def _parse_special_cases(self, cmd, args, arguments, i, li):
if cmd in ('TITL', 'REM'):
args = args.strip()
if args: arg_tuple = (args, )
else: arg_tuple = ()
return (cmd, arg_tuple)
if cmd == 'SYMM':
if args is None:
raise shelx_error("illegal argument '%s'", i, args)
return (cmd, (self.symm_space.sub('', args).upper(), ))
if cmd == 'EQIV':
m = self._eqiv_pat.search(args)
if m is None:
raise shelx_error("illegal argument '%s'", i, args)
try:
idx = int(m.group(1))
except ValueError:
raise shelx_error("illegal argument '%s'", i, m.group(1))
return (cmd, (idx, self.symm_space.sub('', m.group(2)).upper()))
if cmd == 'SFAC':
sfac_args = []
for e in arguments:
try:
sfac_args.append(float(e))
except ValueError:
sfac_args.append(e)
return (cmd, tuple(sfac_args))
return None
def _parse_special_cases(self, cmd, args, arguments, i, li):
if cmd in ("TITL", "REM"):
args = args.strip()
if args:
arg_tuple = (args,)
else:
arg_tuple = ()
return (cmd, arg_tuple)
if cmd == "SYMM":
if args is None:
raise shelx_error("illegal argument '%s'", i, args)
return (cmd, (self.symm_space.sub("", args).upper(),))
if cmd == "EQIV":
m = self._eqiv_pat.search(args)
if m is None:
raise shelx_error("illegal argument '%s'", i, args)
try:
idx = int(m.group(1))
except ValueError:
raise shelx_error("illegal argument '%s'", i, m.group(1))
return (cmd, (idx, self.symm_space.sub("", m.group(2)).upper()))
if cmd == "SFAC":
sfac_args = []
for e in arguments:
try:
sfac_args.append(float(e))
except ValueError:
sfac_args.append(e)
return (cmd, tuple(sfac_args))
return None
def filtered_commands(self):
active_afix = False
for command, line in self.command_stream:
cmd, args = command[0], command[-1]
if cmd in ('AFIX', 'HKLF'):
if cmd == 'AFIX' and not args:
raise shelx_error("too few arguments", line)
if active_afix:
if n_afixed != n_expected_afixed:
raise shelx_error("wrong number of afixed atoms", line)
self.builder.end_geometrical_constraint()
active_afix = False
n_afixed = 0
if cmd == 'HKLF':
yield command, line
continue
mn = args[0]
if mn == 0: continue
m,n = divmod(mn, 10)
d, sof, u = (None,)*3
params = args[1:]
if not params: pass
elif len(params) == 1: d = params[0]
elif len(params) == 3: d, sof = params[0:]
elif len(params) == 4: d, sof, u = params[0:]
else: raise shelx_error("too many arguments", line)
info = self.constraints.get(m)
if info is not None:
constraint_name, pivot_relative_pos = info
constraint_type = self.builder.make_geometrical_constraint_type(
constraint_name)
self.builder.start_geometrical_constraint(
type_=constraint_type,
bond_length=d,
rotating=n in (7, 8),
stretching=n in (4, 8),
pivot_relative_pos=pivot_relative_pos)
n_expected_afixed = constraint_type.n_constrained_sites
active_afix = True
n_afixed = 0
elif cmd == '__ATOM__':
if active_afix:
if sof is not None: args[4] = sof
if u is not None and len(args) == 6: args[-1] = u
n_afixed += 1
yield command, line
else:
yield command, line
def filtered_commands(self):
active_afix = False
for command, line in self.command_stream:
cmd, args = command[0], command[-1]
if cmd in ('AFIX', 'HKLF'):
if cmd == 'AFIX' and not args:
raise shelx_error("too few arguments", line)
if active_afix:
if n_afixed != n_expected_afixed:
raise shelx_error("wrong number of afixed atoms", line)
self.builder.end_geometrical_constraint()
active_afix = False
n_afixed = 0
if cmd == 'HKLF':
yield command, line
continue
mn = args[0]
if mn == 0: continue
m,n = divmod(mn, 10)
d, sof, u = (None,)*3
params = args[1:]
if not params: pass
elif len(params) == 1: d = params[0]
elif len(params) == 3: d, sof = params[0:]
elif len(params) == 4: d, sof, u = params[0:]
else: raise shelx_error("too many arguments", line)
info = self.constraints.get(m)
if info is not None:
constraint_name, pivot_relative_pos = info
constraint_type = self.builder.make_geometrical_constraint_type(
constraint_name)
self.builder.start_geometrical_constraint(
type_=constraint_type,
bond_length=d,
rotating=n in (7, 8),
stretching=n in (4, 8),
pivot_relative_pos=pivot_relative_pos)
n_expected_afixed = constraint_type.n_constrained_sites
active_afix = True
n_afixed = 0
elif cmd == '__ATOM__':
if active_afix:
if sof is not None: args[4] = sof
if u is not None and len(args) == 6: args[-1] = u
n_afixed += 1
yield command, line
else:
yield command, line
