def _ngl_write_structure(elements, positions, cell):
"""
Turns structure information into a NGLView-readable protein-database-formatted string.
Args:
elements (numpy.ndarray/list): Element symbol for each atom.
positions (numpy.ndarray/list): Vector of Cartesian atom positions.
cell (numpy.ndarray/list): Simulation cell Bravais matrix.
Returns:
(str): The PDB-formatted representation of the structure.
"""
from ase.geometry import cell_to_cellpar, cellpar_to_cell
if cell is None or any(np.max(cell, axis=0) < 1e-2):
# Define a dummy cell if it doesn't exist (eg. for clusters)
max_pos = np.max(positions, axis=0)
max_pos[np.abs(max_pos) < 1e-2] = 10
cell = np.eye(3) * max_pos
cellpar = cell_to_cellpar(cell)
exportedcell = cellpar_to_cell(cellpar)
rotation = np.linalg.solve(cell, exportedcell)
pdb_str = _ngl_write_cell(*cellpar)
pdb_str += "MODEL 1\n"
if rotation is not None:
positions = np.array(positions).dot(rotation)
for i, p in enumerate(positions):
pdb_str += _ngl_write_atom(i, elements[i], *p)
pdb_str += "ENDMDL \n"
return pdb_str
def read_mustem(fd):
r"""Import muSTEM input file.
Reads cell, atom positions, etc. from muSTEM xtl file.
The mustem xtl save the root mean square (RMS) displacement, which is
convert to Debye-Waller (in Ų) factor by:
.. math::
B = RMS * 8\pi^2
"""
check_numpy_version()
from ase.geometry import cellpar_to_cell
# Read comment:
fd.readline()
# Parse unit cell parameter
cellpar = [float(i) for i in fd.readline().split()[:3]]
cell = cellpar_to_cell(cellpar)
# beam energy
fd.readline()
# Number of different type of atoms
element_number = int(fd.readline().strip())
# List of numpy arrays:
# length of the list = number of different atom type (element_number)
# length of each array = number of atoms for each atom type
atomic_numbers = []
positions = []
debye_waller_factors = []
occupancies = []
for i in range(element_number):
# Read the element
_ = fd.readline()
line = fd.readline().split()
atoms_number = int(line[0])
atomic_number = int(line[1])
occupancy = float(line[2])
DW = float(line[3]) * 8 * np.pi**2
# read all the position for each element
positions.append(np.genfromtxt(fname=fd, max_rows=atoms_number))
atomic_numbers.append(np.ones(atoms_number, dtype=int) * atomic_number)
occupancies.append(np.ones(atoms_number) * occupancy)
debye_waller_factors.append(np.ones(atoms_number) * DW)
positions = np.vstack(positions)
atoms = Atoms(cell=cell, scaled_positions=positions)
atoms.set_atomic_numbers(np.hstack(atomic_numbers))
atoms.set_array('occupancies', np.hstack(occupancies))
atoms.set_array('debye_waller_factors', np.hstack(debye_waller_factors))
return atoms
def from_string(data):
"""
Reads a Res from a string.
Args:
data (str): string containing Res data.
Returns:
Res object.
"""
abc = []
ang = []
sp = []
coords = []
info = dict()
coord_patt = re.compile("""(\w+)\s+
([0-9]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)""", re.VERBOSE)
lines = data.splitlines()
line_no = 0
while line_no < len(lines):
line = lines[line_no]
tokens = line.split()
if tokens:
if tokens[0] == 'TITL':
try:
info = Res.parse_title(line)
except (ValueError, IndexError):
info = dict()
elif tokens[0] == 'CELL' and len(tokens) == 8:
abc = [float(tok) for tok in tokens[2:5]]
ang = [float(tok) for tok in tokens[5:8]]
elif tokens[0] == 'SFAC':
for atom_line in lines[line_no:]:
if line.strip() == 'END':
break
else:
match = coord_patt.search(atom_line)
if match:
sp.append(match.group(1)) # 1-indexed
cs = match.groups()[2:5]
coords.append([float(c) for c in cs])
line_no += 1 # Make sure the global is updated
line_no += 1
return Res(Atoms(symbols=sp,
scaled_positions=coords,
cell=cellpar_to_cell(list(abc) + list(ang)),
pbc=True, info=info),
info.get('name'),
info.get('pressure'),
info.get('energy'),
info.get('spacegroup'),
info.get('times_found'))
def gen_STO():
a = b = c = 3.94
alpha = beta = theta = 90
atoms = Atoms(symbols='SrTiO3',
scaled_positions=[(0, 0, 0), (0.5, 0.5, 0.5), (0, 0.5, 0.5),
(0.5, 0, 0.5), (0.5, 0.5, 0)],
cell=cellpar_to_cell([a, b, c, alpha, beta, theta]))
atoms = atoms.repeat([1, 1, 2])
#atoms.set_initial_magnetic_moments()
return atoms
def read_res(lines):
"""
Reads a res file from a string
Args:
lines (str): A list of lines containing Res data.
Returns:
System object that is used internally by Dscribe
"""
abc = []
ang = []
species = []
coords = []
info = dict()
coord_patt = re.compile(
r"""(\w+)\s+
([0-9]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)\s+
([0-9\-\.]+)""", re.VERBOSE)
line_no = 0
title_items = []
while line_no < len(lines):
line = lines[line_no]
tokens = line.split()
if tokens:
if tokens[0] == 'TITL':
# Skip the TITLE line, the information is not used
# in this package
title_items = parse_titl(line)
elif tokens[0] == 'CELL' and len(tokens) == 8:
abc = [float(tok) for tok in tokens[2:5]]
ang = [float(tok) for tok in tokens[5:8]]
elif tokens[0] == 'SFAC':
for atom_line in lines[line_no:]:
if line.strip() == 'END':
break
match = coord_patt.search(atom_line)
if match:
species.append(match.group(1)) # 1-indexed
xyz = match.groups()[2:5]
coords.append([float(c) for c in xyz])
line_no += 1 # Make sure the global is updated
line_no += 1
return title_items, Atoms(symbols=species,
scaled_positions=coords,
cell=cellpar_to_cell(list(abc) + list(ang)),
pbc=True,
info=info)
def write_proteindatabank(fileobj, images, write_arrays=True):
"""Write images to PDB-file."""
if isinstance(fileobj, basestring):
fileobj = paropen(fileobj, 'w')
if hasattr(images, 'get_positions'):
images = [images]
rotation = None
if images[0].get_pbc().any():
from ase.geometry import cell_to_cellpar, cellpar_to_cell
currentcell = images[0].get_cell()
cellpar = cell_to_cellpar(currentcell)
exportedcell = cellpar_to_cell(cellpar)
rotation = np.linalg.solve(currentcell, exportedcell)
# ignoring Z-value, using P1 since we have all atoms defined explicitly
format = 'CRYST1%9.3f%9.3f%9.3f%7.2f%7.2f%7.2f P 1\n'
fileobj.write(format % (cellpar[0], cellpar[1], cellpar[2], cellpar[3],
cellpar[4], cellpar[5]))
# 1234567 123 6789012345678901 89 67 456789012345678901234567 890
format = ('ATOM %5d %4s MOL 1 %8.3f%8.3f%8.3f%6.2f%6.2f'
' %2s \n')
# RasMol complains if the atom index exceeds 100000. There might
# be a limit of 5 digit numbers in this field.
MAXNUM = 100000
symbols = images[0].get_chemical_symbols()
natoms = len(symbols)
for n, atoms in enumerate(images):
fileobj.write('MODEL ' + str(n + 1) + '\n')
p = atoms.get_positions()
occupancy = np.ones(len(atoms))
bfactor = np.zeros(len(atoms))
if write_arrays:
if 'occupancy' in atoms.arrays:
occupancy = atoms.get_array('occupancy')
if 'bfactor' in atoms.arrays:
bfactor = atoms.get_array('bfactor')
if rotation is not None:
p = p.dot(rotation)
for a in range(natoms):
x, y, z = p[a]
occ = occupancy[a]
bf = bfactor[a]
fileobj.write(
format %
(a % MAXNUM, symbols[a], x, y, z, occ, bf, symbols[a].upper()))
fileobj.write('ENDMDL\n')
def motif_cell_fromCrystal(crystal, fill_cell=False):
"""
ccdc.crystal.Crystal --> np.array shape (m,3), np.array shape (3,3).
Optional param fill_cell (default False) will expand an asymmetric unit to
fill the unit cell. This is necessary if the .cif contains an asymmetric unit
that needs to be expanded by symmetry operations.
