def mutateElement(self, oldZ, newZ):
if oldZ != newZ:
if type(oldZ) is str:
oldZ = qtk.n2Z(oldZ)
if type(newZ) is str:
newZ = qtk.n2Z(newZ)
newElem = qtk.Z2n(newZ)
indices = []
for i in range(self.N):
if abs(self.Z[i] - oldZ) < 0.1:
indices.append(i)
self.setAtoms(indices, element=newElem)
def mutateElement(self, oldZ, newZ):
if oldZ != newZ:
if type(oldZ) is str:
oldZ = qtk.n2Z(oldZ)
if type(newZ) is str:
newZ = qtk.n2Z(newZ)
newElem = qtk.Z2n(newZ)
indices = []
for i in range(self.N):
if abs(self.Z[i] - oldZ) < 0.1:
indices.append(i)
self.setAtoms(indices, element=newElem)
def get_index(inp_type):
if inp_type:
if type(inp_type) is str:
Z = qtk.n2Z(inp_type)
else:
try:
Z = int(inp_type)
except Exception as err:
qtk.exit("type not reconized with error:%s" % err)
return np.arange(self.N)[np.asarray(self.Z) == Z]
def read_xyz(self, name, **kwargs):
xyz = open(name, 'r')
content = xyz.readlines()
xyz.close()
content = [line.replace('\t', ' ') for line in content]
prop_list = ['charge', 'celldm', 'scale', 'symmetry']
for prop in prop_list:
try:
prop_str = filter(lambda x: prop in x, content)[0]
prop_str = re.sub('.*:', '', prop_str)
prop_data = prop_str.split(' ')
prop_data = filter(None, prop_data)
if len(prop_data) == 1:
try:
setattr(self, prop, float(prop_data[0]))
except Exception as exc:
if prop == 'symmetry':
setattr(self, prop, prop_data[0].strip())
elif len(prop_data) > 1:
setattr(self, prop, [float(_) for _ in prop_data])
except ValueError as exc:
setattr(self, prop, False)
qtk.warning("setting attribute %s with error: %s" % \
(prop, exc))
except:
setattr(self, prop, False)
if not self.charge:
self.charge = 0
if self.celldm or self.scale:
self.periodic = True
if self.celldm:
assert len(self.celldm) == 6
self.N = int(content[0])
coord_list = content[2 : self.N + 2]
coord = [filter(None,[a for a in entry.split(' ')])
for entry in coord_list]
type_list = list(np.array(coord)[:,0])
self.type_list = [str(elem) for elem in type_list]
self.Z = [qtk.n2Z(elem) for elem in self.type_list]
self.Z = np.array(self.Z)
self.R = np.array(coord)[:,1:4].astype(float)
self.box = False
if self.celldm:
self.periodic = True
angle = self.celldm[3:]
angle_sum = sum([abs(entry) for entry in angle])
if angle_sum == 0:
self.box = self.celldm[:3]
self.R_scale = copy.deepcopy(self.R)
self.R = qtk.fractional2xyz(self.R_scale, self.celldm)
def getMolecule(content):
comment_p = re.compile(r'^ *\t*(?!([!#])).*$')
data_p = re.compile(r'^[\ \t0-9\-\. ]+.*$')
dataContent = filter(comment_p.match, content[3:])
dataContent = filter(data_p.match, dataContent)
N = len(dataContent)
R = []
molData = []
for line in dataContent:
data = filter(None, line.replace('\n', '').split(' '))
molLine = [qtk.n2Z(data[3])]
molLine.extend([float(r) for r in data[:3]])
molData.append(molLine)
return molData
def getMolecule(content):
comment_p = re.compile(r'^ *\t*(?!([!#])).*$')
data_p = re.compile(r'^[\ \t0-9\-\. ]+.*$')
dataContent = filter(comment_p.match, content[3:])
dataContent = filter(data_p.match, dataContent)
N = len(dataContent)
R = []
molData = []
for line in dataContent:
data = filter(None, line.replace('\n', '').split(' '))
molLine = [qtk.n2Z(data[3])]
molLine.extend([float(r) for r in data[:3]])
molData.append(molLine)
return molData
def Molecules(file_name, **kwargs):
"""read nested xyz file and return molecule list"""
xyz = open(file_name, 'r')
content = xyz.readlines()
content = [line.replace('\t', ' ') for line in content]
xyz.close()
itr = 0
more_data = True
mols = []
while more_data:
try:
N = int(content[itr])
prop_list = content[itr + 1]
try:
prop_list = np.array(prop_list.split(' ')).astype(float)
except Exception as err:
qtk.warning(str(err))
prop_list = prop_list
coord_list = content[itr + 2:itr + N + 2]
coord = [
filter(None, [a for a in entry.split(' ')])
for entry in coord_list
]
type_list = list(np.array(coord)[:, 0])
type_list = [str(elem) for elem in type_list]
Z = np.array([qtk.n2Z(elem) for elem in type_list])
R = np.array(coord)[:, 1:4].astype(float)
mol_data = {}
for var in ['N', 'type_list', 'Z', 'R']:
mol_data[str(var)] = eval(var)
itr += N + 2
mols.append(qtk.Molecule(molecule_data=mol_data))
except Exception as err:
qtk.progress(
"Molecules", "%d molecules have been loaded with message %s." %
(len(mols), str(err)))
more_data = False
return mols
def read_vasp(chg_file):
with open(chg_file) as cfile:
content = cfile.readlines()
grid_1d = np.fromstring(''.join(content[2:5]), dtype=float, sep=' ')
grid = grid_1d.reshape([3, 3])
type_list = filter(None, content[5].split(' '))[:-1]
n_list = np.fromstring(content[6], dtype=int, sep=' ')
Z = []
for a in range(len(type_list)):
for n in range(n_list[a]):
Z.append(qtk.n2Z(type_list[a]))
Z = np.array(Z)[:, None]
N = sum(n_list)
R_scale = np.fromstring(''.join(content[8:8 + N]), dtype=float,
sep=' ').reshape([N, 3])
R = []
for i in range(N):
r = np.array([0, 0, 0])
for j in range(3):
r = r + R_scale[i, j] * grid.T[j] / 0.529177249
R.append(r)
R = np.array(R)
zcoord = np.hstack([Z, R])
V = np.dot(np.cross(grid[0], grid[1]), grid[2]) / (0.529177249)**3
step = np.fromstring(content[9 + N], dtype=int, sep=' ')
for i in range(3):
grid[i] = grid[i] / (step[i] * 0.529177249)
# numpy array.reshape does not change memory order
# use ravel to unwine new order
data = np.fromstring(''.join(content[10 + N:]), dtype=float, sep=' ')
data = data.reshape(step, order='F') / V
data_rav = data.ravel()
data = data_rav.reshape(step)
step = step[:, None]
grid = np.hstack([step, grid])
grid = np.vstack([[N, 0, 0, 0], grid])
return data, zcoord, grid
def read_vasp(chg_file):
with open(chg_file) as cfile:
content = cfile.readlines()
grid_1d = np.fromstring(''.join(content[2:5]), dtype=float, sep=' ')
grid = grid_1d.reshape([3,3])
