autodoc-ai-paudelnirajan 0.1.1

An AI-powered tool to automatically document, refactor, and format Python code.

AutoDoc AI

AutoDoc AI is a smart command-line developer tool that automates the tedious parts of maintaining high-quality Python code. It leverages Large Language Models (LLMs) to automatically generate docstrings, suggest intelligent refactorings, and integrate seamlessly into a modern development workflow.

This project was built on a foundation of solid Object-Oriented Programming principles and design patterns, including the Visitor, Strategy, Factory, and Adapter patterns.

---

Core Features

• AI-Powered Docstring Generation: Automatically create high-quality, context-aware docstrings for any function or class that's missing them.
• Intelligent Docstring Correction: Uses AI to evaluate existing docstrings and replaces poor-quality or placeholder documentation with new, improved versions.
• AI-Powered Refactoring:
  • Safe Variable Renaming: Automatically renames poorly named local variables within function scope.
  • Intelligent Naming Suggestions: Acts as an AI-powered linter, suggesting better names for classes and functions that are too short or non-descriptive.
• Seamless Git Integration: Use the --diff flag to only process files that have been changed in your current Git branch, making it incredibly fast and efficient for pre-commit hooks and CI pipelines.
• Codebase-Wide Scanning: Run autodoc on a single file or an entire directory. It automatically finds all Python files and respects your .gitignore to avoid processing virtual environments or build artifacts.
• Automatic Formatting: Integrates with the black code formatter to ensure that all generated and refactored code is perfectly styled.
• Highly Configurable: Set project-wide defaults for style, AI provider, and behavior in your pyproject.toml file.

Installation

You can install AutoDoc AI directly from PyPI:

pip install autodoc-ai-paudelnirajan

Configuration

1. Set Up Your AI Provider

AutoDoc requires an API key from an LLM provider. This project is configured to use Groq.

1. Create a .env file in the root of your project.

2. Add your Groq API key and desired model to the file:

   GROQ_API_KEY="gsk_YourActualGroqApiKeyHere"
   GROQ_MODEL_NAME="llama3-8b-8192"
   

2. Configure Project Defaults (Optional)

You can set default behaviors for your project in your pyproject.toml file. AutoDoc will use these settings unless they are overridden by a command-line flag.

# In your pyproject.toml

[tool.autodoc]
# Settings for your AutoDoc tool
strategy = "groq"
style = "google"           # 'google', 'numpy', or 'rst'
overwrite_existing = true  # Regenerate poor-quality docstrings
refactor = true            # Enable AI-powered refactoring

Usage

AutoDoc is a flexible command-line tool.

Get help:

autodoc --help

Run on a specific file (dry run, prints to console):

autodoc path/to/your/file.py

Run on an entire directory and save changes in-place:

autodoc src/ --in-place

Run in "Git mode" to only process changed files (most common use case):

autodoc --diff --in-place

Enable all AI features to perform a full quality pass:

autodoc . --in-place --overwrite-existing --refactor

How It Works

AutoDoc uses a "Surgical Hybrid" architecture that combines the strengths of traditional and modern tooling:

1. Fast Local Analysis: It uses Python's ast (Abstract Syntax Tree) module to rapidly parse code and identify potential issues (like a missing docstring or a short variable name). This is done locally and is extremely fast.
2. Targeted AI Intelligence: Only when a potential issue is found does it send the small, relevant code snippet to an LLM for intelligent analysis (e.g., "Is this a good name?" or "Generate a docstring for this function").
3. Precise Code Modification: The AI's response is used to perform a safe and precise modification of the AST, which is then unparsed back into valid Python code.
4. Toolchain Integration: Finally, it uses black to ensure the final code is perfectly formatted, integrating seamlessly into the existing Python ecosystem.

License

This project is licensed under the MIT License.