Closed robertoostenveld closed 1 month ago
Besides placing it next to the data (with its original name, so as
CC0-1.0.txt
), ithe name of the license file should also be added to the.bidsignore
file to ensure that the dataset remains compliant.
That's not needed if we name it LICENSE
https://bids-specification.readthedocs.io/en/stable/modality-agnostic-files.html#license
🤦♂️
Yes, naming it LICENSE is smart.
What do you think about making a small bidslicense
application that takes a BIDS dataset, checks whether the LICENSE
file is present, if not: checks whether it can be derived from the dataset_description.json
and adds it?
Could that be part of the BIDScrambe application?
Sure, no problem
When doing the input dataset staging (aka downloading), we could explicitly add the license to make that more explicit. As we are (for now) restricting ourselves to BIDS datasets, we know that the license is specified as a string in the dataset_description.json file.
We could parse the dataset description, get the License field, and use the https://spdx.org/licenses/ list and specifically the text versions on https://github.com/spdx/license-list-data/tree/main/text as the source for the license.
It would be relatively simple to make a python command-line tool to fetch the license. Besides placing it next to the data (with its original name, so as
CC0-1.0.txt
), ithe name of the license file should also be added to the.bidsignore
file to ensure that the dataset remains compliant.