Update hugging_face.py

hugging_face.py

@@ -72,7 +72,6 @@ class RDD2020_Dataset(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        # The URL provided must be the direct link to the zip file
         urls_to_download = {
             "dataset": "https://huggingface.co/datasets/ShixuanAn/RDD2020/resolve/main/RDD2020.zip"
         }
@@ -156,3 +155,49 @@ class RDD2020_Dataset(datasets.GeneratorBasedBuilder):
                     "crack_type": crack_type,
                     "crack_coordinates": crack_coordinates,
                 }
+    def _generate_examples(self, base_path, split):
+    # Iterate over each country directory
+         for country_dir in ['Czech', 'India', 'Japan']:
+            images_dir = f"{base_path}/{country_dir}/images"
+            annotations_dir = f"{base_path}/{country_dir}/annotations/xmls" if split == "train" else None
+
+            for image_file in os.listdir(images_dir):
+                if image_file.endswith('.jpg'):
+                    image_id = f"{country_dir}_{image_file.split('.')[0]}"
+                    image_path = os.path.join(images_dir, image_file)
+
+                    # Initialize the lists to store the parsed data
+                    crack_types = []
+                    crack_coordinates = []
+
+                    if annotations_dir:
+                        annotation_file = image_id + '.xml'
+                        annotation_path = os.path.join(annotations_dir, annotation_file)
+                        if os.path.exists(annotation_path):
+                            # Parse the XML file
+                            tree = ET.parse(annotation_path)
+                            root = tree.getroot()
+
+                            # Extract the information
+                            for obj in root.findall('object'):
+                                crack_type = obj.find('name').text
+                                crack_types.append(crack_type)
+                                bndbox = obj.find('bndbox')
+                                coordinates = {
+                                    "x_min": int(bndbox.find('xmin').text),
+                                    "x_max": int(bndbox.find('xmax').text),
+                                    "y_min": int(bndbox.find('ymin').text),
+                                    "y_max": int(bndbox.find('ymax').text),
+                                }
+                                crack_coordinates.append(coordinates)
+  
+    
+                # Yield the example as a dictionary
+                    yield image_id, {
+                        "image_id": image_id,
+                        "country": country_dir,
+                        "type": split,
+                        "image_path": image_path,
+                        "crack_type": crack_types,
+                        "crack_coordinates": crack_coordinates,
+                    }