def read_gaussian_log(fname,
all_mo=False,
spin=None,
orientation='standard',
i_link=-1,
i_geo=-1,
i_ao=-1,
i_mo=-1,
interactive=True,
**kwargs):
'''Reads all information desired from a Gaussian .log file.
**Parameters:**
fname: str, file descriptor
Specifies the filename for the input file.
fname can also be used with a file descriptor instad of a filename.
all_mo : bool, optional
If True, all molecular orbitals are returned.
spin : {None, 'alpha', or 'beta'}, optional
If not None, returns exclusively 'alpha' or 'beta' molecular orbitals.
orientation : string, choices={'input', 'standard'}, optional
Specifies orientation of the molecule in Gaussian nomenclature. [#first]_
i_link : int, default=-1
Selects the file for linked Gaussian jobs.
i_geo : int, default=-1
Selects the geometry section of the output file.
i_ao : int, default=-1
Selects the atomic orbital section of the output file.
i_mo : int, default=-1
Selects the molecular orbital section of the output file.
interactive : bool
If True, the user is asked to select the different sets.
**Returns:**
qc (class QCinfo) with attributes geo_spec, geo_info, ao_spec, ao_spherical, mo_spec, etot :
See :ref:`Central Variables` for details.
.. [#first] Attention: The MOs in the output are only valid for the standard orientation!
'''
if isinstance(fname, str):
filename = fname
fname = descriptor_from_file(filename, index=0)
else:
filename = fname.name
flines = fname.readlines() # Read the WHOLE file into RAM
if isinstance(fname, str):
fname.close() # Leave existing file descriptors alive
# Search the file the specific sections
count = {
'link': 0,
'geometry': 0,
'geometry_input': 0,
'atomic orbitals': 0,
'molecular orbitals': [],
'state': []
}
def check_sel(count, i, interactive=False, default=-1):
if count == 0:
raise IndexError
elif count == 1:
return 0
message = '\tPlease give an integer from 0 to {0} (default: {0}): '.format(
count - 1)
try:
if interactive:
i = raw_input(message)
i = default if i == '' else int(i)
i = range(count)[i]
except (IndexError, ValueError):
raise IOError(message.replace(':', '!'))
else:
display('\tSelecting the %s' %
('last element.' if
(i == count - 1) else 'element %d.' % i))
return i
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
# Check the file for keywords
if ' Entering Link 1' in line:
count['link'] += 1
try:
display('\tFound %d linked GAUSSIAN files.' % count['link'])
i_link = check_sel(count['link'], i_link, interactive=interactive)
except IndexError:
raise IOError('Found no `Entering Link 1` keyword!')
cartesian_basis = True
c_link = 0
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
thisline = line.split() # The current line split into segments
# Check the file for keywords
if ' Entering Link 1' in line:
c_link += 1
if i_link == (c_link - 1):
if ' orientation:' in line:
if '%s orientation:' % orientation in line.lower():
count['geometry'] += 1
if 'input orientation:' in line.lower():
count['geometry_input'] += 1
elif 'Standard basis:' in line or 'General basis read from cards:' in line:
# Check if a cartesian basis has been applied
if '(5D, 7F)' in line:
cartesian_basis = False
elif '(6D, 10F)' not in line:
raise IOError(
'Please apply a Spherical Harmonics (5D, 7F) or ' +
'a Cartesian Gaussian Basis Set (6D, 10F)!')
elif 'AO basis set' in line:
count['atomic orbitals'] += 1
elif 'The electronic state is ' in line:
count['state'].append(thisline[-1][:-1])
elif 'Orbital Coefficients:' in line:
mo_type = thisline[0]
if mo_type != 'Beta':
count['molecular orbitals'].append(mo_type)
else:
count['molecular orbitals'][-1] = 'Alpha&Beta'
display('\nContent of the GAUSSIAN .log file:')
display('\tFound %d geometry section(s). (%s orientation)' %
(count['geometry'], orientation))
try:
i_geo = check_sel(count['geometry'], i_geo, interactive=interactive)
except IndexError:
count['geometry'] = count['geometry_input']
orientation = 'input'
display('\Looking for "Input orientation": \n' +
'\tFound %d geometry section(s). (%s orientation)' %
(count['geometry'], orientation))
try:
i_geo = check_sel(count['geometry'],
i_geo,
interactive=interactive)
except IndexError:
raise IOError('Found no geometry section!' +
' Are you sure this is a GAUSSIAN .log file?')
try:
display('\tFound %d atomic orbitals section(s) %s.' %
(count['atomic orbitals'],
'(6D, 10F)' if cartesian_basis else '(5D, 7F)'))
i_ao = check_sel(count['atomic orbitals'],
i_ao,
interactive=interactive)
except IndexError:
raise IOError('Write GFINPUT in your GAUSSIAN route section to print' +
' the basis set information!')
try:
display('\tFound the following %d molecular orbitals section(s):' %
len(count['molecular orbitals']))
except IndexError:
raise IOError(
'Write IOP(6/7=3) in your GAUSSIAN route section to print\n' +
' all molecular orbitals!')