"""
from ase.geometry import cellpar_to_cell
cell = cellpar_to_cell([*crystal.cell_lengths, *crystal.cell_angles])
mol = crystal.molecule if not fill_cell else crystal.packing(
inclusion='OnlyAtomsIncluded')
motif = np.array([[a.coordinates.x, a.coordinates.y, a.coordinates.z]
for a in mol.atoms])
return motif, cell
def _get_cell(text):
# first check whether there is a lattice definition
cell = np.zeros((3, 3))
lattice = _cell_3d.findall(text)
if lattice:
pbc = [True, True, True]
for i, row in enumerate(lattice[0].strip().split('\n')):
cell[i] = [float(x) for x in row.split()]
return cell, pbc
pbc = [False, False, False]
lengths = [None, None, None]
angles = [None, None, None]
for row in text.strip().split('\n'):
row = row.strip().lower()
for dim, vecname in enumerate(['a', 'b', 'c']):
if row.startswith('lat_{}'.format(vecname)):
pbc[dim] = True
lengths[dim] = float(row.split()[1])
for i, angle in enumerate(['alpha', 'beta', 'gamma']):
if row.startswith(angle):
angles[i] = float(row.split()[1])
if not np.any(pbc):
return None, pbc
for i in range(3):
a, b, c = np.roll(np.array([0, 1, 2]), i)
if pbc[a] and pbc[b]:
assert angles[c] is not None
if angles[c] is not None:
assert pbc[a] and pbc[b]
# The easiest case: all three lattice vectors and angles are specified
if np.all(pbc):
return cellpar_to_cell(lengths + angles), pbc
# Next easiest case: exactly one lattice vector has been specified
if np.sum(pbc) == 1:
dim = np.argmax(pbc)
cell[dim, dim] = lengths[dim]
return cell, pbc
# Hardest case: two lattice vectors are specified.
dim1, dim2 = [dim for dim, ipbc in enumerate(pbc) if ipbc]
angledim = np.argmin(pbc)
cell[dim1, dim1] = lengths[dim1]
cell[dim2, dim2] = lengths[dim2] * np.sin(angles[angledim])
cell[dim2, dim1] = lengths[dim2] * np.cos(angles[angledim])
return cell, pbc
def read_eon(fileobj):
"""Reads an EON reactant.con file. If *fileobj* is the name of a
"states" directory created by EON, all the structures will be read."""
if isinstance(fileobj, str):
if (os.path.isdir(fileobj)):
return read_states(fileobj)
else:
f = open(fileobj)
else:
f = fileobj
comment = f.readline().strip()
f.readline() # 0.0000 TIME (??)
cell_lengths = f.readline().split()
cell_angles = f.readline().split()
# Different order of angles in EON.
cell_angles = [cell_angles[2], cell_angles[1], cell_angles[0]]
cellpar = [float(x) for x in cell_lengths + cell_angles]
f.readline() # 0 0 (??)
f.readline() # 0 0 0 (??)
ntypes = int(f.readline()) # number of atom types
natoms = [int(n) for n in f.readline().split()]
atommasses = [float(m) for m in f.readline().split()]
symbols = []
coords = []
masses = []
fixed = []
for n in range(ntypes):
symbol = f.readline().strip()
symbols.extend([symbol] * natoms[n])
masses.extend([atommasses[n]] * natoms[n])
f.readline() # Coordinates of Component n
for i in range(natoms[n]):
row = f.readline().split()
coords.append([float(x) for x in row[:3]])
fixed.append(bool(int(row[3])))
if isinstance(fileobj, str):
f.close()
atoms = Atoms(symbols=symbols,
positions=coords,
masses=masses,
cell=cellpar_to_cell(cellpar),
constraint=FixAtoms(mask=fixed),
info=dict(comment=comment))
return atoms
def read_mustem(filename):
"""Import muSTEM input file.
Reads cell, atom positions, etc. from muSTEM xtl file
"""
from ase import Atoms
from ase.geometry import cellpar_to_cell
if isinstance(filename, str):
f = open(filename)
else: # Assume it's a file-like object
f = filename
# Read comment:
f.readline()
# Parse unit cell parameter
cellpar = [float(i) for i in f.readline().strip().split()[:3]]
cell = cellpar_to_cell(cellpar)
# beam energy
f.readline()
# Number of different type of atoms
element_number = int(f.readline().strip())
symbols = []
positions = []
for i in range(element_number):
# Read the element
symbol = str(f.readline().strip())
atoms_number = int(f.readline().split()[0])
# read all the position for each element
for j in range(atoms_number):
line = f.readline()
positions.append([float(i) for i in line.strip().split()])
symbols.append(symbol)
f.close()
atoms = Atoms(cell=cell, scaled_positions=positions)
atoms.set_chemical_symbols(symbols)
return atoms
def set_substrate(atoms,
a=None,
b=None,
c=None,
angle_ab=90.0,
all_angle_90=False,
m=1,
fix_volume=False):
"""
set atoms cellpars to fit the substrate cellpars.
a,b , angle_ab are the the inplane cellpars of the substrate
all_angle_90: if True,all the angles will be set to 90
m is the multiplier. a*m b*m
fix_volume: whether to set c so that the volume is unchanged.
Note that this is not always really the case if angles of a-b plane and c is changed.
"""
cellpars = cell_to_cellpar(atoms.get_cell())
print(a, b, c)
a0, b0, c0 = cellpars[0:3]
if a is not None:
cellpars[0] = a * m
if b is not None:
cellpars[1] = b * m
if c is not None:
cellpars[2] = c * m
if angle_ab is not None:
cellpars[5] = angle_ab
if all_angle_90:
cellpars[3:] = [90, 90, 90]
print(cellpars)
if fix_volume:
ab = cellpars[5]
c = (a0 * b0 * c0 * math.sin(math.radians(ab)) /
(a * b * m * m * math.sin(math.radians(angle_ab))))
cellpars[2] = c
cell = cellpar_to_cell(cellpars)
print(cell)
spos = atoms.get_scaled_positions()
atoms.set_cell(cell)
atoms.set_scaled_positions(spos)
return atoms
def write_v_sim(filename, atoms):
"""Write V_Sim input file.
Writes the atom positions and unit cell.
"""
from ase.geometry import cellpar_to_cell, cell_to_cellpar
if isinstance(filename, str):
f = open(filename)
else: # Assume it's a file-like object
f = filename
# Convert the lattice vectors to triangular matrix by converting
# to and from a set of lengths and angles
cell = cellpar_to_cell(cell_to_cellpar(atoms.cell))
dxx = cell[0, 0]
dyx, dyy = cell[1, 0:2]
dzx, dzy, dzz = cell[2, 0:3]
f.write('===== v_sim input file created using the'
' Atomic Simulation Environment (ASE) ====\n')
f.write('{0} {1} {2}\n'.format(dxx, dyx, dyy))
f.write('{0} {1} {2}\n'.format(dzx, dzy, dzz))
# Use v_sim 3.5 keywords to indicate scaled positions, etc.
f.write('#keyword: reduced\n')
f.write('#keyword: angstroem\n')
if np.alltrue(atoms.pbc):
f.write('#keyword: periodic\n')
elif not np.any(atoms.pbc):
f.write('#keyword: freeBC\n')
elif np.array_equiv(atoms.pbc, [True, False, True]):
f.write('#keyword: surface\n')
else:
raise Exception(
'Only supported boundary conditions are full PBC,'
' no periodic boundary, and surface which is free in y direction'
' (i.e. Atoms.pbc = [True, False, True]).')
# Add atoms (scaled positions)
for position, symbol in zip(atoms.get_scaled_positions(),
atoms.get_chemical_symbols()):
f.write('{0} {1} {2} {3}\n'.format(position[0], position[1],
position[2], symbol))
def test_monoclinic():
"""Test band structure from different variations of hexagonal cells."""
import numpy as np
from ase import Atoms
from ase.calculators.test import FreeElectrons
from ase.geometry import (crystal_structure_from_cell, cell_to_cellpar,
cellpar_to_cell)
from ase.dft.kpoints import get_special_points
mc1 = [[1, 0, 0], [0, 1, 0], [0, 0.2, 1]]
par = cell_to_cellpar(mc1)
mc2 = cellpar_to_cell(par)
mc3 = [[1, 0, 0], [0, 1, 0], [-0.2, 0, 1]]
mc4 = [[1, 0, 0], [-0.2, 1, 0], [0, 0, 1]]
path = 'GYHCEM1AXH1'
firsttime = True
for cell in [mc1, mc2, mc3, mc4]:
a = Atoms(cell=cell, pbc=True)
a.cell *= 3
a.calc = FreeElectrons(nvalence=1, kpts={'path': path})
cs = crystal_structure_from_cell(a.cell)
assert cs == 'monoclinic'
r = a.cell.reciprocal()
k = get_special_points(a.cell)['H']
print(np.dot(k, r))
a.get_potential_energy()
bs = a.calc.band_structure()
coords, labelcoords, labels = bs.get_labels()
assert ''.join(labels) == path
e_skn = bs.energies
# bs.plot()
if firsttime:
coords1 = coords
labelcoords1 = labelcoords
e_skn1 = e_skn
firsttime = False
else:
for d in [
coords - coords1, labelcoords - labelcoords1,
e_skn - e_skn1
]:
print(abs(d).max())
assert abs(d).max() < 1e-13, d
def write_v_sim(filename, atoms):
"""Write V_Sim input file.
Writes the atom positions and unit cell.
