qgyd2021 commited on
Commit
c53e62d
1 Parent(s): 20c7594

[update]add dataset.tar.gz to speed up download

Browse files
Files changed (3) hide show
  1. cppe-5.py +64 -47
  2. cppe_5_backup.py +118 -0
  3. data/dataset.tar.gz +0 -0
cppe-5.py CHANGED
@@ -1,20 +1,26 @@
1
- #!/usr/bin/python3
2
- # -*- coding: utf-8 -*-
3
- from glob import glob
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  import json
5
  import os
6
- from pathlib import Path
7
 
8
  import datasets
9
- from PIL import Image
10
 
11
- # _URL = "https://drive.google.com/uc?id=1MGnaAfbckUmigGUvihz7uiHGC6rBIbvr"
12
-
13
- _HOMEPAGE = "https://sites.google.com/view/cppe5"
14
-
15
- _LICENSE = "Unknown"
16
-
17
- _CATEGORIES = ["Coverall", "Face_Shield", "Gloves", "Goggles", "Mask"]
18
 
19
  _CITATION = """\
20
  @misc{dagli2021cppe5,
@@ -33,6 +39,15 @@ to allow the study of subordinate categorization of medical personal protective
33
  which is not possible with other popular data sets that focus on broad level categories.
34
  """
35
 
 
 
 
 
 
 
 
 
 
36
 
37
  class CPPE5(datasets.GeneratorBasedBuilder):
38
  """CPPE - 5 dataset."""
@@ -47,12 +62,12 @@ class CPPE5(datasets.GeneratorBasedBuilder):
47
  "width": datasets.Value("int32"),
48
  "height": datasets.Value("int32"),
49
  "objects": datasets.Sequence(
50
- feature=datasets.Features({
51
  "id": datasets.Value("int64"),
52
  "area": datasets.Value("int64"),
53
  "bbox": datasets.Sequence(datasets.Value("float32"), length=4),
54
  "category": datasets.ClassLabel(names=_CATEGORIES),
55
- })
56
  ),
57
  }
58
  )
@@ -65,54 +80,56 @@ class CPPE5(datasets.GeneratorBasedBuilder):
65
  )
66
 
67
  def _split_generators(self, dl_manager):
68
- """Returns SplitGenerators."""
69
- train_json = dl_manager.download("data/annotations/train.jsonl")
70
- test_json = dl_manager.download("data/annotations/test.jsonl")
71
-
72
  return [
73
  datasets.SplitGenerator(
74
  name=datasets.Split.TRAIN,
75
  gen_kwargs={
76
- "archive_path": train_json,
77
- "dl_manager": dl_manager,
78
  },
79
  ),
80
  datasets.SplitGenerator(
81
  name=datasets.Split.TEST,
82
  gen_kwargs={
83
- "archive_path": test_json,
84
- "dl_manager": dl_manager,
85
  },
86
  ),
87
  ]
88
 
89
- def _generate_examples(self, archive_path, dl_manager):
90
- """Yields examples."""
91
- archive_path = Path(archive_path)
 
 
 
 
 
92
 
 
93
  idx = 0
94
-
95
- with open(archive_path, "r", encoding="utf-8") as f:
96
- for row in f:
97
- sample = json.loads(row)
98
-
99
- file_path = sample["image"]
100
- file_path = dl_manager.download(file_path)
101
-
102
- with open(file_path, "rb") as image_f:
103
- image_bytes = image_f.read()
104
- # image = Image.open(image_f)
105
-
 
 
 
 
106
  yield idx, {
107
- "image_id": sample["image_id"],
108
- "image": {"path": file_path, "bytes": image_bytes},
109
- # "image": image,
110
- "width": sample["width"],
111
- "height": sample["height"],
112
- "objects": sample["objects"],
113
  }
114
  idx += 1
115
-
116
-
117
- if __name__ == '__main__':
118
- pass
 
1
+ # coding=utf-8
2
+ # Copyright 2021 The HuggingFace Datasets Authors and the current dataset script contributor.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+ """CPPE-5 dataset."""
16
+
17
+
18
+ import collections
19
  import json
20
  import os
 
