Merge branch 'main' of https://huggingface.co/datasets/mb23/GraySpectrogram

GraySpectrogram.py

@@ -94,7 +94,7 @@ class GraySpectrogramConfig(datasets.BuilderConfig):
 class GraySpectrogram(datasets.GeneratorBasedBuilder):
 
     # データのサブセットはここで用意
-    data_urls, metadata_urls = get_information()
+    metadata_urls, data_urls = get_information()
     BUILDER_CONFIGS = [
         GraySpectrogramConfig(
             name="data 0-200",
@@ -195,22 +195,23 @@ class GraySpectrogram(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager: DownloadManager):
 
-        metadata_paths = dl_manager.download(self.config.metadata_url)
-        data_paths = dl_manager.download(self.config.data_url)
-
+        train_metadata_path = dl_manager.download_and_extract(self.config.metadata_url["train"])
+        test_metadata_path = dl_manager.download_and_extract(self.config.metadata_url["test"])
+        train_data_path = dl_manager.download(self.config.data_url["train"])
+        test_data_path = dl_manager.download(self.config.data_url["test"])
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 gen_kwargs={
-                    "images": dl_manager.iter_archive(data_paths["train"]),
-                    "metadata_path": metadata_paths["train"],
+                    "images": dl_manager.iter_archive(train_data_path),
+                    "metadata_path": train_metadata_path,
                 }
             ),
             datasets.SplitGenerator(
                 name=datasets.Split.TEST,
                 gen_kwargs={
-                    "images": dl_manager.iter_archive(data_paths["test"]),
-                    "metadata_path": metadata_paths["test"],
+                    "images": dl_manager.iter_archive(test_data_path),
+                    "metadata_path": test_metadata_path,
                 }
             ),
         ]
@@ -292,7 +293,7 @@ class GraySpectrogram(datasets.GeneratorBasedBuilder):
         num_list = list()
         label_list = list()
 
-        with open(metadata_path) as fin:
+        with open(metadata_path, encoding="utf-8") as fin:
             for line in fin:
                 data =  json.loads(line)
                 file_list.append(data["file_name"])