"""
from ase.geometry import cellpar_to_cell, cell_to_cellpar
if isinstance(filename, str):
f = open(filename)
else: # Assume it's a file-like object
f = filename
# Convert the lattice vectors to triangular matrix by converting
# to and from a set of lengths and angles
cell = cellpar_to_cell(cell_to_cellpar(atoms.cell))
dxx = cell[0, 0]
dyx, dyy = cell[1, 0:2]
dzx, dzy, dzz = cell[2, 0:3]
f.write('===== v_sim input file created using the'
' Atomic Simulation Environment (ASE) ====\n')
f.write('{0} {1} {2}\n'.format(dxx, dyx, dyy))
f.write('{0} {1} {2}\n'.format(dzx, dzy, dzz))
# Use v_sim 3.5 keywords to indicate scaled positions, etc.
f.write('#keyword: reduced\n')
f.write('#keyword: angstroem\n')
if np.alltrue(atoms.pbc):
f.write('#keyword: periodic\n')
elif not np.any(atoms.pbc):
f.write('#keyword: freeBC\n')
elif np.array_equiv(atoms.pbc, [True, False, True]):
f.write('#keyword: surface\n')
else:
raise Exception('Only supported boundary conditions are full PBC,'
' no periodic boundary, and surface which is free in y direction'
' (i.e. Atoms.pbc = [True, False, True]).')
# Add atoms (scaled positions)
for position, symbol in zip(atoms.get_scaled_positions(),
atoms.get_chemical_symbols()):
f.write('{0} {1} {2} {3}\n'.format(
position[0], position[1], position[2], symbol))
def read_proteindatabank(fileobj, index=-1):
"""Read PDB files.
The format is assumed to follow the description given in
http://www.wwpdb.org/documentation/format32/sect8.html and
http://www.wwpdb.org/documentation/format32/sect9.html."""
if isinstance(fileobj, str):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
atoms = Atoms()
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [float(word) for word in line[6:54].split()]
atoms.set_cell(cellpar_to_cell(cellpar))
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
pars = [float(word) for word in line[10:55].split()]
orig[c] = pars[:3]
trans[c] = pars[3]
if line.startswith('ATOM') or line.startswith('HETATM'):
try:
# Atom name is arbitrary and does not necessarily
# contain the element symbol. The specification
# requires the element symbol to be in columns 77+78.
symbol = line[76:78].strip().lower().capitalize()
words = line[30:55].split()
position = np.array([float(words[0]),
float(words[1]),
float(words[2])])
position = np.dot(orig, position) + trans
atoms.append(Atom(symbol, position))
except:
pass
if line.startswith('ENDMDL'):
images.append(atoms)
atoms = Atoms()
if len(images) == 0:
images.append(atoms)
return images[index]
def read_pdb(fileobj, index=-1):
"""Read PDB files.
The format is assumed to follow the description given in
http://www.wwpdb.org/documentation/format32/sect8.html and
http://www.wwpdb.org/documentation/format32/sect9.html."""
if isinstance(fileobj, str):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
atoms = Atoms()
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [float(word) for word in line[6:54].split()]
atoms.set_cell(cellpar_to_cell(cellpar))
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
pars = [float(word) for word in line[10:55].split()]
orig[c] = pars[:3]
trans[c] = pars[3]
if line.startswith('ATOM') or line.startswith('HETATM'):
try:
# Atom name is arbitrary and does not necessarily contain the element symbol.
# The specification requires the element symbol to be in columns 77+78.
symbol = line[76:78].strip().lower().capitalize()
words = line[30:55].split()
position = np.array(
[float(words[0]),
float(words[1]),
float(words[2])])
position = np.dot(orig, position) + trans
atoms.append(Atom(symbol, position))
except:
pass
if line.startswith('ENDMDL'):
images.append(atoms)
atoms = Atoms()
if len(images) == 0:
images.append(atoms)
return images[index]
def write_proteindatabank(fileobj, images):
"""Write images to PDB-file."""
if isinstance(fileobj, basestring):
fileobj = paropen(fileobj, 'w')
if hasattr(images, 'get_positions'):
images = [images]
rotation = None
if images[0].get_pbc().any():
from ase.geometry import cell_to_cellpar, cellpar_to_cell
currentcell = images[0].get_cell()
cellpar = cell_to_cellpar(currentcell)
exportedcell = cellpar_to_cell(cellpar)
rotation = np.linalg.solve(currentcell, exportedcell)
# ignoring Z-value, using P1 since we have all atoms defined explicitly
format = 'CRYST1%9.3f%9.3f%9.3f%7.2f%7.2f%7.2f P 1\n'
fileobj.write(format % (cellpar[0], cellpar[1], cellpar[2],
cellpar[3], cellpar[4], cellpar[5]))
# 1234567 123 6789012345678901 89 67 456789012345678901234567 890
format = ('ATOM %5d %4s MOL 1 %8.3f%8.3f%8.3f 1.00 0.00'
' %2s \n')
# RasMol complains if the atom index exceeds 100000. There might
# be a limit of 5 digit numbers in this field.
MAXNUM = 100000
symbols = images[0].get_chemical_symbols()
natoms = len(symbols)
for n, atoms in enumerate(images):
fileobj.write('MODEL ' + str(n + 1) + '\n')
p = atoms.get_positions()
if rotation is not None:
p = p.dot(rotation)
for a in range(natoms):
x, y, z = p[a]
fileobj.write(format % (a % MAXNUM, symbols[a],
x, y, z, symbols[a].upper()))
fileobj.write('ENDMDL\n')
def read_proteindatabank(fileobj, index=-1):
"""Read PDB files."""
if isinstance(fileobj, basestring):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
atoms = Atoms()
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [float(word) for word in line[6:54].split()]
atoms.set_cell(cellpar_to_cell(cellpar))
atoms.pbc = True
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
pars = [float(word) for word in line[10:55].split()]
orig[c] = pars[:3]
trans[c] = pars[3]
if line.startswith('ATOM') or line.startswith('HETATM'):
try:
# Atom name is arbitrary and does not necessarily
# contain the element symbol. The specification
# requires the element symbol to be in columns 77+78.
symbol = line[76:78].strip().lower().capitalize()
words = line[30:55].split()
position = np.array(
[float(words[0]),
float(words[1]),
float(words[2])])
position = np.dot(orig, position) + trans
atoms.append(Atom(symbol, position))
except Exception as ex:
warnings.warn(
'Discarding atom when reading PDB file: {}'.format(ex))
if line.startswith('ENDMDL'):
images.append(atoms)
atoms = Atoms()
if len(images) == 0:
images.append(atoms)
return images[index]
def read_stru_to_ase(filename):
"""This function read the legacy DISCUS stru file format into a ASE
Atoms object.
:param filename: filename of DISCUS stru file
:type filename: str
:return: ASE Atoms object
:rtype: :class:`ase.Atoms`
"""
from ase import Atoms
from ase.geometry import cellpar_to_cell
with open(filename) as f:
lines = f.readlines()
a = b = c = alpha = beta = gamma = 0
reading_atom_list = False
symbols = []
positions = []
for l in lines:
line = l.replace(',', ' ').split()
if not reading_atom_list: # Wait for 'atoms' line before reading atoms
if line[0] == 'cell':
a, b, c, alpha, beta, gamma = [float(x) for x in line[1:7]]
cell = cellpar_to_cell([a, b, c, alpha, beta, gamma])
if line[0] == 'atoms':
if a == 0:
print("Cell not found")
cell = None
reading_atom_list = True
else:
symbol, x, y, z = line[:4]
symbol = symbol.capitalize()
symbols.append(symbol)
positions.append([float(x), float(y), float(z)])
# Return ASE Atoms object
return Atoms(symbols=symbols, scaled_positions=positions, cell=cell)
def read_stru_to_ase(filename):
"""This function read the legacy DISCUS stru file format into a ASE
Atoms object.
:param filename: filename of DISCUS stru file
:type filename: str
:return: ASE Atoms object
:rtype: :class:`ase.Atoms`
"""
from ase import Atoms
from ase.geometry import cellpar_to_cell
with open(filename) as f:
lines = f.readlines()
a = b = c = alpha = beta = gamma = 0
reading_atom_list = False
symbols = []
positions = []
for l in lines:
line = l.replace(',', ' ').split()
if not reading_atom_list: # Wait for 'atoms' line before reading atoms
if line[0] == 'cell':
a, b, c, alpha, beta, gamma = [float(x) for x in line[1:7]]
cell = cellpar_to_cell([a, b, c, alpha, beta, gamma])
if line[0] == 'atoms':
if a == 0:
print("Cell not found")
cell = None
reading_atom_list = True
else:
symbol, x, y, z = line[:4]
symbol = symbol.capitalize()
symbols.append(symbol)
positions.append([float(x), float(y), float(z)])
# Return ASE Atoms object
return Atoms(symbols=symbols, scaled_positions=positions, cell=cell)
def read_proteindatabank(fileobj, index=-1):
"""Read PDB files."""
if isinstance(fileobj, basestring):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
atoms = Atoms()
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [float(word) for word in line[6:54].split()]
atoms.set_cell(cellpar_to_cell(cellpar))
atoms.pbc = True
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
pars = [float(word) for word in line[10:55].split()]
orig[c] = pars[:3]
trans[c] = pars[3]
if line.startswith('ATOM') or line.startswith('HETATM'):
try:
# Atom name is arbitrary and does not necessarily
# contain the element symbol. The specification
# requires the element symbol to be in columns 77+78.
symbol = line[76:78].strip().lower().capitalize()
words = line[30:55].split()
position = np.array([float(words[0]),
float(words[1]),
float(words[2])])
position = np.dot(orig, position) + trans
atoms.append(Atom(symbol, position))
except Exception as ex:
warnings.warn('Discarding atom when reading PDB file: {}'
.format(ex))
if line.startswith('ENDMDL'):
images.append(atoms)
atoms = Atoms()
if len(images) == 0:
images.append(atoms)
return images[index]
def __init__(self, cell_dims, nx=50, ny=50, nz=50, pbc=[True,True,True], pbc_strict_mode=True):
import numpy as np
from ase.geometry import Cell, cellpar_to_cell
if np.shape(cell_dims)==(6,):
cellpar = cellpar_to_cell(cell_dims)
self.Cell = Cell(cellpar)
elif np.shape(cell_dims)==(3,3):
self.Cell = Cell(cell_dims)
else:
raise Exception("Specify cell parameters as a (6,) (x,y,z,alpha,beta,gamma) or (3x3) array.")
