Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
| @@ -0,0 +1,280 @@ | |||
| { | |||
There was a problem hiding this comment.
Line #63. output["generic"]["steps"] = len(vr.ionic_steps)
The definition of steps here is not the total number of ionic steps it is a method (as I understand) that allows the user to do something like: plt.plot(job.output.steps, job.output.energy_tot). In other words it keeps an index for each ionic step. However, I agree that the parsing is wrong (there is the get_steps in the outcar.py which is never called) and then we end up using the arrange(energy) thing in the output.py.
Reply via ReviewNB
| @@ -0,0 +1,280 @@ | |||
| { | |||
There was a problem hiding this comment.
Line #64. # Another grievous offender that pollutes the interface for no good reason...
The structure object in the generic output is only for the first and final structures: job['input/structure/positions'] and job['output/structure/positions']. The positions for the entire trajectory can be obtained only via job['output/generic/positions'] So I don't think we double the storage or save the trajectory twice (as I understand how we currently parse things).
Reply via ReviewNB
2 participants
contains a notebook for parsing with pymatgen - drop in replacement.
Some notes:
pymatgen outcar parsing is limited to last step parsing only; this is problematic for certain properties (magmoms), etc.
in my testing, pymatgen is about 10% slower in real files, (0.3s per parse of my typical output folder) vs pyiron's current parsing.
This might narrow with extremely large vasp runs, but it also might not.
@ahmedabdelkawy can you test with ab-initio md? just replace path_real with the path that you'd like to test.