Added support for Wildcards in huggingface-cli upload (#2868)

* Added support for Wildcards in huggingface-cli upload

* Made some changes in line endings

* some changes in test cli

* some changes in test cli

* Made changes in upload.py based on suggestions

* Update handling of wildcard arguments

* fix wildcard support and improve test

* fix

* Update src/huggingface_hub/commands/upload.py

Co-authored-by: Lucain <[email protected]>

* Update src/huggingface_hub/commands/upload.py

Co-authored-by: Lucain <[email protected]>

---------

Co-authored-by: Celina Hanouti <[email protected]>
Co-authored-by: Lucain <[email protected]>

Author: 3 people

src/huggingface_hub/commands/upload.py

@@ -30,6 +30,9 @@
     # Upload filtered directory (example: tensorboard logs except for the last run)
     huggingface-cli upload my-cool-model ./model/training /logs --include "*.tfevents.*" --exclude "*20230905*"
 
+    # Upload with wildcard
+    huggingface-cli upload my-cool-model "./model/training/*.safetensors"
+
     # Upload private dataset
     huggingface-cli upload Wauplin/my-cool-dataset ./data . --repo-type=dataset --private
 
@@ -69,7 +72,9 @@ def register_subcommand(parser: _SubParsersAction):
             "repo_id", type=str, help="The ID of the repo to upload to (e.g. `username/repo-name`)."
         )
         upload_parser.add_argument(
-            "local_path", nargs="?", help="Local path to the file or folder to upload. Defaults to current directory."
+            "local_path",
+            nargs="?",
+            help="Local path to the file or folder to upload. Wildcard patterns are supported. Defaults to current directory.",
         )
         upload_parser.add_argument(
             "path_in_repo",
@@ -155,7 +160,16 @@ def __init__(self, args: Namespace) -> None:
         repo_name: str = args.repo_id.split("/")[-1]  # e.g. "Wauplin/my-cool-model" => "my-cool-model"
         self.local_path: str
         self.path_in_repo: str
-        if args.local_path is None and os.path.isfile(repo_name):
+
+        if args.local_path is not None and any(c in args.local_path for c in ["*", "?", "["]):
+            if args.include is not None:
+                raise ValueError("Cannot set `--include` when passing a `local_path` containing a wildcard.")
+            if args.path_in_repo is not None and args.path_in_repo != ".":
+                raise ValueError("Cannot set `path_in_repo` when passing a `local_path` containing a wildcard.")
+            self.local_path = "."
+            self.include = args.local_path
+            self.path_in_repo = "."
+        elif args.local_path is None and os.path.isfile(repo_name):
             # Implicit case 1: user provided only a repo_id which happen to be a local file as well => upload it with same name
             self.local_path = repo_name
             self.path_in_repo = repo_name

tests/test_cli.py

@@ -91,6 +91,42 @@ def test_upload_basic(self) -> None:
         self.assertEqual(cmd.api.token, None)
         self.assertEqual(cmd.quiet, False)
 
+    def test_upload_with_wildcard(self) -> None:
+        """Test uploading files using wildcard patterns."""
+        with tmp_current_directory() as cache_dir:
+            # Create test files
+            (Path(cache_dir) / "model1.safetensors").touch()
+            (Path(cache_dir) / "model2.safetensors").touch()
+            (Path(cache_dir) / "model.bin").touch()
+            (Path(cache_dir) / "config.json").touch()
+
+            # Test basic wildcard pattern
+            cmd = UploadCommand(self.parser.parse_args(["upload", DUMMY_MODEL_ID, "*.safetensors"]))
+            self.assertEqual(cmd.local_path, ".")
+            self.assertEqual(cmd.include, "*.safetensors")
+            self.assertEqual(cmd.path_in_repo, ".")
+            self.assertEqual(cmd.repo_id, DUMMY_MODEL_ID)
+
+            # Test wildcard pattern with specific directory
+            subdir = Path(cache_dir) / "subdir"
+            subdir.mkdir()
+            (subdir / "special.safetensors").touch()
+
+            cmd = UploadCommand(self.parser.parse_args(["upload", DUMMY_MODEL_ID, "subdir/*.safetensors"]))
+            self.assertEqual(cmd.local_path, ".")
+            self.assertEqual(cmd.include, "subdir/*.safetensors")
+            self.assertEqual(cmd.path_in_repo, ".")
+
+            # Test error when using wildcard with --include
+            with self.assertRaises(ValueError):
+                UploadCommand(
+                    self.parser.parse_args(["upload", DUMMY_MODEL_ID, "*.safetensors", "--include", "*.json"])
+                )
+
+            # Test error when using wildcard with explicit path_in_repo
+            with self.assertRaises(ValueError):
+                UploadCommand(self.parser.parse_args(["upload", DUMMY_MODEL_ID, "*.safetensors", "models/"]))
+
     def test_upload_with_all_options(self) -> None:
         """Test `huggingface-cli upload my-file to dummy-repo with all options selected`."""
         cmd = UploadCommand(