self.nx, self.ny, self.nz = nx, ny, nz
self.pbc = np.zeros(3, bool)
self.pbc = pbc
self.X = np.linspace(0, 1, self.nx)
self.Y = np.linspace(0, 1, self.ny)
self.Z = np.linspace(0, 1, self.nz)
# Define meshgrid in fractional coordinates.
self.frac_xx, self.frac_yy, self.frac_zz = np.meshgrid(self.X, self.Y, self.Z, indexing='xy')
# Project from fractional coordinates to cartesian.
stack_mesh = np.stack((self.frac_xx,self.frac_yy,self.frac_zz),axis=-1)
stack_mesh = np.einsum('ji,abcj->iabc', self.Cell.array, stack_mesh)
self.xx, self.yy, self.zz = stack_mesh[0], stack_mesh[1], stack_mesh[2]
# Allows easy conversion from cart -> frac coordinates.
self.inverse_cell_array = np.linalg.inv(self.Cell.array)
# Define the Cell parameters as (a,b,c,alpha,beta,gamma) array without calling
# Cell.cellpar() method everytime.
self.cellpar = self.Cell.cellpar()
# Enforces Mesh and Atoms pbc match - if off, ignores mismatches.
self.strict_mode = pbc_strict_mode
def normalize(atoms, set_origin=False):
"""
set the most near 0 to 0
make cell -> cellpar ->cell
make all the scaled postion between 0 and 1
"""
newatoms = atoms.copy()
newatoms = force_near_0(newatoms)
if set_origin:
positions = newatoms.get_positions()
s_positions = sorted(positions, key=np.linalg.norm)
newatoms.translate(-s_positions[0])
positions = newatoms.get_scaled_positions()
newcell = cellpar_to_cell(cell_to_cellpar(newatoms.get_cell()))
newatoms.set_cell(newcell)
for i, pos in enumerate(positions):
positions[i] = np.mod(pos, 1.0)
newatoms.set_scaled_positions(positions)
return newatoms
def read_jsv(f):
"""Reads a JSV file."""
natom = nbond = npoly = 0
symbols = []
labels = []
cellpar = basis = title = bonds = poly = origin = shell_numbers = None
spacegroup = 1
headline = f.readline().strip()
while True:
line = f.readline()
if not line:
break
line = line.strip()
m = re.match(r"^\[([^]]+)\]\s*(.*)", line)
if m is None or not line:
continue
tag = m.groups()[0].lower()
if len(m.groups()) > 1:
args = m.groups()[1].split()
else:
args = []
if tag == "cell":
cellpar = [float(x) for x in args]
elif tag == "natom":
natom = int(args[0])
elif tag == "nbond":
nbond = int(args[0])
# optional margin of the bondlengths
elif tag == "npoly":
npoly = int(args[0])
elif tag == "space_group":
spacegroup = Spacegroup(*tuple(int(x) for x in args))
elif tag == "title":
title = m.groups()[1]
elif tag == "atoms":
symbols = []
basis = np.zeros((natom, 3), dtype=float)
shell_numbers = -np.ones((natom,), dtype=int) # float?
for i in range(natom):
tokens = f.readline().strip().split()
labels.append(tokens[0])
symbols.append(ase.data.chemical_symbols[int(tokens[1])])
basis[i] = [float(x) for x in tokens[2:5]]
if len(tokens) > 5:
shell_numbers[i] = float(tokens[5]) # float?
elif tag == "bonds":
for i in range(nbond):
f.readline()
bonds = NotImplemented
elif tag == "poly":
for i in range(npoly):
f.readline()
poly = NotImplemented
elif tag == "origin":
origin = NotImplemented
else:
raise ValueError('Unknown tag: "%s"' % tag)
if headline == "asymmetric_unit_cell":
atoms = crystal(symbols=symbols, basis=basis, spacegroup=spacegroup, cellpar=cellpar)
elif headline == "full_unit_cell":
atoms = ase.Atoms(symbols=symbols, scaled_positions=basis, cell=cellpar_to_cell(cellpar))
atoms.info["spacegroup"] = Spacegroup(spacegroup)
elif headline == "cartesian_cell":
atoms = ase.Atoms(symbols=symbols, positions=basis, cell=cellpar_to_cell(cellpar))
atoms.info["spacegroup"] = Spacegroup(spacegroup)
else:
raise ValueError('Invalid JSV file type: "%s"' % headline)
atoms.info["title"] = title
atoms.info["labels"] = labels
if bonds is not None:
atoms.info["bonds"] = bonds
if poly is not None:
atoms.info["poly"] = poly
if origin is not None:
atoms.info["origin"] = origin
if shell_numbers is not None:
atoms.info["shell_numbers"] = shell_numbers
return atoms
def crystal(symbols=None, basis=None, spacegroup=1, setting=1,
cell=None, cellpar=None,
ab_normal=(0, 0, 1), a_direction=None, size=(1, 1, 1),
onduplicates='warn', symprec=0.001,
pbc=True, primitive_cell=False, **kwargs):
"""Create an Atoms instance for a conventional unit cell of a
space group.
Parameters:
symbols : str | sequence of str | sequence of Atom | Atoms
Element symbols of the unique sites. Can either be a string
formula or a sequence of element symbols. E.g. ('Na', 'Cl')
and 'NaCl' are equivalent. Can also be given as a sequence of
Atom objects or an Atoms object.
basis : list of scaled coordinates
Positions of the unique sites corresponding to symbols given
either as scaled positions or through an atoms instance. Not
needed if *symbols* is a sequence of Atom objects or an Atoms
object.
spacegroup : int | string | Spacegroup instance
Space group given either as its number in International Tables
or as its Hermann-Mauguin symbol.
setting : 1 | 2
Space group setting.
cell : 3x3 matrix
Unit cell vectors.
cellpar : [a, b, c, alpha, beta, gamma]
Cell parameters with angles in degree. Is not used when `cell`
is given.
ab_normal : vector
Is used to define the orientation of the unit cell relative
to the Cartesian system when `cell` is not given. It is the
normal vector of the plane spanned by a and b.
a_direction : vector
Defines the orientation of the unit cell a vector. a will be
parallel to the projection of `a_direction` onto the a-b plane.
size : 3 positive integers
How many times the conventional unit cell should be repeated
in each direction.
onduplicates : 'keep' | 'replace' | 'warn' | 'error'
Action if `basis` contain symmetry-equivalent positions:
'keep' - ignore additional symmetry-equivalent positions
'replace' - replace
'warn' - like 'keep', but issue an UserWarning
'error' - raises a SpacegroupValueError
symprec : float
Minimum "distance" betweed two sites in scaled coordinates
before they are counted as the same site.
pbc : one or three bools
Periodic boundary conditions flags. Examples: True,
False, 0, 1, (1, 1, 0), (True, False, False). Default
is True.
primitive_cell : bool
Wheter to return the primitive instead of the conventional
unit cell.
Keyword arguments:
All additional keyword arguments are passed on to the Atoms
constructor. Currently, probably the most useful additional
keyword arguments are `info`, `constraint` and `calculator`.
Examples:
Two diamond unit cells (space group number 227)
>>> diamond = crystal('C', [(0,0,0)], spacegroup=227,
... cellpar=[3.57, 3.57, 3.57, 90, 90, 90], size=(2,1,1))
>>> ase.view(diamond) # doctest: +SKIP
A CoSb3 skutterudite unit cell containing 32 atoms
>>> skutterudite = crystal(('Co', 'Sb'),
... basis=[(0.25,0.25,0.25), (0.0, 0.335, 0.158)],
... spacegroup=204, cellpar=[9.04, 9.04, 9.04, 90, 90, 90])
>>> len(skutterudite)
32
"""
sg = Spacegroup(spacegroup, setting)
if (not isinstance(symbols, str) and
hasattr(symbols, '__getitem__') and
len(symbols) > 0 and
isinstance(symbols[0], ase.Atom)):
symbols = ase.Atoms(symbols)
if isinstance(symbols, ase.Atoms):
basis = symbols
symbols = basis.get_chemical_symbols()
if isinstance(basis, ase.Atoms):
basis_coords = basis.get_scaled_positions()
if cell is None and cellpar is None:
cell = basis.cell
if symbols is None:
symbols = basis.get_chemical_symbols()
else:
basis_coords = np.array(basis, dtype=float, copy=False, ndmin=2)
sites, kinds = sg.equivalent_sites(basis_coords,
onduplicates=onduplicates,
symprec=symprec)
symbols = parse_symbols(symbols)
symbols = [symbols[i] for i in kinds]
if cell is None:
cell = cellpar_to_cell(cellpar, ab_normal, a_direction)
info = dict(spacegroup=sg)
if primitive_cell:
info['unit_cell'] = 'primitive'
else:
info['unit_cell'] = 'conventional'
if 'info' in kwargs:
info.update(kwargs['info'])
kwargs['info'] = info
atoms = ase.Atoms(symbols,
scaled_positions=sites,
cell=cell,
pbc=pbc,
**kwargs)
if isinstance(basis, ase.Atoms):
for name in basis.arrays:
if not atoms.has(name):
array = basis.get_array(name)
atoms.new_array(name, [array[i] for i in kinds],
dtype=array.dtype, shape=array.shape[1:])
if primitive_cell:
from ase.build import cut
prim_cell = sg.scaled_primitive_cell
atoms = cut(atoms, a=prim_cell[0], b=prim_cell[1], c=prim_cell[2])
if size != (1, 1, 1):
atoms = atoms.repeat(size)
return atoms
def __getitem__(self, i=-1):
self._open()
if isinstance(i, slice):
return [self[j] for j in range(*i.indices(self._len()))]
N = self._len()
if 0 <= i < N:
# Non-periodic boundaries have cell_length == 0.0
cell_lengths = \
np.array(self.nc.variables[self._cell_lengths_var][i][:])
pbc = np.abs(cell_lengths > 1e-6)
