fix(_transform): guard IndexError for bare dict annotation

[devteamaegis]

What's broken

Calling transform() on a TypedDict whose field is annotated with a bare, unparameterised dict (e.g. metadata: dict) raises an IndexError. The bug affects both the sync and async transform paths.

class TestParams(TypedDict, total=False):
    metadata: dict  # no type parameters

transform({"metadata": {"key": "value"}}, TestParams)
# IndexError: tuple index out of range

Why it happens

_transform_recursive and _async_transform_recursive both unconditionally access get_args(stripped_type)[1] when origin == dict. For a bare dict, get_args(dict) returns an empty tuple, so index 1 does not exist.

Fix

Store args = get_args(stripped_type) and only recurse with args[1] when len(args) >= 2. When no type arguments are present the data is returned unchanged, which is the correct behaviour for an unparameterised dict.

Test

Added test_bare_dict_no_indexerror in tests/test_transform.py: passes {"metadata": {"key": "value"}} through a TypedDict with a bare dict field and asserts the result is returned unchanged without error. Runs in both sync and async variants.

Fixes #3338