close_enough

A cursed Python decorator that generates function implementations using LLMs.

Installation

uv add close_enough

Usage

import close_enough

@close_enough
def calculate_fibonacci(n: int) -> int:
    """Calculate the nth Fibonacci number."""
    ...

result = calculate_fibonacci(10)  # 55

The decorator generates a function implementation based on the signature and docstring. The implementation is generated from a LLM's output lazily on first call and cached for subsequent calls.

Do not use in production code.

Configuration

The default model is github_copilot/gpt-4. To use a different model:

import close_enough
close_enough.MODEL_NAME = "your-model-name"

License

MIT