class cif(DictMixin):
def __init__(self, blocks=None):
if blocks is not None:
self.blocks = OrderedDict(blocks)
else:
self.blocks = OrderedDict()
self.keys_lower = dict([(key.lower(), key)
for key in self.blocks.keys()])
def __setitem__(self, key, value):
assert isinstance(value, block)
if not re.match(tag_re, '_' + key):
raise Sorry("%s is not a valid data block name" % key)
self.blocks[key] = value
self.keys_lower[key.lower()] = key
def get(self, key, default=None):
key_lower = self.keys_lower.get(key.lower())
if (key_lower is None):
return default
return self.blocks.get(key_lower, default)
def __getitem__(self, key):
result = self.get(key)
if (result is None):
raise KeyError('Unknown CIF data block name: "%s"' % key)
return result
def __delitem__(self, key):
del self.blocks[self.keys_lower[key.lower()]]
del self.keys_lower[key.lower()]
def keys(self):
return self.blocks.keys()
def __repr__(self):
return repr(OrderedDict(self.iteritems()))
def __copy__(self):
return cif(self.blocks.copy())
copy = __copy__
def __deepcopy__(self, memo):
return cif(copy.deepcopy(self.blocks, memo))
def deepcopy(self):
return copy.deepcopy(self)
def show(self,
out=None,
indent=" ",
indent_row=None,
data_name_field_width=34,
loop_format_strings=None,
align_columns=True):
if out is None:
out = sys.stdout
for name, block in self.items():
print >> out, "data_%s" % name
block.show(out=out,
indent=indent,
indent_row=indent_row,
data_name_field_width=data_name_field_width,
loop_format_strings=loop_format_strings,
align_columns=align_columns)
def __str__(self):
s = StringIO()
self.show(out=s)
return s.getvalue()
def validate(self,
dictionary,
show_warnings=True,
error_handler=None,
out=None):
if out is None: out = sys.stdout
from iotbx.cif import validation
errors = {}
if error_handler is None:
error_handler = validation.ErrorHandler()
for key, block in self.blocks.iteritems():
error_handler = error_handler.__class__()
dictionary.set_error_handler(error_handler)
block.validate(dictionary)
errors.setdefault(key, error_handler)
if error_handler.error_count or error_handler.warning_count:
error_handler.show(show_warnings=show_warnings, out=out)
return error_handler
def sort(self, recursive=False, key=None, reverse=False):
self.blocks = OrderedDict(
sorted(self.blocks.items(), key=key, reverse=reverse))
if recursive:
for b in self.blocks.values():
b.sort(recursive=recursive, reverse=reverse)
class loop(DictMixin):
def __init__(self, header=None, data=None):
self._columns = OrderedDict()
self.keys_lower = {}
if header is not None:
for key in header:
self.setdefault(key, flex.std_string())
if data is not None:
# the number of data items must be an exact multiple of the number of headers
assert len(data) % len(
header) == 0, "Wrong number of data items for loop"
n_rows = len(data) // len(header)
n_columns = len(header)
for i in range(n_rows):
self.add_row(
[data[i * n_columns + j] for j in range(n_columns)])
elif header is None and data is not None:
assert isinstance(data, dict) or isinstance(data, OrderedDict)
self.add_columns(data)
self.keys_lower = dict([(key.lower(), key)
for key in self._columns.keys()])
def __setitem__(self, key, value):
if not re.match(tag_re, key):
raise Sorry("%s is not a valid data name" % key)
if len(self) > 0:
assert len(value) == self.size()
if not isinstance(value, flex.std_string):
for flex_numeric_type in (flex.int, flex.double):
if isinstance(value, flex_numeric_type):
value = value.as_string()
else:
try:
value = flex_numeric_type(value).as_string()
except TypeError:
continue
else:
break
if not isinstance(value, flex.std_string):
value = flex.std_string(value)
# value must be a mutable type
assert hasattr(value, '__setitem__')
self._columns[key] = value
self.keys_lower[key.lower()] = key
def __getitem__(self, key):
return self._columns[self.keys_lower[key.lower()]]
def __delitem__(self, key):
del self._columns[self.keys_lower[key.lower()]]
del self.keys_lower[key.lower()]
def keys(self):
return self._columns.keys()
def __repr__(self):
return repr(OrderedDict(self.iteritems()))
def name(self):
return common_substring(self.keys()).rstrip('_').rstrip('.')
def size(self):
size = 0
for column in self.values():
size = max(size, len(column))
return size
def n_rows(self):
return self.size()
def n_columns(self):
return len(self.keys())
def add_row(self, row, default_value="?"):
if isinstance(row, dict):
for key in self:
if key in row:
self[key].append(str(row[key]))
else:
self[key].append(default_value)
else:
assert len(row) == len(self)
for i, key in enumerate(self):
self[key].append(str(row[i]))
def add_column(self, key, values):
if self.size() != 0:
assert len(values) == self.size()
self[key] = values
self.keys_lower[key.lower()] = key
def add_columns(self, columns):
assert isinstance(columns, dict) or isinstance(columns, OrderedDict)
for key, value in columns.iteritems():
self.add_column(key, value)
def update_column(self, key, values):
assert type(key) == type(""), "first argument is column key string"
if self.size() != 0:
assert len(
values) == self.size(), "len(values) %d != self.size() %d" % (
len(values),
self.size(),
)
self[key] = values
self.keys_lower[key.lower()] = key
def delete_row(self, index):
assert index < self.n_rows()
for column in self._columns.values():
del column[index]
def __copy__(self):
new = loop()
new._columns = self._columns.copy()
new.keys_lower = self.keys_lower.copy()
return new
copy = __copy__
def __deepcopy__(self, memo):
new = loop()
new._columns = copy.deepcopy(self._columns, memo)
new.keys_lower = copy.deepcopy(self.keys_lower, memo)
return new
def deepcopy(self):
return copy.deepcopy(self)
def show(self,
out=None,
indent=" ",
indent_row=None,
fmt_str=None,
align_columns=True):
assert self.n_rows() > 0 and self.n_columns() > 0, "keys: %s %d %d" % (
self.keys(),
self.n_rows(),
self.n_columns(),
)
if out is None:
out = sys.stdout
if indent_row is None:
indent_row = indent
assert indent.strip() == ""
assert indent_row.strip() == ""
print >> out, "loop_"
for k in self.keys():
print >> out, indent + k
values = self._columns.values()
range_len_values = range(len(values))
if fmt_str is not None:
# Pretty printing:
# The user is responsible for providing a valid format string.
# Values are not quoted - it is the user's responsibility to place
# appropriate quotes in the format string if a particular value may
# contain spaces.
values = copy.deepcopy(values)
for i, v in enumerate(values):
for flex_numeric_type in (flex.int, flex.double):
if not isinstance(v, flex_numeric_type):
try:
values[i] = flex_numeric_type(v)
except ValueError:
continue
else:
break
if fmt_str is None:
fmt_str = indent_row + ' '.join(["%s"] * len(values))
for i in range(self.size()):
print >> out, fmt_str % tuple(
[values[j][i] for j in range_len_values])
elif align_columns:
fmt_str = []
for i, (k, v) in enumerate(self.iteritems()):
for i_v in range(v.size()):
v[i_v] = format_value(v[i_v])
# exclude and semicolon text fields from column width calculation
v_ = flex.std_string(item for item in v if "\n" not in item)
width = v_.max_element_length()
# See if column contains only number, '.' or '?'
# right-align numerical columns, left-align everything else
v = v.select(~((v == ".") | (v == "?")))
