steabert
commented
4 years ago Hi!
First the bad news: I no longer maintain this project. I should have archived it a while ago, which I will do now I guess. It means that someone else is welcome to take over (the repo will still be readable). This might mean that there have been changes to the H5 format produced by OpenMolcas, but it can also be a bug of course.
I'm not sure if the OpenMolcas community is using molpy, but I would recommend you to check over there, and see if someone can help you, maybe someone is using this of maybe they have something even better :)
Hello guys, I have been trying to convert a H5 file generated by OpenMolcas (CASSCF). I am interested in the 2nd root and natural orbitals, i.e. S2. So I did, penny test.h5 -r 2 -n -c molden -f -o root2.molden
Then I want to convert to a wavefunction file (.wfn) using Molden2AIM. When doing so, the M2A fails to check the normalization simply because the numeric MOs occupancies are not the same as the analytical ones (see the error message below)
Reading basis functions... Computing the overlap matrix... Computing the integrated number of electrons... Sum of MO Occupancies = 130.0000000000 Analytically integrated number of electrons = 126.0416192381 Difference = 3.9583807619 Difference per atom = 0.2827414830 ## Warning! Normalization check failed.
I guess I am doing something wrong but most likely Penny does not convert H5 files properly.
I hope someone could help me out with that issue. Thanks in advance!
PS: I know that I can simply take the *rasscf.molden.2 and use it directly in Molden2AIM to get the wavefunction but the MOs are structured as "open-shell", not "closed-shell" as I want.