Add download from URL

hr500k.py

@@ -31,6 +31,7 @@ are encoded as class labels.
 _HOMEPAGE = 'https://www.clarin.si/repository/xmlui/handle/11356/1183#'
 _LICENSE = ''
 
+_URL = 'https://huggingface.co/datasets/classla/hr500k'
 _TRAINING_FILE = 'train_ner.conllu'
 _DEV_FILE = 'dev_ner.conllu'
 _TEST_FILE = 'test_ner.conllu'
@@ -110,15 +111,23 @@ class Hr500K(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        urls_to_download = {
+            'train': f"{_URL}{_TRAINING_FILE}",
+            'dev': f"{_URL}{_DEV_FILE}",
+            'test': f"{_URL}{_TEST_FILE}",
+        }
+        downloaded_files = dl_manager.download_and_extract(urls_to_download)
+
         return [
             datasets.SplitGenerator(
-                name=datasets.Split.TRAIN, gen_kwargs={'filepath': _TRAINING_FILE, 'split': 'train'}
+                name=datasets.Split.TRAIN, gen_kwargs={'filepath': downloaded_files['train'], 'split': 'train'}
             ),
             datasets.SplitGenerator(
-                name=datasets.Split.VALIDATION, gen_kwargs={'filepath': _DEV_FILE, 'split': 'dev'}
+                name=datasets.Split.VALIDATION, gen_kwargs={'filepath': downloaded_files['dev'], 'split': 'dev'}
             ),
             datasets.SplitGenerator(
-                name=datasets.Split.TEST, gen_kwargs={'filepath': _TEST_FILE, 'split': 'test'}
+                name=datasets.Split.TEST, gen_kwargs={'filepath': downloaded_files['test'], 'split': 'test'}
             ),
         ]
 
@@ -178,3 +187,6 @@ class Hr500K(datasets.GeneratorBasedBuilder):
                 'iob_tags': iob_tags
             }
 
+
+dataset = datasets.load_dataset('hr500k.py', split='test')
+print(dataset.info)