LLM输出的内容先通过正则清理多余内容

rag2_0/demo/intent_recognition_example.py

@@ -204,41 +204,7 @@ class QueryRewriteProcessor:
                                                   enable_query_expansion=True,
                                                   use_jieba=True,
                                                   cur_soft_name=current_softname))
-
-                # 提取分类信息
-                classification = result["classification"]
-                original_query = result["rewrite"]["rewrite"]
-                query_list = result["query_expand"]["all"]
-                # 将字典转换为Classification对象
-                classification_obj = Classification(**classification)
-                
-                # 根据enable_retrieval参数决定是否进行文档检索
-                retrieved_doc = None
-        
-
-                retrieved_doc_titles=[]
-                if retrieved_doc:
-                    retrieved_doc_titles=[doc["title"].split("/")[-1] for doc in retrieved_doc]     
-                # 提取槽位填充信息
-                slot_filling = result.get("slot_filling", {})
-                slot_filling_str = ""
-                if slot_filling and "filled_data" in slot_filling:
-                    # 格式化槽位填充结果
-                    slot_filling_str = json.dumps({
-                        "is_complete": slot_filling.get("is_complete", False),
-                        "missing_slots": slot_filling.get("missing_slots", {}),
-                        "filled_data": slot_filling.get("filled_data", {})
-                    }, ensure_ascii=False, indent=2)
-
-                # 处理成功，返回结果
-                return {
-                    "问题": query,
-                    "问题分类": f"{classification['vertical_classification']} - {classification['sub_classification']}",
-                    "问题改写": result["rewrite"]["rewrite"],
-                    "槽位信息": slot_filling_str,
-                    "检索的文档": "\n".join(retrieved_doc_titles),
-                    "检索的内容":  json.dumps(retrieved_doc, ensure_ascii=False, indent=2) if retrieved_doc else "",
-                }
+                return result
             except Exception as e:
                 logging.error(f"处理问题 '{query}' 时出错: ",exc_info=True)
                 retry_count += 1
@@ -424,19 +390,25 @@ def main():
         logging.info(f"所有处理完成，最终结果已保存至: {output_file}")
     else:
         logging.info(f"文档检索功能状态: 已启用")
-        for idx, query in enumerate(examples):
-            if query.strip() == "":
-                continue
-            query="怎么调整报表顺序"
+        nCount=0
+        while True:
+            query="请问怎么在南网版概算工程软件版里面查施工费？"
             conversation_context={
-                "current_softname": "储能计价通C1软件"
+                "current_softname": "电力建设计价通软件"
             }
-            # 在调试模式下使用完整的参数
-            print(json.dumps(processor.process_query(
+            result = processor.process_query(
                 query, 
                 conversation_context=conversation_context,
                 enable_retrieval=True
-            ), ensure_ascii=False, indent=2))
+            )
+            # 在调试模式下使用完整的参数
+            # print(json.dumps(processor.process_query(
+            #     query, 
+            #     conversation_context=conversation_context,
+            #     enable_retrieval=True
+            # ), ensure_ascii=False, indent=2))
+            nCount+=1
+            print("测试数量:  ", nCount)
 
 def setup_logging():
     # 配置日志输出到控制台