type_list = filter(None, content[5].split(' '))[:-1]
n_list = np.fromstring(content[6], dtype=int, sep=' ')
Z = []
for a in range(len(type_list)):
for n in range(n_list[a]):
Z.append(qtk.n2Z(type_list[a]))
Z = np.array(Z)[:, None]
N = sum(n_list)
R_scale = np.fromstring(''.join(content[8:8+N]),
dtype=float, sep=' ').reshape([N, 3])
R = []
for i in range(N):
r = np.array([0,0,0])
for j in range(3):
r = r + R_scale[i, j] * grid.T[j] / 0.529177249
R.append(r)
R = np.array(R)
zcoord = np.hstack([Z, R])
V = np.dot(np.cross(grid[0], grid[1]), grid[2]) / (0.529177249)**3
step = np.fromstring(content[9+N], dtype=int, sep=' ')
for i in range(3):
grid[i] = grid[i] / (step[i] * 0.529177249)
# numpy array.reshape does not change memory order
# use ravel to unwine new order
data = np.fromstring(''.join(content[10+N:]), dtype=float, sep=' ')
data = data.reshape(step, order='F') / V
data_rav = data.ravel()
data = data_rav.reshape(step)
step = step[:, None]
grid = np.hstack([step, grid])
grid = np.vstack([[N, 0, 0, 0], grid])
return data, zcoord, grid
def haveBond(self, type_a, type_b):
result = False
type_a = type_a.title()
type_b = type_b.title()
if '0' not in self.bonds:
self.findBonds()
if qtk.n2Z(type_a) > qtk.n2Z(type_b):
atom_begin = qtk.n2Z(type_b)
atom_end = qtk.n2Z(type_a)
else:
atom_begin = qtk.n2Z(type_a)
atom_end = qtk.n2Z(type_b)
for key in self.bonds:
if self.bonds[key]['atom_begin'] == atom_begin \
and self.bonds[key]['atom_end'] == atom_end:
result = True
return result
def haveBond(self, type_a, type_b):
result = False
type_a = type_a.title()
type_b = type_b.title()
if '0' not in self.bonds:
self.findBonds()
if qtk.n2Z(type_a) > qtk.n2Z(type_b):
atom_begin = qtk.n2Z(type_b)
atom_end = qtk.n2Z(type_a)
else:
atom_begin = qtk.n2Z(type_a)
atom_end = qtk.n2Z(type_b)
for key in self.bonds:
if self.bonds[key]['atom_begin'] == atom_begin \
and self.bonds[key]['atom_end'] == atom_end:
result = True
return result
def setAtoms(self, index, **kwargs):
if type(index) is int:
index = [index]
if 'element' in kwargs or 'Z' in kwargs:
for i in index:
if 'element' in kwargs:
if type(kwargs['element']) is str:
Z = qtk.n2Z(kwargs['element'])
Zn = kwargs['element']
elif type(kwargs['element']) is int\
or type(kwargs['element']) is float:
Z = kwargs['element']
Zn = qtk.Z2n(Z)
elif 'Z' in kwargs:
Z = kwargs['Z']
Zn = qtk.Z2n(Z)
self.Z[i] = Z
self.type_list[i] = Zn
if 'string' in kwargs:
minZ = min(min(self.Z)-1, 0)
for i in index:
self.string[i] = kwargs['string']
self.Z[i] = minZ
def setAtoms(self, index, **kwargs):
if type(index) is int:
index = [index]
if 'element' in kwargs or 'Z' in kwargs:
for i in index:
if 'element' in kwargs:
if type(kwargs['element']) is str:
Z = qtk.n2Z(kwargs['element'])
Zn = kwargs['element']
elif type(kwargs['element']) is int\
or type(kwargs['element']) is float:
Z = kwargs['element']
Zn = qtk.Z2n(Z)
elif 'Z' in kwargs:
Z = kwargs['Z']
Zn = qtk.Z2n(Z)
self.Z[i] = Z
self.type_list[i] = Zn
if 'string' in kwargs:
minZ = min(min(self.Z)-1, 0)
for i in index:
self.string[i] = kwargs['string']
self.Z[i] = minZ
def writeInp(name=None, **setting):
self.setting.update(setting)
self.setting['no_molecule'] = False
inp, molecule = \
super(PlanewaveInput, self).write(name, **self.setting)
molecule.sort()
type_index = molecule.index
type_list = molecule.type_list
self.pp_path = qtk.setting.bigdft_pp
if 'pp_path' in self.setting:
self.pp_path = self.setting['pp_path']
if 'pp_theory' not in self.setting:
self.setting['pp_theory'] = self.setting['theory']
pp_files = []
typat = []
for a in range(len(type_index) - 1):
start = type_index[a]
end = type_index[a + 1]
for i in range(end - start):
typat.append(a + 1)
for i in range(molecule.N):
pp_file = 'psppar.' + molecule.type_list[i]
pp_list = set([pp[1] for pp in pp_files])
if pp_file not in pp_list:
pp_src = PPCheck(self.setting['theory'],
self.setting['pp_theory'], self.pp_path,
molecule.type_list[i])
pp_files.append([pp_src, pp_file])
##################
# scf section #
##################
self.content['scf'] = odict()
# restart check #
if 'restart' in self.setting and self.setting['restart']\
and 'band_scan' not in self.setting\
and 'dos_mesh' not in self.setting:
self.content['scf']['irdwfk'] = 1
self.content['scf']['getwfk'] = -1
if 'restart_density' in self.setting \
and self.setting['restart_density']\
and 'band_scan' not in self.setting\
and 'dos_mesh' not in self.setting:
self.content['scf']['irdden'] = 1
self.content['scf']['getden'] = -1
if 'kmesh' not in self.setting:
self.setting['kmesh'] = [1, 1, 1]
if 'kshift' not in self.setting:
self.setting['kshift'] = [0.0, 0.0, 0.0]
# kpoints check #
if 'kpoints' in self.setting:
self.content['scf']['kptopt'] = 0
self.content['scf']['nkpt'] = len(self.setting['kpoints'])
self.content['scf']['kpt'] = self.setting['kpoints']
else:
if self.setting['full_kmesh']:
self.content['scf']['kptopt'] = 3
self.content['scf']['ngkpt'] = self.setting['kmesh']
if len(np.array(self.setting['kshift']).shape) > 1:
self.content['scf']['nshiftk'] = len(
self.setting['kshift'])
self.content['scf']['shiftk'] = self.setting['kshift']
nbnd = None
if 'ks_states' in self.setting and self.setting['ks_states']:
vs = int(round(molecule.getValenceElectrons() / 2.0))
nbnd = self.setting['ks_states'] + vs
if 'd_shell' in self.setting:
for a in molecule.type_list:
if a in self.setting['d_shell'] and qtk.n2ve(a) < 10:
nbnd += 5
if 'band_scan' not in self.setting \
and 'dos_mesh' not in self.setting:
self.content['scf']['nband'] = nbnd
# system setup #
if molecule.charge != 0:
self.content['scf']['charge='] = molecule.charge
if molecule.multiplicity != 1:
self.content['scf']['nsppol'] = 2
self.content['scf']['occopt'] = 7
if 'save_restart' not in self.setting \
or not self.setting['save_restart']:
