Pol Febrer Calabozo

By the way you need cython 3.0 to compile it, which is still an alpha version, so you can install it with `pip install --pre Cython`

Cython 3.0 has been in alpha for like 2 years I think and it is now in its 11th iteration. I think it is pretty safe to use. Not using...

Ok, it now works with 0.29.32. I didn't know you could mix both syntaxes. Since you can, this is an easy fix that doesn't annoy me too much and is...

1. Yeah I already thought of this but it was a bit scary since I saw method calls inside the loops. 2. Yes, that's a good compromise I think! I...

By the way, I have thought of an approach that I think would be cheaper. But I'm not sure if there would be problems with precision. Probably you have already...

Also if you do this you could cut some extra time by profitting from the simmetries of the spherical harmonics when computing the reference grids (`toGrid`), no? You only need...

(1) The worry about 50% zeros is because of memory? Do you think that would be a problem? Regarding computation time, doing 2x multiplications (when calculating orbital products) seems like...

Ok, for the moment I'm just trying to separate the calculation of the orbital values on the grid from the products as you proposed. Afterwards I will worry about optimization....

I have already a first working implementation of precalculating orbital values in python + computing the product with cython. In terms of computation time, the implementation seems to be between...

> But maybe the sisl.viz module by @pfebrer can now also achieve this? It does not interface with py3Dmol for now, but it is something I want to do! (because...