mkdos.py (DOS)

mkdos is a python module to work with DOSCAR files. It consists of a class “dos” that lets you read and output information. Most important information is stored in self.ens, self.tot, and self.pdos. Information is retrievable using the get_val() methods.

Python usage

Here is an example script to display s,p,d PDOS summed over atoms 1-4, separating spin-up and spin-down:

#!/usr/bin/env python
from mkdos import dos
import matplotlib.pyplot as plt

dd=dos('DOSCAR')
arr,lab=dd.get_perl(range(4))
colors='b g r c m y k'.split()
for ispin,spin in enumerate([1,-1]):
  for l in range(3):
    plt.plot(dd.ens,spin*arr[ispin,l,:],colors[ispin*3+l])
ax=plt.gca()
ax.legend(lab[1]*2)
plt.show()

Which outputs this:

../_images/mkdos_example.png
Short methods descriptions (see docstrings for details):
  • init Input a DOSCAR, optionally specify spin-polarization, and whether to shift fermi energy to zero
  • get_energies returns vector of energies
  • get_peratom returns array of PDOS for each atom. You can specify to sum spin up and spin dn.
  • get_perl returns array of PDOS for each orbital quantum number. You can specify which atoms to sum over, and/or to sum spin-up and spin-dn.
  • get_perm returns array of PDOS for each l_m for a given l. You can specify which atoms and/or summing spin-up and spin-dn.
  • get_total returns vector of total DOS
  • get_totspin returns vector of total DOS spin-up then spin-dn
  • set_pdos shouldn’t be necessary to use; it’s just for reading the DOSCAR
  • set_total shouldn’t be necessary to use; it’s just for reading the DOSCAR
  • shift_fermi shouldn’t be necessary to use after instantiation.
Notes:

  • get_PDOS() methods return a tuple of the form (out,label), where label tells you what it is you got

  • spin-polarized get_PDOS()[0] may return a 3d array, where the first index is for spin up or down.

  • Note that when identifying atoms, indexing is from 0. This is a very easy bug to squish.

  • If you want to plot s,p,d of atoms 1,2, it is easiest to do something like:

    np.concatenate((dd.get_perl(0,True)[0],dd.get_perl(1,True)[0]),axis=0)

Command-line usage

Like boomerang and periodica, there is a command-line interface, which is used via command-line arguments. It is easily piped to gracey.

  • filename can be inserted anywhere in the arguments. Default is DOSCAR, then DOSCAR.static.
  • noshiftfermi: Do not shift energies to set fermi energy to zero
  • sp: Try to read DOSCAR with spin polarization off (sp=1) or on (sp=2); default reads INCAR to find out.
Only one of the following; they are gone through in order:
  • total: Print total DOS (each line: en, totdos)
  • atom: Print per-atom DOS (each line: en, at1, at2, ...)
  • l: Print per-l DOS (each line: en, s, p, d). Note that spin-up and spin-down are added, and there is no option on the command-line to choose only some atoms. Python has that capability.
  • m=int: Print per-m DOS for the given l=int (0,1,2=s,p,d). Output is each line: en,m1,m2,m3.... Note that spin-up and spin-down are added, and there is no option on the command-line to choose only some atoms. Python has that capability. Also the order of m is not what you may expect, but is the same as in DOSCAR. See http://cms.mpi.univie.ac.at/vasp-forum/forum_viewtopic.php?3.306

Here is an example:

mkdos.py DOSCAR -total | gracey.py

Docstrings from mkdos.py

class mkdos.dos(doscar, sp=None, noshiftfermi=False)

Class to interpret DOS from a DOSCAR file.

Developed by Mordechai Kornbluth Apr 2014, based on Chris’ dformat.py

Suggestions / feature requests are welcome

get_energies()

Returns vector of energies. Same as self.ens

get_peratom(sumspins=False)

Returns array of DOS per atom, with labels (e.g. for plotting): return out,label

sumspins = If system is spin-polarized, add up and dn for output

if sp && !sumspins: indices are [spin,atomnum,energy]

else: indices are [atomnum,energy]

get_perl(atoms=None, sumspins=False)

Returns array of DOS for (s,p,d): return out,label

atoms = which atoms to sum over, indexed from 0. Can be list, integer, array, or None (meaning sum over all atoms; this is default)

sumspins = If system is spin-polarized, add up and dn for output

if sp && !sumspins: indices are [spin,l,energy]. Note label is [[‘up’,’dn’],[‘s’,’p’,’d’]]

else: indices are [l,energy]

get_perm(l, atoms=None, sumspins=False)

Returns array of DOS for various m values of the given l: return out,label

l = 0,1,2 (s,p,d)

atoms = which atoms to sum over, indexed from 0. Can be list, integer, array, or None (meaning sum over all atoms; this is default)

sumspins = If system is spin-polarized, add up and dn for output

if sp && !sumspins: indices are [spin,m,energy]. Note label is e.g. [[‘up’,’dn’],[‘py’,’pz’,’px’]]

else: indices are [m,energy]

get_total()

Returns vector of total DOS

get_totspin()

Returns vector of up DOS, then dn DOS

set_pdos(inplines)

Sets input lines (array) to the partial dos as given in DOSCAR

Input format: natoms * (Header line, nedos lines of PDOS info)

Stores in array self.pdos: 1st index = energy; 2nd = atom num; 3rd = orbital; (4th = spin if applicable)

set_total(inplines)

Sets input lines (array) to the total dos as given in DOSCAR

Also reads energies from this information

Input should be array of strings, length nedos .

Stored in self.tot, and (if sp) self.totup and self.totdn

shift_fermi()

Sets Fermi energy to 0