review of example that Arno shared

[robertoostenveld]

On https://drive.google.com/drive/folders/1UepEeZ80wpPxU-OhmiR9oGphomwRR2oX @arnodelorme shared an example derivative EEG dataset. This issue serves to share comments on that draft dataset.

[robertoostenveld]

also tagging @dungscout96 who was involved in the construction of the example.

[robertoostenveld]

the directory structure is

ds004117der
├── ds004117der
│   ├── sub-001
│   │   └── ses-01
│   │       └── eeg
│   ├── sub-002
│   │   └── ses-01
│   │       └── eeg

where I guess the top-level directory is the raw, and the folder inside it is the derivative. However, there should be a derivative folder level in between like this

ds004117raw
├── derivatives
|   └── ds004117der

See https://bids-specification.readthedocs.io/en/stable/02-common-principles.html#source-vs-raw-vs-derived-data

[robertoostenveld]

according to https://bids-specification.readthedocs.io/en/stable/05-derivatives/01-introduction.html#file-naming-conventions I think the derivative data should be distinguished from the raw data by a desc-<label> entity.

[robertoostenveld]

there is no descriptions.tsv file as discussed in #8

[robertoostenveld]

Regarding the directory structure, I am not sure whether my assumption at https://github.com/bids-standard/bep021/issues/9#issuecomment-1725542841 holds. It rather seems that there is a copy of the same directory in the directory, and that both are derived. The metadata seems to be exactly the same, including the GeneratedBy in the dataset_description.json

[arnodelorme]

The folder in the subfolder was an issue associated with copying the data to Google Drive. I have removed the subfolder. I was under the impression that the inclusion of the raw data was not mandatory.

[arnodelorme]

I have updated the data (see the link above) - I have also removed the data itself to make it lightweight.

[effigies]

the inclusion of the raw data was not mandatory.

Correct. If the dataset has "DatasetType": "derivative" in its dataset_description.json, then it would be recognized as a standalone derivative dataset.