def lex_scatterer(self, args, scatterer_index):
name = args[0]
n = int(args[1])
n_vars = len(args) - 2
if n_vars == 5 or (not self.strictly_shelxl and n_vars == 6):
values, behaviours = self.decode_variables(args[2:7],
u_iso_idx=n_vars - 1)
u = values[4]
isotropic = True
elif n_vars == 10:
unit_cell = self.builder.crystal_symmetry.unit_cell()
values, behaviours = self.decode_variables(
args[2:-3] + (args[-1], args[-2], args[-3]), u_iso_idx=None)
u = adptbx.u_cif_as_u_star(unit_cell, values[-6:])
isotropic = False
else:
raise shelx_error("wrong number of parameters for scatterer",
self.line)
site = values[0:3]
occ = values[3]
scattering_type = eltbx.xray_scattering.get_standard_label(
self.label_for_sfac[n], # works thank to (a) above
exact=True)
scatterer = xray.scatterer(label=name,
site=site,
occupancy=occ,
u=u,
scattering_type=scattering_type)
if (not isotropic or not isinstance(behaviours[-1], tuple)
or behaviours[-1][0] != constant_times_u_eq):
self.scatterer_to_bind_u_eq_to = (scatterer, scatterer_index)
return scatterer, behaviours
def _parse_general_case(self, cmd, cmd_residue, arguments, i, li):
toks = []
for j, arg in enumerate(arguments):
try:
toks.append(float(arg))
except ValueError:
if cmd == 'RESI':
toks.append(arg)
else:
m = self._atom_name_pat.match(arg)
if m is None:
raise shelx_error("illegal argument '%s'", i, arg)
name, symmetry, resnum, plus_minus = m.group(
'name', 'symmetry', 'resnum', 'plusminus')
if name is not None:
if resnum is not None: resnum = int(resnum)
if symmetry is not None: symmetry = int(symmetry)
toks.append(
tokens.atomname_token(name=name,
symmetry=symmetry,
residue_number=resnum,
plus_minus=plus_minus))
continue
element, resnum = m.group('element', 'resnum')
if element is not None:
toks.append(
tokens.element_token(element=element,
residue_number=resnum))
continue
if m.group('range') == '>':
toks.append((tokens.forward_range_tok, ))
elif m.group('range') == '<':
toks.append((tokens.backward_range_tok, ))
if cmd_residue:
return (cmd, cmd_residue, tuple(toks))
else:
if cmd not in self.shelx_commands:
if cmd[0] == 'Q':
toks = (int(cmd[1:]), ) + tuple(toks)
cmd = '__Q_PEAK__'
else:
toks = (cmd, ) + tuple(toks)
cmd = '__ATOM__'
else:
toks = tuple(toks)
return (cmd, toks)
def _parse_general_case(self, cmd, cmd_residue, arguments, i, li):
toks = []
for j, arg in enumerate(arguments):
try:
toks.append(float(arg))
except ValueError:
if cmd == "RESI":
toks.append(arg)
else:
m = self._atom_name_pat.match(arg)
if m is None:
raise shelx_error("illegal argument '%s'", i, arg)
name, symmetry, resnum, plus_minus = m.group("name", "symmetry", "resnum", "plusminus")
if name is not None:
if resnum is not None:
resnum = int(resnum)
if symmetry is not None:
symmetry = int(symmetry)
toks.append(
tokens.atomname_token(
name=name, symmetry=symmetry, residue_number=resnum, plus_minus=plus_minus
)
)
continue
element, resnum = m.group("element", "resnum")
if element is not None:
toks.append(tokens.element_token(element=element, residue_number=resnum))
continue
if m.group("range") == ">":
toks.append((tokens.forward_range_tok,))
elif m.group("range") == "<":
toks.append((tokens.backward_range_tok,))
if cmd_residue:
return (cmd, cmd_residue, tuple(toks))
else:
if cmd not in self.shelx_commands:
if cmd[0] == "Q":
toks = (int(cmd[1:]),) + tuple(toks)
cmd = "__Q_PEAK__"
else:
toks = (cmd,) + tuple(toks)
cmd = "__ATOM__"
else:
toks = tuple(toks)
return (cmd, toks)
def process_possible_constrained_occupancy(self,
scatterer_index,
scatterer,
occupancy_behaviour):
if self.builder_does_occupancy_pair_affine_constraint:
try:
kind, coeff, free_var_idx = occupancy_behaviour
except (TypeError, ValueError):
pass
else:
if not self.free_variable:
raise shelx_error("An instruction FVAR needs to appear before "
"this point in the file,", self.line)
coeff /= scatterer.weight_without_occupancy()
r_coeff = scitbx.math.continued_fraction.from_real(
coeff, eps=1e-5).as_rational()
coeff = round(r_coeff.numerator()/r_coeff.denominator(), ndigits=5)
if kind == constant_times_independent_scalar_parameter:
self.occupancies_depending_on_free_variable.setdefault(
free_var_idx, []).append((coeff, 0, scatterer_index))
elif kind == constant_times_independent_scalar_parameter_minus_1:
self.occupancies_depending_on_free_variable.setdefault(
