FCIDUMP support

[jeffhammond]

The FCIDUMP format [1] seems to be rather popular for supporting implementations of exotic methods like FCIQMC. As noted in the HANDE manual, Molpro, Q-Chem, and PSI4 all support this format. PySCF (http://sunqm.github.io/pyscf/tools.html) and Horton (https://en.wikipedia.org/wiki/HORTON_(software)) also support this format.

I have written something along these lines before and will attempt to create a proper implementation unless someone else gets their first. We do not document support for it that I can find, so if it is already supported, it is hidden from users.

[1] P. J. Knowles, N. C. Handy, Comp. Phys. Commun. 1989, 54, 75. A determinant based full configuration interaction program. https://doi.org/10.1016/0010-4655(89)90033-7

[dvorakm1]

Hi,
Did you make any progress with the FCIDUMP for NWChem? I would like to get all the one- and two-body integrals in the MO basis for my own external code. I have been experimenting myself with the rimp2 module to get these but can't quite get it to work.
Thanks,
Marc

[jeffhammond]

No, but please look at #23 (comment), as you might be able to implement it in the TCE module yourself using the eionly method I noted.

[safischer]

Please, see #23 (comment) for a method to dump the one and two electron integrals in the molecular orbital basis to a file.

[edoapra]

A preliminary implementation of FCIDUMP, following @safischer comment, is now available in the master branch.
Documentation at https://nwchemgit.github.io/FCIDUMP.html
Input example at https://github.com/nwchemgit/nwchem/blob/master/QA/tests/fcidump_hf/fcidump_hf.nw