I've been having trouble with a specific library and have been considering just writing something myself as the calculation isn't so complicated (just getting the energy as a function of a reaction coordinate). As the library uses the huge vasprun.xml file to source the energy it takes quite a while to run (I'm doing the analysis locally) and am thinking about just implementing the analysis manually. Is there a good reason why I shouldn't just take the final electronic energies from the OUTCARs (or OSZICAR) along the path and plot them myself?
Also, why isn't this what the library does? Is it simply because the XML file has so much extra information so they just tend to import it so they don't need to ask the user to copy over the other VASP output files? It seems quite wasteful for my current application (changing some coordinates around some minima, doing some single-point calculations along some paths and then importing the energies).
The advantage of vasp.xml is that everything is tagged and designed to be machine-readable. ASE will let you read the OUTCAR or the vasprun.xml. I think pymatgen will read either but can get a lot more info from the vasprun.xml. Looks like vaspkit can read both but I'm not sure what information it extracts from each. I'm a bit surprised that your library is very slow at reading the vasprun.xml--in my experience it usually takes less than 30 seconds.
Anyways, if you prefer to read from OUTCAR or OSZICAR, that should be fine. There are already libraries that do this, but it's easy to write your own function, as you say. Just make sure you're reading the version of the energy that you're interested in.
- Badges
- Science topic