def read_HS_from_skf(fileobj, symboli, symbolj):
"""
Read Hamitonian and overlap data from DFTB-style .skf file.
Parameters:
-----------
fileobj: filename of file-object to read from
symboli: chemical symbol of the first element
symbolj: chemical symbol of the second element
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
# Homoatomic interactions also contain element data
if symboli == symbolj:
# First line contains grid spacing and number of grid points
l = fortran_readline(fileobj)
dx = l[0]
n = l[1]
n = int(n)
# Contains self-energies, spin-polarization energies, Hubbard-U, ...
l = fileobj.readline()
else:
# First line contains grid spacing and number of grid points
dx, n = fortran_readline(fileobj)
n = int(n)
# The n information is sometimes wrong, better count while reading
#x = dx*np.arange(0, n)
HS = []
l = fileobj.readline()
while l and l.strip() != 'Spline':
if l.strip() == 'Spline':
if i != n - 1:
raise RuntimeError('Spline keyword encountered reading tables '
'for %s-%s before reaching the end of the '
'HS table.' % (symboli, symbolj))
else:
HS += [fortran_readline(l)]
l = fileobj.readline()
# don't care if not spline...
#if not l:
# raise RuntimeError('Premature end-of-file: Keyword "Spline" not found '
# 'for %s-%s.' % ( symboli, symbolj ))
HS = np.array(HS)
x = dx * np.arange(0, HS.shape[0])
#return x[0:HS.shape[0]], np.array(HS)
return x, np.array(HS)
def read_HS_from_skf(fileobj, symboli, symbolj):
"""
Read Hamitonian and overlap data from DFTB-style .skf file.
Parameters:
-----------
fileobj: filename of file-object to read from
symboli: chemical symbol of the first element
symbolj: chemical symbol of the second element
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
# Homoatomic interactions also contain element data
if symboli == symbolj:
# First line contains grid spacing and number of grid points
l = fortran_readline(fileobj)
dx = l[0]
n = l[1]
n = int(n)
# Contains self-energies, spin-polarization energies, Hubbard-U, ...
l = fileobj.readline()
else:
# First line contains grid spacing and number of grid points
dx, n = fortran_readline(fileobj)
n = int(n)
# The n information is sometimes wrong, better count while reading
#x = dx*np.arange(0, n)
HS = [ ]
l = fileobj.readline()
while l and l.strip() != 'Spline':
if l.strip() == 'Spline':
if i != n-1:
raise RuntimeError('Spline keyword encountered reading tables '
'for %s-%s before reaching the end of the '
'HS table.' % ( symboli, symbolj ))
else:
HS += [ fortran_readline(l) ]
l = fileobj.readline()
# don't care if not spline...
#if not l:
# raise RuntimeError('Premature end-of-file: Keyword "Spline" not found '
# 'for %s-%s.' % ( symboli, symbolj ))
HS = np.array(HS)
x = dx*np.arange(0, HS.shape[0])
#return x[0:HS.shape[0]], np.array(HS)
return x, np.array(HS)
def read_skf_table(parfile, s1, s2):
""" Read SlaKo tables from .skf parameter file for elements with symbols s1,s2."""
f = open(parfile)
dx, n = fortran_readline(f)
dx, n = float(dx), int(n)
f.readline()
if s1 == s2: #additional line in homonuclear table
f.readline()
table = []
for i in range(n):
line = fortran_readline(f.readline())
if len(line) > 0:
table += [line]
return dx * np.arange(0, n - 1), np.array(table)
def read_skf_table(parfile,s1,s2):
""" Read SlaKo tables from .skf parameter file for elements with symbols s1,s2."""
f=open(parfile)
dx,n = fortran_readline(f)
dx,n = float(dx), int(n)
f.readline()
if s1==s2: #additional line in homonuclear table
f.readline()
table = []
for i in range(n):
line = fortran_readline( f.readline() )
if len(line)>0:
table += [line]
return dx*np.arange(0,n-1), np.array(table)
def read_element_from_skf(fileobj, symbol):
"""
Read element data from DFTB-style .skf file.