free_var_idx, []).append((coeff, -1, scatterer_index))
def process_possible_constrained_occupancy(self,
scatterer_index,
scatterer,
occupancy_behaviour):
if self.builder_does_occupancy_pair_affine_constraint:
try:
kind, coeff, free_var_idx = occupancy_behaviour
except (TypeError, ValueError):
pass
else:
if not self.free_variable:
raise shelx_error("An instruction FVAR needs to appear before "
"this point in the file,", self.line)
coeff /= scatterer.weight_without_occupancy()
r_coeff = scitbx.math.continued_fraction.from_real(
coeff, eps=1e-5).as_rational()
coeff = round(r_coeff.numerator()/r_coeff.denominator(), ndigits=5)
if kind == constant_times_independent_scalar_parameter:
self.occupancies_depending_on_free_variable.setdefault(
free_var_idx, []).append((coeff, 0, scatterer_index))
elif kind == constant_times_independent_scalar_parameter_minus_1:
self.occupancies_depending_on_free_variable.setdefault(
free_var_idx, []).append((coeff, -1, scatterer_index))
def lex_scatterer(self, args, scatterer_index):
name = args[0]
n = int(args[1])
n_vars = len(args) - 2
if n_vars == 5 or (not self.strictly_shelxl and n_vars == 6):
values, behaviours = self.decode_variables(
args[2:7],
u_iso_idx=n_vars-1)
u = values[4]
isotropic = True
elif n_vars == 10:
unit_cell = self.builder.crystal_symmetry.unit_cell()
values, behaviours = self.decode_variables(
args[2:-3] + (args[-1], args[-2], args[-3]),
u_iso_idx=None)
u = adptbx.u_cif_as_u_star(unit_cell, values[-6:])
isotropic = False
else:
raise shelx_error("wrong number of parameters for scatterer",
self.line)
site = values[0:3]
occ = values[3]
scattering_type = eltbx.xray_scattering.get_standard_label(
self.label_for_sfac[n], # works thank to (a) above
exact=True)
scatterer = xray.scatterer(
label = name,
site = site,
occupancy = occ,
u = u,
scattering_type = scattering_type)
if (not isotropic
or not isinstance(behaviours[-1], tuple)
or behaviours[-1][0] != constant_times_u_eq):
self.scatterer_to_bind_u_eq_to = (scatterer, scatterer_index)
return scatterer, behaviours
def filtered_commands(self):
from scitbx.math import continued_fraction
temperature = 20
for command, line in self.command_stream:
cmd, args = command[0], command[-1]
n_args = len(args)
if cmd == 'OMIT':
if n_args == 3 and isinstance(args[0], float):
self.instructions.setdefault('omit_hkl', [])
self.instructions['omit_hkl'].append([int(i) for i in args])
elif n_args == 2 and isinstance(args[0], float):
self.instructions['omit'] = {
's': args[0],
'two_theta': args[1]}
else:
yield command, line
elif cmd == 'SHEL':
if args:
shel = {'lowres': args[0]}
if n_args > 1:
shel['highres'] = args[1]
elif cmd == 'MERG':
if args:
self.instructions['merg'] = args[0]
elif cmd == 'TEMP':
if len(args) > 1:
raise shelx_error("TEMP takes at most 1 argument")
if args: temperature = args[0]
self.instructions['temp'] = temperature
self.builder.temperature_in_celsius = temperature
elif cmd == 'WGHT':
if n_args > 6:
raise shelx_error("Too many argument for %s" % cmd, line)
default_weighting_scheme = { 'a': 0.1,
'b': 0,
'c': 0,
'd': 0,
'e': 0,
'f': 1/3 }
weighting_scheme = dict([
(key, (arg is not None and arg) or default_weighting_scheme[key])
for key, arg in itertools.izip_longest('abcdef', args) ])
self.instructions['wght'] = weighting_scheme
self.builder.make_shelx_weighting_scheme(**weighting_scheme)
elif cmd == 'HKLF':
# only ONE HKLF instruction allowed
assert 'hklf' not in self.instructions
hklf = {'s': 1, 'matrix': sgtbx.rot_mx()}
hklf['n'] = args[0]
if n_args > 1:
hklf['s'] = args[1]
if n_args > 2:
assert n_args > 10
mx = [ continued_fraction.from_real(e, eps=1e-3).as_rational()
for e in args[2:11] ]
den = reduce(rational.lcm, [ r.denominator() for r in mx ])
nums = [ r.numerator()*(den//r.denominator()) for r in mx ]
hklf['matrix'] = sgtbx.rot_mx(nums, den)
if n_args > 11:
hklf['wt'] = args[11]
if n_args == 13:
hklf['m'] = args[12]
self.instructions['hklf'] = hklf
assert not self.builder or ('wt' not in hklf and 'm' not in hklf)
self.builder.create_shelx_reflection_data_source(
format=hklf['n'],
indices_transform=hklf['matrix'],
data_scale=hklf['s'])
if 'basf' in self.instructions and hklf['n'] == 5:
self.builder.make_non_merohedral_twinning_with_transformed_hkl(
fractions=self.instructions['basf'])
