niix2bids

Automatic BIDS architecture. Convert your DICOM with dcm2niix, then use this package as a command line to organize your dataset into BIDS.

Background

BIDS is widely used for sharing neuroimaging datasets. There are several packages that have a semi-automatic approach to organize datasets to BIDS, such as Dcm2Bids. To transform one or a few cohorts from dicom or nifti to BIDS, I would use Dcm2Bids.

However, for a more automatic approach with no prior on the input directory structure and file names, niix2bids can be better. This should be the case for MRI centers that have lots of cohorts from different research protocol. It may also be useful for clinical data with wide variations in naming conventions and parameters.

Limitations

Only works with:
- MRI data
- Siemens magnets
- DICOM converted using dcm2niix
PatientName must be in the JSON file. Use dcm2niix with option -ba n to disable anonymization.

Usage

usage: niix2bids [-h] -i DIR [DIR ...] -o DIR [--symlink | --copyfile] [--logfile | --no-logfile] [-c FILE] [-v]

    Create BIDS architecture from nifti files and .json sidecars.
    This method expects DICOM converted by dcm2niix (https://github.com/rordenlab/dcm2niix)
    

Required arguments:
  -i DIR [DIR ...], --in_dir DIR [DIR ...]
                        Nifti directories that will be parsed and transformed into BIDS architecture.
                        This directory is usually the output directory of dcm2niix.
                        This argument accepts several paths. You can use syntax such as /path/to/nii/2021_*
  -o DIR, --out_dir DIR
                        Output directory, receiving the BIDS architecture.

Optional arguments:
  -h, --help            show this help message and exit
  --symlink             Use symlink (default). The out_dir will contain symlinks, saving disk spaces
  --copyfile            Use copyfile. Original files will be copied in out_dir. !! be careful of disk space !!
  --logfile             Write logfile (default)
  --no-logfile          Disable writing logfile
  -c FILE, --config_file FILE
                        If you want to use non-coded sequences such as new Products, WIP or C2P,
                        you can provide a config file.
                        Default location is ~/niix2bids_config_file/siemens.py
                        If default location is not present, try to use the template file 
                        located in [niix2bids]/config_file/siemens.py
  -v, --version         show program's version number and exit

niix2bids_version==v2.4.0 + bids_version==v1.6.0

Installation

Python version & other dependencies

Python version

python >= 3.6 Tested on 3.6, 3.9, 3.10

Package dependencies

pandas for DataFrame
nibabel to load the nifti file header

How to

Recommended setup

Use conda to create a new python environment.

Standard

conda create --name niix2bids
conda activate niix2bids
pip install git+https://github.com/benoitberanger/niix2bids

Developer

If you want to install in "developer" mode using the Git local repository, clone the repo before, then change the installation line :

cd /path/to/mydir/
git clone https://github.com/benoitberanger/niix2bids
conda create --name niix2bids
conda activate niix2bids
pip install -e niix2bids/

If you want to make a PR, this the recommended strategy, because you can control branching in your local clone of the repository

NOT tested

pip install git+https://github.com/benoitberanger/niix2bids
The installation might crash because of wrong dependency management. Creating a fresh conda environment is the recommended setup. See above.

Notes

pip install niix2bids is not possible yet. I did not register this packaged on https://pypi.org.