for i, j in enumerate(count['molecular orbitals']):
string = '\t\tSection %d: %s Orbitals' % (i, j)
try:
string += ' (electronic state: %s)' % count['state'][i]
except IndexError:
pass
display(string)
i_mo = check_sel(len(count['molecular orbitals']),
i_mo,
interactive=interactive)
if spin is not None:
if spin != 'alpha' and spin != 'beta':
raise IOError('`spin=%s` is not a valid option' % spin)
else:
display('Reading only molecular orbitals of spin %s.' % spin)
# Set a counter for the AOs
basis_count = 0
# Initialize some variables
sec_flag = None
skip = 0
c_link = 0
c_geo = 0
c_ao = 0
c_mo = 0
c_sao = 0
old_ao = -1
orb_sym = []
qc = QCinfo()
index = []
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
thisline = line.split() # The current line split into segments
# Check the file for keywords
if ' Entering Link 1' in line:
c_link += 1
if i_link == (c_link - 1):
if '%s orientation:' % orientation in line.lower():
# The section containing information about
# the molecular geometry begins
if i_geo == c_geo:
qc.geo_info = []
qc.geo_spec = []
sec_flag = 'geo_info'
c_geo += 1
skip = 4
elif 'Standard basis:' in line or 'General basis read from cards:' in line:
# Check if a cartesian basis has been applied
if '(5D, 7F)' in line:
cartesian_basis = False
elif '(6D, 10F)' not in line:
raise IOError(
'Please apply a Spherical Harmonics (5D, 7F) or ' +
'a Cartesian Gaussian Basis Sets (6D, 10F)!')
elif 'AO basis set' in line:
# The section containing information about
# the atomic orbitals begins
if i_ao == c_ao:
qc.ao_spec = []
if not cartesian_basis:
qc.ao_spherical = []
sec_flag = 'ao_info'
c_ao += 1
basis_count = 0
bNew = True # Indication for start of new AO section
elif 'Orbital symmetries:' in line:
sec_flag = 'mo_sym'
add = ''
orb_sym = []
elif 'Orbital Coefficients:' in line:
# The section containing information about
# the molecular orbitals begins
if (i_mo == c_mo):
sec_flag = 'mo_info'
mo_type = count['molecular orbitals'][i_mo]
qc.mo_spec = []
offset = 0
add = ''
orb_spin = []
if orb_sym == []:
if 'Alpha' in mo_type:
add = '_a'
orb_spin = ['alpha'] * basis_count
orb_sym = ['A1' + add] * basis_count
if 'Beta' in mo_type:
add = '_b'
orb_spin += ['beta'] * basis_count
orb_sym += ['A1' + add] * basis_count
for i in range(len(orb_sym)):
# for numpy version < 1.6
c = ((numpy.array(orb_sym[:i + 1]) == orb_sym[i]) !=
0).sum()
# for numpy version >= 1.6 this could be used:
#c = numpy.count_nonzero(numpy.array(orb_sym[:i+1]) == orb_sym[i])
qc.mo_spec.append({
'coeffs': numpy.zeros(basis_count),
'energy': 0.,
'sym': '%d.%s' % (c, orb_sym[i])
})
if orb_spin != []:
qc.mo_spec[-1]['spin'] = orb_spin[i]
if mo_type != 'Beta':
c_mo += 1
bNew = True # Indication for start of new MO section
elif 'E(' in line:
qc.etot = float(line.split('=')[1].split()[0])
else:
# Check if we are in a specific section
if sec_flag == 'geo_info':
if not skip:
qc.geo_info.append(
[thisline[1], thisline[0], thisline[1]])
qc.geo_spec.append([float(ij) for ij in thisline[3:]])
if '-----------' in flines[il + 1]:
sec_flag = None
else:
skip -= 1
if sec_flag == 'ao_info':
# Atomic orbital section
if ' ****' in line:
# There is a line with stars after every AO
bNew = True
# If there is an additional blank line, the AO section is complete
if flines[il + 1].split() == []:
sec_flag = None
elif bNew:
