Update the BIDS-like saving to align with template directory

[ibevers]

This pull request updates the BIDS-like structure of the output directory to more closely follow BIDS. It also adds tests and optimizes performance. Some of the changes included in this PR:

• Adds descriptions for each field to help users of the dataset understand it better.

• Adds an n_cores parameter that determines the number of processes to use when extracting features. When using Pydra, this is also the number of separate model instances that will be created.

• Adds tests for the conversion code.

• Adds transcriptions to the features saved with the option to use different Whisper model sizes.

• Adds a with_sensitive parameter that determines whether to include raw audio files and transcriptions in the output.

• Adds a template BIDS-like directory structure.

• Configure to automatically update CHANGES.md

• Remove references to summer school

Ensure all these are constructed or copied to the bids-like dataset saving directory with correct names:

• CHANGES.md
• README.md
• dataset_description.json
• participants.json
• participants.tsv
• phenotype
• <measurement_tool_name>.json
• <measurement_tool_name>.tsv
• sub-<participant_id>
• ses-<session_id>
• beh
• audio
• sub-<participant_id>_ses-<session_id>_task-<task_name>_audio.wav
• sub-<participant_id>_ses-<session_id>_task-<task_name>_features.pt
• sub-<participant_id>_ses-<session_id>_task-<task_name>_metadata.json
• sub-<participant_id>_ses-<session_id>_task-<task_name>_transcript.txt
• sessions.tsv

[ibevers]

dataset_description.json based on:
https://github.com/bids-standard/bids-examples/blob/master/ds052/dataset_description.json
https://bids-specification.readthedocs.io/en/stable/glossary.html#dataset_description-files

[ibevers]

Need to move all questionnaires in beh to phenotype. We don't have any behavioral experiments in the dataset. Therefore, we should not have a beh dir.

[ibevers]

There should be no beh directory.

• Move all audio related jsons in beh to audio
• Remove beh directory

[ibevers]

• Add flag for not including audios or otherwise make this clear

[alistairewj]

• Add senselab>=0.12.0 to the requirements - the library doesn't work with earlier versions
• Are you sure we want to package the audio data in src? it makes the pip package much bigger.
• Reformatting code & making changes makes it very hard to review changes. in the future, if you plan on bulk reformatting, best practice is to only reformat and do it in its own PR. Then afterward you can make your code changes.

Still in the middle of doing my testing with the latest data dump, but I think it's worth approving this and if I come up with any further improvements I can make a new PR.

[ibevers]

Hi @alistairewj, thank you for reviewing! I didn't realize that about reformatting code and making changes in the same PR. In the future, will do my best to separate them🌱

To-dos:

• add senselab>=0.12.0 to requirements
• move audio data out of src

[ibevers]

• Add license text dataset_description.json