Parameters:
-----------
fileobj: filename of file-object to read from
symbol: chemical symbol of the element
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
# First line contains grid spacing and number of grid points, ignore
fileobj.readline()
# Self-energies, spin-polarization energies, Hubbard-U
eself = [0.0] * 3
U = [0.0] * 3
q = [0.0] * 3
eself[0], eself[1], eself[2], espin, \
U[0], U[1], U[2], q[0], q[1], q[2] = fortran_readline(fileobj)
# Initialize data from database
data = copy(box_data[symbol])
data['epsilon'] = {}
for orbital in data['valence_orbitals']:
if 's' in orbital:
data['epsilon'][orbital] = eself[2]
elif 'p' in orbital:
data['epsilon'][orbital] = eself[1]
elif 'd' in orbital:
data['epsilon'][orbital] = eself[0]
# Apparently, only the last U is used
data['U'] = U[2]
data['FWHM'] = sqrt(8 * log(2.0) / pi) / data['U']
energies = []
for orbital in data['valence_orbitals']:
eps = data['epsilon'][orbital]
if 's' in orbital: n = 1
elif 'p' in orbital: n = 3
elif 'd' in orbital: n = 5
energies.extend([eps] * n)
data['onsite_energies'] = energies
data['nr_basis_orbitals'] = len(energies)
data['valence_energies'] = np.array(energies, dtype=float)
data['comment'] = None
# .skf files do not contain basis information
return data, None
def read_element_from_skf(fileobj, symbol):
"""
Read element data from DFTB-style .skf file.
Parameters:
-----------
fileobj: filename of file-object to read from
symbol: chemical symbol of the element
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
# First line contains grid spacing and number of grid points, ignore
fileobj.readline()
# Self-energies, spin-polarization energies, Hubbard-U
eself = [ 0.0 ]*3
U = [ 0.0 ]*3
q = [ 0.0 ]*3
eself[0], eself[1], eself[2], espin, \
U[0], U[1], U[2], q[0], q[1], q[2] = fortran_readline(fileobj)
# Initialize data from database
data = copy(box_data[symbol])
data['epsilon'] = { }
for orbital in data['valence_orbitals']:
if 's' in orbital:
data['epsilon'][orbital] = eself[2]
elif 'p' in orbital:
data['epsilon'][orbital] = eself[1]
elif 'd' in orbital:
data['epsilon'][orbital] = eself[0]
# Apparently, only the last U is used
data['U'] = U[2]
data['FWHM'] = sqrt(8*log(2.0)/pi)/data['U']
energies=[]
for orbital in data['valence_orbitals']:
eps = data['epsilon'][orbital]
if 's' in orbital: n=1
elif 'p' in orbital: n=3
elif 'd' in orbital: n=5
energies.extend( [eps]*n )
data['onsite_energies'] = energies
data['nr_basis_orbitals'] = len(energies)
data['valence_energies'] = np.array(energies, dtype=float)
data['comment'] = None
# .skf files do not contain basis information
return data, None
def read_repulsion_from_skf(fileobj, rep_x0=0.1, rep_dx=0.005):
"""
Read repulsion from DFTB-style .skf file.
The repulsion in the .skf-file consists of an exponential and a spline
part. Both will be converted to a single table.
Parameters:
-----------
fileobj: filename of file-object to read from
rep_x0: minimum value of the repulsion table
rep_dx: step size for discretization
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
l = fileobj.readline()
while l and l.strip() != 'Spline':
l = fileobj.readline()
if not l:
# no spline
fileobj.seek(0)
dx, n = fortran_readline(fileobj) #ignore
items = fortran_readline(fileobj)
if len(
items
) < 20: # in homonuclear tables one element info came before repulsion
items = fortran_readline(fileobj)
cutoff = items[9]
c = [0, 0] + items[1:9]
x = np.linspace(rep_x0, cutoff, int((cutoff - rep_x0) / rep_dx) + 1)
y = np.zeros_like(x)
for i in range(2, 10):
y = y + c[i] * (cutoff - x)**i
#raise RuntimeError('Could not find "Spline" keyword when reading '
