The current pipeline runs per-patient and only uploads data if all all records (metadata) in the database is marked as being downloaded, i.e., that we have access to the raw recordings. For dreem, this means that if there is one patient and 14 records, but one of those records has not been downloaded, then the undownloaded recording must be modified to be skipped. The pipeline running in this way was designed so that all data would be uploaded in one zip.
We need to resolve this after FS, and could do this in two ways:
Modify the pipeline to skip data that is not downloaded
Allow clinical staff to initiate/trigger the pipeline if something went wrong?
It is likely that (2) would be useful regardless given there won't be much technical support in the CoS.
The current pipeline runs per-patient and only uploads data if all all records (metadata) in the database is marked as being downloaded, i.e., that we have access to the raw recordings. For dreem, this means that if there is one patient and 14 records, but one of those records has not been downloaded, then the undownloaded recording must be modified to be skipped. The pipeline running in this way was designed so that all data would be uploaded in one zip.
We need to resolve this after FS, and could do this in two ways:
It is likely that (2) would be useful regardless given there won't be much technical support in the CoS.
Making this issue now for future discussions.