try:
flex.double(v)
except ValueError:
width *= -1
fmt_str.append("%%%is" % width)
fmt_str = indent_row + " ".join(fmt_str)
for i in range(self.size()):
print >> out, (fmt_str %
tuple([values[j][i]
for j in range_len_values])).rstrip()
else:
for i in range(self.size()):
values_to_print = [
format_value(values[j][i]) for j in range_len_values
]
print >> out, ' '.join([indent] + values_to_print)
def __str__(self):
s = StringIO()
self.show(out=s)
return s.getvalue()
def iterrows(self):
""" Warning! Still super-slow! """
keys = self.keys()
s_values = self.values()
range_len_self = range(len(self))
# range is 1% faster than xrange in this particular place.
# tuple (s_values...) is slightly faster than list
for j in range(self.size()):
yield OrderedDict(
zip(keys, (s_values[i][j] for i in range_len_self)))
def find_row(self, kv_dict):
self_keys = self.keys()
for k in kv_dict.keys():
assert k in self_keys
result = []
s_values = self.values()
range_len_self = range(len(self))
for i in range(self.size()):
goodrow = True
for k, v in kv_dict.iteritems():
if self[k][i] != v:
goodrow = False
break
if goodrow:
result.append(
OrderedDict(
zip(self_keys,
[s_values[j][i] for j in range_len_self])))
return result
def sort(self, key=None, reverse=False):
self._columns = OrderedDict(
sorted(self._columns.items(), key=key, reverse=reverse))
def order(self, order):
def _cmp_key(k1, k2):
for i, o in enumerate(order):
if k1 == o: break
for j, o in enumerate(order):
if k2 == o: break
if k1 < k2: return -1
return 1
keys = self._columns.keys()
keys.sort(_cmp_key)
tmp = OrderedDict()
for o in order:
tmp[o] = self._columns[o]
self._columns = tmp
def __eq__(self, other):
if (len(self) != len(other) or self.size() != other.size()
or self.keys() != other.keys()):
return False
for value, other_value in zip(self.values(), other.values()):
if (value == other_value).count(True) != len(value):
return False
return True
class miller_array_builder(crystal_symmetry_builder):
observation_types = {
'_refln_F_squared': xray.intensity(),
'_refln_intensity': xray.intensity(),
'_refln_F': xray.amplitude(),
'_refln_A': None,
}
def __init__(self, cif_block, base_array_info=None, wavelengths=None):
crystal_symmetry_builder.__init__(self, cif_block)
if base_array_info is not None:
self.crystal_symmetry = self.crystal_symmetry.join_symmetry(
other_symmetry=base_array_info.crystal_symmetry_from_file,
force=True)
self._arrays = OrderedDict()
if (wavelengths is None) :
wavelengths = {}
if base_array_info is None:
base_array_info = miller.array_info(source_type="cif")
refln_containing_loops = self.get_miller_indices_containing_loops()
for self.indices, refln_loop in refln_containing_loops:
self.wavelength_id_array = None
self.crystal_id_array = None
self.scale_group_array = None
wavelength_ids = [None]
crystal_ids = [None]
scale_groups = [None]
for key, value in refln_loop.iteritems():
# need to get these arrays first
if (key.endswith('wavelength_id') or
key.endswith('crystal_id') or
key.endswith('scale_group_code')):
data = as_int_or_none_if_all_question_marks(value, column_name=key)
if data is None:
continue
counts = data.counts()
if key.endswith('wavelength_id'):
wavelength_ids = counts.keys()
if len(counts) == 1: continue
array = miller.array(
miller.set(self.crystal_symmetry, self.indices).auto_anomalous(), data)
if key.endswith('wavelength_id'):
self.wavelength_id_array = array
wavelength_ids = counts.keys()
elif key.endswith('crystal_id'):
self.crystal_id_array = array
crystal_ids = counts.keys()
elif key.endswith('scale_group_code'):
self.scale_group_array = array
scale_groups = counts.keys()
for label, value in sorted(refln_loop.items()):
for w_id in wavelength_ids:
for crys_id in crystal_ids:
for scale_group in scale_groups:
if 'index_' in label: continue
key = label
labels = [label]
wavelength = None
if (key.endswith('wavelength_id') or
key.endswith('crystal_id') or
key.endswith('scale_group_code')):
w_id = None
crys_id = None
scale_group = None
key_suffix = ''
if w_id is not None:
key_suffix += '_%i' %w_id
labels.insert(0, "wavelength_id=%i" %w_id)
wavelength = wavelengths.get(w_id, None)
if crys_id is not None:
key_suffix += '_%i' %crys_id
labels.insert(0, "crystal_id=%i" %crys_id)
if scale_group is not None:
key_suffix += '_%i' %scale_group
labels.insert(0, "scale_group_code=%i" %scale_group)
key += key_suffix
sigmas = None
if key in self._arrays: continue
array = self.flex_std_string_as_miller_array(
value, wavelength_id=w_id, crystal_id=crys_id,
scale_group_code=scale_group)
if array is None: continue
if '_sigma' in key:
sigmas_label = label
key = None
for suffix in ('', '_meas', '_calc'):
if sigmas_label.replace('_sigma', suffix) in refln_loop:
key = sigmas_label.replace('_sigma', suffix) + key_suffix
break
if key is None:
key = sigmas_label + key_suffix
elif key in self._arrays and self._arrays[key].sigmas() is None:
sigmas = array
array = self._arrays[key]
check_array_sizes(array, sigmas, key, sigmas_label)
sigmas = as_flex_double(sigmas, sigmas_label)
array.set_sigmas(sigmas.data())
info = array.info()
array.set_info(
info.customized_copy(labels=info.labels+[sigmas_label],
wavelength=wavelength))
continue
elif 'PHWT' in key:
phwt_label = label
fwt_label = label.replace('PHWT', 'FWT')
if fwt_label not in refln_loop: continue
phwt_array = array
if fwt_label in self._arrays:
array = self._arrays[fwt_label]
check_array_sizes(array, phwt_array, fwt_label, phwt_label)
phases = as_flex_double(phwt_array, phwt_label)
info = array.info()
array = array.phase_transfer(phases, deg=True)
array.set_info(
info.customized_copy(labels=info.labels+[phwt_label]))
self._arrays[fwt_label] = array
continue
elif 'HL_' in key:
hl_letter = key[key.find('HL_')+3]
hl_key = 'HL_' + hl_letter
key = key.replace(hl_key, 'HL_A')
if key in self._arrays:
continue # this array is already dealt with
hl_labels = [label.replace(hl_key, 'HL_'+letter) for letter in 'ABCD']
hl_keys = [key.replace(hl_key, 'HL_'+letter) for letter in 'ABCD']
hl_values = [cif_block.get(hl_key) for hl_key in hl_labels]
if hl_values.count(None) == 0:
selection = self.get_selection(
hl_values[0], wavelength_id=w_id,
crystal_id=crys_id, scale_group_code=scale_group)
hl_values = [as_double_or_none_if_all_question_marks(
hl.select(selection), column_name=lab)
for hl, lab in zip(hl_values, hl_labels)]
array = miller.array(miller.set(
self.crystal_symmetry, self.indices.select(selection)
).auto_anomalous(), flex.hendrickson_lattman(*hl_values))
labels = labels[:-1]+hl_labels
elif '.B_' in key or '_B_' in key:
if '.B_' in key:
key, key_b = key.replace('.B_', '.A_'), key
label, label_b = label.replace('.B_', '.A_'), label
elif '_B_' in key:
key, key_b = key.replace('_B', '_A'), key
label, label_b = label.replace('_B', '_A'), label
if key in refln_loop and key_b in refln_loop:
b_part = array.data()
if key in self._arrays:
info = self._arrays[key].info()
a_part = self._arrays[key].data()
self._arrays[key] = self._arrays[key].array(
data=flex.complex_double(a_part, b_part))
self._arrays[key].set_info(
info.customized_copy(labels=info.labels+[key_b]))
continue
elif ('phase_' in key and not "_meas" in key and
self.crystal_symmetry.space_group() is not None):
alt_key1 = label.replace('phase_', 'F_')
alt_key2 = alt_key1 + '_au'
if alt_key1 in refln_loop:
phase_key = label
key = alt_key1+key_suffix
elif alt_key2 in refln_loop:
phase_key = label
key = alt_key2+key_suffix
else: phase_key = None
if phase_key is not None:
phases = array.data()
if key in self._arrays:
array = self._arrays[key]
array = as_flex_double(array, key)
check_array_sizes(array, phases, key, phase_key)
info = self._arrays[key].info()
self._arrays[key] = array.phase_transfer(phases, deg=True)
self._arrays[key].set_info(
info.customized_copy(labels=info.labels+[phase_key]))
else:
array = self.flex_std_string_as_miller_array(
refln_loop[label], wavelength_id=w_id, crystal_id=crys_id,
scale_group_code=scale_group)
check_array_sizes(array, phases, key, phase_key)
array.phase_transfer(phases, deg=True)
labels = labels+[label, phase_key]
if base_array_info.labels is not None:
labels = base_array_info.labels + labels
def rstrip_substrings(string, substrings):
for substr in substrings:
if substr == '': continue
if string.endswith(substr):
string = string[:-len(substr)]
return string
# determine observation type
stripped_key = rstrip_substrings(
key, [key_suffix, '_au', '_meas', '_calc', '_plus', '_minus'])
if (stripped_key.endswith('F_squared') or
stripped_key.endswith('intensity') or
stripped_key.endswith('.I') or
stripped_key.endswith('_I')) and (
array.is_real_array() or array.is_integer_array()):
array.set_observation_type_xray_intensity()
elif (stripped_key.endswith('F') and (
array.is_real_array() or array.is_integer_array())):
array.set_observation_type_xray_amplitude()
if (array.is_xray_amplitude_array() or
array.is_xray_amplitude_array()):
# e.g. merge_equivalents treats integer arrays differently, so must
# convert integer observation arrays here to be safe
if isinstance(array.data(), flex.int):
array = array.customized_copy(data=array.data().as_double())
array.set_info(base_array_info.customized_copy(labels=labels))
if (array.is_xray_amplitude_array() or
array.is_xray_amplitude_array()):
info = array.info()
array.set_info(info.customized_copy(wavelength=wavelength))
self._arrays.setdefault(key, array)
for key, array in self._arrays.copy().iteritems():
if ( key.endswith('_minus') or '_minus_' in key
or key.endswith('_plus') or '_plus_' in key):
if '_minus' in key:
minus_key = key
plus_key = key.replace('_minus', '_plus')
elif '_plus' in key:
plus_key = key
minus_key = key.replace('_plus', '_minus')
if plus_key in self._arrays and minus_key in self._arrays:
plus_array = self._arrays.pop(plus_key)
minus_array = self._arrays.pop(minus_key)
minus_array = minus_array.customized_copy(
indices=-minus_array.indices()).set_info(minus_array.info())
array = plus_array.concatenate(
minus_array, assert_is_similar_symmetry=False)
array = array.customized_copy(anomalous_flag=True)
array.set_info(minus_array.info().customized_copy(
labels=list(
OrderedSet(plus_array.info().labels+minus_array.info().labels))))
array.set_observation_type(plus_array.observation_type())
self._arrays.setdefault(key, array)
if len(self._arrays) == 0:
raise CifBuilderError("No reflection data present in cif block")
def get_miller_indices_containing_loops(self):
loops = []
for loop in self.cif_block.loops.values():
for key in loop.keys():
if 'index_h' not in key: continue
hkl_str = [loop.get(key.replace('index_h', 'index_%s' %i)) for i in 'hkl']
if hkl_str.count(None) > 0:
raise CifBuilderError(
"Miller indices missing from current CIF block (%s)"
%key.replace('index_h', 'index_%s' %'hkl'[hkl_str.index(None)]))
hkl_int = []
for i,h_str in enumerate(hkl_str):
try:
h_int = flex.int(h_str)
except ValueError, e:
raise CifBuilderError(
"Invalid item for Miller index %s: %s" % ("HKL"[i], str(e)))
hkl_int.append(h_int)
indices = flex.miller_index(*hkl_int)
loops.append((indices, loop))
break
return loops
class loop(DictMixin):
def __init__(self, header=None, data=None):
self._columns = OrderedDict()
self.keys_lower = {}
if header is not None:
for key in header:
self.setdefault(key, flex.std_string())
if data is not None:
# the number of data items must be an exact multiple of the number of headers
assert len(data) % len(header) == 0, "Wrong number of data items for loop"
n_rows = len(data)//len(header)
n_columns = len(header)
for i in range(n_rows):
self.add_row([data[i*n_columns+j] for j in range(n_columns)])
elif header is None and data is not None:
assert isinstance(data, dict) or isinstance(data, OrderedDict)
self.add_columns(data)
self.keys_lower = dict(
[(key.lower(), key) for key in self._columns.keys()])
def __setitem__(self, key, value):
if not re.match(tag_re, key):
raise Sorry("%s is not a valid data name" %key)
if len(self) > 0:
assert len(value) == self.size()
if not isinstance(value, flex.std_string):
for flex_numeric_type in (flex.int, flex.double):
if isinstance(value, flex_numeric_type):
value = value.as_string()
else:
try:
value = flex_numeric_type(value).as_string()
except TypeError:
continue
else:
break
if not isinstance(value, flex.std_string):
value = flex.std_string(value)
# value must be a mutable type
assert hasattr(value, '__setitem__')
self._columns[key] = value
self.keys_lower[key.lower()] = key
def __getitem__(self, key):
return self._columns[self.keys_lower[key.lower()]]
def __delitem__(self, key):
del self._columns[self.keys_lower[key.lower()]]
del self.keys_lower[key.lower()]
def keys(self):
return self._columns.keys()
def __repr__(self):
return repr(OrderedDict(self.iteritems()))
def name(self):
return common_substring(self.keys()).rstrip('_').rstrip('.')
def size(self):
size = 0
for column in self.values():
size = max(size, len(column))
return size
def n_rows(self):
size = 0
for column in self.values():
size = max(size, len(column))
return size
def n_columns(self):
return len(self.keys())
def add_row(self, row, default_value="?"):
if isinstance(row, dict):
for key in self:
if key in row:
self[key].append(str(row[key]))
else:
self[key].append(default_value)
else:
assert len(row) == len(self)
for i, key in enumerate(self):
self[key].append(str(row[i]))
def add_column(self, key, values):
if self.size() != 0:
assert len(values) == self.size()
self[key] = values
self.keys_lower[key.lower()] = key
def add_columns(self, columns):
assert isinstance(columns, dict) or isinstance(columns, OrderedDict)
for key, value in columns.iteritems():
self.add_column(key, value)
def update_column(self, key, values):
assert type(key)==type(""), "first argument is column key string"
if self.size() != 0:
assert len(values) == self.size(), "len(values) %d != self.size() %d" % (
len(values),
self.size(),
)
self[key] = values
self.keys_lower[key.lower()] = key
def delete_row(self, index):
assert index < self.n_rows()
for column in self._columns.values():
del column[index]
def __copy__(self):
new = loop()
new._columns = self._columns.copy()
new.keys_lower = self.keys_lower.copy()
return new
copy = __copy__
def __deepcopy__(self, memo):
new = loop()
new._columns = copy.deepcopy(self._columns, memo)
new.keys_lower = copy.deepcopy(self.keys_lower, memo)
return new
def deepcopy(self):
return copy.deepcopy(self)
def show(self, out=None, indent=" ", indent_row=None, fmt_str=None, align_columns=True):
assert self.n_rows() > 0 and self.n_columns() > 0, "keys: %s %d %d" % (
self.keys(),
self.n_rows(),
self.n_columns(),
)
if out is None:
out = sys.stdout
if indent_row is None:
indent_row = indent
assert indent.strip() == ""
assert indent_row.strip() == ""
print >> out, "loop_"
for k in self.keys():
print >> out, indent + k
values = self._columns.values()
if fmt_str is not None:
