zhuchi76 commited on
Commit
2dda33e
·
verified ·
1 Parent(s): 7f91632

Update script to hub

Browse files
Files changed (1) hide show
  1. Boat_dataset.py +13 -21
Boat_dataset.py CHANGED
@@ -6,22 +6,19 @@ import os
6
 
7
  import datasets
8
 
9
-
10
  _CITATION = """\
11
  @InProceedings{huggingface:dataset,
12
  title = {Boat dataset},
13
- author={Tzu-Chi Chen, Inc.
14
- },
15
  year={2024}
16
  }
17
  """
18
 
19
-
20
  _DESCRIPTION = """\
21
- This dataset is designed to solve object detection task.
22
  """
23
 
24
- _HOMEPAGE = "https://huggingface.co/datasets/zhuchi76/Boat_dataset/resolve/main"
25
 
26
  _LICENSE = ""
27
 
@@ -69,8 +66,8 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
69
  # Download all files and extract them
70
  downloaded_files = dl_manager.download_and_extract(_URLS)
71
 
72
- # Paths to files and archives
73
- image_archive_path = downloaded_files["images"]
74
  classes_file_path = downloaded_files["classes"]
75
  annotations_paths = downloaded_files["anno"]
76
 
@@ -82,7 +79,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
82
  datasets.SplitGenerator(
83
  name=datasets.Split.TRAIN,
84
  gen_kwargs={
85
- "image_archive_path": image_archive_path,
86
  "annotations_file": annotations_paths["train"],
87
  "classes": classes,
88
  "split": "train",
@@ -91,7 +88,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
91
  datasets.SplitGenerator(
92
  name=datasets.Split.VALIDATION,
93
  gen_kwargs={
94
- "image_archive_path": image_archive_path,
95
  "annotations_file": annotations_paths["val"],
96
  "classes": classes,
97
  "split": "val",
@@ -100,7 +97,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
100
  datasets.SplitGenerator(
101
  name=datasets.Split.TEST,
102
  gen_kwargs={
103
- "image_archive_path": image_archive_path,
104
  "annotations_file": annotations_paths["test"],
105
  "classes": classes,
106
  "split": "val_real",
@@ -108,23 +105,18 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
108
  ),
109
  ]
110
 
111
- def _generate_examples(self, dl_manager, image_archive_path, annotations_file, classes, split):
112
- # Open the archive using iter_archive to access files without extracting them
113
- with dl_manager.iter_archive(image_archive_path) as archive:
114
- # Convert path: file-like object pairs to a dict for quick access
115
- images_dict = {os.path.basename(path): file for path, file in archive}
116
-
117
  # Process annotations
118
  with open(annotations_file, encoding="utf-8") as f:
119
  for key, row in enumerate(f):
120
  try:
121
  data = json.loads(row.strip())
122
- file = images_dict.get(data["file_name"])
123
- if not file:
124
- continue # Skip if file is not found in the archive
125
  yield key, {
126
  "image_id": data["image_id"],
127
- "image_path": file, # You can save the path or file object depending on your processing needs
128
  "width": data["width"],
129
  "height": data["height"],
130
  "objects": {
 
6
 
7
  import datasets
8
 
 
9
  _CITATION = """\
10
  @InProceedings{huggingface:dataset,
11
  title = {Boat dataset},
12
+ author={Tzu-Chi Chen, Inc.},
 
13
  year={2024}
14
  }
15
  """
16
 
 
17
  _DESCRIPTION = """\
18
+ This dataset is designed to solve an object detection task with images of boats.
19
  """
20
 
21
+ _HOMEPAGE = "https://huggingface.co/datasets/zhuchi76/Boat_dataset"
22
 
23
  _LICENSE = ""
24
 
 
66
  # Download all files and extract them
67
  downloaded_files = dl_manager.download_and_extract(_URLS)
68
 
69
+ # Extract the image archive
70
+ image_dir = dl_manager.extract(downloaded_files["images"])
71
  classes_file_path = downloaded_files["classes"]
72
  annotations_paths = downloaded_files["anno"]
73
 
 
79
  datasets.SplitGenerator(
80
  name=datasets.Split.TRAIN,
81
  gen_kwargs={
82
+ "image_dir": image_dir,
83
  "annotations_file": annotations_paths["train"],
84
  "classes": classes,
85
  "split": "train",
 
88
  datasets.SplitGenerator(
89
  name=datasets.Split.VALIDATION,
90
  gen_kwargs={
91
+ "image_dir": image_dir,
92
  "annotations_file": annotations_paths["val"],
93
  "classes": classes,
94
  "split": "val",
 
97
  datasets.SplitGenerator(
98
  name=datasets.Split.TEST,
99
  gen_kwargs={
100
+ "image_dir": image_dir,
101
  "annotations_file": annotations_paths["test"],
102
  "classes": classes,
103
  "split": "val_real",
 
105
  ),
106
  ]
107
 
108
+ def _generate_examples(self, image_dir, annotations_file, classes, split):
 
 
 
 
 
109
  # Process annotations
110
  with open(annotations_file, encoding="utf-8") as f:
111
  for key, row in enumerate(f):
112
  try:
113
  data = json.loads(row.strip())
114
+ file_path = os.path.join(image_dir, data["file_name"])
115
+ if not os.path.isfile(file_path):
116
+ continue # Skip if file is not found in the directory
117
  yield key, {
118
  "image_id": data["image_id"],
119
+ "image_path": file_path, # Provide the full path to the image
120
  "width": data["width"],
121
  "height": data["height"],
122
  "objects": {