New loading method

beatbox.py

@@ -56,36 +56,41 @@ class BeatboxDataset(datasets.GeneratorBasedBuilder):
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
         dl_manager.download_config.ignore_url_params = True
-        audio_path = {}
-        local_extracted_archive = {}
-        metadata_path = {}
-        split_type = {
-            "train": datasets.Split.TRAIN,
-            "test": datasets.Split.TEST
-        }
-        for split in split_type:
-            audio_path[split] = dl_manager.download(
-                f"{_DATA_URL}/audio_{split}.tar.gz")
-            local_extracted_archive[split] = dl_manager.extract(
-                audio_path[split]) if not dl_manager.is_streaming else None
-            metadata_path[split] = dl_manager.download_and_extract(
-                f"{_DATA_URL}/metadata_{split}.csv.gz")
-        path_to_clips = "."
+
+        audio_path = dl_manager.download(_DATA_URL)
+        local_extracted_archive = dl_manager.extract(
+            audio_path) if not dl_manager.is_streaming else None
+        path_to_clips = "dataset"
 
         return [
             datasets.SplitGenerator(
-                name=split_type[split],
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "local_extracted_archive":
+                    local_extracted_archive,
+                    "audio_files":
+                    dl_manager.iter_archive(audio_path),
+                    "metadata_path":
+                    dl_manager.download_and_extract(
+                        "dataset/metadata_train.csv.gz"),
+                    "path_to_clips":
+                    path_to_clips,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
                 gen_kwargs={
                     "local_extracted_archive":
-                    local_extracted_archive[split],
+                    local_extracted_archive,
                     "audio_files":
-                    dl_manager.iter_archive(audio_path[split]),
+                    dl_manager.iter_archive(audio_path),
                     "metadata_path":
-                    dl_manager.download_and_extract(metadata_path[split]),
+                    dl_manager.download_and_extract(
+                        "dataset/metadata_test.csv.gz"),
                     "path_to_clips":
                     path_to_clips,
                 },
-            ) for split in split_type
+            ),
         ]
 
     def _generate_examples(
@@ -101,13 +106,12 @@ class BeatboxDataset(datasets.GeneratorBasedBuilder):
         with open(metadata_path, "r", encoding="utf-8") as f:
             reader = csv.DictReader(f)
             for row in reader:
-                if self.config.name == "_all_":
-                    row["path"] = os.path.join(path_to_clips, row["path"])
-                    # if data is incomplete, fill with empty values
-                    for field in data_fields:
-                        if field not in row:
-                            row[field] = ""
-                    metadata[row["path"]] = row
+                row["path"] = os.path.join(path_to_clips, row["path"])
+                # if data is incomplete, fill with empty values
+                for field in data_fields:
+                    if field not in row:
+                        row[field] = ""
+                metadata[row["path"]] = row
         id_ = 0
         for path, f in audio_files:
             if path in metadata: