Prepare tests for hfh 0.14 (#5788)

* test hfh 0.14.0rc1

* fix style

* fix tests for hfh>=0.14

* fix test token

* package version

* Remove CI run on push to test-hfh* branch

---------

Co-authored-by: Quentin Lhoest <lhoest.q@gmail.com>
Co-authored-by: testbot <lucainp@hf.co>
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>

tests/test_load.py

@@ -852,19 +852,19 @@ def test_loading_from_the_datasets_hub():
 
 @pytest.mark.integration
 def test_loading_from_the_datasets_hub_with_use_auth_token():
-    from requests import get
+    true_request = requests.Session().request
 
-    def assert_auth(url, *args, headers, **kwargs):
+    def assert_auth(method, url, *args, headers, **kwargs):
         assert headers["authorization"] == "Bearer foo"
-        return get(url, *args, headers=headers, **kwargs)
+        return true_request(method, url, *args, headers=headers, **kwargs)
 
-    with patch("requests.get") as mock_head:
-        mock_head.side_effect = assert_auth
+    with patch("requests.Session.request") as mock_request:
+        mock_request.side_effect = assert_auth
         with tempfile.TemporaryDirectory() as tmp_dir:
             with offline():
                 with pytest.raises((ConnectionError, requests.exceptions.ConnectionError)):
                     load_dataset(SAMPLE_NOT_EXISTING_DATASET_IDENTIFIER, cache_dir=tmp_dir, use_auth_token="foo")
-        mock_head.assert_called()
+        mock_request.assert_called()
 
 
 @pytest.mark.integration

tests/utils.py

@@ -14,6 +14,7 @@
 
 import pyarrow as pa
 import pytest
+import requests
 from packaging import version
 
 from datasets import config
@@ -322,9 +323,9 @@ def offline(mode=OfflineSimulationMode.CONNECTION_FAILS, timeout=1e-16):
     HF_DATASETS_OFFLINE_SET_TO_1: the HF_DATASETS_OFFLINE environment variable is set to 1.
         This makes the http/ftp calls of the library instantly fail and raise an OfflineModeEmabled error.
     """
-    from requests import request as online_request
+    online_request = requests.Session().request
 
-    def timeout_request(method, url, **kwargs):
+    def timeout_request(session, method, url, **kwargs):
         # Change the url to an invalid url so that the connection hangs
         invalid_url = "https://10.255.255.1"
         if kwargs.get("timeout") is None:
@@ -342,18 +343,16 @@ def timeout_request(method, url, **kwargs):
             e.args = (max_retry_error,)
             raise
 
-    def offline_socket(*args, **kwargs):
-        raise OSError("Offline mode is enabled.")
+    def raise_connection_error(session, prepared_request, **kwargs):
+        raise requests.ConnectionError("Offline mode is enabled.", request=prepared_request)
 
     if mode is OfflineSimulationMode.CONNECTION_FAILS:
-        # inspired from https://stackoverflow.com/a/18601897
-        with patch("socket.socket", offline_socket):
+        with patch("requests.Session.send", raise_connection_error):
             yield
     elif mode is OfflineSimulationMode.CONNECTION_TIMES_OUT:
         # inspired from https://stackoverflow.com/a/904609
-        with patch("requests.request", timeout_request):
-            with patch("requests.api.request", timeout_request):
-                yield
+        with patch("requests.Session.request", timeout_request):
+            yield
     elif mode is OfflineSimulationMode.HF_DATASETS_OFFLINE_SET_TO_1:
         with patch("datasets.config.HF_DATASETS_OFFLINE", True):
             yield