This PR should fix an issue where passing the same truth and test PAFs resulted in precision and recall values less than 1.0. It also refactors the way searches are done, using a defaultdict[sequence name] -> IntervalTree to find records on the query / target and adds several additional match metrics.
This PR should fix an issue where passing the same truth and test PAFs resulted in precision and recall values less than 1.0. It also refactors the way searches are done, using a defaultdict[sequence name] -> IntervalTree to find records on the query / target and adds several additional match metrics.