Installation#

This guide will walk you through the three levels of installation to meet your specific needs:

  1. Basic Installation - For users who just want to get started.

  2. Installation with Optional Dependencies - For those looking for extended functionality.

  3. Developer Installation - For contributors or users who need to dive deep into the codebase.

Pyxu is designed to be accessible and is compatible with Linux, macOS, and Windows, although it has not been extensively tested on the latter.

Basic Installation#

The core of Pyxu is lightweight and straightforward to install. You’ll need Python (>= 3.10, < 3.13) and a few mandatory Python packages. While these dependencies will be automatically installed via pip, we highly recommend installing NumPy and SciPy via conda to benefit from optimized math libraries.

First, to install NumPy and SciPy from conda-forge:

conda install -c conda-forge numpy scipy

And then install Pyxu:

pip install pyxu

That’s it for the basic installation; you’re ready to go!

Installation with Optional Dependencies#

For extended features, you can install Pyxu with its optional dependencies:

pip install pyxu[complete-cpu]  # full CPU-only user install.
pip install pyxu[complete11]    # full CPU/GPU  user install. (CUDA 11.x)
pip install pyxu[complete]      # full CPU/GPU  user install. (CUDA 12.x)

More fine-grained extras can be installed by looking at the project.optional-dependencies field in pyproject.toml at the root of the repository.

Warning

The host system must have CUDA 11.x or 12.x installed to use the GPU. Similarly, some Pyxu operators may require extra system dependencies such as LLVM. If problems arise, we provide Docker receipes to easily create Pyxu user/developer environments.

Developer Installation#

If you’re interested in contributing to Pyxu or experimenting with its codebase, you must clone the repository and install it with developer dependencies:

git clone https://github.com/pyxu-org/pyxu
cd pyxu
pip install -e ".[dev,complete]"
pre-commit install

Pyxu uses tox to automate common operations during the development process. The commands below summarize the main operations:

tox run -e py311-test         # run test suite. (fast subset: ~5[min])
tox run -e py311-test -- all  # run test suite. (full suite; much longer)
tox run -e pre-commit         # run pre-commit hooks.
tox run -e doc                # build HTML docs. (incremental update)
tox run -e doc -- clean       # build HTML docs. (from scratch)
tox run -e dist               # build universal wheels for distribution.

All available tox environments can be viewed by running:

tox list

Interoperation with Deep Learning Frameworks#

If you wish to use Pyxu in combination with deep learning frameworks like JAX and PyTorch, you’ll need to install them separately. For more information, consult the installation guides for JAX and PyTorch.

You’re All Set!#

You are now ready to harness the capabilities of Pyxu for your projects and research. If you have any questions or contributions, we would be happy to hear from you!