libpymcr

libpymcr is a Python module for loading a compiled Matlab ctf archive and running functions within it in Python.

Whilst there is an official Mathworks binding for Python, libpymcr provides a few benefits over the official package:

• It is compatible with versions of Python not supported by Matlab (e.g. 3.10, 3.11, 3.12). Moreover, the supported Python versions are not locked to the Matlab versions, unlike with the official packages.
• When converting data between Matlab and Python it avoids data copies where ever possible by wrapping the underlying arrays in the target type (numpy or Matlab mxArray). In the official bindings, a data copy is required when converting data from Python to Matlab (inputs to functions).
• It provides a simpler syntax, and if you include the provided call.m and call_python.mex files in your compiled package, you will also be able to access Matlab objects transparently in Python, and pass Python callables to Matlab to evaluate (e.g. in a fitting routine).

Getting started

You can install the package using:

pip install libpymcr

You must create a compiled Matlab archive (ctf file) of your program using the Matlab Compiler SDK toolbox, using the mcc command:

mcc -W CTF:your_program_name -U mfile1 mfile2 mfile3

Then in Python, you can load this and call the Matlab functions with:

import libpymcr
m = libpymcr.Matlab('your_program_name.ctf')
m.mfile1()

The functions, mfile1, mfile2 etc. are exposed to Python can can be called as methods of the Matlab() object.