elif cmd == 'TWIN':
# only ONE twin instruction allowed
assert 'twin' not in self.instructions
twin = {}
if n_args > 0:
assert n_args >= 9
twin['matrix'] = sgtbx.rt_mx(
sgtbx.rot_mx([int(i) for i in args[0:9]]))
if n_args > 9:
twin['n'] = args[9]
self.instructions['twin'] = twin
if 'basf' in self.instructions: self.issue_merohedral_twinning()
elif cmd == 'BASF':
self.instructions['basf'] = args
if 'twin' in self.instructions: self.issue_merohedral_twinning()
elif cmd == 'EXTI':
if len(args) == 1:
self.instructions['exti'] = args[0]
else:
yield command, line
else:
# All token have been read without errors or early bailout
assert 'hklf' in self.instructions, "Missing HKLF instruction"
def filtered_commands(self):
self.label_for_sfac = None
overall_scale = None
scatterer_index = 0
conformer_index = 0
sym_excl_index = 0
part_sof = None
current_residue = (None, None)
line_of_scatterer_named = {}
in_the_midst_of_atom_list = False
idx_assigned_by_builder_to_free_var_idx = {}
builder = self.builder
if self.builder_does_occupancy_pair_affine_constraint:
self.occupancies_depending_on_free_variable = {}
for command, line in self.command_stream:
self.line = line
cmd, args = command[0], command[-1]
if cmd == 'SFAC':
builder.make_structure()
if len(args) == 15 and isinstance(args[1], float):
raise NotImplementedError(
'''SFAC label a1 b1 a2 b2 a3 b3 a4 b4 c f' f" mu r wt''')
self.label_for_sfac = ('*',) + args # (a) working around
# ShelXL 1-based indexing
elif cmd == 'FVAR':
if overall_scale is None:
overall_scale = args[0]
self.free_variable = args # (b) ShelXL indexes into the whole array
else:
# ShelXL allows for more than one FVAR instruction
self.free_variable = self.free_variable + args
elif cmd == 'PART':
part_sof = None
conformer_index = 0
sym_excl_index = 0
part_number = 0
if args:
part_number = int(args[0])
if len(args) == 2:
part_sof = self.decode_one_variable(args[1])
if part_number > 0: conformer_index = part_number
elif part_number < 0: sym_excl_index = abs(part_number)
elif cmd == "RESI":
current_residue = args
self.builder.add_residue(*current_residue)
elif cmd == '__ATOM__':
if not in_the_midst_of_atom_list:
if self.label_for_sfac is None:
raise shelx_error("An instruction SFAC needs to appear before "
"this point in the file,", self.line)
in_the_midst_of_atom_list = True
scatterer, behaviour_of_variable = self.lex_scatterer(
args, scatterer_index)
residue_number, residue_class = current_residue
name = scatterer.label.upper()
line_1 = line_of_scatterer_named.get((residue_number, name))
if line_1 is not None:
raise shelx_error("Residue #%i has two scatterers named %s, "
"(with perhaps a difference in letter case)"
"defined at lines %i and %i"
% (residue_number, name, line, line_1),
line=None)
line_of_scatterer_named[(residue_number, name)] = line
if (conformer_index or sym_excl_index) and part_sof:
scatterer.occupancy, behaviour_of_variable[3] = part_sof
builder.add_scatterer(scatterer, behaviour_of_variable,
occupancy_includes_symmetry_factor=True,
conformer_index=conformer_index,
sym_excl_index=sym_excl_index,
residue_number=residue_number,
residue_class=residue_class)
# As the builder accepted the scatterer, we assume that it called
# scatterer.apply_symmetry so that scatterer.multiplicity and dependents
# are correctly set. Hence our fetching the scatterer from the builder
# instead of using the pre-existing local variable.
scatterer = builder.structure.scatterers()[-1]
self.process_possible_constrained_occupancy(scatterer_index,
scatterer,
behaviour_of_variable[3])
self.process_possible_u_iso_constraints(scatterer_index,
behaviour_of_variable[4])
scatterer_index += 1
elif cmd == '__Q_PEAK__':
assert not self.strictly_shelxl,\
"Q-peaks amidst atoms in strict ShelXL model"
builder.add_electron_density_peak(site = args[2:5],
height = args[-1])
else:
yield command, line
self.process_occupancies_depending_on_free_variable()
def raise_parameter_error():
raise shelx_error(
"scatterer parameter #%d '%s'", line, i_cv+1, coded_variable)
def __iter__(self):
"""
Yields the commands in self.file as tuples:
- either (cmd, args) where args is a tuple,
e.g. ('CELL', (0.71073, 1, 2, 3, 90, 90, 90))
- or (cmd, (residue_class, residue_number), args) for those commands
which can be suffixed by a residue, e.g.
