Change lookup label and internal identifier of subjects and und subject state

[spieschnik]

Lookup label: A label that will make it easier to find the specimen, e.g. using the short name of the dataset
Internal identifier: Use ID from file names, if clearly present in the files, if not create your own IDs

Example
Subject:
lookup label: 7t_trt_sub-01
internal identifier: sub-01

Subject-state:
lookup label: 7t_trt_sub-01_state-01
internal identifier: sub-01_state-01

[Peyman-N]

@lzehl @spieschnik
Are the following names reasonable for the following bids example dataset:

eeg_rest_fmri
Dataset name in dataset description file: EEG, fMRI and NODDI at rest'

Subject
internal identifier: sub-32
lookup label: EEG-fMRI-and-NODDI-at-rest_sub-32
SubjectState
internalIdentifier: sub-32_state-01
lookupLabel: EEG-fMRI-and-NODDI-at-rest_sub-32_state-01

ds005
Dataset name in dataset description file: Mixed-gambles task

Subject
internalIdentifier: sub-01
lookupLabel: Mixed-gambles-task_sub-01
SubjectState
internalIdentifier: sub-01_state-01
lookupLabel: Mixed-gambles-task_sub-01_state-01

[lzehl]

I would suggest to not use the DSV name here but the short citation, e.g.,
Zehl-et-al-2003_sub-001_state-01 as alternatives for those datasets that do not have explicitly nice short names defined.
A short citation is also what EBRAINS agreed on as short name for those datasets that do not provided nice short name no?

[Peyman-N]

The problem with that approach is that author list extraction is not 100% reliable, and assigning a publication year to a dataset before it is actually published feels questionable. Moreover, the BIDS documentation does not mention including a publication date.

Should we then use the current year as the publication year and rely on whatever information we have from the authors list?

Perhaps a better solution would be to make the process interactive, so the converter prompts the user to provide a short name.

[lzehl]

@Peyman-N this is a matter of adjusting the short citation to what is available at the time (and the dataset might have been indeed already published). Take whatever is given as first author (and leave out the year) or alternatively (when provided) use the DOI (without prefix). It will/may be overwritten later anyway but you HAVE TO SET the short name if not provided.

We also want that the converter runs as automatically as possible in a first run. I don't like prompts in this case because they force manual interaction during the process. I would suggest making this a variable in the convert function or a function later that can be used to overwrite the default.

This is partially discussed as well in the respective PR #84