def download_sequence_result(dest_file, group_dir, result, accession, subtree,
update_accs, expanded):
ts = time.time()
temp_file_path = os.path.join(group_dir, str(ts) + 'temp.txt')
download_report(utils.SEQUENCE, result, accession, temp_file_path, subtree)
header = True
with open(temp_file_path) as f:
for line in f:
if header:
header = False
continue
data_accession = line.strip()
write_record = False
if result == utils.SEQUENCE_UPDATE_RESULT:
update_accs.append(data_accession)
write_record = True
elif result == utils.SEQUENCE_RELEASE_RESULT:
if data_accession not in update_accs:
write_record = True
if write_record:
sequenceGet.write_record(dest_file, data_accession,
output_format)
dest_file.flush()
os.remove(temp_file_path)
return update_accs
def download_sequence_set(accession_list, mol_type, assembly_dir, output_format, expanded, quiet):
failed_accessions = []
count = 0
sequence_cnt = len(accession_list)
divisor = utils.get_divisor(sequence_cnt)
if sequence_cnt > 0:
if not quiet:
print 'fetching {0} sequences: {1}'.format(sequence_cnt, mol_type)
target_file_path = os.path.join(assembly_dir, utils.get_filename(mol_type, output_format))
target_file = open(target_file_path, 'w')
for accession in accession_list:
success = sequenceGet.write_record(target_file, accession, output_format, expanded)
if not success:
failed_accessions.append(accession)
else:
count += 1
if count % divisor == 0 and not quiet:
print 'downloaded {0} of {1} sequences'.format(count, sequence_cnt)
if not quiet:
print 'downloaded {0} of {1} sequences'.format(count, sequence_cnt)
target_file.close()
elif not quiet:
print 'no sequences: ' + mol_type
if len(failed_accessions) > 0:
print 'Failed to fetch following {0}, format {1}'.format(mol_type, output_format)
print ','.join(failed_accessions)