The GPTFunction Python module provides a decorator for wrapping Python functions to generate an OpenAI GPT function calling schema. This module is particularly useful for those who need to integrate Python functions with GPT in a structured and efficient manner. The decorator transforms a regular Python function into a GPT function with a JSON schema, which can then be utilized by GPT for various purposes.

• Type Hinting and Documentation: The module enforces type hinting for parameters, supporting types like str, int, float, typing.Literal, and subclasses of Enum.
• Docstring Parsing: It parses the function's docstring to extract descriptions, ensuring a detailed and clear schema.
• Automatic Schema Generation: Generates the necessary schema for GPT integration seamlessly.

pip install gptfunction

from gptfunction import gptfunction

@gptfunction
def output_user(name: str, age: int) -> None:
    """
    Outputs a user's name and age to the console.

    :param name: The name of the user.
    :param age: The age of the user.
    """
    print(f"Name: {name}, Age: {age}")

gpt_tools = [output_user.schema()]

from gptfunction import gptfunction
from enum import Enum
from typing import Literal

class Fruit(Enum):
    APPLE = 'apple'
    BANANA = 'banana'

@gptfunction
def favorite_fruit(user_name: str, fruit: Fruit, quantity: Literal[1, 2, 3]) -> str:
    """
    Returns a string stating the user's favorite fruit and quantity.

    :param user_name: Name of the user.
    :param fruit: The preferred fruit.
    :param quantity: The quantity preferred (1, 2, or 3).
    :return: A descriptive string.
    """
    return f"{user_name} likes {quantity} {fruit.value}(s)."

print(favorite_fruit.schema())

• str: For string values.
• int: For integer values.
• float: For floating-point values.
• typing.Literal: For specifying a literal set of values.
• Enum: For enumerated types, with string values as Enum members.

• @gptfunction: This decorator should be used above the function definition. It processes the function and creates a GPT function schema.

• schema(use_required: bool): Returns the JSON schema of the wrapped function.
  • use_required: Indicates whether the schema should specify required parameters (default: True).
• description(): Retrieves the function's description from its docstring.
• name(): Returns the name of the function.

Contributions to improve GPTFunction are welcome. Please follow the standard procedures for submitting issues and pull requests on the project's GitHub repository.

Distributed under the MIT License. See LICENSE for more information.