what kind of calculations are you trying to do? Using linear algebra can be very fast because it can replace for loops. I am using numpy and scipy for best-fitting geometric shapes to vertices. also, if you find example code in other programing languages(C++, Matlab, fortran), ChatGPT is great for converting to python.

@Dead Dead I am in the middle of cleaning up my code, no idea if it is "correct" but it does function. I will post a simplified example to play with since my main code too messy. I am trying to keep the processing independent from the blender side incase I want to reuse it in other applications.

do you use separate computers for version testing? same question for different OS types. it seems like testing can get complex fairly fast. 😒

@Victor Stepanov nice idea, thanks

@Victor Stepanov thanks, this is helpful!

@Aaqil Ilyas , @Victor Stepanov , thanks for the welcome. I am currently working a blender plugin. It's has two purposes. the first is to best fit geometry to 3d scanned data. currently I have plane, polynomial surface, cylinder. I am still adding cone and sphere. these are mostly based on a least squares fit approach. the second is to use the fit geometry to align the scanned data to the global coordinate system. I am working through the logic for this section now. The part can then be exported to parametric cad or rebuilt in blender. my purpose is parametric since I am mor comfortable there. I plan to ask questions soon.