-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_splits_seq.py
executable file
·153 lines (131 loc) · 7.34 KB
/
create_splits_seq.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
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
import pdb
import os
import pandas as pd
from datasets.dataset_generic import Generic_WSI_Classification_Dataset, Generic_MIL_Dataset, save_splits
import argparse
import numpy as np
parser = argparse.ArgumentParser(description='Creating splits for whole slide classification')
parser.add_argument('--label_frac', type=float, default= 1.0,
help='fraction of labels (default: 1)')
parser.add_argument('--seed', type=int, default=1,
help='random seed (default: 1)')
parser.add_argument('--unique_tests', default=False, action='store_true',help='ensure all test sets are unique')
parser.add_argument('--k', type=int, default=10,
help='number of splits (default: 10)')
parser.add_argument('--task', type=str, choices=['task_1_tumor_vs_normal', 'task_2_tumor_subtyping','custom','custom_998','custom_556','custom_714','custom_912_aug','custom_20','nsclc'])
parser.add_argument('--val_frac', type=float, default= 0.1,
help='fraction of labels for validation (default: 0.1)')
parser.add_argument('--test_frac', type=float, default= 0.1,
help='fraction of labels for test (default: 0.1)')
args = parser.parse_args()
if args.task == 'task_1_tumor_vs_normal':
args.n_classes=2
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/tumor_vs_normal_dummy_clean.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'normal_tissue':0, 'tumor_tissue':1},
patient_strat=True,
ignore=[])
elif args.task == 'custom': ## Our first dataset of 655 WSIs
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_655.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'custom_998': ## An expanded dataset of 998 WSIs
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_998.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'custom_556': ## The above 998 WSIs with biopsies removed, leaving only resections
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_556.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'custom_714': ## Smaller set after pathologist review of labels
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_714.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'custom_912_aug': ## The above 912 with augmentations
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_912_aug.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'custom_20': ## 20 WSIs for prototyping/testing
args.n_classes=5
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_all_20.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'high_grade':0,'low_grade':1,'clear_cell':2,'endometrioid':3,'mucinous':4},
patient_strat= True,
ignore=[])
elif args.task == 'nsclc':
args.n_classes=2
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/set_nsclc.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'luad':0,'lusc':1},
patient_strat= True,
ignore=[])
elif args.task == 'task_2_tumor_subtyping':
args.n_classes=3
dataset = Generic_WSI_Classification_Dataset(csv_path = 'dataset_csv/tumor_subtyping_dummy_clean.csv',
shuffle = False,
seed = args.seed,
print_info = True,
label_dict = {'subtype_1':0, 'subtype_2':1, 'subtype_3':2},
patient_strat= True,
patient_voting='maj',
ignore=[])
else:
raise NotImplementedError
num_slides_cls = np.array([len(cls_ids) for cls_ids in dataset.patient_cls_ids])
val_num = np.round(num_slides_cls * args.val_frac).astype(int)
test_num = np.round(num_slides_cls * args.test_frac).astype(int)
if __name__ == '__main__':
if args.label_frac > 0:
label_fracs = [args.label_frac]
else:
label_fracs = [0.1, 0.25, 0.5, 0.75, 1.0]
for lf in label_fracs:
split_dir = 'splits/'+ str(args.task) + '_{}'.format(int(lf * 100))
os.makedirs(split_dir, exist_ok=True)
dataset.create_splits(k = args.k, val_num = val_num, test_num = test_num, label_frac=lf)
used_test_indices=[]
used_val_indices=[]
for i in range(args.k):
dataset.set_splits()
descriptor_df = dataset.test_split_gen(return_descriptor=True)
splits = dataset.return_splits(from_id=True)
used_test_indices=used_test_indices+list(splits[2].slide_data['slide_id'])
used_val_indices=used_val_indices+list(splits[1].slide_data['slide_id'])
save_splits(splits, ['train', 'val', 'test'], os.path.join(split_dir, 'splits_{}.csv'.format(i)))
save_splits(splits, ['train', 'val', 'test'], os.path.join(split_dir, 'splits_{}_bool.csv'.format(i)), boolean_style=True)
descriptor_df.to_csv(os.path.join(split_dir, 'splits_{}_descriptor.csv'.format(i)))
print("total test WSIs: ",len(used_test_indices))
print("Unique test WSIs: ",len(set(used_test_indices)))
print("total val WSIs: ",len(used_val_indices))
print("Unique val WSIs: ",len(set(used_val_indices)))