# Pretty printing:
# The user is responsible for providing a valid format string.
# Values are not quoted - it is the user's responsibility to place
# appropriate quotes in the format string if a particular value may
# contain spaces.
values = copy.deepcopy(values)
for i, v in enumerate(values):
for flex_numeric_type in (flex.int, flex.double):
if not isinstance(v, flex_numeric_type):
try:
values[i] = flex_numeric_type(v)
except ValueError:
continue
else:
break
if fmt_str is None:
fmt_str = indent_row + ' '.join(["%s"]*len(values))
for i in range(self.size()):
print >> out, fmt_str % tuple([values[j][i] for j in range(len(values))])
elif align_columns:
fmt_str = []
for i, (k, v) in enumerate(self.iteritems()):
for i_v in range(v.size()):
v[i_v] = format_value(v[i_v])
# exclude and semicolon text fields from column width calculation
v_ = flex.std_string(item for item in v if "\n" not in item)
width = v_.max_element_length()
# See if column contains only number, '.' or '?'
# right-align numerical columns, left-align everything else
v = v.select(~( (v == ".") | (v == "?") ))
try:
flex.double(v)
except ValueError:
width *= -1
fmt_str.append("%%%is" %width)
fmt_str = indent_row + " ".join(fmt_str)
for i in range(self.size()):
print >> out, (fmt_str %
tuple([values[j][i]
for j in range(len(values))])).rstrip()
else:
for i in range(self.size()):
values_to_print = [format_value(values[j][i]) for j in range(len(values))]
print >> out, ' '.join([indent] + values_to_print)
def __str__(self):
s = StringIO()
self.show(out=s)
return s.getvalue()
def iterrows(self):
keys = self.keys()
for j in range(self.size()):
yield OrderedDict(zip(keys, [self.values()[i][j] for i in range(len(self))]))
def sort(self, key=None, reverse=False):
self._columns = OrderedDict(
sorted(self._columns.items(), key=key, reverse=reverse))
def order(self, order):
def _cmp_key(k1, k2):
for i, o in enumerate(order):
if k1==o: break
for j, o in enumerate(order):
if k2==o: break
if k1<k2: return -1
return 1
keys = self._columns.keys()
keys.sort(_cmp_key)
tmp = OrderedDict()
for o in order:
tmp[o]=self._columns[o]
self._columns = tmp
def __eq__(self, other):
if (len(self) != len(other) or
self.size() != other.size() or
self.keys() != other.keys()):
return False
for value, other_value in zip(self.values(), other.values()):
if (value == other_value).count(True) != len(value):
return False
return True
class cif(DictMixin):
def __init__(self, blocks=None):
if blocks is not None:
self.blocks = OrderedDict(blocks)
else:
self.blocks = OrderedDict()
self.keys_lower = dict([(key.lower(), key) for key in self.blocks.keys()])
def __setitem__(self, key, value):
assert isinstance(value, block)
if not re.match(tag_re, '_'+key):
raise Sorry("%s is not a valid data block name" %key)
self.blocks[key] = value
self.keys_lower[key.lower()] = key
def get(self, key, default=None):
key_lower = self.keys_lower.get(key.lower())
if (key_lower is None):
return default
return self.blocks.get(key_lower, default)
def __getitem__(self, key):
result = self.get(key)
if (result is None):
raise KeyError('Unknown CIF data block name: "%s"' % key)
return result
def __delitem__(self, key):
del self.blocks[self.keys_lower[key.lower()]]
del self.keys_lower[key.lower()]
def keys(self):
return self.blocks.keys()
def __repr__(self):
return repr(OrderedDict(self.iteritems()))
def __copy__(self):
return cif(self.blocks.copy())
copy = __copy__
def __deepcopy__(self, memo):
return cif(copy.deepcopy(self.blocks, memo))
def deepcopy(self):
return copy.deepcopy(self)
def show(self, out=None, indent=" ", indent_row=None,
data_name_field_width=34,
loop_format_strings=None):
if out is None:
out = sys.stdout
for name, block in self.items():
print >> out, "data_%s" %name
block.show(
out=out, indent=indent, indent_row=indent_row,
data_name_field_width=data_name_field_width,
loop_format_strings=loop_format_strings)
def __str__(self):
s = StringIO()
self.show(out=s)
return s.getvalue()
def validate(self, dictionary, show_warnings=True, error_handler=None, out=None):
if out is None: out = sys.stdout
from iotbx.cif import validation
errors = {}
if error_handler is None:
error_handler = validation.ErrorHandler()
for key, block in self.blocks.iteritems():
error_handler = error_handler.__class__()
dictionary.set_error_handler(error_handler)
block.validate(dictionary)
errors.setdefault(key, error_handler)
if error_handler.error_count or error_handler.warning_count:
error_handler.show(show_warnings=show_warnings, out=out)
return error_handler
def sort(self, recursive=False, key=None, reverse=False):
self.blocks = OrderedDict(sorted(self.blocks.items(), key=key, reverse=reverse))
if recursive:
for b in self.blocks.values():
b.sort(recursive=recursive, reverse=reverse)
class miller_array_builder(crystal_symmetry_builder):
observation_types = {
'_refln_F_squared': xray.intensity(),
'_refln_intensity': xray.intensity(),
'_refln_F': xray.amplitude(),
'_refln_A': None,
}
def __init__(self, cif_block, base_array_info=None):
crystal_symmetry_builder.__init__(self, cif_block)
if base_array_info is not None:
self.crystal_symmetry = self.crystal_symmetry.join_symmetry(
other_symmetry=base_array_info.crystal_symmetry_from_file,
force=True)
self._arrays = OrderedDict()
if base_array_info is None:
base_array_info = miller.array_info(source_type="cif")
refln_containing_loops = self.get_miller_indices_containing_loops()
for self.indices, refln_loop in refln_containing_loops:
self.wavelength_id_array = None
self.crystal_id_array = None
self.scale_group_array = None
wavelength_ids = [None]
crystal_ids = [None]
scale_groups = [None]
for key, value in refln_loop.iteritems():
# need to get these arrays first
if (key.endswith('wavelength_id') or key.endswith('crystal_id')
or key.endswith('scale_group_code')):
data = as_int_or_none_if_all_question_marks(
value, column_name=key)
if data is None: continue
counts = data.counts()
if len(counts) == 1: continue
array = miller.array(
miller.set(self.crystal_symmetry,
self.indices).auto_anomalous(), data)
if key.endswith('wavelength_id'):
self.wavelength_id_array = array
wavelength_ids = counts.keys()
elif key.endswith('crystal_id'):
self.crystal_id_array = array
crystal_ids = counts.keys()
elif key.endswith('scale_group_code'):
self.scale_group_array = array
scale_groups = counts.keys()
for label, value in sorted(refln_loop.items()):
for w_id in wavelength_ids:
for crys_id in crystal_ids:
for scale_group in scale_groups:
if 'index_' in label: continue
key = label
labels = [label]
if (key.endswith('wavelength_id')
or key.endswith('crystal_id')
or key.endswith('scale_group_code')):
w_id = None
crys_id = None
