def read_data(self):
r""" Read DOS, as calculated and written by VASP
Returns
-------
E : numpy.ndarray
energy points (in eV)
DOS : numpy.ndarray
DOS points (in 1/eV)
"""
# read first line
line = self.readline() # NIONS, NIONS, JOBPAR_, WDES%INCDIJ
line = self.readline() # AOMEGA, LATT_CUR%ANORM(1:3) *1e-10, POTIM * 1e-15
line = self.readline() # TEMP
line = self.readline() # ' CAR '
line = self.readline() # name
line = self.readline().split()
Emax = float(line[0])
Emin = float(line[1])
NE = int(line[2])
Ef = float(line[3])
E = np.empty([NE], np.float32)
# Determine output
line = arrayf(self.readline().split())
ns = (len(line) - 1) // 2
DOS = np.empty([ns, NE], np.float32)
E[0] = line[0]
DOS[:, 0] = line[1:ns+1]
for ie in range(1, NE):
line = arrayf(self.readline().split())
E[ie] = line[0]
DOS[:, ie] = line[1:ns+1]
return E - Ef, DOS
def _voronoi_hirshfeld_charges():
""" Read output from Voronoi/Hirshfeld charges """
nonlocal pd
# Expecting something like this:
# Voronoi Atomic Populations:
# Atom # dQatom Atom pop S Sx Sy Sz Species
# 1 -0.02936 4.02936 0.00000 -0.00000 0.00000 0.00000 C
# Define the function that parses the charges
def _parse_charge(line):
atom_idx, *vals, symbol = line.split()
# assert that this is a proper line
# this should catch cases where the following line of charge output
# is still parseable
atom_idx = int(atom_idx)
return list(map(float, vals))
# first line is the header
header = (
self.readline().replace("dQatom", "dq") # dQatom in master
.replace(" Qatom", " dq") # Qatom in 4.1
.replace("Atom pop", "e") # not found in 4.1
.split())[2:-1]
# We have found the header, prepare a list to read the charges
atom_charges = []
line = ' '
while line != "":
try:
line = self.readline()
charge_vals = _parse_charge(line)
atom_charges.append(charge_vals)
except:
# We already have the charge values and we reached a line that can't be parsed,
# this means we have reached the end.
break
if pd is None:
# not as_dataframe
return _a.arrayf(atom_charges)
# determine how many columns we have
# this will remove atom indices and species, so only inside
ncols = len(atom_charges[0])
assert ncols == len(header)
# the precision is limited, so no need for double precision
return pd.DataFrame(atom_charges,
columns=header,
dtype=np.float32,
index=pd.RangeIndex(stop=len(atom_charges),
name="atom"))
def _empty_charge():
# return for single value with nan values
return _a.arrayf([[None]])