# Do we have a cell origin?
if self._has_variable(self._cell_origin_var):
origin = np.array(
self.nc.variables[self._cell_origin_var][i][:])
else:
origin = np.zeros([3], dtype=float)
# Do we have an index variable?
if self._has_variable(self.index_var):
index = np.array(self.nc.variables[self.index_var][i][:]) + \
self.index_offset
else:
index = np.arange(self.n_atoms)
# Read element numbers
self.numbers = self._get_data(self._numbers_var, i, exc=False)
if self.numbers is None:
self.numbers = np.ones(self.n_atoms, dtype=int)
else:
self.numbers = np.array(self.numbers[index])
if self.types_to_numbers is not None:
self.numbers = self.types_to_numbers[self.numbers]
self.masses = atomic_masses[self.numbers]
# Read positions
positions = np.array(self._get_data(self._positions_var, i)[index])
# Determine cell size for non-periodic directions
for dim in np.arange(3)[np.logical_not(pbc)]:
origin[dim] = positions[:, dim].min()
cell_lengths[dim] = positions[:, dim].max() - origin[dim]
# Construct cell shape from cell lengths and angles
cell = cellpar_to_cell(
list(cell_lengths) +
list(self.nc.variables[self._cell_angles_var][i])
)
# Compute momenta from velocities (if present)
momenta = self._get_data(self._velocities_var, i, exc=False)
if momenta is not None:
momenta = momenta[index] * self.masses.reshape(-1, 1)
# Fill info dict with additional data found in the NetCDF file
info = {}
for name in self.extra_per_frame_atts:
info[name] = np.array(self.nc.variables[name][i])
# Create atoms object
atoms = ase.Atoms(
positions=positions - origin.reshape(1, -1),
numbers=self.numbers,
cell=cell,
momenta=momenta,
masses=self.masses,
pbc=pbc,
info=info
)
# Attach additional arrays found in the NetCDF file
for name in self.extra_per_frame_vars:
atoms.set_array(name, self.nc.variables[name][i][index])
for name in self.extra_per_file_vars:
atoms.set_array(name, self.nc.variables[name][:])
self._close()
return atoms
i = N + i
if i < 0 or i >= N:
self._close()
raise IndexError('Trajectory index out of range.')
return self[i]
def read_v_sim(fd):
"""Import V_Sim input file.
Reads cell, atom positions, etc. from v_sim ascii file
"""
from ase import Atoms, units
from ase.geometry import cellpar_to_cell
import re
# Read comment:
fd.readline()
line = fd.readline() + ' ' + fd.readline()
box = line.split()
for i in range(len(box)):
box[i] = float(box[i])
keywords = []
positions = []
symbols = []
unit = 1.0
re_comment = re.compile(r'^\s*[#!]')
re_node = re.compile(r'^\s*\S+\s+\S+\s+\S+\s+\S+')
while True:
line = fd.readline()
if line == '':
break # EOF
p = re_comment.match(line)
if p is not None:
# remove comment character at the beginning of line
line = line[p.end():].replace(',', ' ').lower()
if line[:8] == "keyword:":
keywords.extend(line[8:].split())
elif re_node.match(line):
unit = 1.0
if not ("reduced" in keywords):
if (("bohr" in keywords) or ("bohrd0" in keywords)
or ("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
fields = line.split()
positions.append([
unit * float(fields[0]), unit * float(fields[1]),
unit * float(fields[2])
])
symbols.append(fields[3])
if ("surface" in keywords) or ("freeBC" in keywords):
raise NotImplementedError
# create atoms object based on the information
if "angdeg" in keywords:
cell = cellpar_to_cell(box)
else:
unit = 1.0
if (("bohr" in keywords) or ("bohrd0" in keywords)
or ("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
cell = np.zeros((3, 3))
cell.flat[[0, 3, 4, 6, 7, 8]] = box[:6]
cell *= unit
if "reduced" in keywords:
atoms = Atoms(cell=cell, scaled_positions=positions)
else:
atoms = Atoms(cell=cell, positions=positions)
atoms.set_chemical_symbols(symbols)
return atoms
def crystal(symbols=None,
basis=None,
occupancies=None,
spacegroup=1,
setting=1,
cell=None,
cellpar=None,
ab_normal=(0, 0, 1),
a_direction=None,
size=(1, 1, 1),
onduplicates='warn',
symprec=0.001,
pbc=True,
primitive_cell=False,
**kwargs):
"""Create an Atoms instance for a conventional unit cell of a
space group.
Parameters:
symbols : str | sequence of str | sequence of Atom | Atoms
Element symbols of the unique sites. Can either be a string
formula or a sequence of element symbols. E.g. ('Na', 'Cl')
and 'NaCl' are equivalent. Can also be given as a sequence of
Atom objects or an Atoms object.
basis : list of scaled coordinates
Positions of the unique sites corresponding to symbols given
either as scaled positions or through an atoms instance. Not
needed if *symbols* is a sequence of Atom objects or an Atoms
object.
occupancies : list of site occupancies
Occupancies of the unique sites. Defaults to 1.0 and thus no mixed
occupancies are considered if not explicitly asked for. If occupancies
are given, the most dominant species will yield the atomic number.
spacegroup : int | string | Spacegroup instance
Space group given either as its number in International Tables
or as its Hermann-Mauguin symbol.
setting : 1 | 2
Space group setting.
cell : 3x3 matrix
Unit cell vectors.
cellpar : [a, b, c, alpha, beta, gamma]
Cell parameters with angles in degree. Is not used when `cell`
is given.
ab_normal : vector
Is used to define the orientation of the unit cell relative
to the Cartesian system when `cell` is not given. It is the
normal vector of the plane spanned by a and b.
a_direction : vector
Defines the orientation of the unit cell a vector. a will be
parallel to the projection of `a_direction` onto the a-b plane.
size : 3 positive integers
How many times the conventional unit cell should be repeated
in each direction.
onduplicates : 'keep' | 'replace' | 'warn' | 'error'
Action if `basis` contain symmetry-equivalent positions:
'keep' - ignore additional symmetry-equivalent positions
'replace' - replace
'warn' - like 'keep', but issue an UserWarning
'error' - raises a SpacegroupValueError
symprec : float
Minimum "distance" betweed two sites in scaled coordinates
before they are counted as the same site.
pbc : one or three bools
Periodic boundary conditions flags. Examples: True,
False, 0, 1, (1, 1, 0), (True, False, False). Default
is True.
primitive_cell : bool
Wheter to return the primitive instead of the conventional
unit cell.
Keyword arguments:
All additional keyword arguments are passed on to the Atoms
constructor. Currently, probably the most useful additional
keyword arguments are `info`, `constraint` and `calculator`.
