Dear colleagues,

I am now starting to investigate how to extract information automatically from QM code outputs. The tools for which I have found so far for this job are jumbo-converters, cclib, Torbjörn's et al. vasp2cif and, of course, the good old Perl :).

To get some impression of how the inputs and outputs look like, I have run a bunch of open source codes on my computer (so far, abinit, mopac7, mpqc, nwchem, psi3 and quantum-espresso -- in alphabetic order, these are those that could be be installed using the default package repositories on Ubuntu 12.04). I managed to run some examples and tutorials provided with each of these packages (http://www.crystallography.net/archives/2014/software/qchem/), and started to play with the extraction of the data there (http://www.crystallography.net/archives/2014/software/qchem/abinit/make/bin/...) a couple of weeks ago. We have also discussed with Peter that I could also try to make Quantum Espresso log analyser using Jumbo-converters, unless, of course, QE has some log converter already.

I am, however, not an expert in QM computations, so I am not in a position to run a really good optimisation, the one that would be worth depositing to TCOD. May I therefore ask you to send me some samples quality runs of your favourite codes -- preferably the open source ones, and, if possible, the ones which I may expose publicly. Maybe some of your students make standard exercisers which you could share. Quantum-espresso and NWChem are especially interesting in this context here ;) If you would prefer that the files you send me are kept private, please let me know about this.

VASP inputs and outputs would be handy as well; though I can not rerun them I can analyse the logs. And I have an old example from Stefaan using W2K (9008878_W2K_PBE_06.5_00250.tar.gz), but I would appreciate a citation of the command that was used to run it.

I will then try different data extraction methods and see how they can be integrated into the TCOD deposition system.

Sincerely yours, Saulius