Warn if user tries to upload a parquet file to a model repo

src/huggingface_hub/hf_api.py

@@ -3733,6 +3733,15 @@ def create_commit(
                     " new CommitOperationAdd object if you want to create a new commit."
                 )
 
+        if repo_type != "dataset":
+            for addition in additions:
+                if addition.path_in_repo.endswith((".arrow", ".parquet")):
+                    warnings.warn(
+                        f"It seems that you are about to commit a data file ({addition.path_in_repo}) to a {repo_type}"
+                        " repository. You are sure this is intended? If you are trying to upload a dataset, please"
+                        " set `repo_type='dataset'` or `--repo-type=dataset` in a CLI."
+                    )
+
         logger.debug(
             f"About to commit to the hub: {len(additions)} addition(s), {len(copies)} copie(s) and"
             f" {nb_deletions} deletion(s)."

tests/test_hf_api.py

@@ -379,6 +379,33 @@ def test_upload_file_bytesio(self, repo_url: RepoUrl) -> None:
             with open(hf_hub_download(repo_id=repo_id, filename="temp/new_file.md", cache_dir=cache_dir)) as f:
                 self.assertEqual(f.read(), content.getvalue().decode())
 
+    @use_tmp_repo()
+    def test_upload_data_files_to_model_repo(self, repo_url: RepoUrl) -> None:
+        # If a .parquet file is uploaded to a model repo, it should be uploaded correctly but a warning is raised.
+        with self.assertWarns(UserWarning) as cm:
+            self._api.upload_file(
+                path_or_fileobj=b"content",
+                path_in_repo="data.parquet",
+                repo_id=repo_url.repo_id,
+            )
+        assert (
+            cm.warnings[0].message.args[0]
+            == "It seems that you are about to commit a data file (data.parquet) to a model repository. You are sure this is intended? If you are trying to upload a dataset, please set `repo_type='dataset'` or `--repo-type=dataset` in a CLI."
+        )
+
+        # Same for arrow file
+        with self.assertWarns(UserWarning) as cm:
+            self._api.upload_file(
+                path_or_fileobj=b"content",
+                path_in_repo="data.arrow",
+                repo_id=repo_url.repo_id,
+            )
+
+        # Still correctly uploaded
+        files = self._api.list_repo_files(repo_url.repo_id)
+        assert "data.parquet" in files
+        assert "data.arrow" in files
+
     def test_create_repo_return_value(self) -> None:
         REPO_NAME = repo_name("org")
         url = self._api.create_repo(repo_id=REPO_NAME)