-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathconfig_utils.py
174 lines (126 loc) · 8.01 KB
/
config_utils.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
import yaml
from datautils.patch_sampling import get_num_valid_patches
def keys2kwargs(config_dict):
"""
Eg: {'input-modality': 'PET'} --> {'input_modality': 'PET'}
"""
kwargs = {}
for key in config_dict.keys():
kw = key.replace('-', '_')
kwargs[kw] = config_dict[key]
return kwargs
def build_config(cli_args, training=True):
"""
Build a global config dict from the cli args
"""
global_config = {}
# Read YAML config files --
with open(cli_args.data_config_file, 'r') as dc:
yaml_data_config = yaml.safe_load(dc)
with open(cli_args.nn_config_file, 'r') as nnc:
yaml_nn_config = yaml.safe_load(nnc)
if training:
with open(cli_args.trainval_config_file, 'r') as tvc:
yaml_trainval_config = yaml.safe_load(tvc)
else:
with open(cli_args.infer_config_file, 'r') as ic:
yaml_infer_config = yaml.safe_load(ic)
# Handle overrides (Modify the YAML derived config dicts) --
if training:
if cli_args.run_name is not None:
yaml_trainval_config['trainer-kwargs']['logging_config']['run-name'] = cli_args.run_name
# Get individual settings --
global_config['nn-name'] = yaml_nn_config['nn-name']
# Construct kwargs dicts for the data pipeline --
global_config['preprocessor-kwargs'] = yaml_data_config['preprocessor-kwargs']
data_dir = f"{yaml_data_config['data-root-dir']}/{yaml_data_config['dataset-name'].split('-')[1]}_hecktor_nii"
if training:
train_dataset_kwargs = yaml_data_config['patient-dataset-kwargs'].copy()
train_dataset_kwargs['data_dir'] = data_dir
train_dataset_kwargs['patient_id_filepath'] = yaml_data_config['patient-id-filepath']
train_dataset_kwargs['mode'] = yaml_trainval_config['trainer-kwargs']['training_config']['train-subset-name']
train_patch_sampler_kwargs = yaml_data_config['train-patch-sampler-kwargs']
train_patch_sampler_kwargs['patch_size'] = yaml_data_config['patch-size']
train_patch_sampler_kwargs['volume_size'] = yaml_data_config['volume-size']
train_patch_queue_kwargs = yaml_data_config['train-patch-queue-kwargs']
train_patch_loader_kwargs = {'batch_size': yaml_data_config['batch-of-patches-size'],
'shuffle': False,
'num_workers': 0}
val_dataset_kwargs = yaml_data_config['patient-dataset-kwargs'].copy()
val_dataset_kwargs['augment_data'] = False
val_dataset_kwargs['data_dir'] = data_dir
val_dataset_kwargs['patient_id_filepath'] = yaml_data_config['patient-id-filepath']
val_dataset_kwargs['mode'] = yaml_trainval_config['trainer-kwargs']['validation_config']['val-subset-name']
val_patch_sampler_kwargs = yaml_data_config['val-patch-sampler-kwargs']
val_patch_sampler_kwargs['patch_size'] = yaml_data_config['patch-size']
val_patch_sampler_kwargs['volume_size'] = yaml_data_config['volume-size']
val_patch_aggregator_kwargs = yaml_data_config['val-patch-aggregator-kwargs']
val_patch_aggregator_kwargs['patch_size'] = yaml_data_config['patch-size']
val_patch_aggregator_kwargs['volume_size'] = yaml_data_config['volume-size']
# Add into the global config
global_config['train-dataset-kwargs'] = train_dataset_kwargs
global_config['train-patch-sampler-kwargs'] = train_patch_sampler_kwargs
global_config['train-patch-queue-kwargs'] = train_patch_queue_kwargs
global_config['train-patch-loader-kwargs'] = train_patch_loader_kwargs
global_config['val-dataset-kwargs'] = val_dataset_kwargs
global_config['val-patch-sampler-kwargs'] = val_patch_sampler_kwargs
global_config['val-patch-aggregator-kwargs'] = val_patch_aggregator_kwargs
else: # For inference
dataset_kwargs = yaml_data_config['patient-dataset-kwargs'].copy()
dataset_kwargs['augment_data'] = False
dataset_kwargs['data_dir'] = data_dir
dataset_kwargs['patient_id_filepath'] = yaml_data_config['patient-id-filepath']
dataset_kwargs['mode'] = yaml_infer_config['inferer-kwargs']['inference_config']['subset-name']
patch_sampler_kwargs = yaml_data_config['val-patch-sampler-kwargs']
patch_sampler_kwargs['patch_size'] = yaml_data_config['patch-size']
patch_sampler_kwargs['volume_size'] = yaml_data_config['volume-size']
patch_aggregator_kwargs = yaml_data_config['val-patch-aggregator-kwargs']
patch_aggregator_kwargs['patch_size'] = yaml_data_config['patch-size']
patch_aggregator_kwargs['volume_size'] = yaml_data_config['volume-size']
# Add into the global config
global_config['dataset-kwargs'] = dataset_kwargs
global_config['patch-sampler-kwargs'] = patch_sampler_kwargs
global_config['patch-aggregator-kwargs'] = patch_aggregator_kwargs
# Integrate NN kwargs into global config --
global_config['nn-kwargs'] = yaml_nn_config['nn-kwargs']
# Construct the Trainer's or Inferer's kwargs --
val_valid_patches_per_volume = get_num_valid_patches(yaml_data_config['patch-size'],
yaml_data_config['volume-size'],
focal_point_stride=yaml_data_config['val-patch-sampler-kwargs']['focal_point_stride'],
padding=yaml_data_config['val-patch-sampler-kwargs']['padding'])
input_data_config = {}
input_data_config['is-bimodal'] = yaml_data_config['is-bimodal']
if yaml_data_config['is-bimodal']:
input_data_config['input-modality'] = None
input_data_config['input-representation'] = yaml_data_config['patient-dataset-kwargs']['input_representation']
else:
input_data_config['input-modality'] = yaml_data_config['patient-dataset-kwargs']['input_modality']
input_data_config['input-representation'] = None
if training: # Trainer stuff
hardware_config = yaml_trainval_config['trainer-kwargs']['hardware_config']
training_config = yaml_trainval_config['trainer-kwargs']['training_config']
training_config['dataset-name'] = yaml_data_config['dataset-name']
validation_config = yaml_trainval_config['trainer-kwargs']['validation_config']
validation_config['batch-of-patches-size'] = yaml_data_config['batch-of-patches-size']
validation_config['valid-patches-per-volume'] = val_valid_patches_per_volume
logging_config = yaml_trainval_config['trainer-kwargs']['logging_config']
logging_config['wandb-config'] = {}
logging_config['wandb-config']['patch-size'] = yaml_data_config['patch-size']
# Add into the global config
global_config['trainer-kwargs'] = {'hardware_config': hardware_config,
'input_data_config': input_data_config,
'training_config': training_config,
'validation_config': validation_config,
'logging_config': logging_config}
else: # Inferer stuff
hardware_config = yaml_infer_config['inferer-kwargs']['hardware_config']
inference_config = yaml_infer_config['inferer-kwargs']['inference_config']
inference_config['dataset-name'] = yaml_data_config['dataset-name']
inference_config['patient-id-filepath'] = yaml_data_config['patient-id-filepath']
inference_config['batch-of-patches-size'] = yaml_data_config['batch-of-patches-size']
inference_config['valid-patches-per-volume'] = val_valid_patches_per_volume
# Add into the global config
global_config['inferer-kwargs'] = {'hardware_config': hardware_config,
'input_data_config': input_data_config,
'inference_config': inference_config}
return global_config