def _stockholm_to_tabular_msa(fh, constructor=None):
# Checks that user has passed required constructor parameter
if constructor is None:
raise ValueError("Must provide `constructor` parameter indicating the "
"type of sequences in the alignment. `constructor` "
"must be a subclass of `GrammaredSequence` "
"(e.g., `DNA`, `RNA`, `Protein`).")
# Checks that contructor parameter is supported
elif not issubclass(constructor, GrammaredSequence):
raise TypeError("`constructor` must be a subclass of "
"`GrammaredSequence`.")
# Checks that the file isn't empty
try:
line = next(fh)
except StopIteration:
raise StockholmFormatError("File is empty.")
# Checks that the file follows basic format (includes the required header)
if not _is_header(line):
raise StockholmFormatError("File missing required Stockholm header "
"line.")
msa_data = _MSAData()
for line in fh:
if line.isspace():
continue
line = line.rstrip('\n')
if _is_sequence_line(line):
seq_name, seq_data = _parse_sequence_line(line)
msa_data.add_sequence(seq_name, seq_data)
elif line.startswith("#=GF"):
feature_name, feature_data = _parse_gf_line(line)
msa_data.add_gf_metadata(feature_name, feature_data)
elif line.startswith("#=GS"):
seq_name, feature_name, feature_data = _parse_gs_line(line)
msa_data.add_gs_metadata(seq_name, feature_name, feature_data)
elif line.startswith("#=GR"):
seq_name, feature_name, feature_data = _parse_gr_line(line)
msa_data.add_gr_metadata(seq_name, feature_name, feature_data)
elif line.startswith('#=GC'):
feature_name, feature_data = _parse_gc_line(line)
msa_data.add_gc_metadata(feature_name, feature_data)
elif _is_footer(line):
break
else:
raise StockholmFormatError("Unrecognized line: %r" % line)
if not _is_footer(line):
raise StockholmFormatError('Final line does not conform to Stockholm '
'format. Must contain only "//".')
return msa_data.build_tabular_msa(constructor)
def _format_positional_metadata(df, data_type):
# Asserts positional metadata feature names are unique
if not df.columns.is_unique:
num_repeated_columns = len(df.columns) - len(set(df.columns))
raise StockholmFormatError('%s feature names must be unique. '
'Found %d duplicate names.'
% (data_type, num_repeated_columns))
str_df = df.astype(str)
# Asserts positional metadata dataframe items are one character long
for column in str_df.columns:
if (str_df[column].str.len() != 1).any():
raise StockholmFormatError("%s must contain a single character for"
" each position's value. Found value(s)"
" in column %s of incorrect length."
% (data_type, column))
return str_df
def add_gf_metadata(self, feature_name, feature_data):
# Handles first instance of labelled tree
if feature_name == 'TN' and 'NH' not in self._metadata:
self._metadata['NH'] = OrderedDict()
self._metadata['NH'][feature_data] = ''
# Handles second instance of labelled tree
elif feature_name == 'TN' and 'NH' in self._metadata:
if feature_data in self._metadata['NH']:
raise StockholmFormatError("Tree name %r used multiple times "
"in file." % feature_data)
self._metadata['NH'][feature_data] = ''
# Handles extra line(s) of an already created tree
elif feature_name == 'NH' and feature_name in self._metadata:
trees = self._metadata[feature_name]
if isinstance(trees, OrderedDict):
tree_id = next(reversed(trees))
self._metadata[feature_name][tree_id] = (trees[tree_id] +
feature_data)
else:
self._metadata[feature_name] = (self._metadata[feature_name] +
feature_data)
elif feature_name == 'RN':
if feature_name not in self._metadata:
self._metadata[feature_name] = [OrderedDict()]
else:
self._metadata[feature_name].append(OrderedDict())
elif feature_name in _REFERENCE_TAGS:
if 'RN' not in self._metadata:
raise StockholmFormatError("Expected 'RN' tag to precede "
"'%s' tag." % feature_name)
reference_dict = self._metadata['RN'][-1]
if feature_name not in reference_dict:
reference_dict[feature_name] = feature_data
else:
padding = _get_padding(reference_dict[feature_name])
reference_dict[feature_name] += padding + feature_data
elif feature_name in self._metadata:
padding = _get_padding(self._metadata[feature_name][-1])
self._metadata[feature_name] = (self._metadata[feature_name] +
padding + feature_data)
else:
self._metadata[feature_name] = feature_data
def _tabular_msa_to_stockholm(obj, fh):
if not obj.index.is_unique:
raise StockholmFormatError("The TabularMSA's index labels must be"
" unique.")