('HFIX', (parser.residue_number_tok, 1), (23, ))
Notes:
- For atoms, cmd is '__ATOM__' and the name is the first argument
- For so-called Q-peaks, cmd is '__Q_PEAK__' and the number is the
first argument
- In args, floating point items are reported as is whereas any
string comes as (type, value) where type is one of the class
constants defined just above this method (actually, for
the residue_number_tok, the value is an integer stored as a float).
- The atomic refinable variables are not decoded: this is the job of
a parser, not of a lexer.
"""
continued = False
for i, li in enumerate(self.file):
if not li: continue
if continued:
m = self._continuation_pat.search(li)
if m is None:
raise shelx_error("illegal continuation line error", i)
cont_args, continued = m.groups()
arguments.extend(cont_args.split())
else:
m = self._cmd_pat.search(li)
if m is None:
if li[0].isspace(): continue
else:
m = self._include_filename_pat.search(li)
if m is not None:
cmd, filename = m.group(1, 2)
yield (cmd, filename), i
continue
raise shelx_error("illegal command or atom name error", i)
cmd = m.group(1) or m.group(6)
if cmd:
cmd = cmd.upper()
cmd_residue = None
else:
(cmd, cmd_residue_class,
cmd_residue_number) = m.group(3, 4, 5)
cmd = cmd.upper()
if cmd_residue_class:
cmd_residue = (tokens.residue_class_tok,
cmd_residue_class.upper())
elif cmd_residue_number:
cmd_residue = (tokens.residue_number_tok,
int(cmd_residue_number))
else:
cmd_residue = (tokens.all_residues_tok, )
args = m.group(2) or m.group(7) or ""
continued = m.group(8)
arguments = args.split()
if not continued:
result = self._parse_special_cases(cmd, args, arguments, i, li)
if result is None:
result = self._parse_general_case(cmd, cmd_residue,
arguments, i, li)
yield result, i
if cmd == 'HKLF': break
def filtered_commands(self):
from scitbx.math import continued_fraction
temperature = 20
for command, line in self.command_stream:
cmd, args = command[0], command[-1]
n_args = len(args)
if cmd == 'OMIT':
if n_args == 3 and isinstance(args[0], float):
self.instructions.setdefault('omit_hkl', [])
self.instructions['omit_hkl'].append([int(i) for i in args])
elif n_args == 2 and isinstance(args[0], float):
self.instructions['omit'] = {
's': args[0],
'two_theta': args[1]}
else:
yield command, line
elif cmd == 'SHEL':
if args:
shel = {'lowres': args[0]}
if n_args > 1:
shel['highres'] = args[1]
elif cmd == 'MERG':
if args:
self.instructions['merg'] = args[0]
elif cmd == 'TEMP':
if len(args) > 1:
raise shelx_error("TEMP takes at most 1 argument")
if args: temperature = args[0]
self.instructions['temp'] = temperature
self.builder.temperature_in_celsius = temperature
elif cmd == 'WGHT':
if n_args > 6:
raise shelx_error("Too many argument for %s" % cmd, line)
default_weighting_scheme = { 'a': 0.1,
'b': 0,
'c': 0,
'd': 0,
'e': 0,
'f': 1/3 }
weighting_scheme = dict([
(key, (arg is not None and arg) or default_weighting_scheme[key])
for key, arg in itertools.izip_longest('abcdef', args) ])
self.instructions['wght'] = weighting_scheme
self.builder.make_shelx_weighting_scheme(**weighting_scheme)
elif cmd == 'HKLF':
# only ONE HKLF instruction allowed
assert 'hklf' not in self.instructions
hklf = {'s': 1, 'matrix': sgtbx.rot_mx()}
hklf['n'] = args[0]
if n_args > 1:
hklf['s'] = args[1]
if n_args > 2:
assert n_args > 10
mx = [ continued_fraction.from_real(e, eps=1e-3).as_rational()
for e in args[2:11] ]
den = reduce(rational.lcm, [ r.denominator() for r in mx ])
nums = [ r.numerator()*(den//r.denominator()) for r in mx ]
hklf['matrix'] = sgtbx.rot_mx(nums, den)
if n_args > 11:
hklf['wt'] = args[11]
if n_args == 13:
hklf['m'] = args[12]
self.instructions['hklf'] = hklf
assert not self.builder or ('wt' not in hklf and 'm' not in hklf)
self.builder.create_shelx_reflection_data_source(
format=hklf['n'],
indices_transform=hklf['matrix'],
data_scale=hklf['s'])
if 'basf' in self.instructions and hklf['n'] == 5:
self.builder.make_non_merohedral_twinning_with_transformed_hkl(
fractions=self.instructions['basf'])
elif cmd == 'TWIN':
# only ONE twin instruction allowed
assert 'twin' not in self.instructions
twin = {}
if n_args > 0:
assert n_args >= 9
twin['matrix'] = sgtbx.rt_mx(
sgtbx.rot_mx([int(i) for i in args[0:9]]))
if n_args > 9:
twin['n'] = args[9]
self.instructions['twin'] = twin
