feat(ai-rule-engine): 集成冲突检测到 API

- handle_create_rule 返回冲突警告
- 使用 ConflictDetector 检测新规则与现有规则的冲突

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

src/rule_engine/api.py

@@ -11,6 +11,7 @@ from rule_engine.matcher import RuleMatcher
 from rule_engine.executor import RuleExecutor
 from rule_engine.callback import LLMEallback, MockLLMClient
 from rule_engine.compiler import RuleCompiler
+from rule_engine.conflict import ConflictDetector
 
 
 class RuleEngineAPI:
@@ -20,6 +21,7 @@ class RuleEngineAPI:
         self.store = store
         self.executor = RuleExecutor()
         self.matcher = RuleMatcher(store, self.executor)
+        self.conflict_detector = ConflictDetector()
 
         # LLM 回调（可选）
         if enable_callback:
@@ -61,6 +63,17 @@ class RuleEngineAPI:
         if not code:
             code = "def rule(facts):\n    return None"
 
+        # 检查与现有规则的冲突
+        existing_rules = [r.to_dict() for r in self.store.list_rules()]
+        new_rule_dict = {
+            "id": "new_rule",
+            "name": request.name,
+            "code": code,
+            "priority": request.priority,
+            "is_active": True
+        }
+        conflicts = self.conflict_detector.check_rule_with_existing(new_rule_dict, existing_rules)
+
         rule = self.store.create_rule(
             name=request.name,
             condition_template=request.condition_template,
@@ -69,7 +82,11 @@ class RuleEngineAPI:
             priority=request.priority
         )
 
-        return 201, rule.to_dict()
+        result = rule.to_dict()
+        if conflicts:
+            result["conflict_warning"] = conflicts[0] if len(conflicts) == 1 else conflicts
+
+        return 201, result
 
     def handle_get_rule(self, rule_id: str) -> tuple:
         """获取单个规则。"""