# Writes header
fh.write("# STOCKHOLM 1.0\n")
# Writes GF data to file
if obj.has_metadata():
for gf_feature, gf_feature_data in viewitems(obj.metadata):
if gf_feature == 'NH' and isinstance(gf_feature_data, dict):
for tree_id, tree in viewitems(obj.metadata[gf_feature]):
fh.write("#=GF TN %s\n" % tree_id)
fh.write("#=GF NH %s\n" % tree)
else:
fh.write("#=GF %s %s\n" % (gf_feature, gf_feature_data))
unpadded_data = []
# Writes GS data to file, retrieves GR data, and retrieves sequence data
for seq, seq_name in zip(obj, obj.index):
seq_name = str(seq_name)
if seq.has_metadata():
for gs_feature, gs_feature_data in viewitems(seq.metadata):
fh.write("#=GS %s %s %s\n" %
(seq_name, gs_feature, gs_feature_data))
unpadded_data.append((seq_name, str(seq)))
if seq.has_positional_metadata():
df = _format_positional_metadata(
seq.positional_metadata, 'Sequence-specific positional '
'metadata (GR)')
for gr_feature in df.columns:
gr_feature_data = ''.join(df[gr_feature])
gr_string = "#=GR %s %s" % (seq_name, gr_feature)
unpadded_data.append((gr_string, gr_feature_data))
# Retrieves GC data
if obj.has_positional_metadata():
df = _format_positional_metadata(
obj.positional_metadata, 'Multiple sequence alignment '
'positional metadata (GC)')
for gc_feature in df.columns:
gc_feature_data = ''.join(df[gc_feature])
gc_string = "#=GC %s" % gc_feature
unpadded_data.append((gc_string, gc_feature_data))
# Writes GR, GC, and raw data to file with padding
_write_padded_data(unpadded_data, fh)
# Writes footer
fh.write("//\n")
def add_gf_metadata(self, feature_name, feature_data):
# Handles first instance of labelled tree
if feature_name == 'TN' and 'NH' not in self._metadata:
self._metadata['NH'] = OrderedDict()
self._metadata['NH'][feature_data] = ''
# Handles second instance of labelled tree
elif feature_name == 'TN' and 'NH' in self._metadata:
if feature_data in self._metadata['NH']:
raise StockholmFormatError("Tree name %r used multiple times "
"in file." % feature_data)
self._metadata['NH'][feature_data] = ''
# Handles extra line(s) of an already created tree
elif feature_name == 'NH' and feature_name in self._metadata:
trees = self._metadata[feature_name]
tree_id = list(trees.keys())[-1]
self._metadata[feature_name][tree_id] = (trees[tree_id] +
feature_data)
elif feature_name in self._metadata:
self._metadata[feature_name] = (self._metadata[feature_name] +
feature_data)
else:
self._metadata[feature_name] = feature_data
def build_tabular_msa(self, constructor):
if len(self._seqs) != len(self._seq_order):
invalid_seq_names = set(self._seqs) - set(self._seq_order)
raise StockholmFormatError('Found GS or GR metadata for '
'nonexistent sequence(s): %r'
% invalid_seq_names)
seqs = []
for seq_name in self._seq_order:
seqs.append(self._seqs[seq_name].build_sequence(constructor))
positional_metadata = self._positional_metadata
if not positional_metadata:
positional_metadata = None
metadata = self._metadata
if not metadata:
metadata = None
# Constructs TabularMSA
return TabularMSA(seqs, metadata=metadata,
positional_metadata=positional_metadata,
index=self._seq_order)
def _check_for_malformed_line(line, expected_len):
if len(line) != expected_len:
raise StockholmFormatError('Line contains %d item(s). It must '
'contain exactly %d item(s).'
% (len(line), expected_len))
def _raise_duplicate_error(message):
raise StockholmFormatError(message+' Note: If the file being used is in '
'Stockholm interleaved format, this '
'is not supported by the reader.')
def _tabular_msa_to_stockholm(obj, fh):
if not obj.index.is_unique:
raise StockholmFormatError("The TabularMSA's index labels must be"
" unique.")
# Writes header
fh.write("# STOCKHOLM 1.0\n")
# Writes GF data to file
if obj.has_metadata():
for gf_feature, gf_feature_data in obj.metadata.items():
if gf_feature == 'NH' and isinstance(gf_feature_data, dict):
for tree_id, tree in gf_feature_data.items():
fh.write("#=GF TN %s\n" % tree_id)
fh.write("#=GF NH %s\n" % tree)
elif gf_feature == 'RN':
if not isinstance(gf_feature_data, list):
raise StockholmFormatError(
"Expected 'RN' to contain a list of reference "
"dictionaries, got %r." % gf_feature_data)
for ref_num, dictionary in enumerate(gf_feature_data, start=1):
if not isinstance(dictionary, dict):
raise StockholmFormatError(
"Expected reference information to be stored as a "
"dictionary, found reference %d stored as %r." %
(ref_num, type(dictionary).__name__))
fh.write("#=GF RN [%d]\n" % ref_num)
for feature in dictionary:
if feature not in _REFERENCE_TAGS:
formatted_reference_tags = ', '.join(
[tag for tag in _REFERENCE_TAGS])
raise StockholmFormatError(
"Invalid reference tag %r found in reference "
"dictionary %d. Valid reference tags are: %s."
% (feature, ref_num, formatted_reference_tags))
fh.write("#=GF %s %s\n" %
(feature, dictionary[feature]))
else:
fh.write("#=GF %s %s\n" % (gf_feature, gf_feature_data))
unpadded_data = []
# Writes GS data to file, retrieves GR data, and retrieves sequence data
for seq, seq_name in zip(obj, obj.index):
seq_name = str(seq_name)
if seq.has_metadata():
for gs_feature, gs_feature_data in seq.metadata.items():
fh.write("#=GS %s %s %s\n" %
(seq_name, gs_feature, gs_feature_data))
unpadded_data.append((seq_name, str(seq)))
if seq.has_positional_metadata():
df = _format_positional_metadata(
seq.positional_metadata, 'Sequence-specific positional '
'metadata (GR)')
for gr_feature in df.columns:
gr_feature_data = ''.join(df[gr_feature])
gr_string = "#=GR %s %s" % (seq_name, gr_feature)
unpadded_data.append((gr_string, gr_feature_data))
# Retrieves GC data
if obj.has_positional_metadata():
df = _format_positional_metadata(
obj.positional_metadata, 'Multiple sequence alignment '
'positional metadata (GC)')
for gc_feature in df.columns:
gc_feature_data = ''.join(df[gc_feature])
gc_string = "#=GC %s" % gc_feature
unpadded_data.append((gc_string, gc_feature_data))
# Writes GR, GC, and raw data to file with padding
_write_padded_data(unpadded_data, fh)
# Writes footer
fh.write("//\n")