# 'repulsion.')
else:
n, cutoff = fortran_readline(fileobj)
n = int(n)
# Coefficients for the exponential
c1, c2, c3 = fortran_readline(fileobj)
x1, x2, splc1, splc2, splc3, splc4 = fortran_readline(fileobj)
x = np.linspace(rep_x0, cutoff, int((cutoff - rep_x0) / rep_dx) + 1)
y = c3 + np.exp(c2 - c1 * x)
i0 = np.searchsorted(x, x1) + 1
for j in range(n - 1):
if j > 0:
last_x2 = x2
x1, x2, splc1, splc2, splc3, splc4 = fortran_readline(fileobj)
assert x1 == last_x2
i1 = np.searchsorted(x, x2) + 1
y[i0:i1] = \
splc1 + \
splc2 * (x[i0:i1]-x1) + \
splc3 * (x[i0:i1]-x1)**2 + \
splc4 * (x[i0:i1]-x1)**3
i0 = i1
# The last entry is a fifth-order polynomial
last_x2 = x2
x1, x2, splc1, splc2, splc3, splc4, splc5, splc6 = fortran_readline(
fileobj)
assert x1 == last_x2
i1 = np.searchsorted(x, x2) + 1
y[i0:i1] = \
splc1 + \
splc2 * (x[i0:i1]-x1) + \
splc3 * (x[i0:i1]-x1)**2 + \
splc4 * (x[i0:i1]-x1)**3 + \
splc5 * (x[i0:i1]-x1)**4 + \
splc6 * (x[i0:i1]-x1)**5
return x, y
def read_repulsion_from_skf(fileobj, rep_x0=0.1, rep_dx=0.005):
"""
Read repulsion from DFTB-style .skf file.
The repulsion in the .skf-file consists of an exponential and a spline
part. Both will be converted to a single table.
Parameters:
-----------
fileobj: filename of file-object to read from
rep_x0: minimum value of the repulsion table
rep_dx: step size for discretization
"""
if isinstance(fileobj, str):
fileobj = open(fileobj)
l = fileobj.readline()
while l and l.strip() != 'Spline':
l = fileobj.readline()
if not l:
# no spline
fileobj.seek(0)
dx, n = fortran_readline(fileobj) #ignore
items = fortran_readline(fileobj)
if len(items)<20: # in homonuclear tables one element info came before repulsion
items = fortran_readline(fileobj)
cutoff = items[9]
c = [0,0] + items[1:9]
x = np.linspace(rep_x0, cutoff, int((cutoff-rep_x0)/rep_dx)+1)
y = np.zeros_like(x)
for i in range(2,10):
y = y + c[i]*(cutoff-x)**i
#raise RuntimeError('Could not find "Spline" keyword when reading '
# 'repulsion.')
else:
n, cutoff = fortran_readline(fileobj)
n = int(n)
# Coefficients for the exponential
c1, c2, c3 = fortran_readline(fileobj)
x1, x2, splc1, splc2, splc3, splc4 = fortran_readline(fileobj)
x = np.linspace(rep_x0, cutoff, int((cutoff-rep_x0)/rep_dx)+1)
y = c3 + np.exp(c2-c1*x)
i0 = np.searchsorted(x, x1)+1
for j in range(n-1):
if j > 0:
last_x2 = x2
x1, x2, splc1, splc2, splc3, splc4 = fortran_readline(fileobj)
assert x1 == last_x2
i1 = np.searchsorted(x, x2)+1
y[i0:i1] = \
splc1 + \
splc2 * (x[i0:i1]-x1) + \
splc3 * (x[i0:i1]-x1)**2 + \
splc4 * (x[i0:i1]-x1)**3
i0 = i1
# The last entry is a fifth-order polynomial
last_x2 = x2
x1, x2, splc1, splc2, splc3, splc4, splc5, splc6 = fortran_readline(fileobj)
assert x1 == last_x2
i1 = np.searchsorted(x, x2)+1
y[i0:i1] = \
splc1 + \
splc2 * (x[i0:i1]-x1) + \
splc3 * (x[i0:i1]-x1)**2 + \
splc4 * (x[i0:i1]-x1)**3 + \
splc5 * (x[i0:i1]-x1)**4 + \
splc6 * (x[i0:i1]-x1)**5
return x, y