adds download of tar gz to dataset script

README.md

@@ -555,7 +555,7 @@ dataset_info:
   - name: test
     num_bytes: 60898170
     num_examples: 2625
-  download_size: 0
+  download_size: 1984870735
   dataset_size: 2031172375
 ---
 

bird-species-dataset.py

@@ -20,6 +20,8 @@ _DATA_DIR = 'data/'
 
 _VERSION = "0.1.0"
 
+_REPO = "https://huggingface.co/datasets/chriamue/bird-species-dataset"
+
 def _CLASSES() -> list[str]:
 
     classes = """ABBOTTS BABBLER
@@ -584,10 +586,14 @@ class BirdSpeciesDataset(datasets.GeneratorBasedBuilder):
     
     def _split_generators(self, dl_manager):
         data_dir = _DATA_DIR
+        train_archive = dl_manager.download(f"{_REPO}/resolve/main/data/train.tar.gz")
+        valid_archive = dl_manager.download(f"{_REPO}/resolve/main/data/valid.tar.gz")
+        test_archive = dl_manager.download(f"{_REPO}/resolve/main/data/test.tar.gz")
+
         return [
-            SplitGenerator(name=Split.TRAIN, gen_kwargs={"split": "train", "files": dl_manager.iter_archive(os.path.join(data_dir, "train.tar.gz"))}),
-            SplitGenerator(name=Split.VALIDATION, gen_kwargs={"split": "valid", "files": dl_manager.iter_archive(os.path.join(data_dir, "valid.tar.gz"))}),
-            SplitGenerator(name=Split.TEST, gen_kwargs={"split": "test", "files": dl_manager.iter_archive(os.path.join(data_dir, "test.tar.gz"))}),
+            SplitGenerator(name=Split.TRAIN, gen_kwargs={"split": "train", "files": dl_manager.iter_archive(train_archive)}),
+            SplitGenerator(name=Split.VALIDATION, gen_kwargs={"split": "valid", "files": dl_manager.iter_archive(valid_archive)}),
+            SplitGenerator(name=Split.TEST, gen_kwargs={"split": "test", "files": dl_manager.iter_archive(test_archive)}),
         ]
 
     def _generate_examples(self, files, split):