if 'restart' in self.setting and self.setting['restart']\
and ('band_scan' not in self.setting\
and 'dos_mesh' not in self.setting):
self.content['scf']['prtwf'] = 0
#if 'wf_convergence' in self.setting:
if 'band_scan' in self.setting\
or 'dos_mesh' in self.setting:
if 'restart' not in self.setting\
or not self.setting['restart']:
self.content['scf']['tolwfr'] = \
self.setting['wf_convergence']
else:
self.content['scf']['tolwfr'] = \
self.setting['wf_convergence']
# clean up for the case of restart
if 'restart' in self.setting and self.setting['restart']\
and ('band_scan' in self.setting or 'dos_mesh' in self.setting):
keep_lst = [
'nband',
'tolwfr',
'nstep',
'ecut',
'irdwfk',
'irdden',
'getwfk',
'getden',
]
for key in self.content['scf'].iterkeys():
if key not in keep_lst and 'prt' not in key:
self.content['scf'].pop(key)
######################
# additional setting #
######################
if 'abinit_setting' in self.setting:
self.content['additional'] = odict([])
for item in self.setting['abinit_setting']:
self.content['additional'][item] = ' '
def is_strnum(q):
if isinstance(q, Number) or type(q) is str:
return True
else:
return False
#################
# bandstructure #
#################
if 'band_scan' in self.setting and self.setting['band_scan']:
if molecule.symmetry and molecule.symmetry.lower() == 'fcc':
if 'kshift' not in self.setting:
self.setting['kshift'] = [
[0.5, 0.5, 0.5],
[0.5, 0.0, 0.0],
[0.0, 0.5, 0.0],
[0.0, 0.0, 0.5],
]
bnds = self.setting['band_scan']
if len(bnds) != 2 \
or is_strnum(bnds[0]) \
or len(bnds[0]) != len(bnds[1]) - 1:
qtk.exit('band_scan format: [lst_div, lst_coord]')
bnd_content = odict([
('iscf', -2),
('getden', -1),
('kptopt', -len(bnds[0])),
('tolwfr', self.setting['wf_convergence']),
('ndivk', bnds[0]),
('kptbounds', bnds[1]),
])
if nbnd:
bnd_content['nband'] = nbnd
if 'save_restart' not in self.setting \
or not self.setting['save_restart']:
bnd_content['prtwf'] = 0
else:
bnd_content['prtwf'] = 1
if 'save_density' not in self.setting \
or not self.setting['save_density']:
bnd_content['prtden'] = 0
else:
bnd_content['prtden'] = 1
if 'dos_mesh' in self.setting:
bnd_content['prtden'] = 1
if 'restart' in self.setting and self.setting['restart']:
bnd_content['irdden'] = 1
bnd_content['irdwfk'] = 1
bnd_content['getwfk'] = -1
self.content['band_scan'] = bnd_content
#####################
# density of states #
#####################
if 'dos_mesh' in self.setting and self.setting['dos_mesh']:
dos_content = odict([
('iscf', -3),
('ngkpt', self.setting['dos_mesh']),
('shiftk', [0.0, 0.0, 0.0]),
('tolwfr', self.setting['wf_convergence']),
('prtwf', 0),
])
if 'band_scan' in self.setting:
dos_content['getden'] = -2
else:
dos_content['getden'] = -1
if 'smearing' in self.setting:
smr = self.setting['smearing']
smr_dict = {
'fermi': 3,
'marzari': 4,
'marzari_monotonic': 5,
'paxton': 6,
'gaussian': 7,
}
if smr in smr_dict:
dos_content['occopt'] = smr_dict[smr]
dos_content['prtdos'] = 1
else:
qtk.warning("%s for occupation scheme not found" % smr)
if nbnd:
dos_content['nband'] = nbnd
if 'save_restart' not in self.setting \
or not self.setting['save_restart']:
dos_content['prtwf'] = 0
else:
dos_content['prtwf'] = 1
if 'save_density' not in self.setting \
or not self.setting['save_density']:
dos_content['prtden'] = 0
else:
dos_content['prtden'] = 1
if 'restart' in self.setting and self.setting['restart']:
dos_content['irdden'] = 1
dos_content['irdwfk'] = 1
dos_content['getwfk'] = -1
self.content['dos'] = dos_content
################
# cell section #
################
self.content['cell'] = odict([
('ecut', float(self.setting['cutoff'] / 2.)),
('nstep', self.setting['scf_step']),
])
self.content['cell']['acell'] = '3*1.889726124993'
self.celldm2lattice()
if not molecule.symmetry:
self.content['cell']['chkprim'] = 0
self.content['cell']['rprim'] = self.setting['lattice']
#################
# atoms section #
#################
znucl = []
for a in range(len(type_index) - 1):
symb = type_list[type_index[a]]
znucl.append(int(qtk.n2Z(symb)))
self.content['atoms'] = odict([
('natom', molecule.N),
('ntypat', (len(type_index) - 1)),
('typat', typat),
('znucl', znucl),
])
if hasattr(molecule, 'scale') and molecule.scale:
if hasattr(molecule, 'R_scale'):
R_scale = molecule.R_scale.copy()
for i in range(3):
s = molecule.scale[i]
if s > 0: R_scale[:, i] = R_scale[:, i] / s
self.content['atoms']['xred'] = R_scale
else:
qtk.warning('R_scale not found but scale is set')
self.content['atoms']['xangst'] = molecule.R
else:
self.content['atoms']['xangst'] = molecule.R
#########################
# write content to file #
#########################
datasets = 1
for key in ['band_scan', 'dos']:
if key in self.content:
datasets += 1
if self.setting['restart'] and datasets > 1:
datasets -= 1
self.content['datasets']['ndtset'] = datasets
inp.write('# abinit input generated by qctoolkit\n')
ds_itr = 0
ds_str = ''
if self.content['datasets']['ndtset'] == 1:
self.content['datasets'].pop('ndtset')
for section_key in self.content.iterkeys():
if len(self.content[section_key]) > 0:
inp.write('\n# %s section\n' % section_key)
if section_key in ['scf', 'band_scan', 'dos'] and datasets > 1:
ds_itr += 1
ds_str = str(ds_itr)
for key, value in self.content[section_key].iteritems():
if ds_str and section_key in ['scf', 'band_scan', 'dos']:
key = key + ds_str
if is_strnum(value):
inp.write('%s %s\n' % (key, str(value)))
else:
inp.write('%s' % key)
head = ''.join([' ' for _ in key])
if 'xangst' in key\
or 'xcart' in key\
or 'xred' in key:
inp.write('\n\n')
for vec in value:
inp.write('%s' % head)
for v in vec:
inp.write(' %s' % str(v))
inp.write('\n')
elif is_strnum(value[0]):
for v in value:
inp.write(' %s' % str(v))
inp.write('\n')
else:
for v in value[0]:
inp.write(' %s' % str(v))
