-
Notifications
You must be signed in to change notification settings - Fork 0
/
buzzdetect.py
56 lines (46 loc) · 2.8 KB
/
buzzdetect.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
import argparse
import sys
# custom class credit: Steven Bethard
class MyParser(argparse.ArgumentParser):
def error(self, message):
sys.stderr.write('error: %s\n' % message)
self.print_help()
sys.exit(2)
parser = argparse.ArgumentParser(
prog='buzzdetect.py',
description='A program to train and apply machine learning models for bioacoustics',
epilog='github.com/OSU-Bee-Lab/buzzdetect'
)
subparsers = parser.add_subparsers(help='sub-command help', dest='action', required=True)
# analyze
parser_analyze = subparsers.add_parser('analyze', help='analyze all audio files in a directory')
parser_analyze.add_argument('--modelname', help='the name of the directory holding the model data', required=True, type=str)
parser_analyze.add_argument('--cpus', help='the number of parallel processes to create for analysis', required=False, type=int, default=1)
parser_analyze.add_argument('--memory', help='the total memory to allot to analysis, in GB (estimated)', required=False, type=float, default=None)
parser_analyze.add_argument('--dir_raw', help='the directory holding audio files to be analyzed', required=False, default="./audio_in", type=str)
parser_analyze.add_argument('--dir_out', help='the directory to output the analysis results', required=False, default=None, type=str)
parser_analyze.add_argument('--verbosity', help='how detailed the terminal readout should be (integer, 0-2)', required=False, default=1, type=int)
# train
parser_train = subparsers.add_parser('train', help='train a new model')
parser_train.add_argument('--modelname', help='the name for your new model; created as a directory in ./models/', required=True, type=str)
parser_train.add_argument('--metadata', help='the name of the training metadata CSV file in ./training/metadata', required=False, default = 'metadata_raw', type=str)
parser_train.add_argument('--weights', help='the name of the class weights CSV file in ./training/weights', required=False, default=None, type=str)
parser_train.add_argument('--epochs', help='the maximum number of training epochs (early stopping is on)', required=False, default=100, type=int)
args = parser.parse_args()
if args.action == "train":
print(f"training new model {args.modelname}")
from buzzcode.train_model import generate_model
generate_model(modelname=args.modelname, metadata_name=args.metadata, weights_name=args.weights, epochs_in=args.epochs)
elif args.action == "analyze":
print(f"analyzing audio in {args.dir_raw} with model {args.modelname}")
from buzzcode.analyze_directory import analyze_batch
if args.memory is None:
args.memory = args.cpus
analyze_batch(
modelname=args.modelname,
cpus=args.cpus,
memory_allot=args.memory,
dir_raw=args.dir_raw,
dir_out=args.dir_out,
verbosity=args.verbosity,
)