Examples:
Two diamond unit cells (space group number 227)
>>> diamond = crystal('C', [(0,0,0)], spacegroup=227,
... cellpar=[3.57, 3.57, 3.57, 90, 90, 90], size=(2,1,1))
>>> ase.view(diamond) # doctest: +SKIP
A CoSb3 skutterudite unit cell containing 32 atoms
>>> skutterudite = crystal(('Co', 'Sb'),
... basis=[(0.25,0.25,0.25), (0.0, 0.335, 0.158)],
... spacegroup=204, cellpar=[9.04, 9.04, 9.04, 90, 90, 90])
>>> len(skutterudite)
32
"""
sg = Spacegroup(spacegroup, setting)
if (not isinstance(symbols, basestring)
and hasattr(symbols, '__getitem__') and len(symbols) > 0
and isinstance(symbols[0], ase.Atom)):
symbols = ase.Atoms(symbols)
if isinstance(symbols, ase.Atoms):
basis = symbols
symbols = basis.get_chemical_symbols()
if isinstance(basis, ase.Atoms):
basis_coords = basis.get_scaled_positions()
if cell is None and cellpar is None:
cell = basis.cell
if symbols is None:
symbols = basis.get_chemical_symbols()
else:
basis_coords = np.array(basis, dtype=float, copy=False, ndmin=2)
if occupancies is not None:
occupancies_dict = {}
for index, coord in enumerate(basis_coords):
# Compute all distances and get indices of nearest atoms
dist = spatial.distance.cdist(coord.reshape(1, 3), basis_coords)
indices_dist = np.flatnonzero(dist < symprec)
occ = {symbols[index]: occupancies[index]}
# Check nearest and update occupancy
for index_dist in indices_dist:
if index == index_dist:
continue
else:
occ.update({symbols[index_dist]: occupancies[index_dist]})
occupancies_dict[index] = occ.copy()
sites, kinds = sg.equivalent_sites(basis_coords,
onduplicates=onduplicates,
symprec=symprec)
# this is needed to handle deuterium masses
masses = None
if 'masses' in kwargs:
masses = kwargs['masses'][kinds]
del kwargs['masses']
symbols = parse_symbols(symbols)
if occupancies is None:
symbols = [symbols[i] for i in kinds]
else:
# make sure that we put the dominant species there
symbols = [
sorted(occupancies_dict[i].items(), key=lambda x: x[1])[-1][0]
for i in kinds
]
if cell is None:
cell = cellpar_to_cell(cellpar, ab_normal, a_direction)
info = dict(spacegroup=sg)
if primitive_cell:
info['unit_cell'] = 'primitive'
else:
info['unit_cell'] = 'conventional'
if 'info' in kwargs:
info.update(kwargs['info'])
if occupancies is not None:
info['occupancy'] = occupancies_dict
kwargs['info'] = info
atoms = ase.Atoms(symbols,
scaled_positions=sites,
cell=cell,
pbc=pbc,
masses=masses,
**kwargs)
# if all occupancies are 1, no partial occupancy present
if occupancies:
if not all([occ == 1 for occ in occupancies]):
# use tags to identify sites, and in particular the occupancy
atoms.set_tags(kinds)
if isinstance(basis, ase.Atoms):
for name in basis.arrays:
if not atoms.has(name):
array = basis.get_array(name)
atoms.new_array(name, [array[i] for i in kinds],
dtype=array.dtype,
shape=array.shape[1:])
if primitive_cell:
from ase.build import cut
prim_cell = sg.scaled_primitive_cell
# Preserve calculator if present:
calc = atoms.calc
atoms = cut(atoms, a=prim_cell[0], b=prim_cell[1], c=prim_cell[2])
atoms.calc = calc
if size != (1, 1, 1):
atoms = atoms.repeat(size)
return atoms
"""Test band structure from different variations of hexagonal cells."""
import numpy as np
from ase import Atoms
from ase.calculators.test import FreeElectrons
from ase.geometry import crystal_structure_from_cell, cell_to_cellpar, cellpar_to_cell
from ase.dft.kpoints import get_special_points
mc1 = [[1, 0, 0], [0, 1, 0], [0, 0.2, 1]]
par = cell_to_cellpar(mc1)
mc2 = cellpar_to_cell(par)
mc3 = [[1, 0, 0], [0, 1, 0], [-0.2, 0, 1]]
path = 'GYHCEM1AXH1'
firsttime = True
for cell in [mc1, mc2, mc3]:
a = Atoms(cell=cell, pbc=True)
a.cell *= 3
a.calc = FreeElectrons(nvalence=1, kpts={'path': path})
print(crystal_structure_from_cell(a.cell))
r = a.get_reciprocal_cell()
k = get_special_points(a.cell)['H']
print(np.dot(k, r))
a.get_potential_energy()
bs = a.calc.band_structure()
coords, labelcoords, labels = bs.get_labels()
assert ''.join(labels) == path
e_skn = bs.energies
# bs.plot()
if firsttime:
coords1 = coords
labelcoords1 = labelcoords
def read_proteindatabank(fileobj, index=-1, read_arrays=True):
"""Read PDB files."""
if isinstance(fileobj, str):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
occ = []
bfactor = []
residuenames = []
residuenumbers = []
atomtypes = []
symbols = []
positions = []
cell = None
pbc = None
def build_atoms():
atoms = Atoms(symbols=symbols,
cell=cell, pbc=pbc,
positions=positions)
if not read_arrays:
return atoms
info = {'occupancy': occ,
'bfactor': bfactor,
'residuenames': residuenames,
'atomtypes': atomtypes,
'residuenumbers': residuenumbers}
for name, array in info.items():
if len(array) == 0:
pass
elif len(array) != len(atoms):
warnings.warn('Length of {} array, {}, '
'different from number of atoms {}'.
format(name, len(array), len(atoms)))
else:
atoms.set_array(name, np.array(array))
return atoms
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [float(line[6:15]), # a
float(line[15:24]), # b
float(line[24:33]), # c
float(line[33:40]), # alpha
float(line[40:47]), # beta
float(line[47:54])] # gamma
cell = cellpar_to_cell(cellpar)
pbc = True
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
orig[c] = [float(line[10:20]),
float(line[20:30]),
float(line[30:40])]
trans[c] = float(line[45:55])
if line.startswith('ATOM') or line.startswith('HETATM'):
# Atom name is arbitrary and does not necessarily
# contain the element symbol. The specification
# requires the element symbol to be in columns 77+78.
# Fall back to Atom name for files that do not follow
# the spec, e.g. packmol.
# line_info = symbol, name, altloc, resname, coord, occupancy,
# bfactor, resseq
line_info = read_atom_line(line)
try:
symbol = label_to_symbol(line_info[0])
except (KeyError, IndexError):
symbol = label_to_symbol(line_info[1])
position = np.dot(orig, line_info[4]) + trans
atomtypes.append(line_info[1])
residuenames.append(line_info[3])
if line_info[5] is not None:
occ.append(line_info[5])
bfactor.append(line_info[6])
residuenumbers.append(line_info[7])
symbols.append(symbol)
positions.append(position)
if line.startswith('END'):
# End of configuration reached
# According to the latest PDB file format (v3.30),
# this line should start with 'ENDMDL' (not 'END'),
# but in this way PDB trajectories from e.g. CP2K
# are supported (also VMD supports this format).
atoms = build_atoms()
images.append(atoms)
occ = []
bfactor = []
residuenames = []
atomtypes = []
symbols = []
positions = []
cell = None
pbc = None
if len(images) == 0:
atoms = build_atoms()
images.append(atoms)
return images[index]
def read_jsv(f):
"""Reads a JSV file."""
natom = nbond = npoly = 0
symbols = []
labels = []
cellpar = basis = title = bonds = poly = origin = shell_numbers = None
spacegroup = 1
headline = f.readline().strip()
while True:
line = f.readline()
if not line:
break
line = line.strip()
m = re.match(r'^\[([^]]+)\]\s*(.*)', line)
if m is None or not line:
continue
tag = m.groups()[0].lower()
if len(m.groups()) > 1:
args = m.groups()[1].split()
else:
args = []
if tag == 'cell':
cellpar = [float(x) for x in args]
elif tag == 'natom':
natom = int(args[0])
elif tag == 'nbond':
nbond = int(args[0])
# optional margin of the bondlengths
elif tag == 'npoly':
npoly = int(args[0])
elif tag == 'space_group':
spacegroup = Spacegroup(*tuple(int(x) for x in args))
elif tag == 'title':
title = m.groups()[1]
elif tag == 'atoms':
symbols = []
basis = np.zeros((natom, 3), dtype=float)
shell_numbers = -np.ones((natom, ), dtype=int) # float?
for i in range(natom):
tokens = f.readline().strip().split()
labels.append(tokens[0])
symbols.append(ase.data.chemical_symbols[int(tokens[1])])
basis[i] = [float(x) for x in tokens[2:5]]
if len(tokens) > 5:
shell_numbers[i] = float(tokens[5]) # float?
elif tag == 'bonds':
for i in range(nbond):
f.readline()
bonds = NotImplemented
elif tag == 'poly':
for i in range(npoly):
f.readline()
poly = NotImplemented
elif tag == 'origin':
origin = NotImplemented
else:
raise ValueError('Unknown tag: "%s"' % tag)
if headline == 'asymmetric_unit_cell':
atoms = crystal(
symbols=symbols,
basis=basis,
spacegroup=spacegroup,
cellpar=cellpar,
)
elif headline == 'full_unit_cell':
atoms = ase.Atoms(
symbols=symbols,
scaled_positions=basis,
cell=cellpar_to_cell(cellpar),
)
atoms.info['spacegroup'] = Spacegroup(spacegroup)
elif headline == 'cartesian_cell':
atoms = ase.Atoms(
symbols=symbols,
positions=basis,
cell=cellpar_to_cell(cellpar),
)
atoms.info['spacegroup'] = Spacegroup(spacegroup)
else:
raise ValueError('Invalid JSV file type: "%s"' % headline)
atoms.info['title'] = title
atoms.info['labels'] = labels
if bonds is not None:
atoms.info['bonds'] = bonds
if poly is not None:
atoms.info['poly'] = poly
if origin is not None:
atoms.info['origin'] = origin
if shell_numbers is not None:
atoms.info['shell_numbers'] = shell_numbers
return atoms
def __getitem__(self, i=-1):
self._open()
if isinstance(i, slice):
return [self[j] for j in range(*i.indices(self._len()))]
N = self._len()
if 0 <= i < N:
# Non-periodic boundaries have cell_length == 0.0
cell_lengths = \
np.array(self.nc.variables[self._cell_lengths_var][i][:])
pbc = np.abs(cell_lengths > 1e-6)
