Closed alwilson closed 1 year ago
@mballance I was curious what all was done in the swizzler with dists so I made some changes. Thought I'd throw my changes up to see what you think. Full context here: https://github.com/fvutils/pyvsc/issues/191
Hi @alwilson, sorry for the delay in looking at this in-depth. Good catch on the difference between SV expected behavior and what was implemented in PyVSC. Changes look good to me, and appreciate the addition of a test!
This removes the soft constraint created after building the hard constraints for a distribution. It also reverses the unittest meant for testing that distribution non-zero weights can override soft constraints.
This saves off the weight list and a function to calculate them into ConstraintDistScopeModel, in the hopes that at some point recalling the randomization model and swizzler would re-randomize the distribution weight targeted.
This also adds enough type hints to make it easier to track down target_range and distribution weight list usage.