inp.write('\n')
for vec in value[1:]:
inp.write('%s' % head)
for v in vec:
inp.write(' %s' % str(v))
inp.write('\n')
if 'restart_wavefunction_path' in self.setting:
rstwf_path = self.setting['restart_wavefunction_path']
if os.path.exists(rstwf_path):
wf_lst = sorted(
glob.glob(os.path.join(rstwf_path, '*o_*WFK')))
assert len(wf_lst) > 0
wf_src = os.path.abspath(wf_lst[-1])
pp_files.append([wf_src, name + 'i_WFK'])
else:
qtk.warning('%s not found' % rstwf_path)
if 'restart_wavefunction_file' in self.setting:
rst_file = self.setting['restart_wavefunction_file']
if os.path.exists(rst_file):
pp_files.append([rst_file, name + 'i_WFK'])
else:
qtk.warning('%s not found' % rst_file)
if 'restart_density_path' in self.setting:
rstdn_path = self.setting['restart_density_path']
if os.path.exists(rstdn_path):
dn_lst = sorted(
glob.glob(os.path.join(rstdn_path, '*o_*DEN')))
assert len(dn_lst) > 0
dn_src = os.path.abspath(dn_lst[-1])
pp_files.append([dn_src, name + 'i_DEN'])
else:
qtk.warning('%s not found' % rstdn_path)
if 'restart_density_file' in self.setting:
rst_file = self.setting['restart_density_file']
if os.path.exists(rst_file):
pp_files.append([rst_file, name + 'i_DEN'])
else:
qtk.warning('%s not found' % rst_file)
inp.close(dependent_files=pp_files)
if hasattr(inp, 'final_name'):
self.setting['no_molecule'] = True
self.setting['root_dir'] = name
files = \
super(PlanewaveInput, self).write(name, **self.setting)
files.extension = 'files'
files.write(inp.final_name + '\n')
root = os.path.splitext(inp.final_name)[0]
files.write(root + '.out\n')
files.write(root + 'i\n')
files.write(root + 'o\n')
files.write(root + 'x\n')
for pp in pp_files:
files.write(pp[1] + '\n')
files.close(no_cleanup=True)
return inp
def __init__(self, qmout, **kwargs):
PlanewaveOutput.__init__(self, qmout, **kwargs)
out_file = open(qmout)
data = out_file.readlines()
out_file.close()
Et_pattern = re.compile("^.*total energy *=.*$")
f_pattern = re.compile("^.*atom.*type.*force.*=.*$")
Et_list = filter(Et_pattern.match, data)
f_list = filter(f_pattern.match, data)
if len(Et_list) > 0:
Et_str = filter(Et_pattern.match, data)[-1]
Et = float(Et_str.split()[-2])
self.Et, self.unit = qtk.convE(Et, 'Ry-Eh')
out_folder = os.path.split(os.path.abspath(qmout))[0]
save = glob.glob(os.path.join(out_folder, '*.save'))
# extract force information
if len(f_list) > 0:
fstr = [
filter(None,
fstr.split('=')[-1].split(' ')) for fstr in f_list
]
# atomic unit force, HF/au, converted from Ry/au
self.force = 0.5 * np.array([[float(comp) for comp in atm]
for atm in fstr])
# extract band structure from xml files
if save:
save = save[0]
try:
data_xml = os.path.join(save, 'data-file.xml')
xml_file = open(data_xml)
tree = ET.parse(xml_file)
xml_file.close()
self.xml = tree.getroot()
kpoints = []
band = []
# extract celldm
celldm = []
cellVec = []
for i in range(1, 4):
cellvStr = filter(
None,
self.xml[2][4][i].text.replace('\n',
'').split(' '))
cellVec.append([float(v) for v in cellvStr])
self.celldm = qtk.cellVec2celldm(cellVec)
# extract structure
R = []
N = int(self.xml[3][0].text.replace('\n', ''))
Nsp = int(self.xml[3][1].text.replace('\n', ''))
for i in range(N):
RiStr = self.xml[3][5 + Nsp + i].get('tau')
Ri = [float(r) * 0.529177249 for r in RiStr.split(' ')]
ni = self.xml[3][5 + Nsp + i].get('SPECIES')
Z = [qtk.n2Z(ni)]
Z.extend(Ri)
R.append(Z)
self.molecule = qtk.Molecule()
self.molecule.build(R)
# access data for each kpoint
for k in self.xml[-2]:
k_str = k[0].text
coord = [float(c) for c in k_str.split()]
weight = float(k[1].text.split()[0])
coord.append(weight)
kpoints.append(coord)
ev_file = os.path.join(save, k[2].attrib['iotk_link'])
k_xml_file = open(ev_file)
k_xml = ET.parse(k_xml_file)
k_xml_file.close()
ev_k = k_xml.getroot()
ev_str = ev_k[2].text.split()
ev = [qtk.convE(float(entry), 'Eh-eV')[0]\
for entry in ev_str]
band.append(ev)
occ_str = ev_k[3].text.split()
occ = [float(entry) for entry in occ_str]
self.kpoints = np.array(kpoints)
self.mo_eigenvalues = copy.deepcopy(band[0])
self.band = np.array(band)
self.occupation = occ
diff = np.diff(occ)
pos = diff[np.where(abs(diff) > 0.5)]
mask = np.in1d(diff, pos)
ind = np.array(range(len(diff)))
if len(ind[mask]) > 0:
N_state = ind[mask][0]
vb = max(self.band[:, N_state])
cb = min(self.band[:, N_state + 1])
vb_pos = np.argmax(self.band[:, N_state])
cb_pos = np.argmin(self.band[:, N_state + 1])
self.Eg = cb - vb
if vb_pos == cb_pos:
self.Eg_direct = True
else:
self.Eg_direct = False
cell = []
for i in range(1, 4):
vec = filter(
None,
self.xml[2][4][i].text.replace('\n',
'').split(' '))
cell.append([float(v) for v in vec])
self.lattice = np.array(cell) / 1.889726124993
self.celldm = qtk.lattice2celldm(self.lattice)
self.molecule.R_scale = qtk.xyz2fractional(
self.molecule.R, self.celldm)
except IOError:
qtk.warning('xml file of job %s not found' % qmout)
else:
qtk.warning('job %s not finished' % qmout)
def __init__(self, qmout, **kwargs):
PlanewaveOutput.__init__(self, qmout, **kwargs)
out_file = open(qmout)
data = out_file.readlines()
out_file.close()
Et_pattern = re.compile("^.*total energy *=.*$")
f_pattern = re.compile("^.*atom.*type.*force.*=.*$")
Et_list = filter(Et_pattern.match, data)
f_list = filter(f_pattern.match, data)
if len(Et_list) > 0:
Et_str = filter(Et_pattern.match, data)[-1]
Et = float(Et_str.split()[-2])
self.Et, self.unit = qtk.convE(Et, 'Ry-Eh')
out_folder = os.path.split(os.path.abspath(qmout))[0]
save = glob.glob(os.path.join(out_folder, '*.save'))
# extract force information
if len(f_list) > 0:
fstr = [filter(None, fstr.split('=')[-1].split(' '))
for fstr in f_list]
# atomic unit force, HF/au, converted from Ry/au
self.force = 0.5 * np.array(
[[float(comp) for comp in atm] for atm in fstr]
)
