Psi4Integrals::update_orbitals was passed two different sets of orbitals

[fevangelista]

The following input raises the error:

!---------------------------------------------------------------------------!
!                                                                           !
!  Psi4Integrals::update_orbitals was passed two different sets of orbitals !
!   but the integral object assumes restricted orbitals                     !
!                                                                           !
!---------------------------------------------------------------------------!

Input

import forte

molecule Hdimer{
  H
  H 1 0.5999999999999999
}

set globals{
  basis                   cc-pVDZ
  reference               rhf
  scf_type                direct
  d_convergence           6
  e_convergence           12
}

set {
  docc                    [1,0,0,0,0,0,0,0]
  maxiter                 1000
  level_shift             0.1
}

set forte{
  active_space_solver     cas
  correlation_solver      sa-mrdsrg
  corr_level              dsrg-mrpt2
  frozen_docc             [0,0,0,0,0,0,0,0]
  restricted_docc         [0,0,0,0,0,0,0,0]
  active                  [1,0,0,0,0,1,0,0]
  casscf_reference        true
  root_sym                0
  nroot                   1
  dsrg_s                  0.1
  e_convergence           8
  r_convergence           6
}

Emcscf, wfn = energy('scf', return_wfn=True)
Eldsrg2 = energy('forte',ref_wfn=wfn)

[JonathonMisiewicz]

I've done some investigation into this one. At the time the CASSCF completes, the alpha orbitals have been updated to their final values, but the beta orbitals have not. The call to rotate orbitals becomes a call to update orbitals, which raises the error.

The fault is in the CASSCF procedure. The beta orbitals should be updated alongside the alpha orbitals. I leave fixing that to people more familiar with the CAS code.

[fevangelista]

@JonathonMisiewicz: thanks for finding the problem. One more reason to eliminate beta orbitals from Forte.

[JonathonMisiewicz]

Oh, so you want Forte to deal with spin-restricted orbitals only? (RHF and ROHF, ignore UHF?)