determine what to do when techMD detects corrupt files

[andrewjbtw]

[This is a stub for future work.]

Sometimes files are deposited that have some kind of file corruption - missing end-of-file data, bad headers, etc.. This can lead the techMD process to stall because the metadata extraction tools report errors when trying to characterize the files.

We don't have a good process for handling this situation right now. We can skip the techMD step and get the files deposited but it seems like we should have some kind of techMD field that we could use to query for known-corrupt files, and then potentially follow up with users if that's a possibility.

Note that this ticket is specifically about situations where a file is already corrupt prior to deposit, not corruption that might be caused by hardware or software failures during accessioning. Our system should be catching any corruption that happens during accessioning by verifying file checksums as files make their way through the pipeline.