refactor: script

license_plates.py

@@ -66,29 +66,52 @@ class LicensePlates(datasets.GeneratorBasedBuilder):
         data = dl_manager.iter_archive(data)
         return [
             datasets.SplitGenerator(name=datasets.Split.TRAIN,
-                                    gen_kwargs={
-                                        "data": data
-                                    }),
+                                    gen_kwargs={"data": data}),
         ]
 
     def _generate_examples(self, data):
 
         annotations = []
-        images = []
+        images = {}
 
         for idx, (file_path, file) in enumerate(data):
             if file_path.endswith('.csv'):
-                annotations
+                df = pd.read_csv(file_path, sep=',')
+                annotations.append(df)
+
+            else:
+                file_name = file_path.split('/')[-1]
+                images[file_name] = (file_path, file.read())
+
+        if len(annotations) > 1:
+            annotations_df = pd.concat(annotations)
+        else:
+            annotations_df = annotations[0]
+
+        annotations_df.drop(
+            columns=['license_plate.region', 'license_plate.color'],
+            inplace=True)
+
+        for row in annotations_df.itertuples(index=False):
+            image = images[row[0]]
+            name, ext = row[0].split('.')
+            labeled_image = images[f'{name}_labeled.{ext}']
+
             yield idx, {
                 "image": {
-                    "path": image_path,
-                    "bytes": image.read()
+                    "path": image[0],
+                    "bytes": image[1]
                 },
-                "mask": {
-                    "path": mask_path,
-                    "bytes": mask.read()
+                "labeled_image": {
+                    "path": labeled_image[0],
+                    "bytes": labeled_image[1]
                 },
-                'id': annotations_df['id'].iloc[idx],
-                'gender': annotations_df['gender'].iloc[idx],
-                'age': annotations_df['age'].iloc[idx]
+                'bbox': row[1],
+                'license_plate.id': row[2],
+                'license_plate.visibility': row[3],
+                'license_plate.rows_count': row[4],
+                'license_plate.number': row[5],
+                'license_plate.serial': row[6],
+                'license_plate.country': row[7],
+                'license_plate.mask': row[8]
             }