if 'basf' in self.instructions: self.issue_merohedral_twinning()
elif cmd == 'BASF':
self.instructions['basf'] = args
if 'twin' in self.instructions: self.issue_merohedral_twinning()
elif cmd == 'EXTI':
if len(args) == 1:
self.instructions['exti'] = args[0]
else:
yield command, line
else:
# All token have been read without errors or early bailout
assert 'hklf' in self.instructions, "Missing HKLF instruction"
def filtered_commands(self):
self.label_for_sfac = None
overall_scale = None
scatterer_index = 0
conformer_index = 0
sym_excl_index = 0
part_sof = None
current_residue = (None, None)
line_of_scatterer_named = {}
in_the_midst_of_atom_list = False
idx_assigned_by_builder_to_free_var_idx = {}
builder = self.builder
if self.builder_does_occupancy_pair_affine_constraint:
self.occupancies_depending_on_free_variable = {}
for command, line in self.command_stream:
self.line = line
cmd, args = command[0], command[-1]
if cmd == 'SFAC':
builder.make_structure()
if len(args) == 15 and isinstance(args[1], float):
raise NotImplementedError(
'''SFAC label a1 b1 a2 b2 a3 b3 a4 b4 c f' f" mu r wt''')
self.label_for_sfac = ('*',) + args # (a) working around
# ShelXL 1-based indexing
elif cmd == 'FVAR':
if overall_scale is None:
overall_scale = args[0]
self.free_variable = args # (b) ShelXL indexes into the whole array
else:
# ShelXL allows for more than one FVAR instruction
self.free_variable = self.free_variable + args
elif cmd == 'PART':
part_sof = None
conformer_index = 0
sym_excl_index = 0
part_number = 0
if args:
part_number = int(args[0])
if len(args) == 2:
part_sof = self.decode_one_variable(args[1])
if part_number > 0: conformer_index = part_number
elif part_number < 0: sym_excl_index = abs(part_number)
elif cmd == "RESI":
current_residue = args
self.builder.add_residue(*current_residue)
elif cmd == '__ATOM__':
if not in_the_midst_of_atom_list:
if self.label_for_sfac is None:
raise shelx_error("An instruction SFAC needs to appear before "
"this point in the file,", self.line)
in_the_midst_of_atom_list = True
scatterer, behaviour_of_variable = self.lex_scatterer(
args, scatterer_index)
residue_number, residue_class = current_residue
name = scatterer.label.upper()
line_1 = line_of_scatterer_named.get((residue_number, name))
if line_1 is not None:
raise shelx_error("Residue #%i has two scatterers named %s, "
"(with perhaps a difference in letter case)"
"defined at lines %i and %i"
% (residue_number, name, line, line_1),
line=None)
line_of_scatterer_named[(residue_number, name)] = line
if (conformer_index or sym_excl_index) and part_sof:
scatterer.occupancy, behaviour_of_variable[3] = part_sof
builder.add_scatterer(scatterer, behaviour_of_variable,
occupancy_includes_symmetry_factor=True,
conformer_index=conformer_index,
sym_excl_index=sym_excl_index,
residue_number=residue_number,
residue_class=residue_class)
# As the builder accepted the scatterer, we assume that it called
# scatterer.apply_symmetry so that scatterer.multiplicity and dependents
# are correctly set. Hence our fetching the scatterer from the builder
# instead of using the pre-existing local variable.
scatterer = builder.structure.scatterers()[-1]
self.process_possible_constrained_occupancy(scatterer_index,
scatterer,
behaviour_of_variable[3])
self.process_possible_u_iso_constraints(scatterer_index,
behaviour_of_variable[4])
scatterer_index += 1
elif cmd == '__Q_PEAK__':
assert not self.strictly_shelxl,\
"Q-peaks amidst atoms in strict ShelXL model"
builder.add_electron_density_peak(site = args[2:5],
height = args[-1])
else:
yield command, line
self.process_occupancies_depending_on_free_variable()
def __iter__(self):
"""
Yields the commands in self.file as tuples:
- either (cmd, args) where args is a tuple,
e.g. ('CELL', (0.71073, 1, 2, 3, 90, 90, 90))
- or (cmd, (residue_class, residue_number), args) for those commands
which can be suffixed by a residue, e.g.
('HFIX', (parser.residue_number_tok, 1), (23, ))
Notes:
- For atoms, cmd is '__ATOM__' and the name is the first argument
- For so-called Q-peaks, cmd is '__Q_PEAK__' and the number is the
first argument
- In args, floating point items are reported as is whereas any
string comes as (type, value) where type is one of the class
constants defined just above this method (actually, for
the residue_number_tok, the value is an integer stored as a float).
