Objective for today was to verify that the cube files generated using cclib.method.volume matched with the expected values. While doing this, I found a likely bug in the function. When the number of grid points were determined in the function, there was an off-by-one error. After submitting a PR addressing this issue, I generated cube file using Psi4 and tried to compare the values. What I expect is that the values calculated from electrondensity will be comparable to Dt.cube and values calculated from wavefunction will be comparable to Psi_a_1...cube files.

It looks like the wavefunctions are being calculated correctly – there is some discrepencies in the data because cclib uses angstroms and Psi4 uses bohr. Some errors are caused from the fact that the grid itself contains error in 1e-3 level.

For electron density, there is pretty big difference which I should investigate further. I may be making mistakes in how I am supplying the flags to Psi4. I’ll look more into this and hopefully make progress tomorrow.