scale_group = None
key_suffix = ''
if w_id is not None:
key_suffix += '_%i' % w_id
labels.insert(0, "wavelength_id=%i" % w_id)
if crys_id is not None:
key_suffix += '_%i' % crys_id
labels.insert(0, "crystal_id=%i" % crys_id)
if scale_group is not None:
key_suffix += '_%i' % scale_group
labels.insert(
0, "scale_group_code=%i" % scale_group)
key += key_suffix
sigmas = None
if key in self._arrays: continue
array = self.flex_std_string_as_miller_array(
value,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
if array is None: continue
if '_sigma' in key:
sigmas_label = label
key = None
for suffix in ('', '_meas', '_calc'):
if sigmas_label.replace(
'_sigma', suffix) in refln_loop:
key = sigmas_label.replace(
'_sigma', suffix) + key_suffix
break
if key is None:
key = sigmas_label + key_suffix
elif key in self._arrays and self._arrays[
key].sigmas() is None:
sigmas = array
array = self._arrays[key]
check_array_sizes(array, sigmas, key,
sigmas_label)
sigmas = as_flex_double(
sigmas, sigmas_label)
array.set_sigmas(sigmas.data())
info = array.info()
array.set_info(
info.customized_copy(
labels=info.labels +
[sigmas_label]))
continue
elif 'PHWT' in key:
phwt_label = label
fwt_label = label.replace('PHWT', 'FWT')
if fwt_label not in refln_loop: continue
phwt_array = array
if fwt_label in self._arrays:
array = self._arrays[fwt_label]
check_array_sizes(array, phwt_array,
fwt_label, phwt_label)
phases = as_flex_double(
phwt_array, phwt_label)
info = array.info()
array = array.phase_transfer(phases,
deg=True)
array.set_info(
info.customized_copy(
labels=info.labels + [phwt_label]))
self._arrays[fwt_label] = array
continue
elif 'HL_' in key:
hl_letter = key[key.find('HL_') + 3]
hl_key = 'HL_' + hl_letter
key = key.replace(hl_key, 'HL_A')
if key in self._arrays:
continue # this array is already dealt with
hl_labels = [
label.replace(hl_key, 'HL_' + letter)
for letter in 'ABCD'
]
hl_keys = [
key.replace(hl_key, 'HL_' + letter)
for letter in 'ABCD'
]
hl_values = [
cif_block.get(hl_key)
for hl_key in hl_labels
]
if hl_values.count(None) == 0:
selection = self.get_selection(
hl_values[0],
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
hl_values = [
as_double_or_none_if_all_question_marks(
hl.select(selection),
column_name=lab)
for hl, lab in zip(
hl_values, hl_labels)
]
array = miller.array(
miller.set(
self.crystal_symmetry,
self.indices.select(
selection)).auto_anomalous(),
flex.hendrickson_lattman(*hl_values))
labels = labels[:-1] + hl_labels
elif '.B_' in key or '_B_' in key:
if '.B_' in key:
key, key_b = key.replace('.B_', '.A_'), key
label, label_b = label.replace(
'.B_', '.A_'), label
elif '_B_' in key:
key, key_b = key.replace('_B', '_A'), key
label, label_b = label.replace('_B',
'_A'), label
if key in refln_loop and key_b in refln_loop:
b_part = array.data()
if key in self._arrays:
info = self._arrays[key].info()
a_part = self._arrays[key].data()
self._arrays[key] = self._arrays[
key].array(
data=flex.complex_double(
a_part, b_part))
self._arrays[key].set_info(
info.customized_copy(
labels=info.labels + [key_b]))
continue
elif ('phase_' in key and not key.endswith('_meas')
and self.crystal_symmetry.space_group()
is not None):
alt_key1 = label.replace('phase_', 'F_')
alt_key2 = alt_key1 + '_au'
if alt_key1 in refln_loop:
phase_key = label
key = alt_key1 + key_suffix
elif alt_key2 in refln_loop:
phase_key = label
key = alt_key2 + key_suffix
else:
phase_key = None
if phase_key is not None:
phases = array.data()
if key in self._arrays:
array = self._arrays[key]
array = as_flex_double(array, key)
check_array_sizes(
array, phases, key, phase_key)
info = self._arrays[key].info()
self._arrays[
key] = array.phase_transfer(
phases, deg=True)
self._arrays[key].set_info(
info.customized_copy(
labels=info.labels +
[phase_key]))
else:
array = self.flex_std_string_as_miller_array(
refln_loop[label],
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
check_array_sizes(
array, phases, key, phase_key)
array.phase_transfer(phases, deg=True)
labels = labels + [label, phase_key]
if base_array_info.labels is not None:
labels = base_array_info.labels + labels
def rstrip_substrings(string, substrings):
for substr in substrings:
if substr == '': continue
if string.endswith(substr):
string = string[:-len(substr)]
return string
# determine observation type
stripped_key = rstrip_substrings(
key, [
key_suffix, '_au', '_meas', '_calc',
'_plus', '_minus'
])
if (stripped_key.endswith('F_squared')
or stripped_key.endswith('intensity')
or stripped_key.endswith('.I')
or stripped_key.endswith('_I')) and (
array.is_real_array()
or array.is_integer_array()):
array.set_observation_type_xray_intensity()
elif (stripped_key.endswith('F')
and (array.is_real_array()
or array.is_integer_array())):
array.set_observation_type_xray_amplitude()
if (array.is_xray_amplitude_array()
or array.is_xray_amplitude_array()):
# e.g. merge_equivalents treats integer arrays differently, so must
# convert integer observation arrays here to be safe
if isinstance(array.data(), flex.int):
array = array.customized_copy(
data=array.data().as_double())
array.set_info(
base_array_info.customized_copy(labels=labels))
self._arrays.setdefault(key, array)
for key, array in self._arrays.copy().iteritems():
if (key.endswith('_minus') or '_minus_' in key
or key.endswith('_plus') or '_plus_' in key):
if '_minus' in key:
minus_key = key
plus_key = key.replace('_minus', '_plus')
elif '_plus' in key:
plus_key = key
minus_key = key.replace('_plus', '_minus')
if plus_key in self._arrays and minus_key in self._arrays:
plus_array = self._arrays.pop(plus_key)
minus_array = self._arrays.pop(minus_key)
minus_array = minus_array.customized_copy(
indices=-minus_array.indices()).set_info(
minus_array.info())
array = plus_array.concatenate(
minus_array, assert_is_similar_symmetry=False)
array = array.customized_copy(anomalous_flag=True)
array.set_info(
minus_array.info().customized_copy(labels=list(
OrderedSet(plus_array.info().labels +
minus_array.info().labels))))
array.set_observation_type(plus_array.observation_type())
self._arrays.setdefault(key, array)
if len(self._arrays) == 0:
raise CifBuilderError("No reflection data present in cif block")
def get_miller_indices_containing_loops(self):
loops = []
for loop in self.cif_block.loops.values():
for key in loop.keys():
if 'index_h' not in key: continue
hkl_str = [
loop.get(key.replace('index_h', 'index_%s' % i))
for i in 'hkl'
]
if hkl_str.count(None) > 0:
raise CifBuilderError(
"Miller indices missing from current CIF block (%s)" %
key.replace('index_h',
'index_%s' % 'hkl'[hkl_str.index(None)]))
hkl_int = []
for i, h_str in enumerate(hkl_str):
try:
h_int = flex.int(h_str)
