arg-gpt

A Python package that simplifies building GPT interfaces for your Python code. This package helps you expose your Python functions to GPT models, enabling natural language interaction with your code.

Installation

pip install arg-gpt

Or using Poetry:

poetry add arg-gpt

Basic Setup

1. Set up your OpenAI API key in a .env file:

OPENAI_API_KEY=your_api_key_here

2. Use the @ai_func decorator to register functions for GPT:

from arg_gpt.ai_func import ai_func

@ai_func
def calculate_sum(a: int, b: int) -> int:
    """Calculate the sum of two numbers."""
    return a + b

3. Create a CLI interface to interact with your functions:

import openai
import typer
from arg_gpt.ai_func import get_ai_functions
import arg_gpt.prompts as prompts
import arg_gpt.gpt_helpers as gpt_helpers
from dotenv import load_dotenv

load_dotenv()

def run_conversation(prompt: str, functions: list):
    client = openai.Client()
    messages = prompts.request_detailed_result() + prompts.remain_functional() + prompts.user_prompt(prompt)
    response = gpt_helpers.call_gpt_with_function(client, functions, messages)
    messages.extend(gpt_helpers.interpret_response(response, functions))
    return messages[-1].content

if __name__ == "__main__":
    typer.run(lambda x: run_conversation(x, functions=get_ai_functions()))

Now you can interact with your functions using natural language:

python your_script.py "calculate the sum of 5 and 3"

Examples

The package includes two example implementations in the examples directory:

OpenAI Example

Located at examples/example_cmd.py, this example demonstrates using arg-gpt with OpenAI's API.

To run this example:

1. Ensure you have set OPENAI_API_KEY in your .env file
2. Run the example:

python examples/example_cmd.py "your prompt here"

Groq Example

Located at examples/example_groq.py, this example shows integration with Groq's LLM API.

To run this example:

1. Ensure you have set GROQ_API_KEY in your .env file
2. Run the example:

python examples/example_groq.py "your prompt here"

Both examples use a CLI interface where you can provide natural language prompts to interact with the registered functions. The examples will process your prompt and execute the appropriate functions based on your request.

License

This project is licensed under the terms of the LICENSE file included in the repository.

Author

Philip Krejov

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.