def sff_filter(in_file, out_file, iterator_filter, inter):
count = 0
try:
from Bio.SeqIO.SffIO import ReadRocheXmlManifest
except ImportError:
# Prior to Biopython 1.56 this was a private function
from Bio.SeqIO.SffIO import _sff_read_roche_index_xml as ReadRocheXmlManifest
with open(in_file, "rb") as in_handle:
try:
manifest = ReadRocheXmlManifest(in_handle)
except ValueError:
manifest = None
in_handle.seek(0)
with open(out_file, "wb") as out_handle:
writer = SffWriter(out_handle, xml=manifest)
in_handle.seek(0) # start again after getting manifest
if inter:
from itertools import chain
count = writer.write_file(
chain.from_iterable(
iterator_filter(pair(SffIterator(in_handle)))))
assert count % 2 == 0, "Odd number of records? %i" % count
count /= 2
else:
count = writer.write_file(
iterator_filter(SffIterator(in_handle)))
return count
def sff_filter(in_file, pos_file, neg_file, wanted):
"""SFF filter."""
try:
from Bio.SeqIO.SffIO import SffIterator, SffWriter
except ImportError:
sys.exit("SFF filtering requires Biopython 1.54 or later")
try:
from Bio.SeqIO.SffIO import ReadRocheXmlManifest
except ImportError:
# Prior to Biopython 1.56 this was a private function
from Bio.SeqIO.SffIO import _sff_read_roche_index_xml as ReadRocheXmlManifest
in_handle = open(in_file, "rb") # must be binary mode!
try:
manifest = ReadRocheXmlManifest(in_handle)
except ValueError:
manifest = None
# This makes two passes though the SFF file with isn't so efficient,
# but this makes the code simple.
pos_count = neg_count = 0
if pos_file is not None:
out_handle = open(pos_file, "wb")
writer = SffWriter(out_handle, xml=manifest)
in_handle.seek(0) # start again after getting manifest
pos_count = writer.write_file(rec for rec in SffIterator(in_handle)
if clean_name(rec.id) in wanted)
out_handle.close()
if neg_file is not None:
out_handle = open(neg_file, "wb")
writer = SffWriter(out_handle, xml=manifest)
in_handle.seek(0) # start again
neg_count = writer.write_file(rec for rec in SffIterator(in_handle)
if clean_name(rec.id) not in wanted)
out_handle.close()
# And we're done
in_handle.close()
# At the time of writing, Galaxy doesn't show SFF file read counts,
# so it is useful to put them in stdout and thus shown in job info.
return pos_count, neg_count
def test_write(self):
filename = "Roche/E3MFGYR02_random_10_reads.sff"
with open(filename, "rb") as handle:
metadata = ReadRocheXmlManifest(handle)
with open(filename, "rb") as handle:
sff = list(SffIterator(handle))
b_handle = BytesIO()
w = SffWriter(b_handle, xml=metadata)
w.write_file(sff) # list
data = b_handle.getvalue()
# And again with an iterator...
handle = BytesIO()
w = SffWriter(handle, xml=metadata)
w.write_file(iter(sff))
self.assertEqual(data, handle.getvalue())
# Check 100% identical to the original:
with open(filename, "rb") as handle:
original = handle.read()
self.assertEqual(len(data), len(original))
self.assertEqual(data, original)
del data
def run(self, proc_name=None):
sffpath = self.id_str + '.sff'
try:
with open(sffpath, 'wb') as fh:
self.proc_name = proc_name
self.sff_file = SffWriter(fh)
self.sff_file.write_file(self.reads_for_barcode(
self.reads_sff))
logger.info(
"%s reads of %s matched %s" %
(self._matched_reads, self._processed, self.id_str))
except ValueError:
# No reads for barcode so remove the temporary file
os.unlink(sffpath)
def test_no_index(self):
# Does a lot of work to create a no-index SFF file
# (in the process checking this bit of SffWriter works)
records = list(SeqIO.parse(BytesIO(self.good), "sff"))
with BytesIO() as handle:
writer = SffWriter(handle, index=False)
count = writer.write_file(records)
self.assertEqual(count, len(records))
handle.seek(0)
new = list(SeqIO.parse(handle, "sff"))
self.assertEqual(len(records), len(new))
for a, b in zip(records, new):
self.assertEqual(a.id, b.id)
handle.seek(0)
with self.assertRaises(ValueError) as cm:
values = _sff_find_roche_index(handle)
err = str(cm.exception)
self.assertEqual(err, "No index present in this SFF file")
def sff_filter(in_file, out_file, iterator_filter):
count = 0
try:
from Bio.SeqIO.SffIO import SffIterator, SffWriter
except ImportError:
stop_err("SFF filtering requires Biopython 1.54 or later")
try:
from Bio.SeqIO.SffIO import ReadRocheXmlManifest
except ImportError:
#Prior to Biopython 1.56 this was a private function
from Bio.SeqIO.SffIO import _sff_read_roche_index_xml as ReadRocheXmlManifest
with open(in_file, "rb") as in_handle:
try:
manifest = ReadRocheXmlManifest(in_handle)
except ValueError:
manifest = None
in_handle.seek(0)
with open(out_file, "wb") as out_handle:
writer = SffWriter(out_handle, xml=manifest)
in_handle.seek(0) #start again after getting manifest
count = writer.write_file(iterator_filter(SffIterator(in_handle)))
#count = writer.write_file(SffIterator(in_handle))
return count
short_clipped += 1
elif keep_negatives:
if len(seq) >= min_len:
negs += 1
yield record
else:
short_neg += 1
in_handle = open(in_file, "rb")
try:
manifest = ReadRocheXmlManifest(in_handle)
except ValueError:
manifest = None
in_handle.seek(0)
out_handle = open(out_file, "wb")
writer = SffWriter(out_handle, xml=manifest)
writer.write_file(process(SffIterator(in_handle)))
#End of SFF code
elif seq_format.lower().startswith("fastq"):
in_handle = open(in_file, "rU")
out_handle = open(out_file, "w")
reader = fastqReader(in_handle)
writer = fastqWriter(out_handle)
if forward:
for record in reader:
seq = record.sequence.upper()
result = primer.search(seq)
if result:
#Forward primer, take everything after it
cut = result.end()
record.sequence = seq[cut:]
padding = len(index) % 8
if padding:
padding = 8 - padding
index += chr(0) * padding
assert len(index) % 8 == 0
# Ugly bit of code to make a fake index at start
index = ".diy1.00This is a fake index block (DIY = Do It Yourself), which is allowed under the SFF standard.\0"
padding = len(index) % 8
if padding:
padding = 8 - padding
index += chr(0) * padding
with open("Roche/E3MFGYR02_random_10_reads.sff", "rb") as handle:
records = list(SffIterator(handle))
with open("Roche/E3MFGYR02_alt_index_at_start.sff", "w") as out_handle:
w = SffWriter(out_handle, index=False, xml=None)
# Fake the header...
w._number_of_reads = len(records)
w._index_start = 0
w._index_length = 0
w._key_sequence = records[0].annotations["flow_key"]
w._flow_chars = records[0].annotations["flow_chars"]
w._number_of_flows_per_read = len(w._flow_chars)
w.write_header()
w._index_start = out_handle.tell()
w._index_length = len(index)
out_handle.seek(0)
w.write_header() # this time with index info
w.handle.write(index)
for record in records:
w.write_record(record)
