Backup missing sessions and sessions with shorter JSON sidecars to S3

[rbreslow]

The following cases result in files failing to be processed. These need to be accounted for in order for a user to review manually and re-process as needed.

For these studies, DICOMs are not converted to NIfTIs and the entire study fails. These are currently being tracked in a table that is output to a CSV in program/site/files/:

• Missing sessions
• Missing c-ids
• Missing target Flywheel project

NOTE: Need to avoid overwriting output CSV files on s3

For these acquisitions, the NIfTIs+JSONs get separated into the following directories (the remaining acquisitions in the study are processed completely):

• NIfTIs_to_check
• NIfTIs_short_json

[afamiliar]

Note to self: if missing sub/ses label, currently fails at fw ingest (structure_nifti_files outputs to command line NIfTI files created but no subject mapping to use.) - consider exiting earlier