This code can be found also at this link:
At the moment I have a non coordinate based routine for batch selecting face groups...this is probably the most interesting and complex part of the algorithm.
I've extended this now to two dimensions on a given mesh surface, but at the moment selection routines on a given 2d surface are limited to 2x2 polygons(faces).
I may work on a NxN generalized case in time here.
Sort of thinking of this problem also as a base selection routine (as in number theory) extending a 2x2 selection to some number n which is compromised of some multiple of the 2x2 coupled with a remainders algorithm (for the odd case as opposed to the even case). The selection routine is nice since it is non coordinate based, and can arbitrarily use any corner point as a start to the loop algorithm which fills in the rest of the selection grid. It uses in this case a 'sensing' mechanism to sense direction without actually checking coordinates. Logically this 'sensing' is typified by occupancy relations...that is occupancy in so far as neighbor face relations in comparing as set of potentially fuzzy or ambiguous choices. The algorithm is tested mostly for even subdivision meshes, meaning that a 2x2 grid, for instance, evenly divides the surface.
https://github.com/christophermoverton/BlenderPythonScripts/blob/master/insetextrude.py
At the moment I have a non coordinate based routine for batch selecting face groups...this is probably the most interesting and complex part of the algorithm.
I've extended this now to two dimensions on a given mesh surface, but at the moment selection routines on a given 2d surface are limited to 2x2 polygons(faces).
I may work on a NxN generalized case in time here.
Sort of thinking of this problem also as a base selection routine (as in number theory) extending a 2x2 selection to some number n which is compromised of some multiple of the 2x2 coupled with a remainders algorithm (for the odd case as opposed to the even case). The selection routine is nice since it is non coordinate based, and can arbitrarily use any corner point as a start to the loop algorithm which fills in the rest of the selection grid. It uses in this case a 'sensing' mechanism to sense direction without actually checking coordinates. Logically this 'sensing' is typified by occupancy relations...that is occupancy in so far as neighbor face relations in comparing as set of potentially fuzzy or ambiguous choices. The algorithm is tested mostly for even subdivision meshes, meaning that a 2x2 grid, for instance, evenly divides the surface.
https://github.com/christophermoverton/BlenderPythonScripts/blob/master/insetextrude.py
No comments:
Post a Comment