YuxuanZhang888 commited on
Commit
0233002
1 Parent(s): a6625d5

Changed the name of my dataset loading script.

Browse files
Files changed (1) hide show
  1. ColonCancerCTDataset.py +158 -0
ColonCancerCTDataset.py ADDED
@@ -0,0 +1,158 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import pydicom
2
+ from PIL import Image
3
+ import numpy as np
4
+ import io
5
+ import datasets
6
+ import gdown
7
+ import re
8
+ import s3fs
9
+ import random
10
+
11
+ example_manifest_url = "https://drive.google.com/uc?id=1JBkQTXeieyN9_6BGdTF_DDlFFyZrGyU6"
12
+ example_manifest_file = gdown.download(example_manifest_url, 'manifest_file.s5cmd', quiet = False)
13
+ full_manifest_url = "https://drive.google.com/uc?id=1KP6qxcQoPF4MJdEPNwW7J6BlL_sUJ17j"
14
+ full_manifest_file = gdown.download(full_manifest_url, 'full_manifest_file.s5cmd', quiet = False)
15
+ fs = s3fs.S3FileSystem(anon=True)
16
+
17
+ _DESCRIPTION = "This is the description"
18
+ _HOMEPAGE = "https://imaging.datacommons.cancer.gov/"
19
+ _LICENSE = "https://fairsharing.org/FAIRsharing.0b5a1d"
20
+ _CITATION = "National Cancer Institute Imaging Data Commons (IDC) Collections was accessed on DATE from https://registry.opendata.aws/nci-imaging-data-commons"
21
+
22
+
23
+ class ColonCancerCTDataset(datasets.GeneratorBasedBuilder):
24
+ """TODO: Short description of my dataset."""
25
+ VERSION = datasets.Version("1.1.0")
26
+
27
+ BUILDER_CONFIGS = [
28
+ datasets.BuilderConfig(name="example", version=VERSION, description="This is a subset of the full dataset for demonstration purposes"),
29
+ datasets.BuilderConfig(name="full_data", version=VERSION, description="This is the complete dataset"),
30
+ ]
31
+ DEFAULT_CONFIG_NAME = "example"
32
+
33
+ def _info(self):
34
+ return datasets.DatasetInfo(
35
+ description=_DESCRIPTION,
36
+ features=datasets.Features(
37
+ {
38
+ "image": datasets.Image(),
39
+ "ImageType": datasets.Sequence(datasets.Value('string')),
40
+ "StudyDate": datasets.Value('string'),
41
+ "SeriesDate": datasets.Value('string'),
42
+ "Manufacturer": datasets.Value('string'),
43
+ "StudyDescription": datasets.Value('string'),
44
+ "SeriesDescription": datasets.Value('string'),
45
+ "PatientSex": datasets.Value('string'),
46
+ "PatientAge": datasets.Value('string'),
47
+ "PregnancyStatus": datasets.Value('string'),
48
+ "BodyPartExamined": datasets.Value('string'),
49
+ }),
50
+ homepage = _HOMEPAGE,
51
+ license = _LICENSE,
52
+ citation = _CITATION
53
+
54
+ )
55
+
56
+ def _split_generators(self, dl_manager):
57
+ """Returns SplitGenerators."""
58
+ # TODO: This method is tasked with downloading/extracting the data and defining the splits depending on the
59
+ s3_series_paths = []
60
+ s3_individual_paths = []
61
+ if self.config.name == 'example':
62
+ manifest_file = example_manifest_file
63
+ else:
64
+ manifest_file = full_manifest_file
65
+
66
+ with open(manifest_file, 'r') as file:
67
+ for line in file:
68
+ match = re.search(r'cp (s3://[\S]+) .', line)
69
+ if match:
70
+ s3_series_paths.append(match.group(1)[:-2]) # Deleting the '/*' in directories
71
+ for series in s3_series_paths:
72
+ for content in fs.ls(series):
73
+ s3_individual_paths.append(fs.info(content)['Key'])
74
+
75
+ random.shuffle(s3_individual_paths)
76
+
77
+ # Define the split sizes
78
+ train_size = int(0.7 * len(s3_individual_paths))
79
+ val_size = int(0.15 * len(s3_individual_paths))
80
+ # Split the paths into train, validation, and test sets
81
+ train_paths = s3_individual_paths[:train_size]
82
+ val_paths = s3_individual_paths[train_size:train_size + val_size]
83
+ test_paths = s3_individual_paths[train_size + val_size:]
84
+
85
+ return [
86
+ datasets.SplitGenerator(
87
+ name=datasets.Split.TRAIN,
88
+ gen_kwargs={
89
+ "paths": train_paths,
90
+ "split": "train"
91
+ },
92
+ ),
93
+ datasets.SplitGenerator(
94
+ name=datasets.Split.VALIDATION,
95
+ gen_kwargs={
96
+ "paths": val_paths,
97
+ "split": "dev"
98
+ },
99
+ ),
100
+ datasets.SplitGenerator(
101
+ name=datasets.Split.TEST,
102
+ gen_kwargs={
103
+ "paths": test_paths,
104
+ "split": "test"
105
+ },
106
+ ),
107
+ ]
108
+
109
+ def _generate_examples(self, paths, split):
110
+ """Yields examples."""
111
+ # TODO: This method will yield examples, i.e. rows in the dataset.
112
+ for path in paths:
113
+ key = path
114
+ with fs.open(path, 'rb') as f:
115
+ dicom_data = pydicom.dcmread(f)
116
+ pixel_array = dicom_data.pixel_array
117
+ # Adjust for MONOCHROME1 to invert the grayscale values
118
+ if dicom_data.PhotometricInterpretation == "MONOCHROME1":
119
+ pixel_array = np.max(pixel_array) - pixel_array
120
+ # Normalize or scale 16-bit or other depth images to 8-bit
121
+ if pixel_array.dtype != np.uint8:
122
+ pixel_array = (np.divide(pixel_array, np.max(pixel_array)) * 255).astype(np.uint8)
123
+ # Convert to RGB if it is not already (e.g., for color images)
124
+ if len(pixel_array.shape) == 2:
125
+ im = Image.fromarray(pixel_array, mode="L") # L mode is for grayscale
126
+ elif len(pixel_array.shape) == 3 and pixel_array.shape[2] in [3, 4]:
127
+ im = Image.fromarray(pixel_array, mode="RGB")
128
+ else:
129
+ raise ValueError("Unsupported DICOM image format")
130
+ with io.BytesIO() as output:
131
+ im.save(output, format="PNG")
132
+ png_image = output.getvalue()
133
+ # Extracting metadata
134
+ ImageType = dicom_data.get("ImageType", "")
135
+ StudyDate = dicom_data.get("StudyDate", "")
136
+ SeriesDate = dicom_data.get("SeriesDate", "")
137
+ Manufacturer = dicom_data.get("Manufacturer", "")
138
+ StudyDescription = dicom_data.get("StudyDescription", "")
139
+ SeriesDescription = dicom_data.get("SeriesDescription", "")
140
+ PatientSex = dicom_data.get("PatientSex", "")
141
+ PatientAge = dicom_data.get("PatientAge", "")
142
+ PregnancyStatus = dicom_data.get("PregnancyStatus", "")
143
+ if PregnancyStatus == None:
144
+ PregnancyStatus = "None"
145
+ else:
146
+ PregnancyStatus = "Yes"
147
+ BodyPartExamined = dicom_data.get("BodyPartExamined", "")
148
+ yield key, {"image": png_image,
149
+ "ImageType": ImageType,
150
+ "StudyDate": StudyDate,
151
+ "SeriesDate": SeriesDate,
152
+ "Manufacturer": Manufacturer,
153
+ "StudyDescription": StudyDescription,
154
+ "SeriesDescription": SeriesDescription,
155
+ "PatientSex": PatientSex,
156
+ "PatientAge": PatientAge,
157
+ "PregnancyStatus": PregnancyStatus,
158
+ "BodyPartExamined": BodyPartExamined}