Skip to content

Latest commit

 

History

History
91 lines (62 loc) · 2.45 KB

README.md

File metadata and controls

91 lines (62 loc) · 2.45 KB

llmfuncs

llmfuncs is a Python package for documenting, managing, and validating Python functions based on their signature and docstrings. It automatically generates JSON Schema for functions, and provides a way to validate function arguments against the schema.

Features

  • Automatic generation of JSON Schema from Python function signature and docstrings.
  • Support for Python built-in types and many typing module types.
  • Use and validate functions with JSON arguments against their schema.
  • Management of multiple functions through a ToolCollection object.
  • Bulk addition of functions from a module, a package, or a glob pattern.
  • Simple and intuitive API.

Installation

Install llmfuncs using pip:

pip install llmfuncs

Usage

Tool

The Tool class encapsulates a function and its associated schema.

from llmfuncs.tool import Tool

def greet(name: str) -> str:
    """Greet someone.

    Args:
        name: Name of the person to greet.

    Returns:
        A greeting message.
    """
    return f"Hello, {name}!"

tool = Tool(greet)
print(tool.name())  # "greet"
print(tool.schema())

ToolCollection

The ToolCollection class is a container for multiple Tool objects. It provides methods to add tools and use them.

from llmfuncs.tool import Tool, ToolCollection

tool_collection = ToolCollection()
tool_collection.add_tool(tool)
print(len(tool_collection))  # 1

You can also add tools in bulk from a module, a package, or a glob pattern.

tool_collection.add_tools_from_module('some_module')
tool_collection.add_tools_from_package('some_package')
tool_collection.add_tools_from_glob('*.py')

And here's how to use a tool:

json_args = '{"name": "World"}'
result = tool_collection.use_tool("greet", json_args)
print(result)  # "Hello, World!"

For more detailed usage and examples, please check the API documentation and the example scripts in the examples folder.

Creating New Tools

If you have an existing module that you want to work with llmfuncs, you can try passing it through an LLM to generate the type hints and docstrings for any functions missing them. For example:

Please update this module to add type hints to all function parameters and Google style docstrings to each function.

Contribute

We welcome contributions to llmfuncs!

License

This project is licensed under the terms of the MIT license. For more details, see the LICENSE file.