loading script

dusha_emotion_audio.py

@@ -35,16 +35,16 @@ class Dusha(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager):
         metadata = dl_manager.download(_METADATA_URL)
-        archive = dl_manager.download_and_extract(_DATA_URL)
+        archive = dl_manager.download(_DATA_URL)
         return [datasets.SplitGenerator(
             name=datasets.Split.ALL,
             gen_kwargs={
-                "audio_files": archive,
+                "audio_files": dl_manager.iter_archive(archive),
                 "metadata": metadata},
         )]
 
     def _generate_examples(self, audio_files, metadata):
-        examples = list()
+        examples = dict()
 
         with open(metadata, encoding="utf-8") as f:
             csv_reader = csv.reader(f, delimiter=",")
@@ -52,15 +52,14 @@ class Dusha(datasets.GeneratorBasedBuilder):
             for row in csv_reader:
                 audio_path, label = row
                 full_audio_path = os.path.join(audio_files, audio_path)
-                with open(full_audio_path) as audio_file:
-                    audio = {"path": full_audio_path, "bytes": audio_file.read()}
                 res = dict()
                 res["file"] = full_audio_path
-                res["audio"] = audio
                 res["label"] = label
-                examples.append(res)
+                examples[full_audio_path](res)
 
         key = 0
-        for example in examples:
-            yield key, example
-            key += 1
+        for path, f in audio_files:
+            if path in examples:
+                audio = {"path": path, "bytes": f.read()}
+                yield key, {**examples[path], "audio": audio}
+                key += 1