pyMechT

pyMechT is a Python package for simulating the mechanical response of soft biological tissues. The ethos of pyMechT is to create simplified virtual experimental setups, rather than finite element analyses. Thus, varying parameters and running simulations is much faster, making it feasible to perform Bayesian inference and Markov Chain Monte Carlo analyses. A short overview is provided in the video below.

Documentation

Find the full documentation at https://pymecht.readthedocs.io/en/latest/.

Structure

pyMechT is a collection of modules for:

• MatModel: defining material models
• SampleExperiment: simulating experiments, such as uniaxial extension, biaxial extension, and inflation-extension. Simulations can be either disp_controlled or force_controlled
• ParamDict: a custom dictionary class of a datastructure called Param, which facilitates storing/varying/fitting parameters
• ParamFitter: fitting parameters to experimental data
• RandomParameters and MCMC: Bayesian inference by running Monte Carlo (MC) and Markov chain Monte Carlo (MCMC) simulations

This package is developed and maintained by the Computational Biomechanics Research Group at the University of Glasgow.

Required dependencies are:

• matplotlib
• numpy
• pandas
• pyDOE
• scipy
• torch
• sympy
• tqdm

Installation

Step 1 (optional): Create a virtual environment

To create an environment in Anaconda, execute:

conda create -n pymecht

To activate this virtual environment, execute:

conda activate pymecht

This is an option, but recommended step. There are other options for create and managing environments (such as venv or virtualenv)

Step 2: Install via pip

User

pyMechT can be installed directly from PyPI via pip by using:

pip install pymecht

Developer

To install as a devloper, it is recommended to fork from the repo and clone this fork locally.

Step 2.1 Fork from ankushaggarwal/pymecht

To fork a branch, head to the Github repository and click the fork button in the top right-hand corner.

Step 2.2 Clone the forked repo

To clone this repo locally, use the

git clone <repo-address>

where <repo-address> can be replaced by either the https or ssh addresses of the forked repo.

Step 2.3 Install developer version of pyMechT

To install a developer version of pyMechT, navigate to the locally cloned repo and execute:

python setup.py develop

An editable version of pyMechT is now installed. All local changes to the cloned source code files will be reflected when pyMechT is imported.

Step 3: Check installation

Ensure that pyMechT has been installed by executing:

pip list

The package and version should be visible in the resulting list.

Contributing to pyMechT

To contribute to the pyMechT framework, install pyMechT using the developer options. All changes should be made to your forked repo. If there is a new feature or bug fix, raise a pull request. In the case that an additional feature is added, a corresponding example and test should be written in the respective python scripts.