21
 
22
  import datasets
 
23
 
 
 
 
 
 
 
 
24
 
25
  _CITATION = """\
26
  @misc{dagli2021cppe5,
 
39
  which is not possible with other popular data sets that focus on broad level categories.
40
  """
41
 
42
+ _HOMEPAGE = "https://sites.google.com/view/cppe5"
43
+
44
+ _LICENSE = "Unknown"
45
+
46
+ # _URL = "https://drive.google.com/uc?id=1MGnaAfbckUmigGUvihz7uiHGC6rBIbvr"
47
+ _URL = "data/dataset.tar.gz"
48
+
49
+ _CATEGORIES = ["Coverall", "Face_Shield", "Gloves", "Goggles", "Mask"]
50
+
51
 
52
  class CPPE5(datasets.GeneratorBasedBuilder):
53
  """CPPE - 5 dataset."""
 
62
  "width": datasets.Value("int32"),
63
  "height": datasets.Value("int32"),
64
  "objects": datasets.Sequence(
65
+ {
66
  "id": datasets.Value("int64"),
67
  "area": datasets.Value("int64"),
68
  "bbox": datasets.Sequence(datasets.Value("float32"), length=4),
69
  "category": datasets.ClassLabel(names=_CATEGORIES),
70
+ }
71
  ),
72
  }
73
  )
 
80
  )
81
 
82
  def _split_generators(self, dl_manager):
83
+ archive = dl_manager.download(_URL)
 
 
 
84
  return [
85
  datasets.SplitGenerator(
86
  name=datasets.Split.TRAIN,
87
  gen_kwargs={
88
+ "annotation_file_path": "annotations/train.json",
89
+ "files": dl_manager.iter_archive(archive),
90
  },
91
  ),
92
  datasets.SplitGenerator(
93
  name=datasets.Split.TEST,
94
  gen_kwargs={
95
+ "annotation_file_path": "annotations/test.json",
96
+ "files": dl_manager.iter_archive(archive),
97
  },
98
  ),
99
  ]
100
 
101
+ def _generate_examples(self, annotation_file_path, files):
102
+ def process_annot(annot, category_id_to_category):
103
+ return {
104
+ "id": annot["id"],
105
+ "area": annot["area"],
106
+ "bbox": annot["bbox"],
107
+ "category": category_id_to_category[annot["category_id"]],
108
+ }
109
 
110
+ image_id_to_image = {}
111
  idx = 0
112
+ # This loop relies on the ordering of the files in the archive:
113
+ # Annotation files come first, then the images.
114
+ for path, f in files:
115
+ file_name = os.path.basename(path)
116
+ if path == annotation_file_path:
117
+ annotations = json.load(f)
118
+ category_id_to_category = {category["id"]: category["name"] for category in annotations["categories"]}
119
+ image_id_to_annotations = collections.defaultdict(list)
120
+ for annot in annotations["annotations"]:
121
+ image_id_to_annotations[annot["image_id"]].append(annot)
122
+ image_id_to_image = {annot["file_name"]: annot for annot in annotations["images"]}
123
+ elif file_name in image_id_to_image:
124
+ image = image_id_to_image[file_name]
125
+ objects = [
126
+ process_annot(annot, category_id_to_category) for annot in image_id_to_annotations[image["id"]]
127
+ ]
128
  yield idx, {
129
+ "image_id": image["id"],
130
+ "image": {"path": path, "bytes": f.read()},
131
+ "width": image["width"],
132
+ "height": image["height"],
133
+ "objects": objects,
 
