pipeline json files should be created in a subdirectory with unique prefix

[dshean]

Potential issues using same output prefix/directory for multiple runs and conflicts with old pipeline_dsm_####.json files.
Low priority

[scottyhq]

I suggest we require outputting to an empty directory for now. The current workflow does not really have checks to see if files already exist or not and start from intermediate states. That could be nice at some point, but until then it seems just safer to ensure a clean slate when starting.

[ShashankBice]

Yes, I agree that if the re-run is with new options, it should be in a seperate directory.

I will note in readme to do start with a fresh output directory if a users want.

Actually, Scott I have implemented a restart flag to skip already processed tiles, will expose that later today, if a run was cancelled due to some reason (e.g., job running out of scheduled time or memory etc), where we can note that the flags should be the same as the previous run.

[dshean]

The ability to validate existing products and restart is potentially valuable for larger AOIs, but also introduces complications (e.g., corrupt files partially written when interrupted).

I think we can potentially have the users define an output directory, and then can use the filename of the AOI vector file as the "prefix". The ASP-style combined output directory and filename prefix is not the best option, in my opinion.