except ValueError, e:
raise CifBuilderError(
"Invalid item for Miller index %s: %s" %
("HKL"[i], str(e)))
hkl_int.append(h_int)
indices = flex.miller_index(*hkl_int)
loops.append((indices, loop))
break
return loops
class miller_array_builder(crystal_symmetry_builder):
# Changes to this class should pass regression tests:
# cctbx_project\mmtbx\regression\tst_cif_as_mtz_wavelengths.py
# cctbx_project\iotbx\cif\tests\tst_lex_parse_build.py
# phenix_regression\cif_as_mtz\tst_cif_as_mtz.py
observation_types = {
# known types of column data to be tagged as either amplitudes or intensities as per
# https://www.iucr.org/__data/iucr/cifdic_html/2/cif_mm.dic/index.html
'_refln.F_squared': xray.intensity(),
'_refln_F_squared': xray.intensity(),
'_refln.intensity': xray.intensity(),
'_refln.I(+)': xray.intensity(),
'_refln.I(-)': xray.intensity(),
'_refln.F_calc': xray.amplitude(),
'_refln.F_meas': xray.amplitude(),
'_refln.FP': xray.amplitude(),
'_refln.F-obs': xray.amplitude(),
'_refln.Fobs': xray.amplitude(),
'_refln.F-calc': xray.amplitude(),
'_refln.Fcalc': xray.amplitude(),
'_refln.pdbx_F_': xray.amplitude(),
'_refln.pdbx_I_': xray.intensity(),
'_refln.pdbx_anom_difference': xray.amplitude(),
}
def guess_observationtype(self, labl):
for okey in self.observation_types.keys():
if labl.startswith(okey):
return self.observation_types[okey]
return None
def __init__(self, cif_block, base_array_info=None, wavelengths=None):
crystal_symmetry_builder.__init__(self, cif_block)
self._arrays = OrderedDict()
self._origarrays = OrderedDict(
) # used for presenting raw data tables in HKLviewer
basearraylabels = []
if base_array_info is not None:
self.crystal_symmetry = self.crystal_symmetry.join_symmetry(
other_symmetry=base_array_info.crystal_symmetry_from_file,
force=True)
if base_array_info.labels:
basearraylabels = base_array_info.labels
if (wavelengths is None):
wavelengths = {}
if base_array_info is None:
base_array_info = miller.array_info(source_type="cif")
refln_containing_loops = self.get_miller_indices_containing_loops()
for self.indices, refln_loop in refln_containing_loops:
self.wavelength_id_array = None
self.crystal_id_array = None
self.scale_group_array = None
wavelength_ids = [None]
crystal_ids = [None]
scale_groups = [None]
for key, value in six.iteritems(refln_loop):
# Get wavelength_ids, crystal_id, scale_group_code columns for selecting data of other
# columns in self.get_selection() used by self.flex_std_string_as_miller_array()
if (key.endswith('wavelength_id') or key.endswith('crystal_id')
or key.endswith('scale_group_code')):
data = as_int_or_none_if_all_question_marks(
value, column_name=key)
if data is None:
continue
counts = data.counts()
if key.endswith('wavelength_id'):
wavelength_ids = list(counts.keys())
if len(counts) == 1: continue
array = miller.array(
miller.set(self.crystal_symmetry,
self.indices).auto_anomalous(), data)
if key.endswith('wavelength_id'):
self.wavelength_id_array = array
wavelength_ids = list(counts.keys())
elif key.endswith('crystal_id'):
self.crystal_id_array = array
crystal_ids = list(counts.keys())
elif key.endswith('scale_group_code'):
self.scale_group_array = array
scale_groups = list(counts.keys())
labelsuffix = []
wavelbl = []
cryslbl = []
scalegrplbl = []
self._origarrays["HKLs"] = self.indices
alllabels = list(sorted(refln_loop.keys()))
remaininglabls = alllabels[:] # deep copy the list
# Parse labels matching cif column conventions
# https://mmcif.wwpdb.org/dictionaries/mmcif_pdbx_v50.dic/Categories/refln.html
# and extract groups of labels or just single columns.
# Groups corresponds to the map coefficients, phase and amplitudes,
# amplitudes or intensities with sigmas and hendrickson-lattman columns.
phaseamplabls, remaininglabls = self.get_phase_amplitude_labels(
remaininglabls)
mapcoefflabls, remaininglabls = self.get_mapcoefficient_labels(
remaininglabls)
HLcoefflabls, remaininglabls = self.get_HL_labels(remaininglabls)
data_sig_obstype_labls, remaininglabls = self.get_FSigF_ISigI_labels(
remaininglabls)
for w_id in wavelength_ids:
for crys_id in crystal_ids:
for scale_group in scale_groups:
# If reflection data files contain more than one crystal, wavelength or scalegroup
# then add their id(s) as a suffix to data labels computed below. Needed for avoiding
# ambuguity but avoid when not needed to make labels more human readable!
if (len(wavelength_ids) > 1
or len(wavelengths) > 1) and w_id is not None:
wavelbl = ["wavelength_id=%i" % w_id]
if len(crystal_ids) > 1 and crys_id is not None:
cryslbl = ["crystal_id=%i" % crys_id]
if len(scale_groups) > 1 and scale_group is not None:
scalegrplbl = ["scale_group_code=%i" % scale_group]
labelsuffix = scalegrplbl + cryslbl + wavelbl
jlablsufx = ""
if len(labelsuffix):
jlablsufx = "," + ",".join(labelsuffix)
for mapcoefflabl in mapcoefflabls:
A_array = refln_loop[mapcoefflabl[0]]
B_array = refln_loop[mapcoefflabl[1]]
# deselect any ? marks in the two arrays, assuming both A and B have the same ? marks
selection = self.get_selection(
A_array,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
A_array = A_array.select(selection)
B_array = B_array.select(selection)
# form the miller array with map coefficients
data = flex.complex_double(flex.double(A_array),
flex.double(B_array))
millarr = miller.array(
miller.set(self.crystal_symmetry,
self.indices.select(
selection)).auto_anomalous(),
data)
# millarr will be None for column data not matching w_id,crys_id,scale_group values
if millarr is None: continue
labl = basearraylabels + mapcoefflabl + labelsuffix
millarr.set_info(
base_array_info.customized_copy(
labels=labl,
wavelength=wavelengths.get(w_id, None)))
self._arrays[mapcoefflabl[0] + jlablsufx] = millarr
for phaseamplabl in phaseamplabls:
amplitudestrarray = refln_loop[phaseamplabl[0]]
phasestrarray = refln_loop[phaseamplabl[1]]
millarr = self.flex_std_string_as_miller_array(
amplitudestrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
phasesmillarr = self.flex_std_string_as_miller_array(
phasestrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
# millarr will be None for column data not matching w_id,crys_id,scale_group values
if millarr is None or phasesmillarr is None:
continue
phases = as_flex_double(phasesmillarr,
phaseamplabl[1])
millarr = millarr.phase_transfer(phases, deg=True)
labl = basearraylabels + phaseamplabl + labelsuffix
millarr.set_info(
base_array_info.customized_copy(
labels=labl,
wavelength=wavelengths.get(w_id, None)))
self._arrays[phaseamplabl[0] + jlablsufx] = millarr
for datlabl, siglabl, otype in data_sig_obstype_labls:
datastrarray = refln_loop[datlabl]
millarr = self.flex_std_string_as_miller_array(
datastrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
# millarr will be None for column data not matching w_id,crys_id,scale_group values
if millarr is None: continue
millarr = as_flex_double(millarr, datlabl)
datsiglabl = [datlabl]
if siglabl:
sigmasstrarray = refln_loop[siglabl]
sigmas = self.flex_std_string_as_miller_array(
sigmasstrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
sigmas = as_flex_double(sigmas, siglabl)
millarr.set_sigmas(sigmas.data())
datsiglabl = [datlabl, siglabl]
datsiglabl = basearraylabels + datsiglabl + labelsuffix
millarr.set_info(
base_array_info.customized_copy(
labels=datsiglabl,
wavelength=wavelengths.get(w_id, None)))
if otype is not None:
millarr.set_observation_type(otype)
self._arrays[datlabl + jlablsufx] = millarr
for hl_labels in HLcoefflabls:
hl_values = [
cif_block.get(hl_key) for hl_key in hl_labels
]
if hl_values.count(None) == 0:
selection = self.get_selection(
hl_values[0],
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
hl_values = [
as_double_or_none_if_all_question_marks(
hl.select(selection), column_name=lab)
for hl, lab in zip(hl_values, hl_labels)
]
# hl_values will be None for column data not matching w_id,crys_id,scale_group values
if hl_values == [None, None, None, None]:
continue
millarr = miller.array(
miller.set(
self.crystal_symmetry,
self.indices.select(
selection)).auto_anomalous(),
flex.hendrickson_lattman(*hl_values))
hlabels = basearraylabels + hl_labels + labelsuffix
millarr.set_info(
base_array_info.customized_copy(
labels=hlabels,
wavelength=wavelengths.get(w_id,
None)))
self._arrays[hl_labels[0] +
jlablsufx] = millarr
# pick up remaining columns if any that weren't identified above
for label in alllabels:
if "index_" in label:
continue
datastrarray = refln_loop[label]
if label in remaininglabls:
labels = basearraylabels + [label
] + labelsuffix
lablsufx = jlablsufx
millarr = self.flex_std_string_as_miller_array(
datastrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
# millarr will be None for column data not matching w_id,crys_id,scale_group values
if (label.endswith(
'wavelength_id'
) or label.endswith(
'crystal_id'
) or # get full array if any of these labels, not just subsets
label.endswith('scale_group_code')):
millarr = self.flex_std_string_as_miller_array(
datastrarray,
wavelength_id=None,
crystal_id=None,
scale_group_code=None)
lablsufx = ""
labels = basearraylabels + [label]
if millarr is None: continue
otype = self.guess_observationtype(label)
if otype is not None:
millarr.set_observation_type(otype)
millarr.set_info(
base_array_info.customized_copy(
labels=labels,
wavelength=wavelengths.get(w_id,
None)))
self._arrays[label + lablsufx] = millarr
origarr = self.flex_std_string_as_miller_array(
datastrarray,
wavelength_id=w_id,
crystal_id=crys_id,
scale_group_code=scale_group)
newlabel = label.replace("_refln.", "")
newlabel2 = newlabel.replace("_refln_", "")
if origarr: # want only genuine miller arrays
self._origarrays[newlabel2 +
jlablsufx] = origarr.data()
# Convert any groups of I+,I-,SigI+,SigI- (or amplitudes) arrays into anomalous arrays
# i.e. both friedel mates in the same array
for key, array in six.iteritems(self._arrays.copy()):
plus_key = ""
if '_minus' in key:
minus_key = key
plus_key = key.replace('_minus', '_plus')
elif '-' in key:
minus_key = key
plus_key = key.replace('-', '+')
elif '_plus' in key:
plus_key = key
minus_key = key.replace('_plus', '_minus')
elif '+' in key:
plus_key = key
minus_key = key.replace('+', '-')
if plus_key in self._arrays and minus_key in self._arrays:
plus_array = self._arrays.pop(plus_key)
minus_array = self._arrays.pop(minus_key)
minus_array = minus_array.customized_copy(
indices=-minus_array.indices()).set_info(
minus_array.info())
array = plus_array.concatenate(
minus_array, assert_is_similar_symmetry=False)
array = array.customized_copy(anomalous_flag=True)
array.set_info(minus_array.info().customized_copy(labels=list(
OrderedSet(plus_array.info().labels +
minus_array.info().labels))))
array.set_observation_type(plus_array.observation_type())
self._arrays.setdefault(key, array)
if len(self._arrays) == 0:
raise CifBuilderError("No reflection data present in cif block")
# Sort the ordered dictionary to resemble the order of columns in the cif file
# This is to avoid any F_meas arrays accidentally being put adjacent to
# pdbx_anom_difference arrays in the self._arrays OrderedDict. Otherwise these
# arrays may unintentionally be combined into a reconstructed anomalous amplitude
# array when saving as an mtz file due to a problem in the iotbx/mtz module.
# See http://phenix-online.org/pipermail/cctbxbb/2021-March/002289.html
arrlstord = []
arrlst = list(self._arrays)
for arr in arrlst:
for i, k in enumerate(refln_loop.keys()):
if arr.split(",")[0] == k:
arrlstord.append((arr, i))
# arrlstord must have the same keys as in the self._arrays dictionary
assert sorted(arrlst) == sorted([e[0] for e in arrlstord])
sortarrlst = sorted(arrlstord, key=lambda arrord: arrord[1])
self._ordarrays = OrderedDict()
for sortkey, i in sortarrlst:
self._ordarrays.setdefault(sortkey, self._arrays[sortkey])
self._arrays = self._ordarrays
def get_HL_labels(self, keys):
lstkeys = list(keys) # cast into list if not a list
HLquads = []
alllabels = " ".join(lstkeys)
""" Hendrickson-Lattmann labels could look like: 'HLAM', 'HLBM', 'HLCM', 'HLDM'
or like 'HLanomA', 'HLanomB', 'HLanomC', 'HLanomD'
Use a regular expression to group them accordingly
"""
allmatches = re.findall(r"(\S*(HL(\S*)[abcdABCD](\S*)))", alllabels)
HLtagslst = list(set([(e[2], e[3]) for e in allmatches]))
usedkeys = []
for m in HLtagslst:
hllist = []
for hm in allmatches:
if m == (hm[2], hm[3]):
hllist.append((hm[0], hm[1]))
if len(hllist) == 4:
HLquads.append([e[0] for e in hllist])
for e in hllist:
usedkeys.append(e[0])
remainingkeys = []
for e in lstkeys:
if e not in usedkeys:
remainingkeys.append(e)
return HLquads, remainingkeys
def get_mapcoefficient_labels(self, keys):
# extract map coeffficients labels from list of cif column labels
# e.g. ( _refln.A_calc_au _refln.B_calc_au ) , ( _refln.A_calc _refln.B_calc )
lstkeys = list(keys) # cast into list if not a list
remainingkeys = lstkeys[:] # deep copy the list
alllabels = " ".join(lstkeys)
mapcoefflabels = []
A_matches = re.findall(
r"( (\s*_refln[\._]A_)(\S*) )", alllabels, re.VERBOSE
) # [('_refln.PHWT', '_refln.PH', 'WT'), ('_refln.PHDELWT', '_refln.PH', 'DELWT')]
for label in lstkeys:
for m in A_matches:
Blabel = m[1].replace("A_", "B_") + m[2]
if Blabel == label:
mapcoefflabels.append([m[0], label])
remainingkeys.remove(m[0])
remainingkeys.remove(label)
return mapcoefflabels, remainingkeys
