BIDSme is a open-source python tool that converts ("bidsifies") source-level (raw) neuroimaging datasets to BIDS-conformed. Rather then depending on complex or ambiguous programmatic logic for the identification of imaging modalities, BIDSme uses a direct mapping approach to identify and convert the raw source data into BIDS data. The information sources that can be used to map the source data to BIDS are retrieved dynamically from source data headers (DICOM, BrainVision, nifti, etc.) and file structure (file and/or directory names, e.g. number of files).
The retrieved information can be modified/adjusted by a set of plugins. Plugins can also be used to complete the bidsified dataset, for example by parsing log files.
NB: BIDSme support variety of formats including nifty, dicom, BrainVision. Additional formats can be implemented.
The mapping information is stored as key-value pairs in human-readable, widely supported YAML files, generated from a template yaml-file.
Bidsme can be installed using pip:
python3 -m pip install git+https://github.com/CyclotronResearchCentre/bidsme.gitIt will automatically install packages from requirements.txt. When treating specific data formats, additional modules may be required:
- pydicom>=1.4.2 (for DICOM images)
- nibabel>=3.1.0 (for ECAT7 images)
- mne (for various EEG/MEG recordings)
It is recommended to use virtual environment when installing bidsme (more info here and here).
More details on how to install bidsme can be found in INSTALLATION.md
bidsme can be used with command-line interface and within Python3 shell (or script).
A extensive tutorial, aviable there, should provide a step-by-step guidence how to bidsify a complex dataset. The tutorial uses an example/toy dataset aviable here.
Some additional documentation are aviable in doc directory, namely:
Bugs and suggestions can be communicated by opening an issue. More direct contibutions are done using pull requests.
For more informations, please refer to contribution guide.
bidsme started as a fork of bidscoin, which can be used as an easier-to-use alternative to bidsme, focused on MRI datasets.
Development of bidsme was made possible by Fonds National de la Recherche Scientifique (F.R.S.-FNRS, Belgium) and the University of Liège.