Keep extra columns when creating compressed detection events
Current compress_detections drops all non-standard columns and returns the event-by-event data, dropping most standard data columns as well. We could actually keep most of these columns if we either add them to the groupby in the case of columns that shouldn't change within a detection_extract (collectioncode, institutioncode, etc), and additionally, we should provide a way for the user to keep all other non-standard columns. Giving them the choice of groupby.agg() methods (mean, first, etc) for all their extra columns, perhaps.