# extract band structure from xml files
if save:
save = save[0]
try:
data_xml = os.path.join(save, 'data-file.xml')
xml_file = open(data_xml)
tree = ET.parse(xml_file)
xml_file.close()
self.xml = tree.getroot()
kpoints = []
band = []
# extract celldm
celldm = []
cellVec = []
for i in range(1, 4):
cellvStr = filter(None,
self.xml[2][4][i].text.replace('\n', '').split(' ')
)
cellVec.append([float(v) for v in cellvStr])
self.celldm = qtk.cellVec2celldm(cellVec)
# extract structure
R = []
N = int(self.xml[3][0].text.replace('\n', ''))
Nsp = int(self.xml[3][1].text.replace('\n', ''))
for i in range(N):
RiStr = self.xml[3][5+Nsp+i].get('tau')
Ri = [float(r) * 0.529177249 for r in RiStr.split(' ')]
ni = self.xml[3][5+Nsp+i].get('SPECIES')
Z = [qtk.n2Z(ni)]
Z.extend(Ri)
R.append(Z)
self.molecule = qtk.Molecule()
self.molecule.build(R)
# access data for each kpoint
for k in self.xml[-2]:
k_str = k[0].text
coord = [float(c) for c in k_str.split()]
weight = float(k[1].text.split()[0])
coord.append(weight)
kpoints.append(coord)
ev_file = os.path.join(save, k[2].attrib['iotk_link'])
k_xml_file = open(ev_file)
k_xml = ET.parse(k_xml_file)
k_xml_file.close()
ev_k = k_xml.getroot()
ev_str = ev_k[2].text.split()
ev = [qtk.convE(float(entry), 'Eh-eV')[0]\
for entry in ev_str]
band.append(ev)
occ_str = ev_k[3].text.split()
occ = [float(entry) for entry in occ_str]
self.kpoints = np.array(kpoints)
self.mo_eigenvalues = copy.deepcopy(band[0])
self.band = np.array(band)
self.occupation = occ
diff = np.diff(occ)
pos = diff[np.where(abs(diff) > 0.5)]
mask = np.in1d(diff, pos)
ind = np.array(range(len(diff)))
if len(ind[mask]) > 0:
N_state = ind[mask][0]
vb = max(self.band[:, N_state])
cb = min(self.band[:, N_state + 1])
vb_pos = np.argmax(self.band[:, N_state])
cb_pos = np.argmin(self.band[:, N_state + 1])
self.Eg = cb - vb
if vb_pos == cb_pos:
self.Eg_direct = True
else:
self.Eg_direct = False
cell = []
for i in range(1, 4):
vec = filter(
None,
self.xml[2][4][i].text.replace('\n', '').split(' '))
cell.append([float(v) for v in vec])
self.lattice = np.array(cell) / 1.889726124993
self.celldm = qtk.lattice2celldm(self.lattice)
self.molecule.R_scale = qtk.xyz2fractional(
self.molecule.R, self.celldm)
except IOError:
qtk.warning('xml file of job %s not found' % qmout)
else:
qtk.warning('job %s not finished' % qmout)
def getBasis():
######################################
# extract basis function information #
######################################
basis_dict = {"S":0, "P":1, "D":2, "F":3, "G":4, "H":5}
basis_P = re.compile(r" *[0-9]+ [A-Z] [0-9]\.[0-9]{8}")
batom_P = re.compile(r"^ [0-9A-Za-z\-_\.]+ *\([A-Z][a-z]*\)")
bname_P = re.compile(r"\((.*)\)")
coord_P = re.compile(r"^ [0-9A-Za-z\.\-_]+ +[- ][0-9\.]{9,}")
basisStr = filter(basis_P.match, data)
batomStr = filter(batom_P.match, data)
coordStr = filter(coord_P.match, data)
# change 'Sulphur' to 'Sulfur' for NWChem format
# 'Sulphur' 'Sulfur'
def atomNameConv(old, new):
_matched = filter(lambda x: old in x, batomStr)
if _matched:
_matched = _matched[0]
_s = batomStr.index(_matched)
batomStr[_s] = re.sub(old, new, batomStr[_s])
_s = data.index(_matched)
data[_s] = re.sub(old, new, data[_s])
atomNameConv('Sulphur', 'Sulfur')
atomNameConv('Aluminium', 'Aluminum')
_exponents = [float(filter(None, s.split(' '))\
[2]) for s in basisStr]
_coefficients = [float(filter(None, s.split(' '))\
[3]) for s in basisStr]
_N = [int(filter(None, s.split(' '))[0])\
for s in basisStr]
_type = [filter(None, s.split(' '))[1]\
for s in basisStr]
_bfnInd = [data.index(batom) for batom in batomStr]
_bfnEndPtn = re.compile(r" Summary of \"")
_bfnEndStr = filter(_bfnEndPtn.match, data)[0]
_bfnInd.append(data.index(_bfnEndStr))
_ao_keys = [0]
for ind in range(len(_bfnInd)-1):
_key = _ao_keys[-1]
for i in range(_bfnInd[ind]+4, _bfnInd[ind+1]):
if len(data[i]) > 1:
_key = _key + 1
_ao_keys.append(_key)
_atoms = [getattr(pt, bname_P.match(
filter(None, s.split(' '))[1]).group(1).lower()).symbol\
for s in batomStr]
self.type_list = [re.split(r'[\._]',
filter(None, s.split(' '))[0])[0].title()\
for s in coordStr]
self.type_list_unique = list(
collections.OrderedDict.fromkeys(self.type_list)
)
self.R = np.array([filter(None, s.split(' '))[1:4]\
for s in coordStr]).astype(float)
self.N = len(self.R)
self.Z = [qtk.n2Z(e) for e in self.type_list]
self.R_bohr = 1.889725989 * self.R
ZR = []
for i in range(self.N):
vec = [self.Z[i]]
vec.extend(self.R[i])
ZR.append(vec)
self.molecule = qtk.Molecule()
self.molecule.build(ZR)
_N.append(0)
self.basis = []
for i in range(len(self.type_list)):
e = self.type_list[i]
center = self.R_bohr[i]
ind = self.type_list_unique.index(e)
bfn_base = {}
bfn_base['atom'] = e
bfn_base['center'] = center
bfn_base['index'] = i
exp = []
cef = []
for g in range(_ao_keys[ind], _ao_keys[ind+1]):
exp.append(_exponents[g])
cef.append(_coefficients[g])
if _N[g] != _N[g+1] or g+1 >= _ao_keys[ind+1]:
bfn = copy.deepcopy(bfn_base)
bfn['exponents'] = copy.deepcopy(exp)
bfn['coefficients'] = copy.deepcopy(cef)
if _type[g] in basis_dict:
_bfnList = self.basisList(basis_dict[_type[g]])
for bStr in _bfnList:
bfn['type'] = _type[g].lower() + bStr
self.basis.append(copy.deepcopy(bfn))
exp = []
cef = []
def write(self, name=None, **kwargs):
self.setting.update(kwargs)
self.setting['no_molecule'] = False
inp, molecule = \
super(PlanewaveInput, self).write(name, **self.setting)
molecule.sort()
type_index = molecule.index
type_list = molecule.type_list
self.pp_path = qtk.setting.bigdft_pp
if 'pp_path' in self.setting:
self.pp_path = self.setting['pp_path']
if 'pp_theory' not in self.setting:
self.setting['pp_theory'] = self.setting['theory']
pp_files = []
inp.write('# abinit input generated by qctoolkit\n')
# restart section
if 'restart' in self.setting and self.setting['restart']:
