py-cogit 0.1.1

A command-line tool for code analysis.

py-cogit

Your AI senior software engineer, right in your terminal.

PyPI Link

py-cogit is a command-line tool that acts as your personal code reviewer. It analyzes your staged changes (git diff --staged), sends them to Google's AI (Gemini), and provides a detailed analysis on best practices, potential bugs, readability, and even suggests a commit message for you.

Stop committing with fix: stuff. Start committing with confidence.

✨ Key Features

• 🤖 AI-Powered Code Review: Get insights and suggestions from an AI model trained to act as a senior engineer.
• 💡 Structured Analysis: The output is in JSON format, detailing improvement points, potential risks, and observations per file.
• commit Automated Commit Messages: Generates a commit message suggestion following the Conventional Commits standard.
• ⚙️ Highly Configurable: Use the default prompt or completely customize it via a .env file.
• 🔒 Secure and Local: Your code is sent directly to the Google API. Your API key remains safe on your local machine.

🚀 Installation

Installation is easily done via pip.

pip install py-cogit

🛠️ Setup

Before using py-cogit for the first time, you need to set up your Google API key.

1. Get Your API Key: Go to Google AI Studio and create your API key.

2. Create a .env file: In the root of your project (or any parent directory from where you'll run the command), create a file named .env and add the following variables.

   # .env
   
   # Your API key from Google AI Studio
   GOOGLE_API_KEY="PASTE_YOUR_API_KEY_HERE"
   
   # The Gemini model you want to use
   GEMINI_VERSION="gemini-1.5-pro-latest"
   
   # (Optional) You can override the default prompt here
   PROMPT="You are an AI assistant..."
   

Usage

The workflow is simple and integrates seamlessly with Git.

1. Make your code changes as usual.

2. Add your changes to the Git staging area.

   git add .
   

3. Run py-cogit!

   py-cogit
   

📋 Example Output

After running, py-cogit will print a detailed JSON analysis directly to your terminal:

{
  "overall_summary": "This change refactors the script to make the Gemini model version configurable and improves exception handling.",
  "analysis_per_file": [
    {
      "file_name": "main.py",
      "general_observations": [
        "Externalizing the model version to an environment variable is a great practice."
      ],
      "improvement_points": [
        {
          "approximate_line": "22",
          "code_snippet": "model = genai.GenerativeModel(os.getenv(\"GEMINI_VERSION\"))",
          "suggestion": "`os.getenv` will return `None` if the variable is not set, causing an error. Add validation to ensure the variable exists or provide a default value.",
          "category": "Potential Bug",
          "priority": "High"
        }
      ]
    }
  ],
  "potential_risks": [
    "The application now depends on the `GEMINI_VERSION` environment variable. If it is not set, the program will crash."
  ],
  "commit_message_suggestion": {
    "type": "refactor",
    "scope": "main",
    "subject": "configure model version via environment variable"
  }
}

Contributing

Contributions are very welcome! If you have ideas for new features, improvements, or have found a bug, please open an Issue so we can discuss it.

📜 License

This project is licensed under the MIT License. See the LICENSE file for more details.