# Do we have a cell origin?
if self._has_variable(self._cell_origin_var):
origin = np.array(
self.nc.variables[self._cell_origin_var][i][:])
else:
origin = np.zeros([3], dtype=float)
# Do we have an index variable?
if self._has_variable(self.index_var):
index = np.array(self.nc.variables[self.index_var][i][:]) + \
self.index_offset
else:
index = np.arange(self.n_atoms)
# Read element numbers
self.numbers = self._get_data(self._numbers_var, i, exc=False)
if self.numbers is None:
self.numbers = np.ones(self.n_atoms, dtype=int)
else:
self.numbers = np.array(self.numbers[index])
if self.types_to_numbers is not None:
self.numbers = self.types_to_numbers[self.numbers]
self.masses = atomic_masses[self.numbers]
# Read positions
positions = np.array(self._get_data(self._positions_var, i)[index])
# Determine cell size for non-periodic directions
for dim in np.arange(3)[np.logical_not(pbc)]:
origin[dim] = positions[:, dim].min()
cell_lengths[dim] = positions[:, dim].max() - origin[dim]
# Construct cell shape from cell lengths and angles
cell = cellpar_to_cell(
list(cell_lengths) +
list(self.nc.variables[self._cell_angles_var][i]))
# Compute momenta from velocities (if present)
momenta = self._get_data(self._velocities_var, i, exc=False)
if momenta is not None:
momenta = momenta[index] * self.masses.reshape(-1, 1)
# Fill info dict with additional data found in the NetCDF file
info = {}
for name in self.extra_per_frame_atts:
info[name] = np.array(self.nc.variables[name][i])
# Create atoms object
atoms = ase.Atoms(positions=positions - origin.reshape(1, -1),
numbers=self.numbers,
cell=cell,
momenta=momenta,
masses=self.masses,
pbc=pbc,
info=info)
# Attach additional arrays found in the NetCDF file
for name in self.extra_per_frame_vars:
atoms.set_array(name, self.nc.variables[name][i][index])
for name in self.extra_per_file_vars:
atoms.set_array(name, self.nc.variables[name][:])
self._close()
return atoms
i = N + i
if i < 0 or i >= N:
self._close()
raise IndexError('Trajectory index out of range.')
return self[i]
def read_v_sim(filename='demo.ascii'):
"""Import V_Sim input file.
Reads cell, atom positions, etc. from v_sim ascii file
"""
from ase import Atoms, units
from ase.geometry import cellpar_to_cell
import re
if isinstance(filename, str):
f = open(filename)
else: # Assume it's a file-like object
f = filename
# Read comment:
f.readline()
line = f.readline() + ' ' + f.readline()
box = line.split()
for i in range(len(box)):
box[i] = float(box[i])
keywords = []
positions = []
symbols = []
unit = 1.0
re_comment = re.compile('^\s*[#!]')
re_node = re.compile('^\s*\S+\s+\S+\s+\S+\s+\S+')
while(True):
line = f.readline()
if line == '':
break # EOF
p = re_comment.match(line)
if p is not None:
# remove comment character at the beginning of line
line = line[p.end():].replace(',', ' ').lower()
if line[:8] == "keyword:":
keywords.extend(line[8:].split())
elif(re_node.match(line)):
unit = 1.0
if not ("reduced" in keywords):
if (("bohr" in keywords) or ("bohrd0" in keywords) or
("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
fields = line.split()
positions.append([unit*float(fields[0]),
unit*float(fields[1]),
unit*float(fields[2])])
symbols.append(fields[3])
f.close()
if ("surface" in keywords) or ("freeBC" in keywords):
raise NotImplementedError
# create atoms object based on the information
if ("angdeg" in keywords):
cell = cellpar_to_cell(box)
else:
unit = 1.0
if (("bohr" in keywords) or ("bohrd0" in keywords) or
("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
cell = [[unit*box[0], 0.0, 0.0],
[unit*box[1], unit*box[2], 0.0],
[unit*box[3], unit*box[4], unit*box[5]]]
if ("reduced" in keywords):
atoms = Atoms(cell=cell, scaled_positions=positions)
else:
atoms = Atoms(cell=cell, positions=positions)
atoms.set_chemical_symbols(symbols)
return atoms
allH = False
if (len(sys.argv) >= 21):
deltapos[0] = float(sys.argv[12])
deltapos[1] = float(sys.argv[13])
deltapos[2] = float(sys.argv[14])
cella = float(sys.argv[15])
cellb = float(sys.argv[16])
cellc = float(sys.argv[17])
alpha = float(sys.argv[18])
beta = float(sys.argv[19])
gamma = float(sys.argv[20])
if (len(sys.argv) >= 22):
allH = bool(sys.argv[21])
# set unit cell
cell = geometry.cellpar_to_cell([cella, cellb, cellc, alpha, beta, gamma])
invcell = inv(cell)
print(cell)
print(invcell)
# Set centers for wrap_positions
# for when you want to specify how to wrap atoms which are
# positioned exactly on the unit cell edge.
# This is useful when, for example, node2 is shifted (1,0,0)
# and node1 has x position that lies right on UC boundary,
# so you want it to wrap to the higher value (e.g., BAZJET).
centercorrection = np.identity(3) * 0.00001
center1 = np.full(DIM, 0.5) + np.dot(centercorrection, deltapos)
#center2 = np.full(DIM,0.5) - np.dot(centercorrection,deltapos)
# Wrap node positions so they start in the same unit cell
def __getitem__(self, i=-1):
self._open()
if isinstance(i, slice):
return [self[j] for j in range(*i.indices(self._len()))]
N = self._len()
if 0 <= i < N:
# Non-periodic boundaries have cell_length == 0.0
cell_lengths = \
np.array(self.nc.variables[self._cell_lengths_var][i][:])
pbc = np.abs(cell_lengths > 1e-6)
# Do we have a cell origin?
if self._has_variable(self._cell_origin_var):
origin = np.array(
self.nc.variables[self._cell_origin_var][i][:])
else:
origin = np.zeros([3], dtype=float)
# Do we have an index variable?
if (self.index_var is not None and
self._has_variable(self.index_var)):
index = np.array(self.nc.variables[self.index_var][i][:])
# The index variable can be non-consecutive, we here construct
# a consecutive one.
consecutive_index = np.zeros_like(index)
consecutive_index[np.argsort(index)] = np.arange(self.n_atoms)
else:
consecutive_index = np.arange(self.n_atoms)
# Read element numbers
self.numbers = self._get_data(self._numbers_var, i,
consecutive_index, exc=False)
if self.numbers is None:
self.numbers = np.ones(self.n_atoms, dtype=int)
if self.types_to_numbers is not None:
d = set(self.numbers).difference(self.types_to_numbers.keys())
if len(d) > 0:
self.types_to_numbers.update({num: num for num in d})
func = np.vectorize(self.types_to_numbers.get)
self.numbers = func(self.numbers)
self.masses = atomic_masses[self.numbers]
# Read positions
positions = self._get_data(self._positions_var, i,
consecutive_index)
# Determine cell size for non-periodic directions from shrink
# wrapped cell.
for dim in np.arange(3)[np.logical_not(pbc)]:
origin[dim] = positions[:, dim].min()
cell_lengths[dim] = positions[:, dim].max() - origin[dim]
# Construct cell shape from cell lengths and angles
cell = cellpar_to_cell(
list(cell_lengths) +
list(self.nc.variables[self._cell_angles_var][i])
)
# Compute momenta from velocities (if present)
momenta = self._get_data(self._velocities_var, i,
consecutive_index, exc=False)
if momenta is not None:
momenta *= self.masses.reshape(-1, 1)
# Fill info dict with additional data found in the NetCDF file
info = {}
for name in self.extra_per_frame_atts:
info[name] = np.array(self.nc.variables[name][i])
# Create atoms object
atoms = ase.Atoms(
positions=positions,
numbers=self.numbers,
cell=cell,
celldisp=origin,
momenta=momenta,
masses=self.masses,
pbc=pbc,
info=info
)
# Attach additional arrays found in the NetCDF file
for name in self.extra_per_frame_vars:
atoms.set_array(name, self._get_data(name, i,
consecutive_index))
for name in self.extra_per_file_vars:
atoms.set_array(name, self._get_data(name, i,
consecutive_index))
self._close()
return atoms
i = N + i
if i < 0 or i >= N:
self._close()
raise IndexError('Trajectory index out of range.')
return self[i]
def read_v_sim(filename='demo.ascii'):
"""Import V_Sim input file.
