MxlPy

MxlPy (pronounced "em axe el pie") is a Python package for mechanistic learning (Mxl) - the combination of mechanistic modeling and machine learning to deliver explainable, data-informed solutions.

Documentation

You can find extensive documentation directly here on github

Installation

You can install mxlpy using pip: pip install mxlpy.

Due to their sizes, the machine learning packages are optional dependencies. You can install them using

# One of them respectively
pip install mxlpy[torch]
pip install mxlpy[tensorflow]
pip install mxlpy[keras]
pip install mxlpy[jax]
pip install mxlpy[sr]

# together
pip install mxlpy[torch, tensorflow, keras, jax, sr]

If you want access to the sundials solver suite via the assimulo package, we recommend setting up a virtual environment via pixi or mamba / conda using the conda-forge channel.

pixi init
pixi add python assimulo
pixi add --pypi mxlpy

How to cite

If you use this software in your scientific work, please cite this article:

• doi
• bibtex file

Development setup

You have two choices here, using uv (pypi-only) or using pixi (conda-forge, including assimulo)

uv

• Install uv as described in the docs.
• Run uv sync --all-extras --all-groups to install dependencies locally

pixi

• Install pixi as described in the docs
• Run pixi install --frozen

LLMs

We support the llms.txt convention for making documentation available to large language models and the applications that make use of them. It is located at docs/llms.txt

Tool family 🏠

MxlPy is part of a larger family of tools that are designed with a similar set of abstractions. Check them out!

• MxlBricks is built on top of MxlPy to build mechanistic models composed of pre-defined reactions (bricks)
• MxlModels supplies flat, single-file versions of MxlBricks models for easy inspection
• MxlWeb brings simulation of mechanistic models to the browser!
• pysbml simplifies SBML models for import/export with MxlPy
• Parameteriser looks up kinetic parameters from BRENDA and other databases