a-c-r natively supports S3 inputs just like HTTP/S
In short, provide a native S3 version of
http_to_keep. Users should be able to write S3 sources in their inputs, which a-c-r then writes to keep before executing the workflow.
This should support the existing
--defer-downloads option, so the user can decide whether to copy the data when they create the container request, or let it be handled inside the container.
No data to display