python-uml-generator 0.1.1

A LangChain-based AI agent that converts natural language to UML diagrams (PlantUML)

UML Generator AI Agent

A Python-based AI agent that converts natural language descriptions into UML diagrams using a configurable LLM (default: Gemini Pro). The output is PlantUML code, which is rendered as an image.

Features

• Configurable LLM backend (Gemini Pro, OpenAI, etc.)
• Converts human input to PlantUML code
• Renders PlantUML code to PNG image

Installation

1. Install the package from PyPI:

   pip install uml-generator
   

2. Set up your API keys in a .env file in your project directory:

   # Set the model to 'gemini-pro' or an OpenAI model like 'gpt-4'
   UML_AGENT_MODEL=gemini-pro
   
   # Add your API keys
   GEMINI_API_KEY=your_gemini_api_key
   OPENAI_API_KEY=your_openai_api_key
   

CLI Usage

Run the uml-generator command with your prompt.

To print the PlantUML code to the console:

uml-generator "A simple class diagram for a Dog"

To save the generated diagram to a file:

uml-generator "A simple class diagram for a Dog" -o dog_diagram.png

Library Usage

You can also use uml-generator as a library in your own Python code:

import uml_generator

# Generate PlantUML code and save the diagram to a file
prompt = "A class diagram for a simple shopping cart"
plantuml_code = uml_generator.generate(prompt, "shopping_cart.png")

print("--- Generated PlantUML Code ---")
print(plantuml_code)

# Or, generate the code without saving a file
prompt2 = "An activity diagram for making coffee"
plantuml_code_only = uml_generator.generate(prompt2)
print(plantuml_code_only)

Developer Setup

If you want to contribute to the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/milind-nair/uml-generator.git
   cd uml-generator
   

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`
   

3. Install the package in editable mode:

   pip install -e .
   

4. Run the tests:

   python -m unittest discover -s tests
   

Configuration

• Change the model by editing UML_AGENT_MODEL in your .env file.
• Supported models: gemini-pro, openai/gpt-4, etc.

License

MIT