Crawl misses files if multiple studies present for single patient directory

[strixy16]

If patient directory contains multiple studies and an RTSTRUCT references a CT in a different study, crawl function outputs do not show this association (i.e. in the edges.csv)

[skim2257]

This was an intentional design choice. If there's a patient where two different studies have two different series with reference to each other... that's not ideal h a h a... but we can change our sample finding method.