- The atomic refinable variables are not decoded: this is the job of
a parser, not of a lexer.
"""
continued = False
for i, li in enumerate(self.file):
if not li:
continue
if continued:
m = self._continuation_pat.search(li)
if m is None:
raise shelx_error("illegal continuation line error", i)
cont_args, continued = m.groups()
arguments.extend(cont_args.split())
else:
m = self._cmd_pat.search(li)
if m is None:
if li[0].isspace():
continue
else:
m = self._include_filename_pat.search(li)
if m is not None:
cmd, filename = m.group(1, 2)
yield (cmd, filename), i
continue
raise shelx_error("illegal command or atom name error", i)
cmd = m.group(1) or m.group(6)
if cmd:
cmd = cmd.upper()
cmd_residue = None
else:
(cmd, cmd_residue_class, cmd_residue_number) = m.group(3, 4, 5)
cmd = cmd.upper()
if cmd_residue_class:
cmd_residue = (tokens.residue_class_tok, cmd_residue_class.upper())
elif cmd_residue_number:
cmd_residue = (tokens.residue_number_tok, int(cmd_residue_number))
else:
cmd_residue = (tokens.all_residues_tok,)
args = m.group(2) or m.group(7) or ""
continued = m.group(8)
arguments = args.split()
if not continued:
result = self._parse_special_cases(cmd, args, arguments, i, li)
if result is None:
result = self._parse_general_case(cmd, cmd_residue, arguments, i, li)
yield result, i
if cmd == "HKLF":
break
def parse_restraints(self):
for cmd, restraints in sorted(self.cached_restraints.items()):
for line, args in restraints.iteritems():
cmd_residue = args[0]
if cmd_residue is None:
residues = [None]
else:
tok = cmd_residue[0]
if tok == tokens.residue_number_tok:
residues = [cmd_residue[1]]
elif tok == tokens.residue_class_tok:
residues = self.builder.residue_numbers_having_class[cmd_residue[1]]
elif tok == tokens.all_residues_tok:
residues = self.builder.residue_class_of_residue_number.keys()
args = args[1]
floats = []
atoms = []
elements = []
# it seems that shelxl will accept the restraints and target value/sigma
# in any order - we will interpret any floating point values as
# target value (if applicable), sigma, in that order. Any other
# arguments given are atoms
for arg in args:
try:
floats.append(float(arg))
except TypeError, e:
if isinstance(arg, tokens.atomname_token):
atoms.append(arg)
elif isinstance(arg, tokens.element_token):
elements.append(arg)
for residue_number in residues:
try:
i_seqs = self.i_seqs_from_atoms(atoms, residue_number)
if i_seqs is None: continue
restraint_type = self.restraint_types.get(cmd)
if cmd in ('DFIX','DANG'):
assert len(floats) in (1, 2)
distance_ideal = floats[0]
if len(floats) == 2:
sigma = floats[1]
else:
if cmd == 'DFIX': sigma = 0.02
else: sigma = 0.04
assert len(atoms) > 1
weight = 1/(sigma**2)
assert len(atoms) % 2 == 0
for i in range(len(atoms)//2):
atom_pair = atoms[i*2:(i+1)*2]
i_seq_pair = i_seqs[i*2:(i+1)*2]
sym_ops = [
self.symmetry_operations.get(atom.symmetry)
for atom in atom_pair]
self.builder.process_restraint(restraint_type,
distance_ideal=distance_ideal,
weight=weight,
i_seqs=i_seq_pair,
sym_ops=sym_ops)
if cmd == 'SADI':
assert len(floats) <= 1
if len(floats):
sigma = floats[0]
else:
sigma = 0.02
i_seq_pairs = []
sym_ops = []
for i in range(len(atoms)//2):
atom_pair = atoms[i*2:(i+1)*2]
i_seq_pairs.append(i_seqs[i*2:(i+1)*2])
sym_ops.append(
[self.symmetry_operations.get(atom.symmetry)
for atom in atom_pair])
weights = [1/(sigma**2)]*len(i_seq_pairs)
self.builder.process_restraint(restraint_type,
weights=weights,
i_seqs=i_seq_pairs,
sym_ops=sym_ops)
elif cmd == 'FLAT':
assert len(floats) <= 1
if len(floats) == 1:
sigma = floats[0]
else:
sigma = 0.1
assert len(atoms) > 3
sym_ops = [
self.symmetry_operations.get(atom.symmetry) for atom in atoms]
weights = [1/(sigma**2)]*len(i_seqs)
self.builder.process_restraint(restraint_type,
weights=weights,
i_seqs=i_seqs,
sym_ops=sym_ops)
elif cmd == 'CHIV':
pass
elif cmd in ('DELU', 'ISOR', 'SIMU'):
