Yeah that is a massive improvement , congrats ! . Happy to have pushed for OCCT7 . Let's build up something super efficient for FeM out of it now.DeepSOIC wrote:The new tools based on generalFuse have been merged recently!
Since it was already described, I'm just linking to the test request topic:
Here, I only want to say thanks to all who helped me with testing and motivation, especially to bernd!
And to Werner for reviewing, merging, and fixing some of my goof-ups!
EDIT: Forgot to mention sgrogan, who made a libpack with OCC7 that opened up the possibility
Yeah, I forgot to mention you too fixedvejmarie wrote:Happy to have pushed for OCCT7 .
As for Connect on solids, I don't expect much gain. But for connect on shells and wires, the gain can potentially be massive in certain cases.vejmarie wrote:I read your code and was wondering if for some efficiency we could move some part to the C++ code instead of keeping the computation part in Python (easy to say harder to implement)
Do not worry !!DeepSOIC wrote:Yeah, I forgot to mention you too fixed
I will try to run some profiling on my Mac and see if I can get something efficient out of itDeepSOIC wrote:As for Connect on solids, I don't expect much gain. But for connect on shells and wires, the gain can potentially be massive in certain cases.
In any case, I think profiling the actual computation times of the GFA itself and the post-processing Py code is needed to see, how much of an improvement can be achieved, and if it is worth the effort. And probably, only a few small pieces of the code have to be moved to C++.
I am slowly thinking of a general connexity map object. That can be terribly useful for all sorts of stuff. I think OCC itself has something to offer in this regard.
To me it seems symmetrical to Part Common** [[Image:Part XOR.png|32px]] [[Part_XOR|XOR]]: removes space shared by even number of objects (symmetric version of [[Part_Cut|Cut]]) (v0.17)