def convert_multi_to_single(input_file, read_list, output_folder):
'''
Pull the exact read out of the file.
'''
results = [os.path.basename(input_file)]
try:
with MultiFast5File(input_file, 'r') as multi_f5:
read_ids = set(multi_f5.get_read_ids())
for query_read in read_list:
if query_read in read_ids:
try:
read = multi_f5.get_read(query_read)
output_file = os.path.join(
output_folder, "{}.fast5".format(query_read))
create_single_f5(output_file, read)
results.append(os.path.basename(output_file))
except:
traceback.print_exc()
sys.stderr.write(
"{}\n\tFailed to copy read '{}' from {}\n".format(
"convert_multi_to_single", query_read,
input_file))
else:
sys.stderr.write(
"{}\n\tFailed to find read '{}' in {}\n".format(
"convert_multi_to_single", query_read, input_file))
except:
traceback.print_exc()
sys.stderr.write("{}\n\tFailed to copy files from: {}\n".format(
"convert_multi_to_single", input_file))
finally:
return results
def __copy_reads_to__(self, read_ids, output):
if not os.path.exists(output):
os.makedirs(output)
batch_id_files = [
tuple([id] + re.split('(\.fast5|\.tar)\/', self.index_dict[id]))
for id in read_ids if id in self.index_dict
]
batch_id_files.sort(key=lambda x: (x[1], x[2]) if len(x) > 2 else x[1])
for _, id_batch_paths in itertools.groupby(batch_id_files,
key=lambda x: (x[1], x[2])
if len(x) > 2 else x[1]):
fofns = list(id_batch_paths)
if len(fofns) == 1 and len(fofns[0]) == 2:
# single read fast5
id, src_file = fofns[0]
shutil.copy(
os.path.join(os.path.dirname(args.index), src_file),
output)
else:
_, batch_file, batch_ext, _ = fofns[0]
tarFiles = set([x[3] for x in fofns])
# single read fast5 batch in tar archive
if batch_ext == '.tar':
tar_file = os.path.join(os.path.dirname(self.index_file),
batch_file + batch_ext)
with tarfile.open(tar_file) as fp_tar:
tar_members = fp_tar.getmembers()
for tar_member in tar_members:
if any(s in tar_member.name for s in tarFiles):
try:
tar_member.name = os.path.basename(
tar_member.name)
fp_tar.extract(tar_member, path=output)
except:
RuntimeError(
'[ERROR] Could not extract {id} from {batch}.'
.format(id=tar_member.name,
batch=tar_file))
elif batch_ext == '.fast5':
f5_file = os.path.join(os.path.dirname(self.index_file),
batch_file + batch_ext)
with MultiFast5File(f5_file, 'r') as multi_f5:
target_ids = set([x[0] for x in fofns])
for read_id in multi_f5.get_read_ids():
if read_id in target_ids:
try:
read = multi_f5.get_read(read_id)
output_file = os.path.join(
output, "{}.fast5".format(read_id))
multi_to_single_fast5.create_single_f5(
output_file, read)
except:
RuntimeError(
'[ERROR] Could not extract {id} from {batch}.'
.format(id=read_id, batch=f5_file))
else:
pass