Allow sys.argv and args in LightningCLI (#16808)

(cherry picked from commit 0ec44f1)

src/pytorch_lightning/cli.py

@@ -509,9 +509,9 @@ def link_optimizers_and_lr_schedulers(parser: LightningArgumentParser) -> None:
     def parse_arguments(self, parser: LightningArgumentParser, args: ArgsType) -> None:
         """Parses command line arguments and stores it in ``self.config``."""
         if args is not None and len(sys.argv) > 1:
-            raise ValueError(
+            rank_zero_warn(
                 "LightningCLI's args parameter is intended to run from within Python like if it were from the command "
-                "line. To prevent mistakes it is not allowed to provide both args and command line arguments, got: "
+                "line. To prevent mistakes it is not recommended to provide both args and command line arguments, got: "
                 f"sys.argv[1:]={sys.argv[1:]}, args={args}."
             )
         if isinstance(args, (dict, Namespace)):

tests/tests_pytorch/test_cli.py

@@ -1547,8 +1547,6 @@ def test_lightning_cli_with_args_given(args):
     assert cli.model.foo == 456
 
 
-def test_lightning_cli_args_and_sys_argv_exception():
-    with mock.patch("sys.argv", ["", "--model.foo=456"]), pytest.raises(
-        ValueError, match="LightningCLI's args parameter "
-    ):
+def test_lightning_cli_args_and_sys_argv_warning():
+    with mock.patch("sys.argv", ["", "--model.foo=456"]), pytest.warns(Warning, match="LightningCLI's args parameter "):
         LightningCLI(TestModel, run=False, args=["--model.foo=789"])