fix(core): Resolve 'FieldInfo is not JSON serializable' error in Pydantic v2

[jitokim]

Fixes TypeError when generating JSON schemas: "Object of type 'FieldInfo'
is not JSON serializable". This occurs when model_json_schema() includes
internal Pydantic metadata in the output.

Solution: Add mode='serialization' parameter to model_json_schema() to
ensure only JSON-serializable data structures are generated.

Changes:

• Modified _convert_pydantic_to_openai_function() to pass mode='serialization'
• Added test_pydantic_fieldinfo_serialization() to verify correct behavior
• Resolves nested Pydantic v2 model conversion issues

Impact:

• Fixes schema generation for nested Pydantic v2 models
• No breaking changes (Pydantic v1 code path unchanged)
• Previously failing test_convert_to_openai_function_nested_v2 now passes

[codspeed-hq]

CodSpeed Performance Report

Merging #33932 will degrade performances by 26%

_{Comparing jitokim:fix/pydantic-fieldinfo-serialization (d1dafab) with master (83c078f)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

❌ 13 regressions
⏩ 21 skipped¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Mode	Benchmark	BASE	HEAD	Change
❌	WallTime	test_async_callbacks_in_sync	18.3 ms	24.7 ms	-26%
❌	WallTime	test_import_time[BaseChatModel]	456.1 ms	521.9 ms	-12.61%
❌	WallTime	test_import_time[CallbackManager]	397 ms	463 ms	-14.25%
❌	WallTime	test_import_time[ChatPromptTemplate]	506.9 ms	602.4 ms	-15.85%
❌	WallTime	test_import_time[Document]	164.9 ms	189.5 ms	-13%
❌	WallTime	test_import_time[HumanMessage]	235.8 ms	269.2 ms	-12.43%
❌	WallTime	test_import_time[InMemoryRateLimiter]	157.7 ms	175.3 ms	-10.05%
❌	WallTime	test_import_time[InMemoryVectorStore]	531.6 ms	619.3 ms	-14.15%
❌	WallTime	test_import_time[LangChainTracer]	387.8 ms	439.1 ms	-11.69%
❌	WallTime	test_import_time[PydanticOutputParser]	455.8 ms	530 ms	-14%
❌	WallTime	test_import_time[RunnableLambda]	433.6 ms	495 ms	-12.41%
❌	WallTime	test_import_time[Runnable]	427.1 ms	488.5 ms	-12.57%
❌	WallTime	test_import_time[tool]	445.4 ms	530 ms	-15.97%

Footnotes

1. 21 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[eyurtsev]

can we make this test dumb and check result against a dict with a single assert?

We want to be able to see immediately what the expected value of result is instead of having to read through different assertions

[jitokim]

@eyurtsev Thanks for reivew