def get_phase_amplitude_labels(self, keys):
# extract phase and amplitudes labels from list of cif column labels
# e.g. ( _refln.F_calc _refln.phase_calc ) , ( _refln.FC_ALL _refln.PHIC_ALL ), ( _refln.FWT _refln.PHWT )
lstkeys = list(keys) # cast into list if not a list
remainingkeys = lstkeys[:] # deep copy the list
alllabels = " ".join(lstkeys)
phase_amplitudelabels = []
PHmatches = re.findall(
r"((\S*PH)([^I]\S*))", alllabels
) # [('_refln.PHWT', '_refln.PH', 'WT'), ('_refln.PHDELWT', '_refln.PH', 'DELWT')]
for label in lstkeys:
for m in PHmatches:
PFlabel = m[1].replace("PH", "F") + m[2]
Flabel = m[1].replace("PH", "") + m[2]
if Flabel == label or PFlabel == label:
phase_amplitudelabels.append([label, m[0]])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
alllabels = " ".join(remainingkeys)
PHImatches = re.findall(
r"((\S*PHI)(\S*))", alllabels
) # [('_refln.PHIC', '_refln.PHI', 'C'), ('_refln.PHIC_ALL', '_refln.PHI', 'C_ALL')]
for label in lstkeys:
for m in PHImatches:
PFlabel = m[1].replace("PHI", "F") + m[2]
Flabel = m[1].replace("PHI", "") + m[2]
if Flabel == label or PFlabel == label:
phase_amplitudelabels.append([label, m[0]])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
alllabels = " ".join(remainingkeys)
PHDELmatches = re.findall(
r"(((\S*)PH)([^I]\S*(WT)))", alllabels
) # [('_refln.PHDELWT', '_refln.PH', '_refln.', 'DELWT', 'WT')]
for label in lstkeys:
for m in PHDELmatches:
Flabel = m[2] + m[3].replace("WT", "FWT")
if Flabel == label:
phase_amplitudelabels.append([label, m[0]])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
alllabels = " ".join(remainingkeys)
phase_matches = re.findall(
r"((\S*[\._])phase(\S*))",
alllabels) # [('_refln.phase_calc', '_refln.', '')]
for label in lstkeys:
for m in phase_matches:
phaselabel = m[0]
Flabl = m[1] + m[2]
Flabel = m[1] + "F" + m[2]
Faulabel = m[1] + "F" + m[2] + "_au"
if Flabl in label or Flabel in label or Faulabel in label: # in case of _refln.F_calc_au and _refln.phase_calc
if label in remainingkeys and m[
0] in remainingkeys: # in case
if (Flabel + "_sigma_au") in remainingkeys or (
Flabel + "_sigma") in remainingkeys:
continue # give priority to F_meas, F_meas_sigma or F_meas_au, F_meas_sigma_au
phase_amplitudelabels.append([label, m[0]])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
return phase_amplitudelabels, remainingkeys
def get_FSigF_ISigI_labels(self, keys):
# extract amplitudea, sigmas or intensitiy, sigmas labels from list of cif column labels
# e.g. ( _refln.F_meas_sigma_au _refln.F_meas), ( _refln.intensity_sigma _refln.intensity ) ,
# ( _refln.pdbx_I_plus_sigma _refln.pdbx_I_plus )
lstkeys = list(keys) # cast into list if not a list
remainingkeys = lstkeys[:] # deep copy the list
alllabels = " ".join(lstkeys)
labelpairs = []
sigma_matches = re.findall(
r"((\S*[\._])SIG(\S*))",
alllabels) # catch label pairs like F(+),SIGF(+)
for label in lstkeys:
for m in sigma_matches:
FIlabel = m[1] + m[2]
if FIlabel == label:
labelpairs.append(
[label, m[0],
self.guess_observationtype(label)])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
alllabels = " ".join(remainingkeys)
sigma_matches = re.findall(
r"((\S*)_sigma(_*\S*))", alllabels
) # [('_refln.F_meas_sigma_au', '_refln.F_meas', '_au'), ('_refln.intensity_sigma', '_refln.intensity', ''), ('_refln.pdbx_I_plus_sigma', '_refln.pdbx_I_plus', '')]
for label in lstkeys:
for m in sigma_matches:
FIlabel = m[1] + m[2]
if FIlabel == label:
labelpairs.append(
[label, m[0],
self.guess_observationtype(label)])
remainingkeys.remove(label)
remainingkeys.remove(m[0])
alllabels = " ".join(remainingkeys)
# catch generic meas and sigma labels
anymeas_matches = re.findall(r"((\S*)_meas(\S*))",
alllabels) + re.findall(
r"((\S*)_calc(\S*))", alllabels)
anysigma_matches = re.findall(r"((\S*)_sigma(\S*))", alllabels)
for mmatch in anymeas_matches:
for smatch in anysigma_matches:
if mmatch[1] == smatch[1] and mmatch[2] == smatch[2]:
remainingkeys.remove(mmatch[0])
if smatch[
0] in remainingkeys: # in case of say F_squared_calc, F_squared_meas, F_squared_sigma all being present
remainingkeys.remove(smatch[0])
labelpairs.append([
mmatch[0], smatch[0],
self.guess_observationtype(mmatch[0])
])
else:
labelpairs.append([
mmatch[0], None,
self.guess_observationtype(mmatch[0])
])
return labelpairs, remainingkeys
def get_miller_indices_containing_loops(self):
loops = []
for loop in self.cif_block.loops.values():
for key in loop.keys():
if 'index_h' not in key: continue
hkl_str = [
loop.get(key.replace('index_h', 'index_%s' % i))
for i in 'hkl'
]
if hkl_str.count(None) > 0:
raise CifBuilderError(
"Miller indices missing from current CIF block (%s)" %
key.replace('index_h',
'index_%s' % 'hkl'[hkl_str.index(None)]))
hkl_int = []
for i, h_str in enumerate(hkl_str):
try:
h_int = flex.int(h_str)
except ValueError as e:
raise CifBuilderError(
"Invalid item for Miller index %s: %s" %
("HKL"[i], str(e)))
hkl_int.append(h_int)
indices = flex.miller_index(*hkl_int)
loops.append((indices, loop))
break
return loops
def get_selection(self,
value,
wavelength_id=None,
crystal_id=None,
scale_group_code=None):
selection = ~((value == '.') | (value == '?'))
if self.wavelength_id_array is not None and wavelength_id is not None:
selection &= (self.wavelength_id_array.data() == wavelength_id)
if self.crystal_id_array is not None and crystal_id is not None:
selection &= (self.crystal_id_array.data() == crystal_id)
if self.scale_group_array is not None and scale_group_code is not None:
selection &= (self.scale_group_array.data() == scale_group_code)
return selection
def flex_std_string_as_miller_array(self,
value,
wavelength_id=None,
crystal_id=None,
scale_group_code=None):
# Create a miller_array object of only the data and indices matching the
# wavelength_id, crystal_id and scale_group_code submitted or full array if these are None
selection = self.get_selection(value,
wavelength_id=wavelength_id,
crystal_id=crystal_id,
scale_group_code=scale_group_code)
data = value.select(selection)
#if not isinstance(data, flex.double):
try:
data = flex.int(data)
indices = self.indices.select(selection)
except ValueError:
try:
data = flex.double(data)
indices = self.indices.select(selection)
except ValueError:
# if flex.std_string return all values including '.' and '?'
data = value
indices = self.indices
if data.size() == 0: return None
return miller.array(
miller.set(self.crystal_symmetry, indices).auto_anomalous(), data)
def arrays(self):
return self._arrays
def origarrays(self):
"""
return dictionary of raw data found in cif file cast into flex.double arrays
or just string arrays as a fall back.
"""
return self._origarrays