feat: 添加 NapCat 适配器的入站消息编解码功能，增强插件配置更新逻辑和数据库交互测试

src/learners/expression_learner.py

@@ -461,25 +461,43 @@ class ExpressionLearner:
     def _find_similar_expression(
         self, situation: str, similarity_threshold: float = 0.75
     ) -> Optional[Tuple[MaiExpression, float]]:
-        """在数据库中查找相似的表达方式"""
+        """在数据库中查找相似的表达方式。
+
+        Args:
+            situation: 当前待匹配的情景描述。
+            similarity_threshold: 认定为相似表达方式的最低相似度阈值。
+
+        Returns:
+            Optional[Tuple[MaiExpression, float]]: 若找到最相似的表达方式，则返回
+            ``(表达方式对象, 相似度)``；否则返回 ``None``。
+        """
         try:
-            with get_db_session() as session:
+            with get_db_session(auto_commit=False) as session:
                 statement = select(Expression).filter_by(session_id=self.session_id)
                 expressions = session.exec(statement).all()
 
-            best_match: Optional[Expression] = None
-            best_similarity = 0.0
+                best_match: Optional[MaiExpression] = None
+                best_similarity = 0.0
+
+                for db_expression in expressions:
+                    expression = MaiExpression.from_db_instance(db_expression)
+                    candidate_situations = [expression.situation, *expression.content]
+                    for candidate_situation in candidate_situations:
+                        normalized_candidate_situation = candidate_situation.strip()
+                        if not normalized_candidate_situation:
+                            continue
+                        similarity = difflib.SequenceMatcher(
+                            None,
+                            situation,
+                            normalized_candidate_situation,
+                        ).ratio()
+                        if similarity > similarity_threshold and similarity > best_similarity:
+                            best_similarity = similarity
+                            best_match = expression
 
-            for expr in expressions:
-                content_list = json.loads(expr.content_list)
-                for situation in content_list:
-                    similarity = difflib.SequenceMatcher(None, situation, expr.situation).ratio()
-                    if similarity > similarity_threshold and similarity > best_similarity:
-                        best_similarity = similarity
-                        best_match = expr
             if best_match:
-                logger.debug(f"找到相似表达方式情景 [ID: {best_match.id}]，相似度: {best_similarity:.2f}")
-                return MaiExpression.from_db_instance(best_match), best_similarity
+                logger.debug(f"找到相似表达方式情景 [ID: {best_match.item_id}]，相似度: {best_similarity:.2f}")
+                return best_match, best_similarity
 
         except Exception as e:
             logger.error(f"查找相似表达方式失败: {e}")

src/learners/jargon_miner.py

@@ -199,7 +199,7 @@ class JargonMiner:
 
     async def process_extracted_entries(
         self, entries: List[JargonEntry], person_name_filter: Optional[Callable[[str], bool]] = None
-    ):
+    ) -> None:
         """
         处理已提取的黑话条目（从 expression_learner 路由过来的）
 
@@ -230,7 +230,7 @@ class JargonMiner:
             content = entry["content"]
             raw_content_set = entry["raw_content"]
             try:
-                with get_db_session() as session:
+                with get_db_session(auto_commit=False) as session:
                     jargon_items = session.exec(select(Jargon).filter_by(content=content)).all()
             except Exception as e:
                 logger.error(f"查询黑话 '{content}' 失败: {e}")
@@ -306,7 +306,13 @@ class JargonMiner:
                 removed_content, _ = self.cache.popitem(last=False)
                 logger.debug(f"缓存已满，移除最旧的黑话: {removed_content}")
 
-    def _update_jargon(self, db_jargon: Jargon, raw_content_set: Set[str]):
+    def _update_jargon(self, db_jargon: Jargon, raw_content_set: Set[str]) -> None:
+        """更新已有黑话记录并写回数据库。
+
+        Args:
+            db_jargon: 已命中的黑话 ORM 对象。
+            raw_content_set: 本次新增的原始上下文集合。
+        """
         db_jargon.count += 1
         existing_raw_content: List[str] = []
         if db_jargon.raw_content:
@@ -328,7 +334,17 @@ class JargonMiner:
 
         try:
             with get_db_session() as session:
-                session.add(db_jargon)
+                if db_jargon.id is None:
+                    raise ValueError("黑话记录缺少 id，无法更新数据库")
+                statement = select(Jargon).filter_by(id=db_jargon.id).limit(1)
+                if persisted_jargon := session.exec(statement).first():
+                    persisted_jargon.count = db_jargon.count
+                    persisted_jargon.raw_content = db_jargon.raw_content
+                    persisted_jargon.session_id_dict = db_jargon.session_id_dict
+                    persisted_jargon.is_global = db_jargon.is_global
+                    session.add(persisted_jargon)
+                else:
+                    logger.warning(f"黑话 ID {db_jargon.id} 在数据库中未找到，无法更新")
         except Exception as e:
             logger.error(f"更新黑话 '{db_jargon.content}' 失败: {e}")