Eh, big issue.
Naming and metadata are your friends, IMMHO. BUt here we are again... what creates metadata?
- If they name the files\folders in a certain way (eg, timestamp_filename_version), they can upload a zip to a single location and it would be relatively easy to read the names and automatically move things where they do belong.
1.1 pros: they just have to upload one zip file; if they mis-name stuff, that's on them (but only if naming rules are simple)
1.2 cons: you have to "build a manual" to let them be comfy; if naming is not for folders, naming files can be quite the PITA
- If they don't want to name things, then they should know where to put stuff.
2.1 pros: they could orientate themselves between folders and fix mistakes OTF
2.2 cons: they could orientate themselves between folders and delete them accidentally, they can move stuff instead of copying it, exposing the material to the risk to be lost forever. Other general huge mistakes. You should instruct them as well as for point 1 ("teaching" is not really a con... but it's on you in both cases, so I put them here).
Point 1 (\concept0\metadata0_metadata1_metadataN) is a pattern I used to rename very different data from different databases, prior to feed it to vvvv, so that would be simple to disassemble filenames. All in all I got to the point that I just need to input the "name" of a supergroup of data, to process them as expected. But I understand that for me was easy to give significant names to data (I knew the source, so renaming pattern was just a matter of understanding how I wanted to do things)
Hope it's of some help.