# The following AOs are for which atom?
bNew = False
at_num = int(thisline[0]) - 1
ao_num = 0
elif len(thisline) == 4:
# AO information section
# Initialize a new dict for this AO
ao_num = 0 # Initialize number of atomic orbiatls
ao_type = thisline[0].lower() # Type of atomic orbital
pnum = int(thisline[1]) # Number of primatives
for i_ao in ao_type:
# Calculate the degeneracy of this AO and increase basis_count
basis_count += l_deg(
lquant[i_ao], cartesian_basis=cartesian_basis)
qc.ao_spec.append({
'atom': at_num,
'type': i_ao,
'pnum': pnum,
'coeffs': numpy.zeros((pnum, 2))
})
else:
# Append the AO coefficients
coeffs = numpy.array(line.replace('D', 'e').split(),
dtype=numpy.float64)
for i_ao in range(len(ao_type)):
qc.ao_spec[-len(ao_type) +
i_ao]['coeffs'][ao_num, :] = [
coeffs[0], coeffs[1 + i_ao]
]
ao_num += 1
if sec_flag == 'mo_sym':
if 'electronic state' in line:
sec_flag = None
else:
info = line[18:].replace('(', '').replace(')',
'').split()
if 'Alpha' in line:
add = '_a'
elif 'Beta' in line:
add = '_b'
for i in info:
orb_sym.append(i + add)
if sec_flag == 'mo_info':
# Molecular orbital section
info = line[:21].split()
if info == []:
coeffs = line[21:].split()
if bNew:
index = [offset + i for i in range(len(coeffs))]
bNew = False
else:
for i, j in enumerate(index):
qc.mo_spec[j]['occ_num'] = int(
'O' in coeffs[i])
if mo_type not in 'Alpha&Beta':
qc.mo_spec[j]['occ_num'] *= 2
elif 'Eigenvalues' in info:
coeffs = line[21:].replace('-', ' -').split()
if mo_type == 'Natural':
key = 'occ_num'
else:
key = 'energy'
for i, j in enumerate(index):
qc.mo_spec[j][key] = float(coeffs[i])
else:
coeffs = line[21:].replace('-', ' -').split()
if not cartesian_basis and offset == 0:
if old_ao != line[:14].split()[-1] or len(
line[:14].split()) == 4:
old_ao = line[:14].split()[-1]
c_sao += 1
i = c_sao - 1
l = lquant[line[13].lower()]
m = line[14:21].replace(' ', '').lower()
p = 'yzx'.find(m) if len(m) == 1 else -1
if p != -1:
m = p - 1
elif m == '':
m = 0
else:
m = int(m)
qc.ao_spherical.append([i, (l, m)])
for i, j in enumerate(index):
qc.mo_spec[j]['coeffs'][int(info[0]) - 1] = float(
coeffs[i])
if int(info[0]) == basis_count:
bNew = True
offset = index[-1] + 1
if index[-1] + 1 == len(orb_sym):
sec_flag = None
orb_sym = []
# Are all MOs requested for the calculation?
if not all_mo:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['occ_num'] < 0.0000001:
del qc.mo_spec[i]
if spin is not None:
if orb_spin == []:
raise IOError(
'You requested `%s` orbitals, but None of them are present.' %
spin)
else:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['spin'] != spin:
del qc.mo_spec[i]
# Convert geo_info and geo_spec to numpy.ndarrays
qc.format_geo(is_angstrom=True)
return qc
def read_gaussian_fchk(fname, all_mo=False, spin=None, **kwargs):
'''Reads all information desired from a Gaussian FChk file.
**Parameters:**
fname: str, file descriptor
Specifies the filename for the input file.
fname can also be used with a file descriptor instad of a filename.
all_mo : bool, optional
If True, all molecular orbitals are returned.
**Returns:**
qc (class QCinfo) with attributes geo_spec, geo_info, ao_spec, mo_spec, etot :
See :ref:`Central Variables` for details.
'''
if isinstance(fname, str):
filename = fname
fname = descriptor_from_file(filename, index=0)
else:
filename = fname.name
flines = fname.readlines() # Read the WHOLE file into RAM
if isinstance(fname, str):
fname.close() # Leave existing file descriptors alive
# Is this an unrestricted calculation?
has_beta = False
is_6D = False
is_10F = False
for line in flines:
if 'beta mo coefficients' in line.lower():
has_beta = True
if 'Pure/Cartesian d shells' in line:
is_6D = int(line.split()[-1]) == 1
if 'Pure/Cartesian f shells' in line:
is_10F = int(line.split()[-1]) == 1
cartesian_basis = (is_6D and is_10F)
if ((not is_6D) and is_10F) or (is_6D and (not is_10F)):
raise IOError('Please apply a Spherical Harmonics (5D, 7F) or '+
'a Cartesian Gaussian Basis Set (6D, 10F)!')
if spin is not None:
if spin != 'alpha' and spin != 'beta':
raise IOError('`spin=%s` is not a valid option' % spin)
elif has_beta:
display('Reading only molecular orbitals of spin %s.' % spin)
else:
raise IOError('The keyword `spin` is only supported for unrestricted calculations.')
