feat(engine): add faction system for social emergence

- Add Factions model (optimists/fearful/neutral)
- Implement classify_faction() based on stance thresholds
- Add update_factions() to track faction distribution
- Add apply_faction_influence() for faction→mood feedback
- Integrate faction system into tick flow

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

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

engine-python/app/factions.py

@@ -0,0 +1,40 @@
+"""派系系统 - 基于立场分类角色并影响世界"""
+from typing import Dict
+from .models import WorldState, AgentState, Factions
+
+# 派系分类阈值
+OPTIMIST_THRESHOLD = 0.6
+FEARFUL_THRESHOLD = 0.6
+
+
+def classify_faction(agent: AgentState) -> str:
+    """根据 stance 分类角色所属派系"""
+    if agent.stance.optimism > OPTIMIST_THRESHOLD:
+        return "optimists"
+    elif agent.stance.fear > FEARFUL_THRESHOLD:
+        return "fearful"
+    else:
+        return "neutral"
+
+
+def update_factions(state: WorldState) -> None:
+    """统计各派系人数并更新 world_state"""
+    counts = {"optimists": 0, "fearful": 0, "neutral": 0}
+
+    for agent in state.agents.values():
+        faction = classify_faction(agent)
+        counts[faction] += 1
+
+    state.factions = Factions(**counts)
+
+
+def apply_faction_influence(state: WorldState) -> None:
+    """派系分布影响世界情绪"""
+    optimists = state.factions.optimists
+    fearful = state.factions.fearful
+
+    if optimists > fearful:
+        state.town_mood = min(10, state.town_mood + 1)
+    elif fearful > optimists:
+        state.town_mood = max(-10, state.town_mood - 1)
+    # 平局时不变化