优化DifyCompareTest和WorkorderToDify模块，调整日志记录格式，修复API密钥获取方式，增强工单处理流程，添加元数据管理功能，改进并发上传逻辑，更新文档处理方式。

rag2_0/dify/DifyCompareTest.py

@@ -26,7 +26,7 @@ if not os.path.exists(log_dir):
     os.makedirs(log_dir)
 
 # 生成带时间戳的日志文件名
-log_file = os.path.join(log_dir, f'dify_compare_{datetime.now().strftime("%Y%m%d_%H%M%S")}.log')
+log_file = os.path.join(log_dir, f'dify_compare_{datetime.now().strftime("%Y%m%d")}.log')
 
 
 import logging
@@ -43,7 +43,7 @@ logging.basicConfig(
 class DifyCompareTest:
     def __init__(self):
         # 词条与工单同时检索
-        self.both_wiki_worker_client = ChatClient(api_key="app-CPoOMaGDsLRPAe9TW7Xjhszy", base_url=os.getenv("DIFY_BSAE_URL"))
+        self.both_wiki_worker_client = ChatClient(api_key=os.getenv("DIFY_APP_KEY"), base_url=os.getenv("DIFY_BSAE_URL"))
         self.llm = OpenAiLLM(base_url=os.getenv("OPENAI_API_BASE"), model=os.getenv("MODEL_NAME"))
 
     def llm_judge_answer(self, old_answer: str, now_answer: str):
@@ -100,22 +100,31 @@ class DifyCompareTest:
                 answer = result.get('answer', "")
                 if len(answer) == 0:
                     raise Exception(f"回答为空: {result}")
-                judge_result = self.llm_judge_answer(old_answer=old_answer, now_answer=answer)
-                return answer, judge_result
+                if old_answer:
+                    judge_result = self.llm_judge_answer(old_answer=old_answer, now_answer=answer)
+                else:
+                    judge_result=""
+                # 只取回答的前半部分
+                answer = answer.split("----------------------------------------")[0].strip()
+                message_id = result.get('message_id', "")
+                return answer, judge_result, message_id
             except Exception as e:
                 retry_count += 1
                 if retry_count >= max_retries:
                     logging.error(f"词条与工单同时检索调用失败 (尝试 {max_retries} 次后): {e}")
-                    return '', ''
+                    return '', '', ''
                 else:
                     import time
-                    time.sleep(1)  # 等待1秒后重试
+                    time.sleep(10)  # 等待1秒后重试
 
     def process_single_row(self, index, row):
         """处理单行数据的方法"""
         try:
             query = row["提问"]
-            old_answer = row["回答"]
+            if "回答" in row:
+                old_answer = row["回答"]
+            else:
+                old_answer = ""
             current_software = row["当前软件"]
             
             inputs = {
@@ -124,7 +133,7 @@ class DifyCompareTest:
             }
             
             # 调用词条与工单同时检索工作流
-            answer, judge_result = self.process_workflow(
+            answer, judge_result, message_id = self.process_workflow(
                 self.both_wiki_worker_client,
                 inputs,
                 query,
@@ -133,17 +142,17 @@ class DifyCompareTest:
             
             # 构建结果
             result_row = row.copy()
-            result_row["词条与工单同时回答"] = answer
-            result_row["词条与工单同时回答对比"] = judge_result
-            
+            result_row["message_id"] = message_id
+            result_row["回答"] = answer
+            # result_row["词条与工单同时回答对比"] = judge_result
             logging.info(f"成功处理第 {index + 1} 行数据")
             return index, result_row
             
         except Exception as e:
             logging.error(f"处理第 {index + 1} 行数据时出错: {e}")
             result_row = row.copy()
-            result_row["词条与工单同时回答"] = ''
-            result_row["词条与工单同时回答对比"] = ''
+            result_row["回答"] = ''
+            result_row["message_id"] = ''
             return index, result_row
 
       
@@ -166,7 +175,7 @@ class DifyCompareTest:
             logging.info(f"成功读取Excel文件: {excel_path}, 共 {len(df)} 行数据")
             
             # 验证必要的列是否存在
-            required_columns = ["提问", "回答", "当前软件"]
+            required_columns = ["提问", "当前软件"]
             missing_columns = [col for col in required_columns if col not in df.columns]
             if missing_columns:
                 logging.error(f"Excel文件缺少必要的列: {missing_columns}")
@@ -199,8 +208,6 @@ class DifyCompareTest:
                             logging.error(f"线程执行失败 (行{original_index + 1}): {e}")
                             # 添加失败的行
                             result_row = df.iloc[original_index].copy()
-                            result_row["词条与工单同时回答"] = '线程执行失败'
-                            result_row["词条与工单同时回答对比"] = '线程执行失败'
                             results[original_index] = result_row
                             pbar.update(1)
             
@@ -223,7 +230,7 @@ if __name__ == "__main__":
         # 处理第一个文件
         excel_files = [
             # ("data/excel/5月.xlsx", "data/excel/5月问答对比.xlsx"),
-            ("data/excel/其他月.xlsx", "data/excel/其他月问答对比.xlsx")
+            ("data/excel/第四轮问题-Part2.xlsx", "data/excel/第四轮问题-Part2-问答测试.xlsx")
         ]
         
         for excel_path, save_path in excel_files: