toy loading script

Birdsnap.py

@@ -45,7 +45,7 @@ _HOMEPAGE = ""
 # TODO: Add the licence for the dataset here if you can find it
 _LICENSE = ""
 
-_URL = {"images": "images/{archive_name}", "annotations": "annotations.tar.gz"}
+_URL = {"images_test": "images/Acadian_Flycatcher.tar.gz", "images": "images/{archive_name}", "annotations": "annotations.tar.gz"}
 # "http://thomasberg.org/datasets/birdsnap/1.1/birdsnap.tgz"
 
 _IMAGES_ARCHIVES = [
@@ -3002,12 +3002,13 @@ class BirdsnapDataset(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        _URL["images"] = [
+        _URL["images"] = {archive_name:
             _URL["images"].format(archive_name=archive_name)
-            for archive_name in _IMAGES_ARCHIVES
-        ]
+            for archive_name in _IMAGES_ARCHIVES[:4] #TODO change, reduced hare to play with it
+        }
 
-        data_dir = dl_manager.download_and_extract(_URL)
+        data_dir = dl_manager.download(_URL)
+        data_dir = dl_manager.extract(data_dir)
         images_dir = Path(data_dir["images"]) / "images"
         annotations_dir = Path(data_dir["annotations"])
         with open(annotations_dir / "test_images.txt", "r") as f: