fix(memory): route Qwen thinking through chat template kwargs

[LeslieWylie]

Summary

• route Qwen enable_thinking through chat_template_kwargs when an OpenAI-compatible endpoint opts into that transport
• preserve other endpoint-specific chat-template kwargs
• retain existing top-level behavior for DashScope and other endpoints

Validation

• npx vitest run Memory/tests/llm-thinking.test.ts (19 tests passed)
• live SGLang replay verified top-level enable_thinking=false was ignored while chat_template_kwargs.enable_thinking=false produced zero reasoning tokens and valid JSON

[Wang-Daoji]

Thanks for working on this! I suggest adding chat_template_kwargs later as part of dedicated SGLang/vLLM providers. Adding it to the generic OpenAI-compatible path now could affect existing cloud-provider calls and introduce unnecessary risk.