Debug dataset loading script

xd-violence.py

@@ -76,7 +76,7 @@ class XDViolence(datasets.GeneratorBasedBuilder):
         if self.config.name == "rgb":
             features = datasets.Features(
                 {
-                    "video_id": datasets.Value("string"),
+                    "id": datasets.Value("string"),
                     "rgb_feats": datasets.Array3D(
                         shape=(None, 10, 2048),
                         dtype="float32",  # (num_frames, num_crops, feature_dim) use 10 crops by default as of now
@@ -108,8 +108,8 @@ class XDViolence(datasets.GeneratorBasedBuilder):
         else:  # default = "video"
             features = datasets.Features(
                 {
-                    "video_id": datasets.Value("string"),
-                    "video_path": datasets.Value("string"),
+                    "id": datasets.Value("string"),
+                    "path": datasets.Value("string"),
                     "binary_target": datasets.ClassLabel(
                         names=["Non-violence", "Violence"]
                     ),
@@ -164,8 +164,8 @@ class XDViolence(datasets.GeneratorBasedBuilder):
                     header=None,
                     sep=" ",
                     usecols=[0],
-                    names=["video_id"],
-                )["video_id"]
+                    names=["id"],
+                )["id"]
                 .apply(
                     lambda x: urllib.parse.quote(
                         urllib.parse.urljoin(_URL, f"video/{x}.mp4"), safe=":/"
@@ -177,8 +177,8 @@ class XDViolence(datasets.GeneratorBasedBuilder):
                     header=None,
                     sep=" ",
                     usecols=[0],
-                    names=["video_id"],
-                )["video_id"]
+                    names=["id"],
+                )["id"]
                 .apply(
                     lambda x: urllib.parse.quote(
                         urllib.parse.urljoin(_URL, f"video/test_videos/{x}.mp4"),
@@ -220,19 +220,15 @@ class XDViolence(datasets.GeneratorBasedBuilder):
     def _generate_examples(self, annotation_path, video_paths, annotation_reader):
         ann_data = annotation_reader(annotation_path)
 
-        for key, (video_path, annotation) in enumerate(zip(video_paths, ann_data)):
-            video_id = annotation["video_id"]
+        for key, (path, annotation) in enumerate(zip(video_paths, ann_data)):
+            id = annotation["id"]
             frame_annotations = annotation.get("frame_annotations", [])
 
-            try:
-                binary, multilabel = self.extract_labels(video_id)
-            except Exception as e:
-                print(f"Error processing video {video_id}.")
-                raise e
+            binary, multilabel = self.extract_labels(id)
 
             yield key, {
-                "video_id": video_id,
-                "video_path": video_path,
+                "id": id,
+                "path": path,
                 "binary_target": binary,
                 "multilabel_targets": multilabel,
                 "frame_annotations": frame_annotations,
@@ -242,10 +238,8 @@ class XDViolence(datasets.GeneratorBasedBuilder):
     def _read_train_list(path):
         """Reads the train_list.txt file and returns a list of video ids."""
 
-        train_list = pd.read_csv(
-            path, header=None, sep=" ", usecols=[0], names=["video_id"]
-        )
-        train_list["video_id"] = train_list["video_id"].apply(lambda x: x.split("/")[1])
+        train_list = pd.read_csv(path, header=None, sep=" ", usecols=[0], names=["id"])
+        train_list["id"] = train_list["id"].apply(lambda x: x.split("/")[1])
         return train_list.to_dict("records")
 
     @staticmethod
@@ -265,7 +259,7 @@ class XDViolence(datasets.GeneratorBasedBuilder):
 
                     annotations.append(
                         {
-                            "video_id": parts[0],
+                            "id": parts[0],
                             "frame_annotations": [
                                 {"start": parts[start_idx], "end": parts[start_idx + 1]}
                                 for start_idx in range(1, len(parts), 2)
@@ -279,7 +273,7 @@ class XDViolence(datasets.GeneratorBasedBuilder):
 
                     annotations.append(
                         {
-                            "video_id": parts[0],
+                            "id": parts[0],
                             "frame_annotations": [
                                 {"start": parts[start_idx], "end": parts[start_idx + 1]}
                                 for start_idx in range(1, len(parts), 2)