assert len(floats) <= 3
s1 = s2 = dmax = None
if len(floats) > 0:
s1 = floats[0]
if len(floats) > 1:
s2 = floats[1]
if len(floats) > 2:
dmax = floats[2]
if len(i_seqs) == 0:
i_seqs = None
if s1 is None:
if cmd == 'SIMU': s1 = 0.04
elif cmd == 'DELU': s1 = 0.01
else: s1 = 0.1
if cmd == 'DELU':
self.builder.process_restraint(restraint_type,
sigma_12=s1,
sigma_13=s2,
i_seqs=i_seqs)
else:
self.builder.process_restraint(restraint_type,
sigma=s1,
sigma_terminal=s2,
i_seqs=i_seqs)
else:
pass
except (TypeError, AssertionError):
raise shelx_error("Invalid %s instruction" %cmd, line)
def parse_restraints(self):
for cmd, restraints in sorted(self.cached_restraints.items()):
for line, args in restraints.iteritems():
cmd_residue = args[0]
if cmd_residue is None:
residues = [None]
else:
tok = cmd_residue[0]
if tok == tokens.residue_number_tok:
residues = [cmd_residue[1]]
elif tok == tokens.residue_class_tok:
residues = self.builder.residue_numbers_having_class[cmd_residue[1]]
elif tok == tokens.all_residues_tok:
residues = self.builder.residue_class_of_residue_number.keys()
args = args[1]
floats = []
atoms = []
elements = []
# it seems that shelxl will accept the restraints and target value/sigma
# in any order - we will interpret any floating point values as
# target value (if applicable), sigma, in that order. Any other
# arguments given are atoms
for arg in args:
try:
floats.append(float(arg))
except TypeError, e:
if isinstance(arg, tokens.atomname_token):
atoms.append(arg)
elif isinstance(arg, tokens.element_token):
elements.append(arg)
for residue_number in residues:
try:
i_seqs = self.i_seqs_from_atoms(atoms, residue_number)
if i_seqs is None: continue
restraint_type = self.restraint_types.get(cmd)
if cmd in ('DFIX','DANG'):
assert len(floats) in (1, 2)
distance_ideal = floats[0]
if len(floats) == 2:
sigma = floats[1]
else:
if cmd == 'DFIX': sigma = 0.02
else: sigma = 0.04
assert len(atoms) > 1
weight = 1/(sigma**2)
assert len(atoms) % 2 == 0
for i in range(len(atoms)//2):
atom_pair = atoms[i*2:(i+1)*2]
i_seq_pair = i_seqs[i*2:(i+1)*2]
sym_ops = [
self.symmetry_operations.get(atom.symmetry)
for atom in atom_pair]
self.builder.process_restraint(restraint_type,
distance_ideal=distance_ideal,
weight=weight,
i_seqs=i_seq_pair,
sym_ops=sym_ops)
if cmd == 'SADI':
assert len(floats) <= 1
if len(floats):
sigma = floats[0]
else:
sigma = 0.02
i_seq_pairs = []
sym_ops = []
for i in range(len(atoms)//2):
atom_pair = atoms[i*2:(i+1)*2]
i_seq_pairs.append(i_seqs[i*2:(i+1)*2])
sym_ops.append(
[self.symmetry_operations.get(atom.symmetry)
for atom in atom_pair])
weights = [1/(sigma**2)]*len(i_seq_pairs)
self.builder.process_restraint(restraint_type,
weights=weights,
i_seqs=i_seq_pairs,
sym_ops=sym_ops)
elif cmd == 'FLAT':
assert len(floats) <= 1
if len(floats) == 1:
sigma = floats[0]
else:
sigma = 0.1
assert len(atoms) > 3
sym_ops = [
self.symmetry_operations.get(atom.symmetry) for atom in atoms]
weights = [1/(sigma**2)]*len(i_seqs)
self.builder.process_restraint(restraint_type,
weights=weights,
i_seqs=i_seqs,
sym_ops=sym_ops)
elif cmd == 'CHIV':
pass
elif cmd in ('DELU', 'RIGU', 'ISOR', 'SIMU'):
assert len(floats) <= 3
s1 = s2 = dmax = None
if len(floats) > 0:
s1 = floats[0]
if len(floats) > 1:
s2 = floats[1]
if len(floats) > 2:
dmax = floats[2]
if len(i_seqs) == 0:
i_seqs = None
if s1 is None:
if cmd == 'SIMU': s1 = 0.04
elif cmd == 'DELU': s1 = 0.01
elif cmd == 'RIGU': s1 = 0.004
else: s1 = 0.1
if cmd in ('DELU', 'RIGU'):
self.builder.process_restraint(restraint_type,
sigma_12=s1,
sigma_13=s2,
i_seqs=i_seqs)
else:
self.builder.process_restraint(restraint_type,
sigma=s1,
sigma_terminal=s2,
i_seqs=i_seqs)
else:
pass
except (TypeError, AssertionError):
raise shelx_error("Invalid %s instruction" %cmd, line)
def raise_parameter_error():
raise shelx_error(
"scatterer parameter #%d '%s'", line, i_cv+1, coded_variable)
