I have started the day (and the week) by making changes suggested by my mentors on the PRs #884 and #879. Apart from fixing the code so that it fits in the naming conventions, learning the existence of numpy function asanyarray was interesting. It converts any array-like object into a numpy array and simply passes the input itself if it is already a numpy array. This effectively replaces some of some ugly type checking code.

Then, I started to work on QTAIM charges. The first step is quite simple (in a theoretical sense). I have started drafting the code that evaluates charge densities from the information that is available in ccData’s parsed data. ccData has attribute gbasis that stores the coefficients associated with the Contracted Gaussian Basis Sets. To reconstruct ψ from this information, using G=N e-αr2 xa yb zc does the job. a, b, and c are determined based on the orbital type (s, p, d, f). Note also that r in the exponent, as well as x, y, and z needs to be relative to the center of each atom.

This is kind of tragic and funny – while writing this blog post, I’ve just noticed PR #757 in cclib and the existing code in cclib.method.volume that seems to be almost complete but with a TODO comment for restricted calculations. I should talk about this with my mentors, but it looks like starting off from this existing code would be a better and more effective way to go.