-
Notifications
You must be signed in to change notification settings - Fork 12
/
run_infer_wsi.py
executable file
·139 lines (116 loc) · 5.86 KB
/
run_infer_wsi.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
"""run_infer_wsi.py
Usage:
run_infer_wsi.py [--gpu=<id>] [--model=<path>] [--nr_inference_workers=<n>] \
[--nr_post_proc_workers=<n>] [--batch_size=<n>] [--tile_shape=<n>] [--chunk_shape=<n>] \
[--ambiguous_size=<int>] [--wsi_proc_mag=<n>] [--wsi_file_ext=<str>] [--cache_path=<path>] \
[--logging_dir=<path>] [--input_dir=<path>] [--msk_dir=<path>] [--output_dir=<path>] [--patch_input_shape=<n>] \
[--patch_output_shape=<n>] [--wsi_bulk_idx=<n>] [--wsi_proc_step=<n>] [--save_thumb] [--save_mask]
run_infer_wsi.py (-h | --help)
run_infer_wsi.py --version
Options:
-h --help Show this string.
--version Show version.
--gpu=<id> GPU list. [default: 0]
--model=<path> Path to saved checkpoint.
--nr_inference_workers=<n> Number of workers during inference. [default: 0]
--nr_post_proc_workers=<n> Number of workers during post-processing. [default: 0]
--batch_size=<n> Batch size. [default: 30]
--tile_shape=<n> Shape of tile for processing. [default: 2048]
--chunk_shape=<n> Shape of tile for processing. [default: 15000]
--ambiguous_size=<int> Define ambiguous region along tiling grid to perform re-post processing. [default: 64]
--wsi_proc_mag=<n> Microns per pixel used for WSI processing. [default: 0.5]
--wsi_file_ext=<str> File extension of WSIs to process. [default: .svs]
--cache_path=<path> Path for cache. Should be placed on SSD with at least 100GB. [default: cache/]
--logging_dir=<path> Path for python logging. [default: logging/]
--input_dir=<path> Path to input data directory. Assumes the files are not nested within directory.
--msk_dir=<path> Path to directory containing tissue masks. Should have the same name as corresponding WSIs.
--output_dir=<path> Path to output data directory. Will create automtically if doesn't exist. [default: output/]
--patch_input_shape=<n> Shape of input patch to the network- Assume square shape. [default: 448]
--patch_output_shape=<n> Shape of network output- Assume square shape. [default: 144]
--wsi_bulk_idx=<n> Index for batch processing. Indexing is from 0 to n-1. [default: 1]
--wsi_proc_step=<n> Increments for batch WSI processing. [default: 10]
--save_thumb Whether to save the slide thumbnail
--save_mask Whether to save the slide mask
"""
import os
import yaml
import glob
import numpy as np
from docopt import docopt
from misc.utils import rm_n_mkdir
# -------------------------------------------------------------------------------------------------------
if __name__ == "__main__":
args = docopt(__doc__, version="CoBi Gland Inference")
if args["--gpu"]:
os.environ["CUDA_VISIBLE_DEVICES"] = args["--gpu"]
input_dir = args["--input_dir"]
output_dir = args["--output_dir"]
logging_dir = args["--logging_dir"]
cache_path = args["--cache_path"] + args["--wsi_bulk_idx"]
wsi_file_ext = args["--wsi_file_ext"]
# create output directory
if not os.path.exists(output_dir):
rm_n_mkdir(output_dir)
# create logging directory
if not os.path.exists(logging_dir):
rm_n_mkdir(logging_dir)
wsi_file_list = glob.glob(f"{input_dir}/*{wsi_file_ext}")
wsi_file_list.sort()
wsi_list = []
mask_list = []
for wsi_filename in wsi_file_list:
wsi_basename = os.path.basename(wsi_filename)
wsi_basename = wsi_basename[:-5]
# only consider if mask exists! However, if mask directory not supplied, process all!
if not args["--msk_dir"]:
wsi_list.append(wsi_filename)
mask_list.append(None)
elif os.path.isfile(args["--msk_dir"] + wsi_basename + ".png"):
wsi_list.append(wsi_filename)
mask_list.append(args["--msk_dir"] + wsi_basename + ".png")
proc_list = []
for idx, entry in enumerate(wsi_list):
proc_list.append(idx)
step = int(args["--wsi_proc_step"])
start_idx = (int(args["--wsi_bulk_idx"]) - 1) * step
end_idx = int(args["--wsi_bulk_idx"]) * step
proc_list = proc_list[start_idx:end_idx]
wsi_list = np.array(wsi_list)[proc_list].tolist()
mask_list = np.array(mask_list)[proc_list].tolist()
print('Number of WSIs in list:', len(wsi_list))
run_root_dir = args["--model"]
checkpoint_path = "%s/weights.tar" % run_root_dir
with open("%s/settings.yml" % (run_root_dir)) as fptr:
run_paramset = yaml.full_load(fptr)
target_list = ["gland", "lumen", "nuclei", "patch-class"] # all tasks considered by Cerberus
run_args = {
"nr_inference_workers": int(args["--nr_inference_workers"]),
"nr_post_proc_workers": int(args["--nr_post_proc_workers"]),
"batch_size": int(args["--batch_size"]),
"input_list": wsi_list,
"mask_list": mask_list,
"output_dir": output_dir,
"patch_input_shape": int(args["--patch_input_shape"]),
"patch_output_shape": int(args["--patch_output_shape"]),
"save_thumb": args["--save_thumb"],
"save_mask": args["--save_mask"],
"mask_dir": args["--msk_dir"],
"postproc_list": target_list,
}
wsi_run_args = {
"msk_dir": args["--msk_dir"],
"tile_shape": int(args["--tile_shape"]),
"chunk_shape": int(args["--chunk_shape"]),
"ambiguous_size": int(args["--ambiguous_size"]),
"cache_path": cache_path,
"logging_dir": logging_dir,
"wsi_proc_mag": float(args["--wsi_proc_mag"]),
}
run_args.update(wsi_run_args)
from infer.wsi import InferManager
infer = InferManager(
checkpoint_path=checkpoint_path,
decoder_dict=run_paramset["dataset_kwargs"]["req_target_code"],
model_args=run_paramset["model_kwargs"],
)
infer.process_wsi_list(run_args)