rawa-bots 0.2.4

A Python library to easily build Gemini-powered chatbots and generate images.

rawa_bots

A Python library to easily build Gemini-powered chatbots and generate images with minimal code.

Features

• Easy Gemini API integration
• Conversational and single-turn modes
• Automatic conversation history
• Custom generation parameters (temperature, top_p, etc.)
• Graceful error handling with custom exceptions
• Image generation support (with save and preview options)
• Ready for PyPI packaging

Installation

pip install rawa_bots

Note: Requires Python 3.8+ and a valid Gemini API key. The google-genai package will be installed automatically.

Usage Example

from rawa_bots import GeminiBot

bot = GeminiBot(
    api_key="YOUR_API_KEY",
    # model_name is optional; defaults to 'gemini-2.0-flash' for text, 'gemini-2.0-flash-preview-image-generation' for images
    system_prompt="You are helpful.",
    conversational=True,
    allow_image_gen=True  # Enable image generation
)

# Text chat
print(bot.ask("Hello!"))

# Image generation (returns URL or file path)
img_path = bot.generate_image(
    "A dragon flying over a castle at night",
    output_path="dragon.png",  # Save to file (optional)
    show=True                   # Open image after saving (optional)
)
print("Image saved to:", img_path)

Image Generation API

• allow_image_gen: Set to True to enable image generation.
• generate_image(prompt, n=1, output_path=None, show=False)
  • prompt: The image prompt.
  • n: Number of images to generate.
  • output_path: If set, saves the image(s) to this path (if a directory, images are saved inside it; if a filename, image is saved at that path; if not set, images are saved in the current working directory as 'gemini_image.png', 'gemini_image_1.png', etc.).
  • show: If True, opens the image(s) after saving.
  • Returns: file path(s) or URL(s) if available.

By default, images are saved in the current working directory as gemini_image.png (or gemini_image_1.png, etc. for multiple images), unless you specify a different path or directory.

Versioning

This library uses a single version number for all files and features. There is no need to specify or manage multiple versions.

Listing Available Models

You can list all available Gemini models with:

from google import genai
client = genai.Client(api_key="YOUR_API_KEY")
print([m.name for m in client.models.list()])

Error Handling

• GeminiError: Base exception for all GeminiBot errors
• AuthenticationError: Raised for authentication issues
• InvalidModelError: Raised if the model name is invalid
• ImageGenNotAllowedError: Raised if image generation is not enabled
• ImageGenAPIError: Raised for image generation API errors

Testing

Run tests with:

python -m unittest discover tests

License

MIT License