更新专业术语索引文件，优化意图识别逻辑，添加后缀项更新功能，调整重排序参数以提高相关性，同时修正文档中的描述信息。

rag2_0/intent_recognition/IntentRecognition.py

@@ -184,7 +184,7 @@ class IntentRecognizer:
         except Exception as e:
             raise RuntimeError(f"无法解析LLM关键词提取响应: {e}") from e
     
-    def _rerank_matched_terms(self, query_key: str, matched_terms: set, top_k: int = 2) -> List[Term]:
+    def _rerank_matched_terms(self, query_key: str, matched_terms: set, top_k: int = 2, rerank_score:float = 0.6) -> List[Term]:
         """
         对召回的专业术语进行重排序，按与用户查询的相关性排序
         
@@ -198,10 +198,14 @@ class IntentRecognizer:
         """
         if not matched_terms:
             return []
-            
+        
+        if len(matched_terms) <= top_k:
+            return list(matched_terms)
+
         try:
             # 将每个术语转换为可用于重排序的文本表示
-            term_texts = ["名称:" + term.name + "|" + "同义词:" + ";".join(term.synonymous) + "|" + "描述:" + term.description for term in matched_terms]
+            # term_texts = ["名称:" + term.name + "|" + "同义词:" + ";".join(term.synonymous) + "|" + "描述:" + term.description for term in matched_terms]
+            term_texts = ["名称:" + term.name + "|" + "同义词:" + ";".join(term.synonymous) for term in matched_terms]
             
             # 使用重排序模型
             xinference_reranker = SiliconFlowReRankerModel()
@@ -211,7 +215,7 @@ class IntentRecognizer:
             matched_terms_list = list(matched_terms)
             
             # 根据重排序结果获取排序后的术语列表
-            reranked_terms = [matched_terms_list[result["index"]] for result in rerank_results if result["score"] >= 0.6]
+            reranked_terms = [matched_terms_list[result["index"]] for result in rerank_results if result["score"] >= rerank_score]
             
             return reranked_terms
             
@@ -279,7 +283,8 @@ class IntentRecognizer:
             改写结果
         """
         # 准备问题改写提示
-        terms_dict = [term.model_dump(exclude={"description"}) for term in keywords.terms] 
+        # terms_dict = [term.model_dump(exclude={"description"}) for term in keywords.terms] 
+        terms_dict = [term.model_dump() for term in keywords.terms] 
         keywords_str = json.dumps(terms_dict, ensure_ascii=False)
         query_rewrite_parser = PydanticOutputParser(pydantic_object=QueryRewrite)
         # formatted_prompt = query_rewrite_prompt.format(query=query, 
@@ -369,7 +374,7 @@ class IntentRecognizer:
         )
         
         # 步骤3: 进行意图识别和槽位填充
-        result = self._process_intent_and_slot(query, conversation_context, chat_history, previous_slots)
+        result = self._process_intent_and_slot(rewrite.rewrite, conversation_context, chat_history, previous_slots)
         result.update({"keywords": keywords_terms.model_dump(),
                               "rewrite": rewrite.model_dump(),
                               "query_keys": query_keys})