def _parse_phylip_raw(fh):
"""Raw parser for PHYLIP files.
Returns a list of raw (seq, id) values. It is the responsibility of the
caller to construct the correct in-memory object to hold the data.
"""
# Note: this returns the full data instead of yielding each sequence,
# because the header specifies the number of sequences, so the file cannot
# be validated until it's read completely.
# File should have a single header on the first line.
try:
header = next(_line_generator(fh))
except StopIteration:
raise PhylipFormatError("This file is empty.")
n_seqs, seq_len = _validate_header(header)
# All following lines should be ID+sequence. No blank lines are allowed.
data = []
for line in _line_generator(fh):
data.append(_validate_line(line, seq_len))
if len(data) != n_seqs:
raise PhylipFormatError("The number of sequences is not %s " % n_seqs +
"as specified in the header.")
return data
def _tabular_msa_to_phylip(obj, fh):
sequence_count = obj.shape.sequence
if sequence_count < 1:
raise PhylipFormatError(
"TabularMSA can only be written in PHYLIP format if there is at "
"least one sequence in the alignment.")
sequence_length = obj.shape.position
if sequence_length < 1:
raise PhylipFormatError(
"TabularMSA can only be written in PHYLIP format if there is at "
"least one position in the alignment.")
chunk_size = 10
labels = [str(label) for label in obj.index]
for label in labels:
if len(label) > chunk_size:
raise PhylipFormatError(
"``TabularMSA`` can only be written in PHYLIP format if all "
"sequence index labels have %d or fewer characters. Found "
"sequence with index label '%s' that exceeds this limit. Use "
"``TabularMSA.reassign_index`` to assign shorter index labels."
% (chunk_size, label))
fh.write('{0:d} {1:d}\n'.format(sequence_count, sequence_length))
fmt = '{0:%d}{1}\n' % chunk_size
for label, seq in zip(labels, obj):
chunked_seq = chunk_str(str(seq), chunk_size, ' ')
fh.write(fmt.format(label, chunked_seq))
def _alignment_to_phylip(obj, fh):
if obj.is_empty():
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if there is at "
"least one sequence in the alignment.")
sequence_length = obj.sequence_length()
if sequence_length == 0:
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if there is at "
"least one position in the alignment.")
chunk_size = 10
for id_ in obj.ids():
if len(id_) > chunk_size:
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if all "
"sequence IDs have %d or fewer characters. Found sequence "
"with ID '%s' that exceeds this limit. Use "
"Alignment.update_ids to assign shorter IDs." %
(chunk_size, id_))
sequence_count = obj.sequence_count()
fh.write('{0:d} {1:d}\n'.format(sequence_count, sequence_length))
fmt = '{0:%d}{1}\n' % chunk_size
for seq in obj:
chunked_seq = chunk_str(str(seq), chunk_size, ' ')
fh.write(fmt.format(seq.metadata['id'], chunked_seq))
def _validate_line(line, seq_len):
if not line:
raise PhylipFormatError("Empty lines are not allowed.")
ID = line[:10].strip()
seq = line[10:].replace(' ', '')
if len(seq) != seq_len:
raise PhylipFormatError(
"The length of sequence %s is not %s as specified in the header." %
(ID, seq_len))
return (seq, ID)
def _validate_header(header):
header_vals = header.split()
try:
n_seqs, seq_len = [int(x) for x in header_vals]
if n_seqs < 1 or seq_len < 1:
raise PhylipFormatError(
'The number of sequences and the length must be positive.')
except ValueError:
raise PhylipFormatError(
'Found non-header line when attempting to read the 1st record '
'(header line should have two space-separated integers): '
'"%s"' % header)
return n_seqs, seq_len
def _alignment_to_phylip(obj, fh):
if not obj.is_valid():
# TODO update this error message when #670 is resolved
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if all sequences "
"are of equal length and contain only valid characters within "
"their character sets.")
if obj.is_empty():
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if there is at "
"least one sequence in the alignment.")
sequence_length = obj.sequence_length()
if sequence_length == 0:
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if there is at "
"least one position in the alignment.")
chunk_size = 10
for id_ in obj.ids():
if len(id_) > chunk_size:
raise PhylipFormatError(
"Alignment can only be written in PHYLIP format if all "
"sequence IDs have %d or fewer characters. Found sequence "
"with ID '%s' that exceeds this limit. Use "
"Alignment.update_ids to assign shorter IDs." %
(chunk_size, id_))
sequence_count = obj.sequence_count()
fh.write('{0:d} {1:d}\n'.format(sequence_count, sequence_length))
fmt = '{0:%d}{1}\n' % chunk_size
for seq in obj:
chunked_seq = _chunk_str(str(seq), chunk_size)
fh.write(fmt.format(seq.id, chunked_seq))