Open rileyvickers opened 1 year ago
@rileyvicker Is this resolved now?
This and all subsequent additions work correctly when nSubprocesses = 1. However, they all rely on globalID being set, which is done via the call of sets.updateSetPathID in skeletonize.py. This call is what was giving an array size mismatch because there is a call within that does not handle a case when there are no true boundaries, as happens when the entire domain is owned by the same subprocess. I was able to check this by commenting out the offending call on line 686 of sets.pyx
Copy. So I think this is a simple fix where localID is just equal to globalID with the exception of periodic boundary conditions where we match IDs on the opposite faces. Ill add this to my list.
Pull#19 resolves this issue, will close when it is merged.
Don't close this, a single dangler is ruining our good time
medialAxis.trimSets() currently nonfunctional for single processor solves, adding here b/c it is a low priority issue