upd: sharding problem try

rwth_phoenix_weather_2014.py

@@ -97,9 +97,8 @@ class RWTHPhoenixWeather2014(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager: datasets.DownloadManager):
-        example_ids = {}
-        annotations = {}
         frames = {}
+        other_data = {}
 
         dataDirMapper = {
             datasets.Split.TRAIN: "train",
@@ -119,12 +118,12 @@ class RWTHPhoenixWeather2014(datasets.GeneratorBasedBuilder):
 
             df = pd.read_csv(data_csv, sep='|')
 
-            example_ids[split] = df['id']
-            annotations[split] = df['annotation']
+            example_ids = df['id']
+            annotations = df['annotation']
 
             frame_archive_urls = dl_manager.download([
                 f"{base_url}/features/fullFrame-210x260px/{dataDirMapper[split]}/{id}.tar"
-                for id in example_ids[split]
+                for id in example_ids
             ])
 
             frames[split] = [
@@ -132,13 +131,22 @@ class RWTHPhoenixWeather2014(datasets.GeneratorBasedBuilder):
                 for url in frame_archive_urls
             ]
 
+            other_data_split = {}
+
+            for frame, idx, annotation, in zip(frames[split], example_ids, annotations):
+                other_data_split[frame] = {
+                    "id": idx,
+                    "annotation": annotation,
+                }
+
+            other_data[split] = other_data_split
+
         return [
             datasets.SplitGenerator(
                 name=split,
                 gen_kwargs={
-                    "example_ids": example_ids[split],
-                    "annotations": annotations[split],
-                    "frames": frames[split],
+                    "frame_archives": frames[split],
+                    "other_data": other_data[split],
                 },
             )
             for split in [
@@ -148,22 +156,21 @@ class RWTHPhoenixWeather2014(datasets.GeneratorBasedBuilder):
             ]
         ]
 
-    def _generate_examples(self, example_ids, annotations, frames):
-        print(f"example_ids len: {len(example_ids)}")
-        print(f"annotations len: {len(annotations)}")
-        print(f"frames len: {len(frames)}")
+    def _generate_examples(self, frame_archives: list, other_data: dict):
+        print(f"frames len: {len(frame_archives)}")
+        print(f"other_data: {len(other_data)}")
+
+        for key, frames in enumerate(frame_archives):
+            ex = other_data[frames]
 
-        for key, (idx, annotation, frames_list) in enumerate(zip(example_ids, annotations, frames)):
             result = {
-                "id": idx,
-                "transcription": annotation,
+                "id": ex['id'],
+                "transcription": ex['annotation'],
             }
 
-            print(frames[0])
-
             if self.config.name != 'pre-training':
                 result["frames"] = [
-                    {"path": p, "bytes": im.read()} for p, im in frames_list
+                    {"path": p, "bytes": im.read()} for p, im in frames
                 ]
 
             yield key, result