prepare fastapi root layout for server deployment

app/schemas/llm.py

@@ -0,0 +1,20 @@
+from pydantic import BaseModel
+
+
+class LLMTestRequest(BaseModel):
+    """LLM 测试入参：用于快速模型和 reason 模型耗时验证。"""
+
+    message: str = "请用一句话说明医疗问诊训练 Demo 的用途。"
+
+
+class LLMTestResponse(BaseModel):
+    """LLM 测试响应：返回模型名、首 token 时间和总耗时。"""
+
+    model: str
+    first_token_ms: int | None = None
+    total_latency_ms: int
+    stream: bool
+    mock_mode: bool = False
+    fallback_used: bool = False
+    thinking_enabled: bool | None = None
+    reasoning_effort: str | None = None