restricted = (not has_beta)
sec_flag = None
el_num = [0,0]
mo_i0 = {'alpha': 0, 'beta': 0}
what = 'alpha'
index = 0
at_num = 0
ao_num = 0
ao_sp_coeffs = {}
switch = 0
qc = QCinfo()
qc.geo_info = [[],[],[]]
if not cartesian_basis:
qc.ao_spherical = []
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
thisline = line.split() # The current line split into segments
# Check the file for keywords
if 'Number of alpha electrons' in line:
el_num[0] = int(thisline[5])
elif 'Number of beta electrons' in line:
el_num[1] = int(thisline[5])
elif 'Number of basis functions' in line:
basis_number = int(thisline[5])
elif 'Atomic numbers' in line:
sec_flag = 'geo_info'
index = 0
at_num = int(thisline[-1])
count = 0
qc.geo_info[1] = list(range(1,at_num+1))
elif 'Nuclear charges' in line:
sec_flag = 'geo_info'
index = 2
at_num = int(thisline[-1])
count = 0
elif 'Total Energy' in line:
qc.etot = float(thisline[3])
elif 'Current cartesian coordinates' in line:
at_num = int(thisline[-1])/3
sec_flag = 'geo_pos'
qc.geo_spec = []
count = 0
xyz = []
elif 'Shell types' in line:
sec_flag = 'ao_info'
index = 'type'
ao_num = int(thisline[-1])
count = 0
if qc.ao_spec == []:
for ii in range(ao_num):
qc.ao_spec.append({})
elif 'Number of primitives per shell' in line:
sec_flag = 'ao_info'
index = 'pnum'
ao_num = int(thisline[-1])
count = 0
if qc.ao_spec == []:
for ii in range(ao_num):
qc.ao_spec.append({})
elif 'Shell to atom map' in line:
sec_flag = 'ao_info'
index = 'atom'
ao_num = int(thisline[-1])
count = 0
if qc.ao_spec == []:
for ii in range(ao_num):
qc.ao_spec.append({})
elif 'Primitive exponents' in line:
sec_flag = 'ao_coeffs'
ao_num = int(thisline[-1])
count = 0
switch = 0
index = 0
if qc.ao_spec == []:
raise IOError('Shell types need to be defined before the AO exponents!')
if not 'coeffs' in qc.ao_spec[0].keys():
for ii in range(len(qc.ao_spec)):
pnum = qc.ao_spec[ii]['pnum']
qc.ao_spec[ii]['coeffs'] = numpy.zeros((pnum, 2))
elif 'Contraction coefficients' in line:
if 'P(S=P)' not in line:
sec_flag = 'ao_coeffs'
else:
sec_flag = 'ao_sp_coeffs'
ao_sp_coeffs = {0: []}
ao_num = int(thisline[-1])
count = 0
switch = 1
index = 0
if qc.ao_spec == []:
raise IOError('Shell types need to be defined before the AO exponents!')
if not 'coeffs' in qc.ao_spec[0].keys():
for ii in range(len(qc.ao_spec)):
pnum = qc.ao_spec[ii]['pnum']
qc.ao_spec[ii]['coeffs'] = numpy.zeros((pnum, 2))
elif 'Orbital Energies' in line:
sec_flag = 'mo_eorb'
mo_num = int(thisline[-1])
mo_i0[thisline[0].lower()] = len(qc.mo_spec)
if restricted:
if el_num[0] == el_num[1]:
i = el_num[0]
occ = 2
else:
i = el_num[0 if 'Alpha' in line else 1]
occ = 1
else:
i = el_num[0 if 'Alpha' in line else 1]
occ = 1
for ii in range(mo_num):
qc.mo_spec.append({'coeffs': numpy.zeros(basis_number),
'energy': 0.0,
'occ_num': float(occ if ii < i else 0),
'sym': '%i.1' % (ii+1),
'spin':thisline[0].lower()
})
elif 'MO coefficients' in line:
sec_flag = 'mo_coeffs'
count = 0
index = 0
mo_num = int(thisline[-1])
what = thisline[0].lower()
else:
# Check if we are in a specific section
if sec_flag == 'geo_info':
for ii in thisline:
qc.geo_info[index].append(ii)
count += 1
if count == at_num:
sec_flag = None
elif sec_flag == 'geo_pos':
for ii in thisline:
xyz.append(float(ii))
if len(xyz) == 3:
qc.geo_spec.append(xyz)
xyz = []
count += 1
if count == at_num:
sec_flag = None
elif sec_flag == 'ao_info':
for ii in thisline:
ii = int(ii)
if index is 'type':
ii = orbit[abs(ii)]
l = lquant[ii]
if not cartesian_basis:
for m in (range(0,l+1) if l != 1 else [1,0]):
qc.ao_spherical.append([count,(l,m)])
if m != 0:
qc.ao_spherical.append([count,(l,-m)])
elif index is 'atom':
ii -= 1
qc.ao_spec[count][index] = ii
count += 1
if count == ao_num:
sec_flag = None
elif sec_flag == 'ao_coeffs':
for ii in thisline:
qc.ao_spec[index]['coeffs'][count,switch] = float(ii)
count += 1
ao_num -= 1
if count == qc.ao_spec[index]['pnum']:
index += 1
count = 0
if not ao_num:
sec_flag = None
elif sec_flag == 'ao_sp_coeffs':
for ii in thisline:
ao_sp_coeffs[index].append(float(ii))
count += 1
ao_num -= 1
if count == qc.ao_spec[index]['pnum']:
index += 1
ao_sp_coeffs[index] = []
count = 0
if not ao_num:
sec_flag = None
elif sec_flag == 'mo_eorb':
for ii in thisline:
qc.mo_spec[count]['energy'] = float(ii)
count += 1
if index != 0 and not count % basis_number:
sec_flag = None
elif sec_flag == 'mo_coeffs':
for ii in thisline:
qc.mo_spec[mo_i0[what]+index]['coeffs'][count] = float(ii)
count += 1
if count == basis_number:
count = 0
index += 1
if index != 0 and not index % basis_number:
sec_flag = None
# Look for SP atomic orbitals
if ao_sp_coeffs:
ao_new = []
for i,ao in enumerate(qc.ao_spec):
if ao['type'] == 'p' and sum(numpy.abs(ao_sp_coeffs[i])) > 0:
ao_new.append(copy.deepcopy(ao))
ao_new[-1]['type'] = 's'
ao_new.append(ao)
ao_new[-1]['type'] = 'p'
ao_new[-1]['coeffs'][:,1] = numpy.array(ao_sp_coeffs[i])
else:
ao_new.append(ao)
