Closed jessb0t closed 2 years ago
Example: erp-core-flanker-dataset
@georgebuzzell : Yanbin brought up a really good point to me awhile back. When we are creating a repo to house a non-NDCLab dataset on the HPC, many of our standard folders for a dataset project are irrelevant. I believe that these include: data-monitoring
, literature
, and materials
. I'm uncertain about derivatives
, since our instructions are explicitly to download the data to sourcedata
(but we are ostensibly downloading a derivative, preprocessed by someone else)?
Please share your thoughts. Thanks!
good question @jessb0t
when we download the data elsewhere, we will drop it into source data, yes. but, if already in bids format, or after conversion to bids, we will populate derivatives and also materials. so, we should keep those.
Not sure yet about data-monitoring and literature. Data-monitoring is likely not needed, so it can be deleted and added back if needed. I think we should keep literature though, because if nothing else, it should hold the reference publication for the data.
Thanks, @georgebuzzell . I logged these details in issue 150 of the wiki for a future update, likely once we have some lab members actively creating such a dataset. I also updated the lab mgmt manual for future repo creation instructions. In the immediacy, I will share this back with Yanbin!
For instance: