I like this. Was not aware argparse had this option. We need to make sure this doesn't cause any unexpected errors and is compatible with some pending unreleased updates though.

This will be a welcome improvement! On several systems, I have run into the "too many command-line arguments" limitation, and there is no other solution than splitting the dataset in several batches to pick separately (pretty annoying to do, and it makes the job graph look more complicated than it really is).

This would be great to have! We frequently run into this with large datasets, necessitating splitting into smaller batches. This isn't much of an issue for experienced users, but it trips up a lot of first time Topaz users in my experience.