qc.ao_spec = ao_new
# Are all MOs requested for the calculation?
if not all_mo:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['occ_num'] < 0.0000001:
del qc.mo_spec[i]
# Only molecular orbitals of one spin requested?
if spin is not None:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['spin'] != spin:
del qc.mo_spec[i]
if restricted:
# Closed shell calculation
for mo in qc.mo_spec:
del mo['spin']
else:
# Rename MOs according to spin
for mo in qc.mo_spec:
mo['sym'] += '_%s' % mo['spin'][0]
# Check for natural orbital occupations
energy_sum = sum([abs(i['energy']) for i in qc.mo_spec])
if energy_sum < 0.0000001:
display('Attention!\n\tThis FChk file contains natural orbitals. '+
'(There are no energy eigenvalues.)\n\t' +
'In this case, Gaussian does not print the respective natural' +
'occupation numbers!' )
qc.geo_info = numpy.array(qc.geo_info).T
# Convert geo_info and geo_spec to numpy.ndarrays
qc.format_geo(is_angstrom=False)
return qc
def convert_json(jData, all_mo=False, spin=None):
'''Converts a scanlog JSON data instance to an instance of
orbkit's QCinfo class.
**Parameters:**
jData : class
Contains the input JSON data.
all_mo : bool, optional
If True, all molecular orbitals are returned.
spin : {None, 'alpha', or 'beta'}, optional
If not None, returns exclusively 'alpha' or 'beta' molecular orbitals.
**Returns:**
qc (class QCinfo) with attributes geo_spec, geo_info, ao_spec, mo_spec, etot :
See :ref:`Central Variables` for details.
'''
aa_to_au = 1/0.52917720859
# Initialize the variables
qc = QCinfo()
# Converting all information concerning atoms and geometry
qc.geo_spec = numpy.array(jData['results']['geometry']['elements_3D_coords_converged']).reshape((-1, 3)) * aa_to_au
for ii in range(jData["molecule"]['nb_atoms']):
symbol = get_atom_symbol(atom=jData["molecule"]['atoms_Z'][ii])
qc.geo_info.append([symbol,str(ii+1),str(jData["molecule"]['atoms_Z'][ii])])
# Convert geo_info and geo_spec to numpy.ndarrays
qc.format_geo()
# Converting all information about atomic basis set
from pickle import loads
gbasis = loads(bytes(jData['comp_details']['general']['basis_set'], 'utf-8'))
for ii in range(jData["molecule"]['nb_atoms']):
for jj in range(len(gbasis[ii])):
pnum = len(gbasis[ii][jj][1])
qc.ao_spec.append({'atom': ii,
'type': str(gbasis[ii][jj][0]).lower(),
'pnum': pnum,
'coeffs': numpy.zeros((pnum, 2))
})
for kk in range(pnum):
qc.ao_spec[-1]['coeffs'][kk][0] = gbasis[ii][jj][1][kk][0]
qc.ao_spec[-1]['coeffs'][kk][1] = gbasis[ii][jj][1][kk][1]
if "ao_names" in jData['comp_details']['general']:
# Reconstruct exponents list for ao_spec
aonames = jData['comp_details']['general']['ao_names']
cartesian_basis = True
for i in aonames:
if '+' in i or '-' in i:
cartesian_basis = False
# There is a problem here with the 6D 7F basis sets, that are a mixture of cartesian and spherical basis sets.
if not cartesian_basis:
qc.ao_spherical = []
count = 0
for i,ao in enumerate(qc.ao_spec):
l = l_deg(lquant[ao['type']],cartesian_basis=cartesian_basis)
if cartesian_basis:
ao['exp_list'] = []
for ll in range(l):
if cartesian_basis:
ao['exp_list'].append((aonames[count].lower().count('x'),
aonames[count].lower().count('y'),
aonames[count].lower().count('z')))
else:
m = aonames[count].lower().split('_')[-1]
m = m.replace('+',' +').replace('-',' -').replace('s','s 0').split(' ')
p = 'yzx'.find(m[0][-1])
if p != -1:
m = p - 1
else:
m = int(m[-1])
qc.ao_spherical.append([i,(lquant[ao['type']],m)])
count += 1
# Converting all information about molecular orbitals
ele_num = numpy.sum(jData["molecule"]['atoms_Z']) - numpy.sum(jData['comp_details']['general']['core_electrons_per_atoms']) - jData['molecule']['charge']
ue = (jData['molecule']['multiplicity']-1)
# Check for natural orbitals and occupation numbers
is_natorb = False
#if hasattr(ccData,'nocoeffs'):
# if not hasattr(ccData,'nooccnos'):
# raise IOError('There are natural orbital coefficients (`nocoeffs`) in the cclib' +
# ' ccData, but no natural occupation numbers (`nooccnos`)!')
# is_natorb = True
restricted = (len(jData['results']['wavefunction']['MO_energies']) == 1)
if spin is not None:
if spin != 'alpha' and spin != 'beta':
raise IOError('`spin=%s` is not a valid option' % spin)
elif restricted:
raise IOError('The keyword `spin` is only supported for unrestricted calculations.')
