def _parse_fasta_raw(fh, data_parser, format_label):
"""Raw parser for FASTA or QUAL files.
Returns raw values (seq/qual, id, description). It is the responsibility of
the caller to construct the correct in-memory object to hold the data.
"""
line = next(fh)
# header check inlined here and below for performance
if line.startswith('>'):
id_, desc = _parse_fasta_like_header(line)
else:
raise FASTAFormatError(
"Found line without a header in %s-formatted file:\n%s" %
(format_label, line))
data_chunks = []
for line in fh:
if line.startswith('>'):
# new header, so yield current record and reset state
yield data_parser(data_chunks), id_, desc
data_chunks = []
id_, desc = _parse_fasta_like_header(line)
else:
line = line.strip()
if line:
data_chunks.append(line)
else:
raise FASTAFormatError(
"Found blank or whitespace-only line in %s-formatted "
"file." % format_label)
# yield last record in file
yield data_parser(data_chunks), id_, desc
def _fasta_to_generator(fh, qual=FileSentinel, constructor=BiologicalSequence):
if qual is None:
for seq, id_, desc in _parse_fasta_raw(fh, _parse_sequence_data,
'FASTA'):
yield constructor(seq, id=id_, description=desc)
else:
fasta_gen = _parse_fasta_raw(fh, _parse_sequence_data, 'FASTA')
qual_gen = _parse_fasta_raw(qual, _parse_quality_scores, 'QUAL')
for fasta_rec, qual_rec in zip_longest(fasta_gen, qual_gen,
fillvalue=None):
if fasta_rec is None:
raise FASTAFormatError(
"QUAL file has more records than FASTA file.")
if qual_rec is None:
raise FASTAFormatError(
"FASTA file has more records than QUAL file.")
fasta_seq, fasta_id, fasta_desc = fasta_rec
qual_scores, qual_id, qual_desc = qual_rec
if fasta_id != qual_id:
raise FASTAFormatError(
"IDs do not match between FASTA and QUAL records: %r != %r"
% (fasta_id, qual_id))
if fasta_desc != qual_desc:
raise FASTAFormatError(
"Descriptions do not match between FASTA and QUAL "
"records: %r != %r" % (fasta_desc, qual_desc))
# sequence and quality scores lengths are checked in constructor
yield constructor(fasta_seq, id=fasta_id, description=fasta_desc,
quality=qual_scores)
def _parse_quality_scores(chunks):
if not chunks:
raise FASTAFormatError("Found QUAL header without quality scores.")
qual_str = ' '.join(chunks)
try:
return np.asarray(qual_str.split(), dtype=int)
except ValueError:
raise FASTAFormatError(
"Could not convert quality scores to integers:\n%s" % qual_str)
def _fasta_to_generator(fh, qual=FileSentinel, constructor=Sequence, **kwargs):
if qual is None:
for seq, id_, desc in _parse_fasta_raw(fh, _parse_sequence_data,
FASTAFormatError):
yield constructor(seq,
metadata={
'id': id_,
'description': desc
},
**kwargs)
else:
fasta_gen = _parse_fasta_raw(fh, _parse_sequence_data,
FASTAFormatError)
qual_gen = _parse_fasta_raw(qual, _parse_quality_scores,
QUALFormatError)
for fasta_rec, qual_rec in zip_longest(fasta_gen,
qual_gen,
fillvalue=None):
if fasta_rec is None:
raise FASTAFormatError(
"QUAL file has more records than FASTA file.")
if qual_rec is None:
raise FASTAFormatError(
"FASTA file has more records than QUAL file.")
fasta_seq, fasta_id, fasta_desc = fasta_rec
qual_scores, qual_id, qual_desc = qual_rec
if fasta_id != qual_id:
raise FASTAFormatError(
"IDs do not match between FASTA and QUAL records: %r != %r"
% (fasta_id, qual_id))
if fasta_desc != qual_desc:
raise FASTAFormatError(
"Descriptions do not match between FASTA and QUAL "
"records: %r != %r" % (fasta_desc, qual_desc))
# sequence and quality scores lengths are checked in constructor
yield constructor(fasta_seq,
metadata={
'id': fasta_id,
'description': fasta_desc
},
positional_metadata={'quality': qual_scores},
**kwargs)
def _sniffer_data_parser(chunks):
data = _parse_sequence_data(chunks)
try:
_parse_quality_scores(chunks)
except QUALFormatError:
return data
else:
# used for flow control within sniffer, user should never see this
# message
raise FASTAFormatError('Data appear to be quality scores.')
def _parse_sequence_data(chunks):
if not chunks:
raise FASTAFormatError("Found header without sequence data.")
return ''.join(chunks)
