Local setup
This page explains how to set up PARS on your own computer, workstation, or institutional machine.
If you do not want to install PARS and FSL locally, use GitHub Codespaces setup instead.
After completing either setup route, continue to Using PARS. The input files, notebook settings, and run order are explained there only.
Local prerequisites
FSL
The image-processing notebook uses FSL commands such as FLIRT, FAST, BET, and BETSURF. Install FSL and make sure FSL commands are available in the same terminal or environment used to launch Jupyter.
Check this with:
fslversion
which flirt
which fast
which bet
If these commands are not found, PARS will not run correctly.
FSL licence
FSL is third-party software and is licensed separately from PARS. Users are responsible for ensuring that their use of FSL complies with the applicable FSL licence.
Python
PARS requires Python 3.9 or newer. A dedicated environment is recommended.
Using venv:
python3 -m venv .venv
source .venv/bin/activate
Using Conda:
conda create -n pars python=3.11
conda activate pars
Optional external tools
Depending on the downstream workflow, you may also need:
- FreeSurfer or access to FreeSurfer-derived MRI outputs;
- a C++ compiler if you rebuild the mesh-smoothing program;
- LS-PrePost to inspect the generated finite-element model; and
- LS-DYNA to run simulations using the generated model.
These tools are not the same as the core PARS notebook workflow.
Download PARS
Clone the repository and move into its root directory:
git clone https://github.com/HEADLabIC/PARS.git
cd PARS
Install the PARS Python package
From the root of the repository, run:
pip install -e .
This installs the Python code in src/ so it can be imported by the notebooks.
Start Jupyter
Install Jupyter if it is not already available:
pip install jupyter
Start Jupyter from the PARS repository root:
jupyter lab
Open the notebooks from the notebooks/ folder when following Using PARS.
Next step
Continue to Using PARS for the required subject files, notebook settings, mesh-smoother choice, and output checks.