else:
display('Converting only molecular orbitals of spin %s.' % spin)
import scipy.sparse
sym = {}
shape = (jData['results']['wavefunction']['MO_number_kept'], jData['comp_details']['general']['basis_set_size'])
pre_mocoeffs = jData['results']["wavefunction"]["MO_coefs"]
if restricted:
add = ['']
orb_sym = [None]
mocoeffs = [numpy.asarray(scipy.sparse.csr_matrix(tuple([numpy.asarray(d) for d in pre_mocoeffs[0]]), shape=shape).todense())]
else:
add = ['_a','_b']
orb_sym = ['alpha','beta']
mocoeffs = [numpy.asarray(scipy.sparse.csr_matrix(tuple([numpy.asarray(d) for d in pre_mocoeffs[0]]), shape=shape).todense()),
numpy.asarray(scipy.sparse.csr_matrix(tuple([numpy.asarray(d) for d in pre_mocoeffs[1]]), shape=shape).todense())]
nmo = jData['results']['wavefunction']['MO_number'] if "nmo" in jData['results']['wavefunction'] else len(mocoeffs[0])
for ii in range(nmo):
for i,j in enumerate(add):
a = '%s%s' % (jData['results']['wavefunction']['MO_sym'][i][ii],j)
if a not in sym.keys(): sym[a] = 1
else: sym[a] += 1
#if is_natorb:
# occ_num = ccData.nooccnos[ii]
if not restricted:
occ_num = 1.0 if ii <= jData['results']['wavefunction']['homo_indexes'][i] else 0.0
elif ele_num > ue:
occ_num = 2.0
ele_num -= 2.0
elif ele_num > 0.0 and ele_num <= ue:
occ_num = 1.0
ele_num -= 1.0
ue -= 1.0
else:
occ_num = 0.0
qc.mo_spec.append({'coeffs': mocoeffs[i][ii],
'energy': jData['results']['wavefunction']['MO_energies'][i][ii],
'occ_num': occ_num,
'sym': '%d.%s' %(sym[a],a)
})
if orb_sym[i] is not None:
qc.mo_spec[-1]['spin'] = orb_sym[i]
if spin is not None and spin != orb_sym[i]:
del qc.mo_spec[-1]
# Use default order for atomic basis functions if aonames is not present
if 'ao_names' not in jData['comp_details']['general']:
display('The attribute `aonames` is not present in the parsed data.')
display('Using the default order of basis functions.')
# Check which basis functions have been used
c_cart = sum([l_deg(l=ao['type'], cartesian_basis=True) for ao in qc.ao_spec])
c_sph = sum([l_deg(l=ao['type'], cartesian_basis=False) for ao in qc.ao_spec])
c = create_mo_coeff(qc.mo_spec,'').shape[-1]
if c != c_cart and c == c_sph: # Spherical basis
qc.ao_spherical = get_ao_spherical(qc.ao_spec,p=[0,1])
elif c != c_cart:
display('Warning: The basis set type does not match with pure spherical ' +
'or pure Cartesian basis!')
display('Please specify qc.mo_spec["exp_list"] and/or qc.ao_spherical by your self.')
# Are all MOs requested for the calculation?
if not all_mo:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['occ_num'] < 0.0000001:
del qc.mo_spec[i]
return qc
def read_molden(fname,
all_mo=False,
spin=None,
i_md=-1,
interactive=True,
**kwargs):
'''Reads all information desired from a molden file.
**Parameters:**
fname: str, file descriptor
Specifies the filename for the input file.
fname can also be used with a file descriptor instad of a filename.
all_mo : bool, optional
If True, all molecular orbitals are returned.
spin : {None, 'alpha', or 'beta'}, optional
If not None, returns exclusively 'alpha' or 'beta' molecular orbitals.
i_md : int, default=-1
Selects the `[Molden Format]` section of the output file.
interactive : bool
If True, the user is asked to select the different sets.
**Returns:**
qc (class QCinfo) with attributes geo_spec, geo_info, ao_spec, mo_spec, etot :
See :ref:`Central Variables` for details.
