fix(types): expand FunctionCallOutput.output to accept str or list

[anishesg]

• I understand that this repository is auto-generated and my pull request may not be merged

Changes being requested

The FunctionCallOutput.output field in src/openai/types/responses/response_input_item.py was typed as str only, which is inconsistent with the OpenAI API documentation that specifies the field accepts either a string or a list of content objects. The root cause was an upstream OpenAPI spec gap (openai/openai-openapi#488), but the Python library needs to reflect the documented behavior. The field type has been updated to Union[str, ResponseFunctionCallOutputItemList] (i.e. str | list[ResponseInputTextContent | ResponseInputImageContent | ResponseInputFileContent]) to match the API docs. A regression test in tests/test_models.py verifies that FunctionCallOutput correctly parses both string and list values for output.

Additional context & links

• Issue: FunctionCallOutput is inconsistent with OpenAI API docs #2668
• API reference: https://platform.openai.com/docs/api-reference/responses/list#responses/list-data-function-tool-call-output-output
• Upstream spec issue: FunctionCallOutput is inconsistent with OpenAI API docs openai-openapi#488

Fixes #2668