graham
If you want a pure Python approach there isn't much you can really do except write code that manually loops over ever primitive and vertex and constructs a data structure of points and their neighbors. There have been various discussions about this on this forum, as well as odforce. If you can compile HDK code, you can write a c++ function using inlinecpp.py that will allow you to very quickly return to Python a list of connected points. There is the pointneighbours() expression function that will give you a space separated list of points that share primitives with a point, though sharing primitives is not the same as directly connected with an edge. People have also used VOPs to build lists of connected points as well.
Graham thanks for the speedy help here.
Odforce was down so I could not get anything from there and most of the point promiximity issues were related to point clouds in these forums.
I guess for the starter I will embed pointneighbours() in Python code to see if I get anywhere with that. It probably will be cheaper than trying to create an assocation dabatase in Python. I have also seen examples of the inlinecpp, which sounds like a great idea. Time to dig into hdk for real this time I guess
edit: I just checked the help for the inlinecpp and impressive documentation there. Kudos
k