fix

[hyeongyun0916]

Purpose

Fix /v1/chat/completions/render endpoint to properly serialize Harmony model messages.

The render endpoint was failing with TypeError: Object of type Message is not JSON serializable when used with Harmony models (e.g., openai/gpt-oss-20b). This is because Harmony models return OpenAIMessage Pydantic objects in the conversation, which cannot be directly serialized to JSON.

This PR:

Adds proper serialization logic in api_router.py to handle both Pydantic models (via model_dump()) and TypedDict objects
Updates test assertions to handle both standard message format (role key) and Harmony format (nested author dict)

Test Plan

pytest tests/entrypoints/openai/test_render_no_inference.py -v
pytest tests/entrypoints/openai/test_render.py -v

Test Result

tests/entrypoints/openai/test_render_no_inference.py::test_health_check PASSED
tests/entrypoints/openai/test_render_no_inference.py::test_chat_completion_render_works_without_inference PASSED
tests/entrypoints/openai/test_render_no_inference.py::test_completion_render_works_without_inference PASSED
tests/entrypoints/openai/test_render_no_inference.py::test_inference_disabled PASSED
tests/entrypoints/openai/test_render.py::test_completion_render_basic PASSED
tests/entrypoints/openai/test_render.py::test_chat_completion_render_basic PASSED
tests/entrypoints/openai/test_render.py::test_completion_render_multiple_prompts PASSED
tests/entrypoints/openai/test_render.py::test_chat_completion_render_multi_turn PASSED
tests/entrypoints/openai/test_render.py::test_completion_render_error_invalid_model PASSED
tests/entrypoints/openai/test_render.py::test_chat_completion_render_error_invalid_model PASSED
tests/entrypoints/openai/test_render.py::test_completion_render_no_generation PASSED

================== 11 passed ==================

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.