ez-prompt 0.1.0

An easy-to-use library for creating and seinding prompts to LLMs.

ezprompt

An easy-to-use library for creating and sending prompts to various LLMs.

Features (Planned)

• Simple prompt definition using Markdown and Jinja templating.
• Automatic input validation against prompt templates.
• Model selection from a wide range of providers.
• Context length validation with suggestions for suitable models.
• Cost estimation before sending prompts.
• Up-to-date model information (context size, pricing).
• Asynchronous support for concurrent prompt execution.

Installation

pip install ezprompt # Not yet available on PyPI

Basic Usage

import asyncio
from ezprompt import Prompt

async def main():
    # Define a prompt template (details TBD)
    prompt_template = """
    Translate the following text from {{ source_lang }} to {{ target_lang }}:

    {{ text }}
    """

    # Initialize the prompt
    my_prompt = Prompt(
        template=prompt_template,
        inputs={"source_lang": "English", "target_lang": "French", "text": "Hello, world!"},
        model="gpt-3.5-turbo"
    )

    # Check for potential issues and estimate cost
    issues, cost = await my_prompt.check()
    if issues:
        print(f"Issues found: {issues}")
    else:
        print(f"Estimated cost: ${cost:.6f}")

        # Send the prompt
        response = await my_prompt.send()
        print(f"Model response: {response}")

if __name__ == "__main__":
    # Required environment variables (e.g., OPENAI_API_KEY)
    # Set them according to the model provider you use.
    asyncio.run(main())

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License.