At the moment, in order to get certain algorithms working unweighed trees are treated as weighed with each branch given a weight of 1. While it's been making the major data structure shifts easier it's not so good in terms of memory use. The bipartition table itself needs to have a weighed flag, and when weighs are requested for unweighed trees, it'll still return 1 or some other default... but that one doesn't need to be stored for each branch like it currently is. That'll cut down some memory usage on the larger sets.
At the moment, in order to get certain algorithms working unweighed trees are treated as weighed with each branch given a weight of 1. While it's been making the major data structure shifts easier it's not so good in terms of memory use. The bipartition table itself needs to have a weighed flag, and when weighs are requested for unweighed trees, it'll still return 1 or some other default... but that one doesn't need to be stored for each branch like it currently is. That'll cut down some memory usage on the larger sets.