Installation Guide

Prerequisites

Before installing Udility, ensure your system meets the following requirements:

• Python: Version 3.6 or higher.
• System Libraries: Since Udility relies on cairosvg for image processing, you must have the Cairo graphics library installed on your operating system.

Installing Cairo

• macOS: brew install cairo
• Linux (Ubuntu/Debian): sudo apt-get install libcairo2
• Windows: Cairo is typically bundled with the cairosvg Python wheel, but if you encounter errors, you may need to install the GTK+ stack.

Installation

Install the Udility package directly from PyPI using pip:

pip install Udility

Dependencies

The following Python libraries are automatically resolved and installed during the process:

• openai (API communication)
• cairosvg (SVG to PNG conversion)
• Pillow (Image handling)
• matplotlib (Image visualization)

Configuration

Udility Diffuser uses OpenRouter to access Meta Llama-3.5 models. You must provide a valid API key to authenticate requests.

1. Obtain an API Key

Go to OpenRouter to generate a free or paid API key.

2. Set Environment Variables

The library looks for an environment variable named OPENROUTER_API_KEY. You can set this in two ways:

Option A: Persistent (Recommended)

Add this to your shell profile (e.g., .bashrc or .zshrc):

export OPENROUTER_API_KEY='your_api_key_here'

Option B: Runtime (Python script)

Set the variable at the beginning of your script before calling any Udility functions:

import os

os.environ['OPENROUTER_API_KEY'] = 'your_api_key_here'

Verifying Installation

To ensure everything is set up correctly, run a quick test script to generate a simple illustration:

from Udility import diffuser

try:
    # This will generate an image, save it as 'test.png', and display it.
    diffuser.generate_image_from_text("A simple flowchart of a water cycle.")
    print("Installation successful!")
except Exception as e:
    print(f"Installation check failed: {e}")

If the installation is correct, a window will pop up displaying the generated SVG illustration, and an output.png file will be created in your working directory.