feat(copilot): MCTS dynamic strategy engine with reward model and rollout simulation

- Add MCTSConfig, MCTSNode, StrategyRecommendation data structures
- Add RewardModel with cosine similarity scoring (R = W1·Match + W2·Safe - W3·Risk)
- Add SimulationEngine with 3-level degradation rollout
- Add MCTSEngine with PUCT selection, LLM expansion, backpropagation
- Integrate MCTS into copilot WebSocket session (feature-flagged, off by default)
- Add 11 mcts_* settings to config and rollout LLM provider
- Add user-facing docs and 39 unit tests

Author: ldemon2333

backend/config.py

@@ -36,6 +36,19 @@ class Settings(BaseSettings):
     nls_access_key_id: str = ""
     nls_access_key_secret: str = ""
 
+    # MCTS 动态策略引擎
+    mcts_enabled: bool = False
+    mcts_iterations: int = 8
+    mcts_branch_factor: int = 3
+    mcts_rollout_depth: int = 2
+    mcts_c_puct: float = 1.4
+    mcts_max_tree_nodes: int = 200
+    mcts_merge_threshold: float = 0.8
+    mcts_search_timeout: float = 25.0
+    mcts_rollout_api_base: str = ""
+    mcts_rollout_api_key: str = ""
+    mcts_rollout_model: str = ""
+
     # Copilot — Tavily Web Search
     tavily_api_key: str = ""
 

backend/copilot/mcts_config.py

@@ -0,0 +1,84 @@
+"""MCTS 相关配置和共享数据结构。"""
+from __future__ import annotations
+
+import uuid
+from dataclasses import dataclass, field
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from backend.config import Settings
+
+
+@dataclass
+class MCTSConfig:
+    """MCTS 搜索参数配置。"""
+    enabled: bool = True
+    iterations: int = 8
+    branch_factor: int = 3
+    rollout_depth: int = 2
+    c_puct: float = 1.4
+    max_tree_nodes: int = 200
+    merge_threshold: float = 0.8
+    search_timeout: float = 25.0
+
+    @classmethod
+    def from_settings(cls, s: Settings) -> MCTSConfig:
+        return cls(
+            enabled=s.mcts_enabled,
+            iterations=s.mcts_iterations,
+            branch_factor=s.mcts_branch_factor,
+            rollout_depth=s.mcts_rollout_depth,
+            c_puct=s.mcts_c_puct,
+            max_tree_nodes=s.mcts_max_tree_nodes,
+            merge_threshold=s.mcts_merge_threshold,
+            search_timeout=s.mcts_search_timeout,
+        )
+
+
+@dataclass
+class MCTSNode:
+    """博弈树节点。"""
+    id: str
+    parent_id: str | None = None
+    children: list[str] = field(default_factory=list)
+
+    # 状态
+    actor: str = "candidate"  # "candidate" | "hr"
+    action: str = ""
+    action_embedding: list[float] | None = None
+
+    # 上下文
+    topic: str = ""
+    conversation_snapshot: list[dict] = field(default_factory=list)
+
+    # MCTS 统计量
+    visit_count: int = 0
+    total_reward: float = 0.0
+    prior: float = 0.5
+
+    # 元数据
+    strategy_tree_node_id: str | None = None
+    risk_level: str = "safe"
+    depth: int = 0
+
+    @property
+    def q_value(self) -> float:
+        if self.visit_count == 0:
+            return 0.0
+        return self.total_reward / self.visit_count
+
+    @staticmethod
+    def make_id() -> str:
+        return uuid.uuid4().hex[:10]
+
+
+@dataclass
+class StrategyRecommendation:
+    """MCTS 搜索结果，推送给前端的推荐。"""
+    optimal_response_strategy: str = ""
+    predicted_followups: list[dict] = field(default_factory=list)
+    danger_zones: list[str] = field(default_factory=list)
+    win_rate: float = 0.5
+    best_path: list[dict] = field(default_factory=list)
+    confidence: float = 0.0
+    iterations_completed: int = 0