'''
molden_regex = re.compile(r"\[[ ]{,}[Mm]olden[ ]+[Ff]ormat[ ]{,}\]")
if isinstance(fname, str):
filename = fname
fname = descriptor_from_file(filename, index=0)
else:
filename = fname.name
flines = fname.readlines() # Read the WHOLE file into RAM
if isinstance(fname, str):
fname.close() # Leave existing file descriptors alive
def check_sel(count, i, interactive=False):
if count == 0:
raise IndexError
elif count == 1:
return 0
message = '\tPlease give an integer from 0 to {0}: '.format(count - 1)
try:
if interactive:
i = int(raw_input(message))
i = range(count)[i]
except (IndexError, ValueError):
raise IOError(message.replace(':', '!'))
else:
display('\tSelecting the %s' %
('last element.' if
(i == count - 1) else 'element %d.' % i))
return i
has_alpha = []
has_beta = []
restricted = []
cartesian_basis = []
mixed_warning = []
by_orca = []
count = 0
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
# Check the file for keywords
if molden_regex.search(line):
count += 1
has_alpha.append(False)
has_beta.append(False)
restricted.append(False)
cartesian_basis.append(True)
mixed_warning.append(False)
by_orca.append(False)
if 'orca' in line.lower():
by_orca[-1] = True
if '[5d]' in line.lower() or '[5d7f]' in line.lower():
cartesian_basis[-1] = False
if '[5d10f]' in line.lower():
mixed_warning[-1] = '5D, 10F'
cartesian_basis[-1] = False
if '[7f]' in line.lower():
mixed_warning[-1] = '6D, 7F'
cartesian_basis[-1] = True
if 'Spin' in line and 'alpha' in line.lower():
has_alpha[-1] = True
if 'Spin' in line and 'beta' in line.lower():
has_beta[-1] = True
if 'Occup' in line:
restricted[-1] = restricted[-1] or (float(line.split('=')[1]) >
1. + 1e-4)
if count == 0:
raise IOError('The input file %s is no valid molden file!\n\nIt does' %
filename + ' not contain the keyword: [Molden Format]\n')
else:
if count > 1:
display('\nContent of the molden file:')
display('\tFound %d [Molden Format] keywords, i.e., ' % count +
'this file contains %d molden files.' % count)
i_md = check_sel(count, i_md, interactive=interactive)
if spin is not None:
if restricted[i_md]:
raise IOError(
'The keyword `spin` is only supported for unrestricted calculations.'
)
if spin != 'alpha' and spin != 'beta':
raise IOError('`spin=%s` is not a valid option' % spin)
elif spin == 'alpha' and has_alpha[i_md]:
display('Reading only molecular orbitals of spin alpha.')
elif spin == 'beta' and has_beta[i_md]:
display('Reading only molecular orbitals of spin beta.')
elif (not has_alpha[i_md]) and (not has_beta[i_md]):
raise IOError(
'Molecular orbitals in `molden` file do not contain `Spin=` keyword'
)
elif ((spin == 'alpha' and not has_alpha[i_md])
or (spin == 'beta' and not has_beta[i_md])):
raise IOError(
'You requested `%s` orbitals, but None of them are present.' %
spin)
# Set a counter for the AOs
basis_count = 0
sym = {}
# Declare synonyms for molden keywords
synonyms = {
'Sym': 'sym',
'Ene': 'energy',
'Occup': 'occ_num',
'Spin': 'spin'
}
MO_keys = synonyms.keys()
count = 0
max_l = 0
start_reading = False
# Go through the file line by line
for il in range(len(flines)):
line = flines[il] # The current line as string
thisline = line.split() # The current line split into segments
# Check the file for keywords
if '[molden format]' in line.lower():
# A new file begins
# Initialize the variables
if i_md == count:
qc = QCinfo()
sec_flag = False # A Flag specifying the current section
start_reading = True # Found the selected section
else:
start_reading = False
count += 1
continue
if start_reading:
if '_ENERGY=' in line:
try:
qc.etot = float(thisline[1])
except IndexError:
pass
elif '[atoms]' in line.lower():
# The section containing information about
# the molecular geometry begins
sec_flag = 'geo_info'
if 'Angs' in line:
# The length are given in Angstroem
# and have to be converted to Bohr radii --
aa_to_au = 1 / 0.52917720859
else:
# The length are given in Bohr radii
aa_to_au = 1.0
elif '[gto]' in line.lower():
# The section containing information about
# the atomic orbitals begins
sec_flag = 'ao_info'
bNew = True # Indication for start of new AO section
elif '[mo]' in line.lower():
# The section containing information about
# the molecular orbitals begins
sec_flag = 'mo_info'
bNew = True # Indication for start of new MO section
elif '[sto]' in line.lower():
# The orbkit does not support Slater type orbitals
raise IOError('orbkit does not work for STOs!\nEXIT\n')
elif '[' in line:
sec_flag = None
else:
# Check if we are in a specific section
if sec_flag == 'geo_info' and thisline != []:
# Geometry section
qc.geo_info.append(thisline[0:3])
qc.geo_spec.append(
[float(ii) * aa_to_au for ii in thisline[3:]])
if sec_flag == 'ao_info':
# Atomic orbital section
def check_int(i):
try:
int(i)
return True
except ValueError:
return False
if thisline == []:
# There is a blank line after every AO
bNew = True
elif bNew:
