orca.cis is not parsed when TDDFT core excitations are specified

[zacmathe]

I am interested in using TheoDORE to analyze X-ray absorption transitions calculated using TDDFT XAS in ORCA. In this procedure, excitations are allowed only from user-specified core orbitals.

TheoDORE is not able to read the orca.cis file when core excitations are specified. I have made two minimal TDDFT calculations, one including all excitations (the default; for e.g. UVVis), and one specifying a donor orbital (for XAS). The former can be analyzed using analyze_tden.py as expected, but the latter cannot:

...
Reading CI vectors from binary ORCA file orca.cis
Number of vectors: 5
! Reading ALPHA coefficients ...
  nmo: 33 , nocc: 1 , nact: 1 , nvir: 32
Traceback (most recent call last):
  File "/Users/zm/science/theodore/TheoDORE_2.4/bin/analyze_tden.py", line 56, in <module>
    tdena.read_dens()
  File "/Users/zm/science/theodore/TheoDORE_2.4/theodore/dens_ana_base.py", line 95, in read_dens
    self.state_list = ccli.read(self.mos)
  File "/Users/zm/science/theodore/TheoDORE_2.4/theodore/cclib_interface.py", line 76, in read
    self.tden_orca(state_list, mos, rect_dens)
  File "/Users/zm/science/theodore/TheoDORE_2.4/theodore/cclib_interface.py", line 258, in tden_orca
    raise error_handler.MsgError("Inconsistent number of MOs")
theodore.error_handler.MsgError:

  ERROR: Inconsistent number of MOs

I have attached a zip containing all files from the two orca calculations, plus the piped TheoDORE output. For convenience, here is the orca input file used for both calculations. It should run in a couple seconds on a laptop.

! RKS BP86 Def2-TZVP Def2/J

%tddft 
    OrbWin[0] = 0, 0, -1, -1 # Allow excitations only from orbital 0 for XAS. Omit this line for UVVis.  
    NRoots 5
    end

* xyz 0 1
F   0   0   0
H   0   0   0.9168
*

Addionally, I had to make one trivial modification to TheoDORE_2.4/cclib/parser/orcaparser.py, since we are using ORCA 5 and the output syntax has changed slightly since ORCA 4. Line 782 was changed from: self.etenergies.append(float(broken[-2])) to self.etenergies.append(float(broken[-5])) We are using ORCA 5 because it is a major upgrade to the program. For example, TDDFT calculations on large transition metal complexes run multiple times faster.

Please let me know if I can provide further information.

HF.zip

[felixplasser]

ok! This looks like this may just be a trivial error in terms of the dimension of the expected matrix. I'll see what can be done.

In terms of Orca 5, yes there seem to be some changes. I am not quite sure if that was fixed in cclib yet (which is the external library used to parse Orca).

[RagnarB83]

Hi, any update on this issue? I would also be interested in this. Theodore looks like it would be quite useful for analyzing XAS spectra using ORCA so it would be great if this could be fixed.

[felixplasser]

If either of you wants to fix it, then I can tell you where to look. Otherwise maybe in a month or so ...

[felixplasser]

Sorry, I do not have the time to work on general feature requests.

We can think about it, if you want to turn this into a joint publication.

[RagnarB83]

I will try to find time to take a look at this myself.