type=int,
required=True)
parser.add_argument("--text_count",
help="Text count",
type=int,
required=True)
parser.add_argument("--qpi",
help="Queries per iteration",
type=int,
required=True)
parser.add_argument("--log_file",
help="Whether to save logging into file as well.",
default=None)
parser.add_argument("--preset", help="Some presets for certain systems.")
parser.add_argument("--preset_info",
help="Extra information for a given preset.")
args = parser.parse_args()
# logging
logger = get_logger(args.log_file)
args.logger = logger
if args.preset == "taito":
run_taito(args)
elif args.preset == "taito-timelimit":
run_taito_timelimit(args)
elif args.preset == "kale-timelimit":
run_kale_timelimit(args)
else:
run_normal(args)
help="Minimum length of hits. Default = 0",
default=0,
type=int)
parser.add_argument("--max_length",
help="Maximum length of hits. Default = 100000",
default=100000,
type=int)
parser.add_argument("--split_size", default=0, type=int)
parser.add_argument("--language", help="Language to use.", default="FIN")
parser.add_argument(
"--log_file",
help="Whether to save all logging into a file as well.",
default=None)
args = parser.parse_args()
logger = text_logging.get_logger()
dp = DataPreparer(data_location=args.data_folder,
output_folder=args.output_folder,
language=args.language,
threads=args.threads,
split_size=args.split_size,
logger=logger)
dp.prepare_data()
text_count = dp.get_text_count()
runner = SingleBlastRunner(data=args.data_folder,
output_folder=args.output_folder,
e_value=args.e_value,
word_size=args.word_size,
threads=args.threads,
text_count=text_count,
if __name__ == "__main__":
parser = argparse.ArgumentParser(
description=
"Preparing the data. This should only be run if you are planning on running the software in batches."
)
parser.add_argument("--threads",
help="Number of threads to use",
default=1,
type=int)
parser.add_argument("--data_location",
help="Location of the data files",
required=False)
parser.add_argument(
"--output_folder",
help="A folder where all data will be stored in the end.",
required=True)
parser.add_argument("--language", help="Encoding language", default="FIN")
parser.add_argument(
"--split_size",
type=int,
help="If needed to split the data prior to entering it into the DB",
default=-1)
args = parser.parse_args()
logger = get_logger()
dp = DataPreparer(args.data_location, args.output_folder, args.threads,
args.language, args.split_size, logger)
dp.prepare_data()