Syntax sugar for ctypes that provides a more confortable way to load and setup shared libraries for your projects. The goal of this simple package is to reduce boilerplate and replace tedious syntax with something more intuitive and fancier.

• Load shared libraries and only interface the desired procedures.
• Parse argument and result types together through a dedicated parser function.
• Multiple type aliases available (e.g. real64 / float64 for 64-bit floating point numbers)
• Improved syntax to define C struct and union types
• Tools to work with NumPy arrays

All ctypes types are compatible even if FancyTypes does not provide its own specific class for them. For example, char support is limited since the main focus of the package is numerical computing, so syntax improvements to pass strings around are limited to turning them into NumPy arrays and not much else. Typically, strings and I/O can be handled on the Python side.

The only dependency is NumPy for array handling. The NumPy package is pretty much part of the standard library amongst the Python numerical computing ecosystem, so any project that could potentially use this will most likely already depend on NumPy on its own.

Read the documentation and find out if this package could be useful to you.

The package can be installed from PyPI

pip install fancy-ctypes

Alternatively, download this repository and run

pip install .

The package is imported as

import fancytypes

# For confort
import fancytypes as ft

The package is intended to be platform agnostic, although shared libraries are not. This is a tool that interfaces those libraries with Python, but users still need to build them first however they want.

An interface to C sources was built for the KVP package using FancyTypes. See kvp/lib/init.py and kvp/callers.py in particular.