Re: Errors in nonSCF steps..

[Donny Winston]

Are you able to see any warnings? From reading the pymatgen.symmetry.bandstructure module, it seems that kpath is set based on an inference of lattice type that can fail for a number of reasons, each emitting a separate warning (but not raising an exception) in the code.

On Sunday, November 8, 2015 at 5:21:41 PM UTC-8, Hark Lee wrote:

So pretty much all recent simulations that I submitted fizzled in "GGA_band_structure" steps, 

where nonSCF electronic calculations are attempted with charge densities provided from previous 

CHGCARs. Specifically, the error occurs while setting up new KPOINTS files from the results of the 

previous ionic structure relaxations. (FYI, I am just following the standard Matgen workflows, the 

material is HCP Mg + a few atoms of alloying metal elements)

As shown below, HighSymmpath.kpath from pymatgen/symmetry/bandstructure.py

to locate the high symmetry path in the brillouin zone and write KPOINTS appears to fail. 

Don't know if this is related to the other decoding issues brought up recently, but would really 

appreciate any help or suggestions!

Traceback (most recent call last):

  File "/global/u1/h/harklee/hark_mg/codes/fireworks/fireworks/core/rocket.py", line 211, in run

    m_action = t.run_task(my_spec)

  File "/global/u1/h/harklee/hark_mg/codes/MPWorks/mpworks/firetasks/vasp_setup_tasks.py", line 91, in run_task

    user_incar_settings=user_incar_settings, kpoints_line_density=self.kpoints_line_density)

  File "/global/u1/h/harklee/hark_mg/codes/pymatgen/pymatgen/io/vasp/sets.py", line 1264, in from_previous_vasp_run

    mpnscfvip.write_input(structure, output_dir, make_dir_if_not_present)

  File "/global/u1/h/harklee/hark_mg/codes/pymatgen/pymatgen/io/vasp/sets.py", line 161, in write_input

    for k, v in self.get_all_vasp_input(structure).items():

  File "/global/u1/h/harklee/hark_mg/codes/pymatgen/pymatgen/io/vasp/sets.py", line 129, in get_all_vasp_input

    kpoints = self.get_kpoints(structure)

  File "/global/u1/h/harklee/hark_mg/codes/pymatgen/pymatgen/io/vasp/sets.py", line 1142, in get_kpoints

    coords_are_cartesian=False)

  File "/global/u1/h/harklee/hark_mg/codes/pymatgen/pymatgen/symmetry/bandstructure.py", line 178, in get_kpoints

    for b in self.kpath['path']:

TypeError: 'NoneType' object has no attribute '__getitem__'

[Hark Lee]

Dear Donny (and Patrick), 

Thank you so much for the help, the original post was made out of desperation in the hopes that

there may be some others who have relevant experiences or are going through similar issues.

It turns out that it is not limited to the band structure calculations. Some of workflows fail at the DB 

insertion step right after GGA+U ionic relaxation, although the results don't look particularly suspicious. 

The only common issue here is that it happens usually after a not-so-perfect run, as I mostly see

a failed first attempt due to bad convergence.

I am compiling a list of errors I've encountered so far, I will pay you a visit (or post here) after

digging deeper. I deleted the original message since it may confuse others, as the issue here

is not necessarily limited to band structure calculations.

Again, thank you for the feedback!

Best, 

Hark