fix: handle bare dict annotation in _transform_recursive

[gavin913-lss]

What breaks

When a TypedDict field is annotated with a bare, unparameterised dict (e.g. metadata: dict instead of metadata: dict[str, str]), calling transform() raises IndexError.

from typing import TypedDict
from openai._utils._transform import transform

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

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

Fixes #3338

Root cause

Both _transform_recursive and _async_transform_recursive unconditionally do get_args(stripped_type)[1] when origin == dict. For bare dict, get_args(dict) returns an empty tuple, so the index access crashes.

Fix

Add a length check before indexing. When get_args() returns fewer than 2 elements (bare dict), return the data as-is since there are no type parameters to recurse into.

args = get_args(stripped_type)
if len(args) < 2:
    return cast(object, data)
items_type = args[1]

Tests

Added test_bare_dict_in_typeddict() covering:

• Bare dict with nested values
• Empty dict

Both sync and async paths pass. All existing test_transform.py tests continue to pass.

[marcuswood-oai]

Thank you for the contribution. The missing type arguments for bare dict annotations are now handled in sync/async request transformation by #3760, which is merged into main.

Closing this PR because the merged change resolves the same crash. Please use a release containing #3760 when available; the fix is not included in v3.11.0.