This [starting_index option] seems helpful for resuming a download given the last successful index. Perhaps in the future you could add an option for ending index or number of items. That way a user could download chunks of a CSV in multiple processes.
This would go nicely with a default that checks the provided image folder (if it exists) to see the contents and automatically determine starting index (or just filenames to remove since we force it to be unique).
_Originally posted by @johnbradley in https://github.com/Imageomics/cautious-robot/pull/1#discussion_r1627695112_