Open
Description
Context
Reported by @brian-mott on sci-data-eng
It would be very helpful to have anndata checks done first and to return early/exit if these exist, and then move onto fragment file validation. Currently with a real world dataset with ~60k cells, this takes about 30 mins to validate, and I wouldn't know that obs['is_primary_data'] is not all True until the end of that time. With early returns/exits, I could see that's an issue, quickly fix, then make better use of all the compute/space needs for fragment validation
Activity