-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_valid_set.py
executable file
·89 lines (68 loc) · 2.61 KB
/
create_valid_set.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
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
import io
import os
import sys
import threading
import bson
from skimage.data import imread
from random import randint
from skimage.io import imsave
NUM_EXAMPLES = 12371293
VALID_SET_SIZE = 100
NUM_THREADS = 16
train_set_prods = set()
valid_set_prods = set()
valid_set_classes = set()
class WorkerThread(threading.Thread):
def __init__(self, args, id):
super(WorkerThread, self).__init__()
self.id = id
self.args = args
self.data = bson.decode_file_iter(open(args.train_bson, 'rb'))
def run(self):
for i, d in enumerate(self.data):
if i % NUM_THREADS == self.id:
if self.id == NUM_THREADS - 1:
sys.stderr.write('{}\r'.format(i))
category_id = int(d['category_id'])
prod_id = int(d['_id'])
if (category_id not in valid_set_classes) or (randint(0, VALID_SET_SIZE) == 1):
path_ = args.valid_dir
valid_set_prods.add((prod_id, category_id))
valid_set_classes.add(category_id)
else:
path_ = args.train_dir
train_set_prods.add((prod_id, category_id))
for j, pic in enumerate(d['imgs']):
pic = imread(io.BytesIO(pic['picture']))
img_name = '{}_{}.jpg'.format(prod_id, j)
imsave(os.path.join(path_, img_name), pic)
def main(args):
if not os.path.exists(args.train_dir):
os.makedirs(args.train_dir)
if not os.path.exists(args.valid_dir):
os.makedirs(args.valid_dir)
threads = []
for i in range(NUM_THREADS):
t = WorkerThread(args, i)
t.start()
threads.append(t)
for t in threads:
t.join()
with open(os.path.join(args.train_dir, 'meta.csv'), 'w') as meta_file:
meta_file.write('_id, category_id\n')
for prod, cls in train_set_prods:
meta_file.write('{},{}\n'.format(prod, cls))
with open(os.path.join(args.valid_dir, 'meta.csv'), 'w') as meta_file:
meta_file.write('_id, category_id\n')
for prod, cls in valid_set_prods:
meta_file.write('{},{}\n'.format(prod, cls))
if __name__ == '__main__':
from argparse import ArgumentParser
parser = ArgumentParser()
# Data handling parameters
parser.add_argument('train_bson', type=str, default=None, help='train bson file')
parser.add_argument('train_dir', type=str, default=None, help='train dir')
parser.add_argument('valid_dir', type=str, default=None, help='valid dir')
args = parser.parse_args()
main(args)
exit(0)