Neuro-iX Tools

Common Tools for the Neuro-iX Lab

Getting Started

Installation

You will need an environment with at least Python 3.11, then run:

pip install neuro-ix-tools

Alternatively, you can clone the repository and use:

python cli.py

instead of neuro-ix.

Setup

If you are using the package, simply run:

neuro-ix init

This provides sensible defaults for Narval. The configuration file is stored in your .config folder.

If you are using the repository directly, we recommend using a local .env file. A template is available at .example.env.

Usage

Inside this environment, you have access to the neuro-ix command, which currently exposes one main tool.

FreeSurfer recon-all on SLURM Cluster

We provide a pipeline that simplifies the usage of FreeSurfer on the Narval SLURM cluster. The main command is:

neuro-ix freesurfer recon-all

This allows users to process all subjects in either a BIDS or CAPS (Clinica) dataset with FreeSurfer, using one SLURM job per subject.

Arguments:

• --bids-dataset: Path to the root of a BIDS-compliant dataset
• --clinica-dataset: Path to the root of a Clinica-compliant dataset (CAPS)
• --cortical-stats: Flag to store only FreeSurfer's stats files
• --start-from: Used when there are more than 1000 subjects due to Narval’s job limit. Allows the user to resume processing from a specific subject index.

Example:

neuro-ix freesurfer recon-all --bids-dataset /path/to/dataset

If your dataset includes more than 1000 subjects (e.g., 1500), once the first batch is done, run:

neuro-ix freesurfer recon-all --bids-dataset /path/to/dataset --start-from 1000

Library

As a library, the neuro_ix package exposes:

• Classes to interact with and query BIDS and CAPS datasets for T1-weighted MRIs
• Extendable command classes

Contributing

Setup

Once the repository is cloned, install the development dependencies with:

pip install -r dev_requirements.txt

Tests

Test Tools

We use:

• pytest for unit tests
• pytest-cov for coverage reports
  Run tests via:

pytest --cov

• ruff for linting and formatting (automatically applied via pre-commit)
• Additional tools for code quality: ssort, pydocstyle, mypy, and pylint

Test Data

All test data are extracted from MR-ART:

Nárai, Á., Hermann, P., Auer, T. et al. Movement-related artefacts (MR-ART) dataset of matched motion-corrupted and clean structural MRI brain scans. Sci Data 9, 630 (2022). https://doi.org/10.1038/s41597-022-01694-8

Deployment

Build Package using :

python -m build

And deploy to PyPI with :

twine upload sit/*