Link the currently used version of NeuroBlueprint for each datashuttle version

[JoeZiminski]

At points, breaking changes may be made to the NeuroBlueprint spec (e.g. soon, we will rename behav to bev for better a alignment with BIDS). This has the potential to be very disruptive to datashuttle users, and it must be possible for users to switch between NeuroBlueprint backends (this machinery could be extended to support other specs as backends).

We discussed this on the community call and the approach agreed on was as such:

• On NeuroBlueprint, the spec for each version is encoded in a yaml file. Initially, this can be quite straightforward (what is to be expected at each folder level, and the names of the datatypes). These can live on the NeuroBlueprint repository.
• Datashuttle should read these config files and use them for validation, and to populate fields in the TUI (e.g. allowed narrow / broad datatype names). Some aspects are fairly rigid (sub, then ses, then datatype) but there is flexibility in naming. Initially, these will most likely to be used to hold changes in datatype names or the addition of new datatypes.
• Each project should be associated with a specific NeuroBlueprint config. In this way, existing projects can retain their NeuroBlueprint config, but new projects can be (by default) bumped to the most recent version. Of course, users will have to option to change the spec used on a per-project basis.