# The following AOs are for which atom?
bNew = False
at_num = int(thisline[0]) - 1
ao_num = 0
elif len(thisline) == 3 and check_int(thisline[1]):
# AO information section
# Initialize a new dict for this AO
ao_num = 0 # Initialize number of atomic orbiatls
ao_type = thisline[
0] # Which type of atomic orbital do we have
pnum = int(thisline[1]) # Number of primatives
# Calculate the degeneracy of this AO and increase basis_count
for i_ao in ao_type:
# Calculate the degeneracy of this AO and increase basis_count
basis_count += l_deg(
lquant[i_ao],
cartesian_basis=cartesian_basis[i_md])
max_l = max(max_l, lquant[i_ao])
qc.ao_spec.append({
'atom':
at_num,
'type':
i_ao,
'pnum':
-pnum if by_orca[i_md] else pnum,
'coeffs':
numpy.zeros((pnum, 2))
})
else:
# Append the AO coefficients
coeffs = numpy.array(line.replace('D', 'e').split(),
dtype=numpy.float64)
for i_ao in range(len(ao_type)):
qc.ao_spec[-len(ao_type) +
i_ao]['coeffs'][ao_num, :] = [
coeffs[0], coeffs[1 + i_ao]
]
ao_num += 1
if sec_flag == 'mo_info':
# Molecular orbital section
if '=' in line:
# MO information section
if bNew:
# Create a numpy array for the MO coefficients and
# for backward compability create a simple counter for 'sym'
qc.mo_spec.append({
'coeffs':
numpy.zeros(basis_count),
'sym':
'%d.1' % (len(qc.mo_spec) + 1)
})
bNew = False
# Append information to dict of this MO
info = line.replace('\n', '').replace(' ', '')
info = info.split('=')
if info[0] in MO_keys:
if info[0] == 'Spin':
info[1] = info[1].lower()
elif info[0] != 'Sym':
info[1] = float(info[1])
elif not '.' in info[1]:
from re import search
try:
a = search(r'\d+', info[1]).group()
if a == info[1]:
info[1] = '%s.1' % a
elif info[1].startswith(a):
info[1] = info[1].replace(
a, '%s.' % a, 1)
else:
raise AttributeError
except AttributeError:
if info[1] not in sym.keys():
sym[info[1]] = 1
else:
sym[info[1]] += 1
info[1] = '%d.%s' % (sym[info[1]], info[1])
qc.mo_spec[-1][synonyms[info[0]]] = info[1]
else:
if ('[' or ']') in line:
# start of another section that is not (yet) read
sec_flag = None
else:
# Append the MO coefficients
bNew = True # Reset bNew
index = int(thisline[0]) - 1
try:
# Try to convert coefficient to float
qc.mo_spec[-1]['coeffs'][index] = float(
thisline[1])
except ValueError:
# If it cannot be converted print error message
raise ValueError(
'Error in coefficient %d of MO %s!' %
(index, qc.mo_spec[-1]['sym']) +
'\nSetting this coefficient to zero...')
# Spherical basis?
if not cartesian_basis[i_md]:
qc.ao_spherical = get_ao_spherical(qc.ao_spec, p=[1, 0])
if max_l > 2 and mixed_warning[i_md]:
raise IOError('The input file %s contains ' % filename +
'mixed spherical and Cartesian function (%s).' %
mixed_warning[i_md] +
'ORBKIT does not support these basis functions yet. ' +
'Pleas contact us, if you need this feature!')
# Are all MOs requested for the calculation?
if not all_mo:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['occ_num'] < 0.0000001:
del qc.mo_spec[i]
# Only molecular orbitals of one spin requested?
if spin is not None:
for i in range(len(qc.mo_spec))[::-1]:
if qc.mo_spec[i]['spin'] != spin:
del qc.mo_spec[i]
if restricted[i_md]:
# Closed shell calculation
for mo in qc.mo_spec:
del mo['spin']
else:
# Rename MOs according to spin
for mo in qc.mo_spec:
mo['sym'] += '_%s' % mo['spin'][0]
# Orca uses for all molecular orbitals the same name
sym = [i['sym'] for i in qc.mo_spec]
if sym[1:] == sym[:-1]:
sym = sym[0].split('.')[-1]
for i in range(len(qc.mo_spec)):
qc.mo_spec[i]['sym'] = '%d.%s' % (i + 1, sym)
# Convert geo_info and geo_spec to numpy.ndarrays
qc.format_geo()
# Check the normalization
from orbkit.analytical_integrals import get_ao_overlap, get_lxlylz
norm = numpy.diagonal(get_ao_overlap(qc.geo_spec, qc.geo_spec, qc.ao_spec))
if sum(numpy.abs(norm - 1.)) > 1e-8:
display(
'The atomic orbitals are not normalized correctly, renormalizing...\n'
)
if not by_orca[i_md]:
j = 0
for i in range(len(qc.ao_spec)):
qc.ao_spec[i]['coeffs'][:, 1] /= numpy.sqrt(norm[j])
for n in range(
l_deg(lquant[qc.ao_spec[i]['type']],
cartesian_basis=True)):
j += 1
else:
qc.ao_spec[0]['N'] = 1 / numpy.sqrt(norm[:, numpy.newaxis])
if cartesian_basis[i_md]:
from orbkit.cy_overlap import ommited_cca_norm
cca = ommited_cca_norm(get_lxlylz(qc.ao_spec))
for mo in qc.mo_spec:
mo['coeffs'] *= cca
return qc