ShixuanAn commited on
Commit
0a107d3
1 Parent(s): 3f895a0

Update RDD_2020.py

Browse files
Files changed (1) hide show
  1. RDD_2020.py +20 -23
RDD_2020.py CHANGED
@@ -75,54 +75,51 @@ class RDD2020_Dataset(datasets.GeneratorBasedBuilder):
75
  )
76
 
77
  def _split_generators(self, dl_manager):
78
- data_dir = dl_manager.download_and_extract(_URLS["dataset"])
 
 
 
 
79
  return [
80
  datasets.SplitGenerator(
81
  name=datasets.Split.TRAIN,
82
  gen_kwargs={
83
- "images_dir": os.path.join(data_dir, "train"),
84
- "annotations_dir": os.path.join(data_dir, "train", "annotations"),
85
  "split": "train",
86
  },
87
  ),
88
  datasets.SplitGenerator(
89
  name="test1",
90
  gen_kwargs={
91
- "images_dir": os.path.join(data_dir, "test1"),
92
- "annotations_dir": None, # No annotations for test1
93
  "split": "test1",
94
  },
95
  ),
96
  datasets.SplitGenerator(
97
  name="test2",
98
  gen_kwargs={
99
- "images_dir": os.path.join(data_dir, "test2"),
100
- "annotations_dir": None, # No annotations for test2
101
  "split": "test2",
102
  },
103
  ),
104
  ]
105
 
106
- def _generate_examples(self, images_dir, annotations_dir, split):
107
- """Yields examples as (key, example) tuples."""
108
- for country_dir in sorted(os.listdir(images_dir)): # Sort to ensure consistent order
109
- country_images_dir = os.path.join(images_dir, country_dir, "images")
110
-
111
- # Check if we are in a test split without annotations
112
- if annotations_dir is None:
113
- country_annotations_dir = None
114
- else:
115
- country_annotations_dir = os.path.join(annotations_dir, country_dir, "xmls")
116
 
117
- for image_file in sorted(os.listdir(country_images_dir)): # Sort to ensure consistent order
 
118
  if not image_file.endswith('.jpg'):
119
  continue
120
 
121
  image_id = f"{country_dir}_{image_file.split('.')[0]}"
122
- image_path = os.path.join(country_images_dir, image_file)
123
- if country_annotations_dir:
124
  annotation_file = image_id + '.xml'
125
- annotation_path = os.path.join(country_annotations_dir, annotation_file)
126
  if not os.path.exists(annotation_path):
127
  continue
128
  tree = ET.parse(annotation_path)
@@ -142,7 +139,7 @@ class RDD2020_Dataset(datasets.GeneratorBasedBuilder):
142
  else:
143
  crack_type = []
144
  crack_coordinates = []
145
-
146
  image_resolution = {"width": 600, "height": 600, "depth": 3}
147
  yield image_id, {
148
  "image_id": image_id,
@@ -153,5 +150,5 @@ class RDD2020_Dataset(datasets.GeneratorBasedBuilder):
153
  "crack_type": crack_type,
154
  "crack_coordinates": crack_coordinates,
155
  }
156
-
157
 
 
75
  )
76
 
77
  def _split_generators(self, dl_manager):
78
+ # Assuming the URLs are the GitHub raw URLs to the country directories within each split
79
+ train_dir = "https://github.com/ShixuanAn/RDD_2020/tree/main/train"
80
+ test1_dir = "https://github.com/ShixuanAn/RDD_2020/tree/main/test/test/test1"
81
+ test2_dir = "https://github.com/ShixuanAn/RDD_2020/tree/main/test/test/test2"
82
+
83
  return [
84
  datasets.SplitGenerator(
85
  name=datasets.Split.TRAIN,
86
  gen_kwargs={
87
+ "base_path": train_dir,
 
88
  "split": "train",
89
  },
90
  ),
91
  datasets.SplitGenerator(
92
  name="test1",
93
  gen_kwargs={
94
+ "base_path": test1_dir,
 
95
  "split": "test1",
96
  },
97
  ),
98
  datasets.SplitGenerator(
99
  name="test2",
100
  gen_kwargs={
101
+ "base_path": test2_dir,
 
102
  "split": "test2",
103
  },
104
  ),
105
  ]
106
 
107
+ def _generate_examples(self, base_path, split):
108
+ # Iterate over each country directory
109
+ for country_dir in ['Czech', 'India', 'Japan']:
110
+ images_dir = f"{base_path}/{country_dir}/images"
111
+ annotations_dir = f"{base_path}/{country_dir}/annotations/xmls" if split == "train" else None
 
 
 
 
 
112
 
113
+ # Iterate over each image in the country's image directory
114
+ for image_file in os.listdir(images_dir):
115
  if not image_file.endswith('.jpg'):
116
  continue
117
 
118
  image_id = f"{country_dir}_{image_file.split('.')[0]}"
119
+ image_path = os.path.join(images_dir, image_file)
120
+ if annotations_dir:
121
  annotation_file = image_id + '.xml'
122
+ annotation_path = os.path.join(annotations_dir, annotation_file)
123
  if not os.path.exists(annotation_path):
124
  continue
125
  tree = ET.parse(annotation_path)
 
139
  else:
140
  crack_type = []
141
  crack_coordinates = []
142
+
143
  image_resolution = {"width": 600, "height": 600, "depth": 3}
144
  yield image_id, {
145
  "image_id": image_id,
 
150
  "crack_type": crack_type,
151
  "crack_coordinates": crack_coordinates,
152
  }
153
+
154