feat: AI Town MVP - 完整三层架构实现

- Python FastAPI 引擎：世界状态模拟、全局能量条、世界事件系统
- Node.js WebSocket 服务器：实时通信、事件队列批处理
- 前端仪表盘：世界状态可视化、行动日志、事件展示

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

engine-python/app/main.py

@@ -0,0 +1,37 @@
+from fastapi import FastAPI
+from .models import StepRequest, StepResponse
+from .state import load_state, save_state
+from .engine import process_events, generate_actions, check_and_trigger_global_event
+
+app = FastAPI(title="AI Town Engine", version="0.1.0")
+
+
+@app.post("/step", response_model=StepResponse)
+def step(request: StepRequest) -> StepResponse:
+    """执行一个模拟步骤"""
+    # 加载当前状态
+    state = load_state()
+
+    # 处理事件并更新状态
+    state = process_events(state, request.events)
+
+    # 检查并触发世界级事件
+    state, global_event_result = check_and_trigger_global_event(state)
+
+    # 生成 agent 行动
+    actions = generate_actions(state)
+
+    # 保存状态
+    save_state(state)
+
+    return StepResponse(
+        world_state=state,
+        actions=actions,
+        global_event=global_event_result
+    )
+
+
+@app.get("/health")
+def health():
+    """健康检查"""
+    return {"status": "ok"}