Fix huggingface filesystem repo_type not forwarded (#1791)

huggingface · Oct 31, 2023 · e0e2c16 · e0e2c16
1 parent 7b50c1b
commit e0e2c16
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 11 deletions.
diff --git a/src/huggingface_hub/hf_file_system.py b/src/huggingface_hub/hf_file_system.py
@@ -424,6 +424,7 @@ def _fetch_range(self, start: int, end: int) -> bytes:
             repo_id=self.resolved_path.repo_id,
             revision=self.resolved_path.revision,
             filename=self.resolved_path.path_in_repo,
+            repo_type=self.resolved_path.repo_type,
             endpoint=self.fs.endpoint,
         )
         r = http_backoff("GET", url, headers=headers)

diff --git a/tests/test_hf_file_system.py b/tests/test_hf_file_system.py
@@ -6,11 +6,10 @@
 import fsspec
 import pytest
 
-from huggingface_hub.constants import REPO_TYPES_URL_PREFIXES
 from huggingface_hub.hf_file_system import HfFileSystem
 from huggingface_hub.utils import RepositoryNotFoundError, RevisionNotFoundError
 
-from .testing_constants import ENDPOINT_STAGING, TOKEN, USER
+from .testing_constants import ENDPOINT_STAGING, TOKEN
 from .testing_utils import repo_name
 
 
@@ -22,36 +21,37 @@ def setUpClass(cls):
             fsspec.register_implementation(HfFileSystem.protocol, HfFileSystem)
 
     def setUp(self):
-        self.repo_id = f"{USER}/{repo_name()}"
-        self.repo_type = "dataset"
-        self.hf_path = REPO_TYPES_URL_PREFIXES.get(self.repo_type, "") + self.repo_id
         self.hffs = HfFileSystem(endpoint=ENDPOINT_STAGING, token=TOKEN)
         self.api = self.hffs._api
 
         # Create dummy repo
-        self.api.create_repo(self.repo_id, repo_type=self.repo_type)
+        repo_url = self.api.create_repo(repo_name(), repo_type="dataset")
+        self.repo_id = repo_url.repo_id
+        self.hf_path = f"datasets/{self.repo_id}"
+
+        # Upload files
         self.api.upload_file(
             path_or_fileobj=b"dummy binary data on pr",
             path_in_repo="data/binary_data_for_pr.bin",
             repo_id=self.repo_id,
-            repo_type=self.repo_type,
+            repo_type="dataset",
             create_pr=True,
         )
         self.api.upload_file(
             path_or_fileobj="dummy text data".encode("utf-8"),
             path_in_repo="data/text_data.txt",
             repo_id=self.repo_id,
-            repo_type=self.repo_type,
+            repo_type="dataset",
         )
         self.api.upload_file(
             path_or_fileobj=b"dummy binary data",
             path_in_repo="data/binary_data.bin",
             repo_id=self.repo_id,
-            repo_type=self.repo_type,
+            repo_type="dataset",
         )
 
     def tearDown(self):
-        self.api.delete_repo(self.repo_id, repo_type=self.repo_type)
+        self.api.delete_repo(self.repo_id, repo_type="dataset")
 
     def test_glob(self):
         self.assertEqual(
@@ -141,7 +141,7 @@ def test_modified_time(self):
 
     def test_initialize_from_fsspec(self):
         fs, _, paths = fsspec.get_fs_token_paths(
-            f"hf://{self.repo_type}s/{self.repo_id}/data/text_data.txt",
+            f"hf://datasets/{self.repo_id}/data/text_data.txt",
             storage_options={
                 "endpoint": ENDPOINT_STAGING,
                 "token": TOKEN,