-
Notifications
You must be signed in to change notification settings - Fork 327
/
make_dataset_txt.py
125 lines (107 loc) · 4.34 KB
/
make_dataset_txt.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
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
from collections import defaultdict
from itertools import chain
from os.path import join, split, exists
import numpy as np
import os
import pandas as pd
from deep_utils import DirUtils
from argparse import ArgumentParser
from joblib import Parallel, delayed
from sklearn.model_selection import train_test_split
from tqdm import tqdm
parser = ArgumentParser()
parser.add_argument("--split", action="store_true")
parser.add_argument("--name", default="datasets", type=str)
parser.add_argument("--n_jobs", default=10, type=int)
parser.add_argument("--data", default=".npz", type=str)
parser.add_argument("--train", action="store_true")
parser.add_argument("--nnunet",
default="/media/aicvi/11111bdb-a0c7-4342-9791-36af7eb70fc0/NNUNET_OUTPUT/nnunet_preprocessed/")
args = parser.parse_args()
seed = 1234
def chain(lst: list[list]):
out = []
for l in lst:
out.extend(l)
return out
def npz_csv():
datasets_config = {
# 'CT_CORONARY': {
# 'data_dir': f'{args.nnunet}/Dataset002_china_narco/nnUNetPlans_2d',
# 'num_classes': 3 + 1, # plus background
# 'predict_head': 1
# },
'MRI_MM': {
'data_dir': f'{args.nnunet}/Dataset001_mm/nnUNetPlans_2d',
'num_classes': 3 + 1, # plus background
'predict_head': 0
},
}
samples = []
columns = ["data_dir", "predict_head", "n_classes"]
for dataset_name, config in datasets_config.items():
data_files = DirUtils.list_dir_full_path(config['data_dir'], interest_extensions=args.data)
split_path = config['data_dir'] + "_split"
if exists(split_path):
data = DirUtils.list_dir_full_path(split_path, return_dict=True, interest_extensions=".npz")
seg_img_samples = dict()
for key, val in tqdm(data.items(), desc="getting data"):
item = key.replace("_seg", "").replace("_img", "")
seg_img_samples[item] = val
file_samples = defaultdict(list)
for key, val in tqdm(seg_img_samples.items(), desc="Getting final data"):
item = "_".join(k for k in key.split("_")[:-1])
file_samples[item].append(val)
else:
file_samples = []
if args.split:
split_path = DirUtils.split_extension(config['data_dir'], suffix="_split")
os.makedirs(split_path, exist_ok=True)
else:
split_path = None
print("Getting ready for the data splitting!")
samples_ = Parallel(n_jobs=args.n_jobs)(
delayed(process_file)(config, split_path, filepath, file_samples) for filepath in tqdm(data_files))
samples.extend(samples_)
train, val = train_test_split(samples)
csv_file_path = f'./lists/{args.name}/'
train = chain(train)
val = chain(val)
os.makedirs(os.path.dirname(csv_file_path), exist_ok=True)
pd.DataFrame(train, columns=columns).to_csv(csv_file_path + "/train.txt", index=False)
pd.DataFrame(val, columns=columns).to_csv(csv_file_path + "/val.txt", index=False)
def process_file(config, split_path, filepath, file_samples):
filename = split(filepath)[-1].replace(".npz", "")
if split_path and filename not in file_samples:
# print(filename)
samples = []
file_data = np.load(filepath)
img = file_data['data']
seg = file_data['seg']
for z_index in range(img.shape[1]):
img_ = img[:, z_index, ...]
seg_ = seg[:, z_index, ...]
img_path = join(split_path,
f"{DirUtils.split_extension(split(filepath)[-1], suffix=f'_{z_index:04}')}")
# seg_path = join(split_path,
# f"{DirUtils.split_extension(split(filepath)[-1], suffix=f'_{z_index:04}_seg')}")
if not exists(img_path):
seg_ = seg_.squeeze(0)
seg_[seg_ < 0] = 0
np.savez(img_path, image=img_.squeeze(0), label=seg_)
samples.append(
[img_path,
config['predict_head'],
config['num_classes'],
]
)
# np.savez(seg_path, seg_)
else:
samples = [[
filepath,
config['predict_head'],
config['num_classes'],
]]
return samples
if __name__ == '__main__':
npz_csv()