Support push_to_hub canonical datasets (huggingface#6519)

* Support push_to_hub canonical datasets

* Fix test

Author: albertvillanova

src/datasets/arrow_dataset.py

@@ -5352,14 +5352,13 @@ def push_to_hub(
 
         api = HfApi(endpoint=config.HF_ENDPOINT, token=token)
 
-        repo_url = api.create_repo(
+        _ = api.create_repo(
             repo_id,
             token=token,
             repo_type="dataset",
             private=private,
             exist_ok=True,
         )
-        repo_id = repo_url.repo_id
 
         if revision is not None:
             api.create_branch(repo_id, branch=revision, token=token, repo_type="dataset", exist_ok=True)

src/datasets/dataset_dict.py

@@ -1679,14 +1679,13 @@ def push_to_hub(
 
         api = HfApi(endpoint=config.HF_ENDPOINT, token=token)
 
-        repo_url = api.create_repo(
+        _ = api.create_repo(
             repo_id,
             token=token,
             repo_type="dataset",
             private=private,
             exist_ok=True,
         )
-        repo_id = repo_url.repo_id
 
         if revision is not None:
             api.create_branch(repo_id, branch=revision, token=token, repo_type="dataset", exist_ok=True)

tests/test_upstream_hub.py

@@ -13,6 +13,7 @@
 import numpy as np
 import pytest
 from huggingface_hub import DatasetCard, HfApi
+from huggingface_hub.utils import RepositoryNotFoundError
 
 from datasets import (
     Audio,
@@ -70,16 +71,9 @@ def test_push_dataset_dict_to_hub_name_without_namespace(self, temporary_repo):
         local_ds = DatasetDict({"train": ds})
 
         with temporary_repo() as ds_name:
-            local_ds.push_to_hub(ds_name.split("/")[-1], token=self._token)
-            hub_ds = load_dataset(ds_name, download_mode="force_redownload")
-
-            assert local_ds.column_names == hub_ds.column_names
-            assert list(local_ds["train"].features.keys()) == list(hub_ds["train"].features.keys())
-            assert local_ds["train"].features == hub_ds["train"].features
-
-            # Ensure that there is a single file on the repository that has the correct name
-            files = sorted(self._api.list_repo_files(ds_name, repo_type="dataset"))
-            assert files == [".gitattributes", "README.md", "data/train-00000-of-00001.parquet"]
+            # cannot create a repo without namespace
+            with pytest.raises(RepositoryNotFoundError):
+                local_ds.push_to_hub(ds_name.split("/")[-1], token=self._token)
 
     def test_push_dataset_dict_to_hub_datasets_with_different_features(self, cleanup_repo):
         ds_train = Dataset.from_dict({"x": [1, 2, 3], "y": [4, 5, 6]})