inp.write('\n# restart, reading wavefunction from file\n')
inp.write('irdwfk 1\n')
inp.write('getwfk -1\n')
# cell definition
inp.write('\n# cell definition\n')
# cell specified by Bohr
if not molecule.symmetry:
inp.write('# NOTE: cell defined by lattice vector, ')
inp.write('NOT supported by abinit spacegroup detector!\n')
inp.write('acell 3*1.889726124993\n')
if 'lattice' not in self.setting:
self.celldm2lattice()
lattice_vec = self.setting['lattice']
inp.write('chkprim 0\n')
elif molecule.symmetry.lower() == 'fcc':
inp.write("# fcc primitive cell\n")
a0 = molecule.celldm[0] * 1.889726124993
inp.write('acell 1 1 1\n')
lattice_vec = 0.5 * a0 * np.array([
[0, 1, 1],
[1, 0, 1],
[1, 1, 0],
])
elif molecule.symmetry.lower() == 'bcc':
inp.write("# bcc primitive cell\n")
inp.write('acell 1 1 1\n')
a0 = molecule.celldm[0] * 1.889726124993
lattice_vec = 0.5 * a0 * np.array([
[-1, 1, 1],
[ 1,-1, 1],
[ 1, 1,-1],
])
strList = ['rprim', '', '']
for i in range(3):
vec = lattice_vec[i]
inp.write('%5s % 11.6f % 11.6f % 11.6f\n' % (
strList[i], vec[0], vec[1], vec[2],
))
# atom definition
inp.write("\n# atom definition\n")
inp.write('ntypat %d\n' % (len(type_index) - 1))
inp.write('znucl')
for a in range(len(type_index)-1):
symb = type_list[type_index[a]]
Z = qtk.n2Z(symb)
inp.write(' %d' % Z)
inp.write('\n')
# atom position
inp.write("\n# atom position\n")
inp.write("natom %d\n" % molecule.N)
inp.write('typat')
for a in range(len(type_index)-1):
start = type_index[a]
end = type_index[a+1]
for i in range(end - start):
inp.write(' %d' % (a+1))
inp.write('\nxangst\n\n')
for i in range(molecule.N):
inp.write(' ')
for j in range(3):
inp.write(' % 11.6f' % molecule.R[i][j])
inp.write('\n')
# construct pp files depandency
pp_file = 'psppar.' + molecule.type_list[i]
pp_list = set([pp[1] for pp in pp_files])
if pp_file not in pp_list:
pp_src = PPCheck(self.setting['theory'],
self.setting['pp_theory'],
self.pp_path,
molecule.type_list[i])
pp_files.append([pp_src, pp_file])
inp.write('\n')
# system setting
inp.write("\n# system settings\n")
# cutoff in Hartree
inp.write('ecut %.2f\n' % float(self.setting['cutoff']/2.))
if 'kmesh' not in self.setting:
self.setting['kmesh'] = [1,1,1]
if self.setting['full_kmesh']:
inp.write('kptopt 3\n')
inp.write('ngkpt')
for k in self.setting['kmesh']:
inp.write(' %d' % k)
if 'kshift' not in self.setting:
inp.write('\nshiftk 0.0 0.0 0.0')
if 'ks_states' in self.setting and self.setting['ks_states']:
vs = int(round(molecule.getValenceElectrons() / 2.0))
nbnd = self.setting['ks_states'] + vs
if 'd_shell' in self.setting:
for a in molecule.type_list:
if a in self.setting['d_shell'] and qtk.n2ve(a) < 10:
nbnd += 5
inp.write('\nnband %d' % nbnd)
inp.write('\nnstep %d\n' % self.setting['scf_step'])
if 'wf_convergence' in self.setting:
inp.write('toldfe %.2E\n' % self.setting['wf_convergence'])
if molecule.charge != 0:
inp.write('charge=%d\n' % molecule.charge)
if molecule.multiplicity != 1:
inp.write('nsppol 2 # for spin polarized\n')
inp.write('occopt 7 # for relaxed occupation\n')
if 'save_restart' in self.setting and self.setting['save_restart']:
pass
else:
inp.write('prtwf 0\n')
for item in self.content:
inp.write(item)
inp.close(dependent_files=pp_files)
if hasattr(inp, 'final_name'):
self.setting['no_molecule'] = True
self.setting['root_dir'] = name
files = \
super(PlanewaveInput, self).write(name, **self.setting)
files.extension = 'files'
files.write(inp.final_name + '\n')
root = os.path.splitext(inp.final_name)[0]
files.write(root + '.out\n')
files.write(root + 'i\n')
files.write(root + 'o\n')
files.write(root + 'x\n')
for pp in pp_files:
files.write(pp[1] + '\n')
files.close(no_cleanup = True)
return inp
def getChild():
child = {}
for key in parent1.iterkeys():
if key == 'mutation':
# implementation for element_count constraints
if self.element_count:
constraint_list = []
constraint_keys = []
# tool to constuct list
def _list_append(mcoord, i_list, Zc):
for g in range(len(mcoord)):
grp = mcoord[g]
for i in range(len(grp)):
Z_p = grp[i]
if Z_p == Zc:
i_list.append((g, i))
# construct mutation list
for elem in self.element_count.iterkeys():
i_list = []
_Z = qtk.n2Z(elem)
_list_append(parent1['mutation'], i_list, _Z)
_list_append(parent2['mutation'], i_list, _Z)
constraint_list.append(i_list)
constraint_keys.append(elem)
while True:
selected_list = []
selected_coord = {}
consistant = True
for i in range(len(constraint_list)):
ind_location = constraint_list[i]
random.shuffle(ind_location)
elem = qtk.n2Z(constraint_keys[i])
_count = self.element_count[
constraint_keys[i]][0]
_end = _count
while len(set(ind_location[:_end])) < _count:
_end += 1
_new = list(set(ind_location[:_end]))
to_select = copy.deepcopy(selected_list)
to_select.extend(_new)
if len(set(to_select)) == \
len(selected_list) + len(_new):
selected_list.extend(_new)
for coord in _new:
selected_coord.update({coord: elem})
consistant = consistant & True
else:
consistant = False
child_ind = []
for g in range(len(parent1['mutation'])):
child_list = []
for i in range(len(parent1['mutation'][g])):
if selected_coord.has_key((g, i)):
child_list.append(selected_coord[(g,
i)])
else:
candidate = []
A1 = qtk.Z2n(parent1['mutation'][g][i])
A2 = qtk.Z2n(parent2['mutation'][g][i])
if not self.element_count.has_key(A1):
candidate.append(qtk.n2Z(A1))
if not self.element_count.has_key(A2):
candidate.append(qtk.n2Z(A2))
if len(candidate) == 0:
consistant = False
else:
child_list.append(
random.choice(candidate))
child_ind.append(child_list)
if consistant:
break
for g in range(len(parent1['mutation'])):
for i in range(len(parent1['mutation'][g])):
if not selected_coord.has_key((g, i)):
if random.random() < mutation_rate:
child_ind[g][i] = random.choice(\
self.mutation_target[g])
elem = qtk.Z2n(child_ind[g][i])
if self.element_count.has_key(elem):
coord = random.choice([tupl for tupl,targ
in selected_coord.iteritems()\
if targ == qtk.n2Z(elem)])
del selected_coord[coord]
selected_coord.update({
(g, i):
qtk.n2Z(elem)
})
[gc, ic] = list(coord)
new_targ = random.choice([z for z in \
self.mutation_target[gc]\
if not self.element_count\
.has_key(qtk.Z2n(z))])
child_ind[gc][ic] = new_targ
child.update({'mutation': child_ind})
else:
child_mut = []
for i in range(len(parent1[key])):
# mix two parent
grp = parent1[key][i]
index = range(len(grp))
random.shuffle(index)
ind1 = index[:len(index) / 2]
if len(index) % 2 == 0:
ind2 = index[len(index) / 2:]
append = -1
else:
ind2 = index[len(index) / 2:-1]
append = index[-1]
new = [parent1[key][i][ind] for ind in ind1]
new.extend([parent2[key][i][ind] for ind in ind2])
if append >= 0:
if random.random() >= 0.5:
new.append(parent1[key][i][append])
else:
new.append(parent2[key][i][append])
# mutation, loop through each element
for j in range(len(new)):
if random.random() < mutation_rate:
mut_target = [tar \
for tar in self.mutation_target[i]\
if tar != new[j]]
new[j] = random.choice(mut_target)
child_mut.append(new)
child.update({'mutation': child_mut})
else:
qtk.exit("ccs mate function is not yet implemented for "\
+ key)
return child
def random_coord(self, mode):
# determin coordinate type
if mode == 'mutation':
_dtype = int
_list = copy.deepcopy(self.mutation_list)
_targ = copy.deepcopy(self.mutation_target)
# algorithm for element_count constraint
# 1. construct mapping from flattened to nested mutation list
# 2. select all candidate indices for constrainted element
# 3. remove constrained element from target lists
# 4. random choice of candidate for constrained mutation
# 5. loop through all other indices without constrained elem
# single constraint implementation
if self.element_count:
new_list = []
new_targ = []
constraint = []
selected = []
for elem in self.element_count.iterkeys():
itr = 0
count = 0
_candidate = []
for sublist in _targ:
if qtk.n2Z(elem) in sublist:
_candidate.extend(
range(itr, itr + len(_list[count])))
_targ[count].remove(qtk.n2Z(elem))
itr += len(_list[count])
count += 1
_candidate = [x for x in _candidate\
if not x in flatten(selected)]
selection = sorted(random.sample(
_candidate,
random.choice(self.element_count[elem])),
reverse=True)
selected.append(selection)
for ind in selection:
i, j = map2list(ind, _list)
constraint.append([(i, j), qtk.n2Z(elem)])
constraint_ind = [a[0] for a in constraint]
# end of constraint setting
elif mode == 'stretching':
_list = self.stretching_list
_targ = self.stretching_range
_dtype = float
elif mode == 'rotation':
_list = self.rotation_list
_targ = self.rotation_range
_dtype = float
# !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
# main routine to generate a random coordinate
_vec = []
ind = 0
for _group, _targp in zip(_list, _targ):
_subvec = []
for _atom in _group:
if _dtype == int:
# element_count constraint
if self.element_count:
i, j = map2list(ind, _list)
if (i, j) in constraint_ind:
index = constraint_ind.index((i, j))
_subvec.append(constraint[index][1])
else:
_subvec.append(random.choice(_targp))
# without constraint
else:
_subvec.append(random.choice(_targp))
elif _dtype == float:
_subvec.append(random.uniform(_targp[0], _targp[1]))
ind += 1
_vec.append(_subvec)
if len(_vec) > 0:
return _vec
def getBasis():
######################################
# extract basis function information #
######################################
basis_dict = {"S": 0, "P": 1, "D": 2, "F": 3, "G": 4, "H": 5}
basis_P = re.compile(r" *[0-9]+ [A-Z] [0-9]\.[0-9]{8}")
batom_P = re.compile(r"^ [0-9A-Za-z\-_\.]+ *\([A-Z][a-z]*\)")
bname_P = re.compile(r"\((.*)\)")
coord_P = re.compile(r"^ [0-9A-Za-z\.\-_]+ +[- ][0-9\.]{9,}")
basisStr = filter(basis_P.match, data)
batomStr = filter(batom_P.match, data)
coordStr = filter(coord_P.match, data)
# change 'Sulphur' to 'Sulfur' for NWChem format
# 'Sulphur' 'Sulfur'
def atomNameConv(old, new):
_matched = filter(lambda x: old in x, batomStr)
if _matched:
_matched = _matched[0]
_s = batomStr.index(_matched)
batomStr[_s] = re.sub(old, new, batomStr[_s])
_s = data.index(_matched)
data[_s] = re.sub(old, new, data[_s])
atomNameConv('Sulphur', 'Sulfur')
atomNameConv('Aluminium', 'Aluminum')
_exponents = [float(filter(None, s.split(' '))\
[2]) for s in basisStr]
_coefficients = [float(filter(None, s.split(' '))\
[3]) for s in basisStr]
_N = [int(filter(None, s.split(' '))[0])\
for s in basisStr]
_type = [filter(None, s.split(' '))[1]\
for s in basisStr]
_bfnInd = [data.index(batom) for batom in batomStr]
_bfnEndPtn = re.compile(r" Summary of \"")
_bfnEndStr = filter(_bfnEndPtn.match, data)[0]
_bfnInd.append(data.index(_bfnEndStr))
_ao_keys = [0]
for ind in range(len(_bfnInd) - 1):
_key = _ao_keys[-1]
for i in range(_bfnInd[ind] + 4, _bfnInd[ind + 1]):
if len(data[i]) > 1:
_key = _key + 1
_ao_keys.append(_key)
_atoms = [getattr(pt, bname_P.match(
filter(None, s.split(' '))[1]).group(1).lower()).symbol\
for s in batomStr]
self.type_list = [re.split(r'[\._]',
filter(None, s.split(' '))[0])[0].title()\
for s in coordStr]
self.type_list_unique = list(
collections.OrderedDict.fromkeys(self.type_list))
self.R = np.array([filter(None, s.split(' '))[1:4]\
for s in coordStr]).astype(float)
self.N = len(self.R)
self.Z = [qtk.n2Z(e) for e in self.type_list]
self.R_bohr = 1.889725989 * self.R
ZR = []
for i in range(self.N):
vec = [self.Z[i]]
vec.extend(self.R[i])
ZR.append(vec)
self.molecule = qtk.Molecule()
self.molecule.build(ZR)
_N.append(0)
self.basis = []
for i in range(len(self.type_list)):
e = self.type_list[i]
center = self.R_bohr[i]
ind = self.type_list_unique.index(e)
bfn_base = {}
bfn_base['atom'] = e
