

• Help: Display available commands

  revcd --help

• Review: Review code files and changes

  revcd review [options]

  Options:

  • -f, --file <file>: Review only a specific file
  • -l, --lines <range>: Review only a specific line range in format L:n-m (e.g., L:10-20) (must be used with --file)
  • -h, --hunks-only: Review only the added hunks/lines instead of entire files
  • -s, --staged: Review staged changes instead of unstaged changes (works with --hunks-only)
  • -a, --all: Review all files, not just code files
  • -m, --model <model>: Specify the model to use for review

• Hunks: Show only the changed parts (hunks) of unstaged/uncommitted files

  revcd hunks [options]

  Options:

  • -f, --file <file>: Show hunks for a specific file only
  • -s, --staged: Show hunks for staged changes instead of unstaged

• Generate Env: Copy the example environment file to your current directory

  revcd genv [options]

  Options:

  • -f, --force: Overwrite existing .env file if it exists

Environment Variables

The application uses environment variables for configuration. You can set these in a .env file in the root directory or as system environment variables.

LLM Configuration

# API base URL (supports any OpenAI-compatible endpoint)
# Examples:
#   - OpenAI: https://api.openai.com/v1
#   - Ollama: http://localhost:11434
#   - LM Studio: http://localhost:1234/v1
LLM_API_BASE_URL=http://localhost:11434

# API key (required for some providers like OpenAI, optional for local providers)
LLM_API_KEY=

# Model configurations
LLM_DEFAULT_MODEL=codellama:7b

# Generation parameters
LLM_DEFAULT_TEMPERATURE=0.7
LLM_DEFAULT_MAX_TOKENS=2048
LLM_REQUEST_TIMEOUT=60000

# Rate limiting (milliseconds between requests)
LLM_MIN_REQUEST_INTERVAL=500

# Code review parameters
CODE_FILE_EXTENSIONS=ts,js,py,go

A .env.example file is provided as a template.

Development

Prerequisites

• Node.js (v16 or higher)
• An OpenAI-compatible API endpoint
• Yarn

Setup

# Install dependencies
yarn install

# Build the project
yarn build

# Run in development mode
yarn dev

Project Structure

• src/index.ts: Main entry point for the CLI
• src/services/: Service modules
  • fileService.ts: File operations
  • gitService.ts: Git operations for retrieving commit changes
  • llmService.ts: LLM API operations for code review
  • index.ts: Service exports
• src/config/: Configuration modules
  • codeReview.config.ts: Code review settings
  • llm.config.ts: LLM API configuration
  • index.ts: Configuration exports
• src/utils/: Utility modules
  • languageMap.ts: Programming language detection and mapping
  • index.ts: Utility exports
• src/tests/: Test modules
  • languageMap.test.ts: Tests for language mapping functionality

License

ISC

Author

Akeel Nazir