projectq-framework / fermilib-plugin-psi4 Goto Github PK

Hi!

First, I think this question also applies to OpenFermion, given that the plugins have very similar structure. This might be a very trivial question, but how are charges assigned to atoms for a given molecule?

For instance, for the molecule is Li_2 with charge = 2. Then, I'm guessing each lithium atom gets one positive charge, right? What if I want to specify that one lithium atom has charge +2, while the other is a neutral atom? Is that possible? If I have a chain of say 5 lithium atoms, and set my total charge to +5, does each atom get a +1 charge?

For molecules with atoms with large electronegativity difference, I'm sure it is straightforward, but what about molecules like the ones described here?

Thanks!

I was trying to run FermiLib-plugin-psi4 for new molecules but I keep getting this error that .hdf5 file does not exist. The plugin runs fine for hydrogen molecule, for which data is already present in the data directory. Could you help me figure out what is going on?

Hi,

I'm getting negative one body integrals and negative overlap matrices when I run a simple calculation on hydrogen. I created a molecule by supplying geometry, basis (sto-3g), charge and multiplicity. I ran a SCF calculation (which converged with a reasonable energy). However, when I load this molecule and print out the overlap matrix and one-body tensor, they are diagonal matrices with all negative values on the diagonal. The two-body tensor seems to be okay.

Why is this happening? And if the stored overlap matrix had negative eigenvalues, how did the SCF calculation converge nicely?

Thank you!

It seems that we have accidentally included the GPL (as opposed to L-GPL) headers in some of the files in this repository. We should fix this. Since this might be construed as changing the license, we need explicitly permission from all contributors:
@jarrodmcc
@damiansteiger
@thomashaener
@idk3

Does everybody approve? Please just say "yes" if you do. Thanks.

Hi!

I had a few questions on how MolecularData accesses information and how onebody integrals are constructed.

1. How is molecule.one_body_integrals constructed? Is this from PSI4 or independent of it? I believe these integrals are in MO (since ao_kinetic+ao_potential != molecule.one_body_integrals), but PSI4 normally doesn't keep track of molecular orbital information, so where does this array come from?

2. When molecular.one_body_integrals is already in the MO basis, why does the template file convert the AO orbitals from PSI4 and transform them to the MO, only to set it equal to

molecule.one_body_integrals = one_body_basis_change(
            numpy.asarray(mints.ao_kinetic()), molecule.canonical_orbitals)
        molecule.one_body_integrals += one_body_basis_change(
numpy.asarray(mints.ao_potential()), molecule.canonical_orbitals) 

3. Finally, how does the plugin read the information generated by PSI4 (ie, how are the attributes of MolecularData objects filled)?

Sorry for bothering you with so many questions, I really appreciate any help!

Thanks!

It would be great to have some information about how to install the plugin in README.rst