def run(self, inputs, limit_to_record, output_gbk, output_tsv,
prodigal_meta_mode, protein):
first_output = output_gbk or output_tsv
if not first_output:
raise ValueError(
'Specify at least one of --output-gbk or --output-tsv')
tmp_dir_path = first_output + '.tmp'
logging.debug('Using TMP dir: %s', tmp_dir_path)
if not os.path.exists(tmp_dir_path):
os.mkdir(tmp_dir_path)
prepare_step = DeepBGCAnnotator(tmp_dir_path=tmp_dir_path,
prodigal_meta_mode=prodigal_meta_mode)
writers = []
if output_gbk:
writers.append(GenbankWriter(out_path=output_gbk))
if output_tsv:
writers.append(PfamTSVWriter(out_path=output_tsv))
num_records = 0
for i, input_path in enumerate(inputs):
logging.info('Processing input file %s/%s: %s', i + 1, len(inputs),
input_path)
with util.SequenceParser(input_path, protein=protein) as parser:
for record in parser.parse():
if limit_to_record and record.id not in limit_to_record:
logging.debug(
'Skipping record %s not matching filter %s',
record.id, limit_to_record)
continue
prepare_step.run(record)
for writer in writers:
writer.write(record)
num_records += 1
logging.debug('Removing TMP directory: %s', tmp_dir_path)
shutil.rmtree(tmp_dir_path)
prepare_step.print_summary()
for writer in writers:
writer.close()
logging.info('Saved %s fully annotated records to %s', num_records,
first_output)
def run(self, inputs, output_gbk, output_tsv):
first_output = output_gbk or output_tsv
if not first_output:
raise ValueError(
'Specify at least one of --output-gbk or --output-tsv')
tmp_dir_path = first_output + '.tmp'
logging.debug('Using TMP dir: %s', tmp_dir_path)
if not os.path.exists(tmp_dir_path):
os.mkdir(tmp_dir_path)
prepare_step = DeepBGCAnnotator(tmp_dir_path=tmp_dir_path)
writers = []
if output_gbk:
writers.append(GenbankWriter(out_path=output_gbk))
if output_tsv:
writers.append(PfamTSVWriter(out_path=output_tsv))
num_records = 0
for input_path in inputs:
fmt = util.guess_format(input_path)
if not fmt:
raise NotImplementedError(
"Sequence file type not recognized: {}, ".format(
input_path),
"Please provide a GenBank or FASTA sequence "
"with an appropriate file extension.")
records = SeqIO.parse(input_path, fmt)
for record in records:
prepare_step.run(record)
for writer in writers:
writer.write(record)
num_records += 1
logging.debug('Removing TMP directory: %s', tmp_dir_path)
shutil.rmtree(tmp_dir_path)
prepare_step.print_summary()
for writer in writers:
writer.close()
logging.info('Saved %s fully annotated records to %s', num_records,
first_output)
def run(self, inputs, output, detectors, no_detector, labels, classifiers,
no_classifier, is_minimal_output, limit_to_record, score,
classifier_score, merge_max_protein_gap, merge_max_nucl_gap,
min_nucl, min_proteins, min_domains, min_bio_domains,
prodigal_meta_mode, protein):
if not detectors:
detectors = ['deepbgc']
if not classifiers:
classifiers = ['product_class', 'product_activity']
if not output:
# if not specified, set output path to name of first input file without extension
output, _ = os.path.splitext(
os.path.basename(os.path.normpath(inputs[0])))
if not os.path.exists(output):
os.mkdir(output)
# Save log to LOG.txt file
logger = logging.getLogger('')
logger.addHandler(
logging.FileHandler(os.path.join(output, self.LOG_FILENAME)))
# Define report dir paths
tmp_path = os.path.join(output, self.TMP_DIRNAME)
evaluation_path = os.path.join(output, self.PLOT_DIRNAME)
output_file_name = os.path.basename(os.path.normpath(output))
steps = []
steps.append(
DeepBGCAnnotator(tmp_dir_path=tmp_path,
prodigal_meta_mode=prodigal_meta_mode))
if not no_detector:
if not labels:
labels = [None] * len(detectors)
elif len(labels) != len(detectors):
raise ValueError(
'A separate label should be provided for each of the detectors: {}'
.format(detectors))
for detector, label in zip(detectors, labels):
steps.append(
DeepBGCDetector(
detector=detector,
label=label,
score_threshold=score,
merge_max_protein_gap=merge_max_protein_gap,
merge_max_nucl_gap=merge_max_nucl_gap,
min_nucl=min_nucl,
min_proteins=min_proteins,
min_domains=min_domains,
min_bio_domains=min_bio_domains))
writers = []
writers.append(
GenbankWriter(out_path=os.path.join(output, output_file_name +
'.full.gbk')))
writers.append(
AntismashJSONWriter(
out_path=os.path.join(output, output_file_name +
'.antismash.json')))
is_evaluation = False
if not is_minimal_output:
writers.append(
BGCGenbankWriter(
out_path=os.path.join(output, output_file_name +
'.bgc.gbk')))
writers.append(
ClusterTSVWriter(
out_path=os.path.join(output, output_file_name +
'.bgc.tsv')))
writers.append(
PfamTSVWriter(
out_path=os.path.join(output, output_file_name +
'.pfam.tsv')))
is_evaluation = True
writers.append(
PfamScorePlotWriter(
out_path=os.path.join(evaluation_path, output_file_name +
'.score.png')))
writers.append(
BGCRegionPlotWriter(
out_path=os.path.join(evaluation_path, output_file_name +
'.bgc.png')))
writers.append(
ROCPlotWriter(
out_path=os.path.join(evaluation_path, output_file_name +
'.roc.png')))
writers.append(
PrecisionRecallPlotWriter(
out_path=os.path.join(evaluation_path, output_file_name +
'.pr.png')))
writers.append(
ReadmeWriter(out_path=os.path.join(output, 'README.txt'),
root_path=output,
writers=writers))
if not no_classifier:
for classifier in classifiers:
steps.append(
DeepBGCClassifier(classifier=classifier,
score_threshold=classifier_score))
# Create temp and evaluation dir
if not os.path.exists(tmp_path):
os.mkdir(tmp_path)
if is_evaluation:
if not os.path.exists(evaluation_path):
os.mkdir(evaluation_path)
record_idx = 0
for i, input_path in enumerate(inputs):
logging.info('Processing input file %s/%s: %s', i + 1, len(inputs),
input_path)
with util.SequenceParser(input_path, protein=protein) as parser:
for record in parser.parse():
if limit_to_record and record.id not in limit_to_record:
logging.debug(
'Skipping record %s not matching filter %s',
record.id, limit_to_record)
continue
record_idx += 1
logging.info('=' * 80)
logging.info('Processing record #%s: %s', record_idx,
record.id)
for step in steps:
step.run(record)
logging.info('Saving processed record %s', record.id)
for writer in writers:
writer.write(record)
logging.info('=' * 80)
for step in steps:
step.print_summary()
for writer in writers:
writer.close()
logging.info('=' * 80)
logging.info('Saved DeepBGC result to: {}'.format(output))