134
  }
135
  idx += 1
 
 
 
 
cppe_5_backup.py ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/python3
2
+ # -*- coding: utf-8 -*-
3
+ from glob import glob
4
+ import json
5
+ import os
6
+ from pathlib import Path
7
+
8
+ import datasets
9
+ from PIL import Image
10
+
11
+ # _URL = "https://drive.google.com/uc?id=1MGnaAfbckUmigGUvihz7uiHGC6rBIbvr"
12
+
13
+ _HOMEPAGE = "https://sites.google.com/view/cppe5"
14
+
15
+ _LICENSE = "Unknown"
16
+
17
+ _CATEGORIES = ["Coverall", "Face_Shield", "Gloves", "Goggles", "Mask"]
18
+
19
+ _CITATION = """\
20
+ @misc{dagli2021cppe5,
21
+ title={CPPE-5: Medical Personal Protective Equipment Dataset},
22
+ author={Rishit Dagli and Ali Mustufa Shaikh},
23
+ year={2021},
24
+ eprint={2112.09569},
25
+ archivePrefix={arXiv},
26
+ primaryClass={cs.CV}
27
+ }
28
+ """
29
+
30
+ _DESCRIPTION = """\
31
+ CPPE - 5 (Medical Personal Protective Equipment) is a new challenging dataset with the goal
32
+ to allow the study of subordinate categorization of medical personal protective equipments,
33
+ which is not possible with other popular data sets that focus on broad level categories.
34
+ """
35
+
36
+
37
+ class CPPE5(datasets.GeneratorBasedBuilder):
38
+ """CPPE - 5 dataset."""
39
+
40
+ VERSION = datasets.Version("1.0.0")
41
+
42
+ def _info(self):
43
+ features = datasets.Features(
44
+ {
45
+ "image_id": datasets.Value("int64"),
46
+ "image": datasets.Image(),
47
+ "width": datasets.Value("int32"),
48
+ "height": datasets.Value("int32"),
49
+ "objects": datasets.Sequence(
50
+ feature=datasets.Features({
51
+ "id": datasets.Value("int64"),
52
+ "area": datasets.Value("int64"),
53
+ "bbox": datasets.Sequence(datasets.Value("float32"), length=4),
54
+ "category": datasets.ClassLabel(names=_CATEGORIES),
55
+ })
56
+ ),
57
+ }
58
+ )
59
+ return datasets.DatasetInfo(
60
+ description=_DESCRIPTION,
61
+ features=features,
62
+ homepage=_HOMEPAGE,
63
+ license=_LICENSE,
64
+ citation=_CITATION,
65
+ )
66
+
67
+ def _split_generators(self, dl_manager):
68
+ """Returns SplitGenerators."""
69
+ train_json = dl_manager.download("data/annotations/train.jsonl")
70
+ test_json = dl_manager.download("data/annotations/test.jsonl")
71
+
72
+ return [
73
+ datasets.SplitGenerator(
74
+ name=datasets.Split.TRAIN,
75
+ gen_kwargs={
76
+ "archive_path": train_json,
77
+ "dl_manager": dl_manager,
78
+ },
79
+ ),
80
+ datasets.SplitGenerator(
81
+ name=datasets.Split.TEST,
82
+ gen_kwargs={
83
+ "archive_path": test_json,
84
+ "dl_manager": dl_manager,
85
+ },
86
+ ),
87
+ ]
88
+
89
+ def _generate_examples(self, archive_path, dl_manager):
90
+ """Yields examples."""
91
+ archive_path = Path(archive_path)
92
+
93
+ idx = 0
94
+
95
+ with open(archive_path, "r", encoding="utf-8") as f:
96
+ for row in f:
97
+ sample = json.loads(row)
98
+
99
+ file_path = sample["image"]
100
+ file_path = dl_manager.download(file_path)
101
+
102
+ with open(file_path, "rb") as image_f:
103
+ image_bytes = image_f.read()
104
+ # image = Image.open(image_f)
105
+
106
+ yield idx, {
107
+ "image_id": sample["image_id"],
108
+ "image": {"path": file_path, "bytes": image_bytes},
109
+ # "image": image,
110
+ "width": sample["width"],
111
+ "height": sample["height"],
112
+ "objects": sample["objects"],
113
+ }
114
+ idx += 1
115
+
116
+
117
+ if __name__ == '__main__':
118
+ pass
data/dataset.tar.gz ADDED
File without changes