-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathmake_nnunet_nobe_task.py
266 lines (217 loc) · 11.9 KB
/
make_nnunet_nobe_task.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
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
# -*- coding: utf-8 -*-
"""Builds nnunet dataset with DICOM-Glioma-Seg images.
"""
import os
import pickle
import shutil
import nibabel as nib
import pandas as pd
import SimpleITK as sitk
from batchgenerators.utilities.file_and_folder_operations import *
from collections import OrderedDict
from dotenv import find_dotenv, load_dotenv
from nipype.interfaces.fsl import Reorient2Std
from pathlib import Path
from brats.nnunet.datasets import copy_BraTS_segmentation_and_convert_labels
from brats.utils import dcm2nifti
def build_nnunet_nobet(seg_list, data_dir, task_name):
"""From nnUNet's repo.
"""
# load environment variables
load_dotenv(find_dotenv())
Path(os.environ['nnUNet_raw_data_base']).mkdir(exist_ok=True)
Path(os.environ['nnUNet_preprocessed']).mkdir(exist_ok=True)
Path(os.environ['RESULTS_FOLDER']).mkdir(exist_ok=True)
nnUNet_raw_data = Path(os.environ['nnUNet_raw_data_base'], 'nnUNet_raw_data')
nnUNet_raw_data.mkdir(exist_ok=True)
# make base dirs
target_base = Path(nnUNet_raw_data, task_name)
target_imagesTr = Path(target_base, "imagesTr")
target_imagesVal = Path(target_base, "imagesVal")
target_imagesTs = Path(target_base, "imagesTs")
target_labelsTr = Path(target_base, "labelsTr")
target_base.mkdir(exist_ok=True)
target_imagesTr.mkdir(exist_ok=True)
target_imagesVal.mkdir(exist_ok=True)
target_imagesTs.mkdir(exist_ok=True)
target_labelsTr.mkdir(exist_ok=True)
tmpdir = Path('.tmpdir')
tmpdir.mkdir(exist_ok=True)
print('loading metadta')
with open(data_dir/'train_val_metadata_df.pkl', 'rb') as f:
metadata = pickle.load(f)
brats_ids = pd.read_csv(data_dir/'MICCAI_BraTS2020_TrainingData/name_mapping.csv').dropna(axis=0, subset=['TCGA_TCIA_subject_ID']).set_index('TCGA_TCIA_subject_ID')['BraTS_2020_subject_ID']
mod_num = {
'flair': '0000',
't1': '0001',
't1ce': '0002',
't2': '0003',
}
print('creating segmentation files and linking mri images')
patient_names = list()
for seg_fpath in seg_list:
dcm_mod = seg_fpath.parent.name
sid = seg_fpath.parent.parent.name
mod = dcm_mod.split('Glioma')[0][len('300.000000-'):][:-1]
ref_fpath = next(data_dir.glob(f'TCGA/*/Pre*/{sid}/*t1*'))
mod_img_metadata = metadata.loc[sid,dcm_mod]
mod_label = mod_img_metadata['Modality']
mod_dcm_dir = data_dir/mod_img_metadata['Filepath']
brats_id = brats_ids.loc[sid]
# TODO: WIP
seg_dst_fpath = target_labelsTr/f"{brats_id}.nii.gz"
mod_dst_fpath = target_imagesTr/f"{brats_id}_{mod_num[mod_label.lower()]}.nii.gz"
if not mod_dst_fpath.exists():
mod_fpath = dcm2nifti(mod_dcm_dir, tmpdir)
mod_id = mod_dcm_dir.name.split('-')[-1]
if sid.startswith('TCGA-FG-7643') and isinstance(mod_fpath, list) and mod_label == 'T1CE':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-HT-8114') and isinstance(mod_fpath, list) and mod_label == 'T1CE':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-CS-494') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-CS-5395') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-HT-768') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-HT-769') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-HT-785') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-HT-7882') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-HT-8105') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-A6S6') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-7298') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-7014') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-7010') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-DU-7008') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-DU-6410') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-6401') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-6408') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-DU-640') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-DU-639') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-DU-58') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-0119') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-0128') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-013') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-014') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-0154') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-0158') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-0162') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-016') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-0177') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-017') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-018') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-0190') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-0213') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[0]
if sid.startswith('TCGA-06-0240') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-1084') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-06-5408') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-27-183') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-27-2526') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-76-4932') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-76-6191') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-76-6193') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid.startswith('TCGA-76-628') and isinstance(mod_fpath, list) and mod_label == 'T2':
mod_fpath = mod_fpath[1]
if sid == 'TCGA-06-0646' and mod_label == 'T1CE':
# fix T1CE for CaPTk
mod = nib.load(mod_fpath)
fixed_mod = nib.Nifti1Image(mod.get_fdata()[:,:,:,0], mod.affine, mod.header)
mod_fpath = str(mod_fpath).replace('.nii.gz', '_fixed.nii.gz')
nib.save(fixed_mod, mod_fpath)
mod_fpath = Path(mod_fpath)
# make seg labels continuous
seg_tmp_fpath = tmpdir/('converted_'+seg_fpath.name)
copy_BraTS_segmentation_and_convert_labels(str(seg_fpath), str(seg_tmp_fpath))
# transform processed images back to original orientation and spacing
transform_fpath = next((data_dir/'DICOM_Glioma_SEG_Metadata').glob(f"*/{sid}/Registration_Transforms/*{mod_id}.txt"))
mod_img = sitk.ReadImage(str(mod_fpath), imageIO="NiftiImageIO")
seg_img = sitk.ReadImage(str(seg_tmp_fpath), imageIO="NiftiImageIO")
ref_img = sitk.ReadImage(str(ref_fpath), imageIO="NiftiImageIO")
transform = sitk.ReadTransform(str(transform_fpath))
transf_mod_img = sitk.Resample(mod_img, referenceImage=ref_img, transform=transform.GetInverse())
transf_seg_img = sitk.Resample(seg_img, referenceImage=ref_img, transform=transform.GetInverse())
mod_tmp_fpath = tmpdir/'mod.nii.gz'
sitk.WriteImage(transf_mod_img, str(mod_tmp_fpath))
# reorient and save images
reor = Reorient2Std()
reor.inputs.in_file = mod_tmp_fpath
reor.inputs.out_file = str(mod_dst_fpath)
_ = reor.run()
if not seg_dst_fpath.exists():
patient_names.append(brats_id)
seg_tmp_fpath = tmpdir/'seg.nii.gz'
sitk.WriteImage(transf_seg_img, str(seg_tmp_fpath))
reor = Reorient2Std()
reor.inputs.in_file = seg_tmp_fpath
reor.inputs.out_file = str(seg_dst_fpath)
_ = reor.run()
# clear temporary files
shutil.rmtree(tmpdir, ignore_errors=True)
tmpdir.mkdir(exist_ok=True)
print('creating dataset json')
json_dict = OrderedDict()
json_dict['name'] = task_name[8:]
json_dict['description'] = "nothing"
json_dict['tensorImageSize'] = "4D"
json_dict['reference'] = "see TCGA"
json_dict['licence'] = "see TCGA license"
json_dict['release'] = "0.0"
json_dict['modality'] = {
"0": "FLAIR",
"1": "T1",
"2": "T1c",
"3": "T2",
}
json_dict['labels'] = {
"0": "background",
"1": "edema",
"2": "non-enhancing",
"3": "enhancing",
}
json_dict['numTraining'] = len(patient_names)
json_dict['numTest'] = 0
json_dict['training'] = [{'image': "./imagesTr/%s.nii.gz" % i, "label": "./labelsTr/%s.nii.gz" % i} for i in
patient_names]
json_dict['test'] = []
save_json(json_dict, join(target_base, "dataset.json"))
if __name__ == '__main__':
project_dir = Path(__file__).resolve().parents[0]
data_dir = Path(project_dir, 'data', 'raw')
seg_list = list((data_dir/'DICOM-Glioma-SEG').glob('**/seg.nii.gz'))
build_nnunet_nobet(seg_list, data_dir, task_name='Task108_TCGA_DICOM_nobe')