Reads cell, atom positions, etc. from v_sim ascii file
"""
from ase import Atoms, units
from ase.geometry import cellpar_to_cell
import re
if isinstance(filename, str):
f = open(filename)
else: # Assume it's a file-like object
f = filename
# Read comment:
f.readline()
line = f.readline() + ' ' + f.readline()
box = line.split()
for i in range(len(box)):
box[i] = float(box[i])
keywords = []
positions = []
symbols = []
unit = 1.0
re_comment = re.compile(r'^\s*[#!]')
re_node = re.compile(r'^\s*\S+\s+\S+\s+\S+\s+\S+')
while (True):
line = f.readline()
if line == '':
break # EOF
p = re_comment.match(line)
if p is not None:
# remove comment character at the beginning of line
line = line[p.end():].replace(',', ' ').lower()
if line[:8] == "keyword:":
keywords.extend(line[8:].split())
elif (re_node.match(line)):
unit = 1.0
if not ("reduced" in keywords):
if (("bohr" in keywords) or ("bohrd0" in keywords)
or ("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
fields = line.split()
positions.append([
unit * float(fields[0]), unit * float(fields[1]),
unit * float(fields[2])
])
symbols.append(fields[3])
f.close()
if ("surface" in keywords) or ("freeBC" in keywords):
raise NotImplementedError
# create atoms object based on the information
if ("angdeg" in keywords):
cell = cellpar_to_cell(box)
else:
unit = 1.0
if (("bohr" in keywords) or ("bohrd0" in keywords)
or ("atomic" in keywords) or ("atomicd0" in keywords)):
unit = units.Bohr
cell = [[unit * box[0], 0.0, 0.0], [unit * box[1], unit * box[2], 0.0],
[unit * box[3], unit * box[4], unit * box[5]]]
if ("reduced" in keywords):
atoms = Atoms(cell=cell, scaled_positions=positions)
else:
atoms = Atoms(cell=cell, positions=positions)
atoms.set_chemical_symbols(symbols)
return atoms
def fix_cell(cell,eps=5e-3):
cellpar=cell_to_cellpar(cell)
for i in [3,4,5]:
if abs(cellpar[i]/90.0*np.pi/2-np.pi/2)<eps:
cellpar[i]=90.0
return cellpar_to_cell(cellpar)
def niggli_reduce_cell(cell, epsfactor=None):
from ase.geometry import cellpar_to_cell
if epsfactor is None:
epsfactor = 1e-5
eps = epsfactor * abs(np.linalg.det(cell))**(1. / 3.)
cell = np.asarray(cell)
I3 = np.eye(3, dtype=int)
I6 = np.eye(6, dtype=int)
C = I3.copy()
D = I6.copy()
g0 = np.zeros(6, dtype=float)
g0[0] = np.dot(cell[0], cell[0])
g0[1] = np.dot(cell[1], cell[1])
g0[2] = np.dot(cell[2], cell[2])
g0[3] = 2 * np.dot(cell[1], cell[2])
g0[4] = 2 * np.dot(cell[0], cell[2])
g0[5] = 2 * np.dot(cell[0], cell[1])
g = np.dot(D, g0)
def lt(x, y, eps=eps):
return x < y - eps
def gt(x, y, eps=eps):
return lt(y, x, eps)
def eq(x, y, eps=eps):
return not (lt(x, y, eps) or gt(x, y, eps))
for _ in range(10000):
if (gt(g[0], g[1]) or (eq(g[0], g[1]) and gt(abs(g[3]), abs(g[4])))):
C = np.dot(C, -I3[[1, 0, 2]])
D = np.dot(I6[[1, 0, 2, 4, 3, 5]], D)
g = np.dot(D, g0)
continue
elif (gt(g[1], g[2]) or (eq(g[1], g[2]) and gt(abs(g[4]), abs(g[5])))):
C = np.dot(C, -I3[[0, 2, 1]])
D = np.dot(I6[[0, 2, 1, 3, 5, 4]], D)
g = np.dot(D, g0)
continue
lmn = np.array(gt(g[3:], 0, eps=eps / 2), dtype=int)
lmn -= np.array(lt(g[3:], 0, eps=eps / 2), dtype=int)
if lmn.prod() == 1:
ijk = lmn.copy()
for idx in range(3):
if ijk[idx] == 0:
ijk[idx] = 1
else:
ijk = np.ones(3, dtype=int)
if np.any(lmn != -1):
r = None
for idx in range(3):
if lmn[idx] == 1:
ijk[idx] = -1
elif lmn[idx] == 0:
r = idx
if ijk.prod() == -1:
ijk[r] = -1
C *= ijk[np.newaxis]
D[3] *= ijk[1] * ijk[2]
D[4] *= ijk[0] * ijk[2]
D[5] *= ijk[0] * ijk[1]
g = np.dot(D, g0)
if (gt(abs(g[3]), g[1]) or (eq(g[3], g[1]) and lt(2 * g[4], g[5]))
or (eq(g[3], -g[1]) and lt(g[5], 0))):
s = np.int(np.sign(g[3]))
A = I3.copy()
A[1, 2] = -s
C = np.dot(C, A)
B = I6.copy()
B[2, 1] = 1
B[2, 3] = -s
B[3, 1] = -2 * s
B[4, 5] = -s
D = np.dot(B, D)
g = np.dot(D, g0)
elif (gt(abs(g[4]), g[0]) or (eq(g[4], g[0]) and lt(2 * g[3], g[5]))
or (eq(g[4], -g[0]) and lt(g[5], 0))):
s = np.int(np.sign(g[4]))
A = I3.copy()
A[0, 2] = -s
C = np.dot(C, A)
B = I6.copy()
B[2, 0] = 1
B[2, 4] = -s
B[3, 5] = -s
B[4, 0] = -2 * s
D = np.dot(B, D)
g = np.dot(D, g0)
elif (gt(abs(g[5]), g[0]) or (eq(g[5], g[0]) and lt(2 * g[3], g[4]))
or (eq(g[5], -g[0]) and lt(g[4], 0))):
s = np.int(np.sign(g[5]))
A = I3.copy()
A[0, 1] = -s
C = np.dot(C, A)
B = I6.copy()
B[1, 0] = 1
B[1, 5] = -s
B[3, 4] = -s
B[5, 0] = -2 * s
D = np.dot(B, D)
g = np.dot(D, g0)
elif (
lt(g[[0, 1, 3, 4, 5]].sum(), 0) or
(eq(g[[0, 1, 3, 4, 5]].sum(), 0) and gt(2 *
(g[0] + g[4]) + g[5], 0))):
A = I3.copy()
A[:, 2] = 1
C = np.dot(C, A)
B = I6.copy()
B[2, :] = 1
B[3, 1] = 2
B[3, 5] = 1
B[4, 0] = 2
B[4, 5] = 1
D = np.dot(B, D)
g = np.dot(D, g0)
else:
break
else:
raise RuntimeError('Niggli reduction not done in 10000 steps!\n'
'cell={}\n'
'operation={}'.format(cell.tolist(), C.tolist()))
abc = np.sqrt(g[:3])
# Prevent division by zero e.g. for cell==zeros((3, 3)):
abcprod = max(abc.prod(), 1e-100)
cosangles = abc * g[3:] / (2 * abcprod)
angles = 180 * np.arccos(cosangles) / np.pi
newcell = np.array(cellpar_to_cell(np.concatenate([abc, angles])),
dtype=float)
return newcell, C
def read_proteindatabank(fileobj, index=-1, read_arrays=True):
"""Read PDB files."""
if isinstance(fileobj, basestring):
fileobj = open(fileobj)
images = []
orig = np.identity(3)
trans = np.zeros(3)
atoms = Atoms()
occ = []
bfactor = []
for line in fileobj.readlines():
if line.startswith('CRYST1'):
cellpar = [
float(line[6:15]), # a
float(line[15:24]), # b
float(line[24:33]), # c
float(line[33:40]), # alpha
float(line[40:47]), # beta
float(line[47:54])
] # gamma
atoms.set_cell(cellpar_to_cell(cellpar))
atoms.pbc = True
for c in range(3):
if line.startswith('ORIGX' + '123'[c]):
orig[c] = [
float(line[10:20]),
float(line[20:30]),
float(line[30:40])
]
trans[c] = float(line[45:55])
if line.startswith('ATOM') or line.startswith('HETATM'):
try:
# Atom name is arbitrary and does not necessarily
# contain the element symbol. The specification
# requires the element symbol to be in columns 77+78.
# Fall back to Atom name for files that do not follow
# the spec, e.g. packmol.
try:
symbol = label_to_symbol(line[76:78].strip())
except (KeyError, IndexError):
symbol = label_to_symbol(line[12:16].strip())
# Don't use split() in case there are no spaces
position = np.array([
float(line[30:38]), # x
float(line[38:46]), # y
float(line[46:54])
]) # z
try:
occ.append(float(line[54:60]))
bfactor.append(float(line[60:66]))
except (IndexError, ValueError):
pass
position = np.dot(orig, position) + trans
atoms.append(Atom(symbol, position))
except Exception as ex:
warnings.warn(
'Discarding atom when reading PDB file: {}\n{}'.format(
line.strip(), ex))
if line.startswith('END'):
# End of configuration reached
# According to the latest PDB file format (v3.30),
# this line should start with 'ENDMDL' (not 'END'),
# but in this way PDB trajectories from e.g. CP2K
# are supported (also VMD supports this format).
if read_arrays and len(occ) == len(atoms):
atoms.set_array('occupancy', np.array(occ))
if read_arrays and len(bfactor) == len(atoms):
atoms.set_array('bfactor', np.array(bfactor))
images.append(atoms)
atoms = Atoms()
occ = []
bfactor = []
if len(images) == 0:
# Single configuration with no 'END' or 'ENDMDL'
if read_arrays and len(occ) == len(atoms):
atoms.set_array('occupancy', np.array(occ))
if read_arrays and len(bfactor) == len(atoms):
atoms.set_array('bfactor', np.array(bfactor))
images.append(atoms)
return images[index]