bfn_base['center'] = center
bfn_base['index'] = i
exp = []
cef = []
for g in range(_ao_keys[ind], _ao_keys[ind + 1]):
exp.append(_exponents[g])
cef.append(_coefficients[g])
if _N[g] != _N[g + 1] or g + 1 >= _ao_keys[ind + 1]:
bfn = copy.deepcopy(bfn_base)
bfn['exponents'] = copy.deepcopy(exp)
bfn['coefficients'] = copy.deepcopy(cef)
if _type[g] in basis_dict:
_bfnList = self.basisList(basis_dict[_type[g]])
for bStr in _bfnList:
bfn['type'] = _type[g].lower() + bStr
self.basis.append(copy.deepcopy(bfn))
exp = []
cef = []
def getEt(self, name):
out = open(name, 'r')
data = out.readlines()
out.close()
E_str = filter(lambda x: 'TOTAL ENERGY' in x, data)[-1]
Et_report = float(E_str.split()[-2])
rst_str = filter(lambda x: 'TOTAL ENERGY' in x, data)[-1]
r_scf_str = filter(
lambda x: 'MAXIMUM NUMBER OF ITERATIONS FOR SC' in x, data)[-1]
r_scf = int(filter(None, r_scf_str.split(' '))[-2])
p1 = re.compile('^.*\d \d\.\d{3}E-\d\d.*$')
scf_str = filter(p1.match, data)
if scf_str:
self.scf_step = len(scf_str)
R_str = filter(lambda x: 'ATOMIC COORDINATES' in x, data)[-1]
R_ind = len(data) - data[::-1].index(R_str)
self.molecule.N = 1
R_more = True
ZR = []
while R_more:
ind = R_ind + self.molecule.N
zr_lst = filter(None, data[ind].split(' '))
if len(zr_lst) == 5:
zr = [qtk.n2Z(zr_lst[1])]
zr.extend([float(r) for r in zr_lst[2:]])
ZR.append(zr)
self.molecule.N = self.molecule.N + 1
else:
R_more = False
self.molecule.build(ZR, unit='bohr')
Et_scf = float(filter(None, scf_str[-1].split(' '))[3])
if abs(Et_report - Et_scf) > 1E-6 :
qtk.exit("%s is not finished" % qmout)
else:
self.Et = Et_report
if self.scf_step == 1 and r_scf > 1:
self.Et = nan
detail = []
for scf in scf_str:
detail.append(filter(None, scf.split(' ')))
report_str = filter(lambda x: 'A.U.' in x, data)
report = report_str[:len(report_str)/2]
self.detail = {}
for s in report:
entry = s.split()
if filter(lambda x: '(' in x, entry):
entry_txt = ' '.join(entry[1:-3])
else:
entry_txt = ' '.join(entry[:-3])
energy = float(entry[-2])
self.detail[entry_txt] = energy
self.detail['scf'] = detail
ev_str1 = filter(lambda x: '(EV)' in x, data)
ev_str2 = filter(lambda x: ' EV\n' in x, data)
if ev_str1 and ev_str2:
self.mo_eigenvalues = []
self.occupation = []
ev_start = data.index(ev_str1[0]) + 1
ev_end = data.index(ev_str2[0])
for i in range(ev_start, ev_end):
s = data[i].split()
if len(s) == 6:
ev = [float(s[1]), float(s[4])]
occ = [float(s[2]), float(s[5])]
self.mo_eigenvalues.extend(ev)
self.occupation.extend(occ)
elif len(s) == 3:
ev = float(s[1])
occ = float(s[2])
self.mo_eigenvalues.append(ev)
self.occupation.append(occ)
diff = np.diff(np.array(self.occupation))
pos = diff[np.where(abs(diff) > 1)]
mask = np.in1d(diff, pos)
ind = np.array(range(len(diff)))
N_state = ind[mask][0]
self.Eg = self.mo_eigenvalues[N_state + 1] -\
self.mo_eigenvalues[N_state]
def read_xyz(self, name, **kwargs):
xyz = open(name, 'r')
content = xyz.readlines()
xyz.close()
content = [line.replace('\t', ' ') for line in content]
prop_list = ['charge', 'celldm', 'scale', 'symmetry', 'isolated']
for prop in prop_list:
try:
prop_str = filter(lambda x: prop in x, content)[0]
prop_str = re.sub('.*:', '', prop_str)
prop_data = prop_str.split(' ')
prop_data = filter(None, prop_data)
if len(prop_data) == 1:
try:
setattr(self, prop, float(prop_data[0]))
except Exception as exc:
if prop == 'symmetry' or prop == 'isolated':
setattr(self, prop, prop_data[0].strip())
elif len(prop_data) > 1:
setattr(self, prop, [float(_) for _ in prop_data])
except ValueError as exc:
setattr(self, prop, False)
qtk.warning("setting attribute %s with error: %s" % \
(prop, exc))
except:
setattr(self, prop, False)
# convert celldm acoording to proper symmetry is still not working
# universally due to abinit and vasp k-point sampling
# require explicity reciprocal coordinate
#if hasattr(self, 'symmetry') and self.symmetry in ['bcc', 'fcc']:
# new_dm = copy.deepcopy(self.celldm)
# if np.linalg.norm(np.array(new_dm[3:])) < 1E-5:
# lattice_cube = np.array(new_dm[:3]) * np.eye(3)
# if self.symmetry == 'fcc':
# vec1 = (lattice_cube[1] + lattice_cube[2])/2
# vec2 = (lattice_cube[0] + lattice_cube[2])/2
# vec3 = (lattice_cube[0] + lattice_cube[1])/2
# elif self.symmetry == 'bcc':
# vec1 = (-lattice_cube[0] + lattice_cube[1] + lattice_cube[2])/2
# vec2 = ( lattice_cube[0] - lattice_cube[1] + lattice_cube[2])/2
# vec3 = ( lattice_cube[0] + lattice_cube[1] - lattice_cube[2])/2
# nv1 = vec1 / np.linalg.norm(vec1)
# nv2 = vec2 / np.linalg.norm(vec2)
# nv3 = vec3 / np.linalg.norm(vec3)
# new_dm = [
# np.linalg.norm(vec1),
# np.linalg.norm(vec2),
# np.linalg.norm(vec3),
# np.dot(nv2, nv3),
# np.dot(nv1, nv3),
# np.dot(nv1, nv2),
# ]
# qtk.warning(
# "symmetry is %s but celldm is set to cubic, reset from %s to %s"\
# % (self.symmetry, str(self.celldm), str(new_dm))
# )
# self.celldm = new_dm
if not self.charge:
self.charge = 0
if self.celldm or self.scale:
self.periodic = True
if self.celldm:
assert len(self.celldm) == 6
self.N = int(content[0])
prop_list = content[1]
try:
self.prop_list = np.array(prop_list.split(' ')).astype(float)
except:
self.prop_list = prop_list
coord_list = content[2 : self.N + 2]
coord = [filter(None,[a for a in entry.replace("\n", '').split(' ')])
for entry in coord_list]
type_list = list(np.array(coord)[:,0])
self.type_list = [str(elem) for elem in type_list]
self.Z = [qtk.n2Z(elem) for elem in self.type_list]
self.Z = np.array(self.Z)
self.R = np.array(coord)[:,1:4].astype(float)
self.box = False
if self.celldm:
self.periodic = True
angle = self.celldm[3:]
angle_sum = sum([abs(entry) for entry in angle])
if angle_sum == 0:
self.box = self.celldm[:3]
if not self.isolated:
self.R_scale = copy.deepcopy(self.R)
self.R = qtk.fractional2xyz(self.R_scale, self.celldm)