refactor(embedding/reranker): 重构模型工具类使用环境变量配置

更新.gitignore文件，添加新的数据库文件
在.env中添加EMBEDDING_MODEL_NAME和XINFERENCE_URL配置
重构SiliconFlowEmbeddings和XinferenceReRankerModel使用环境变量
优化DifyCompareTest异常处理和输入验证
修改测试文件路径和并发工作数

rag2_0/dify/DifyCompareTest.py

@@ -121,26 +121,38 @@ class DifyCompareTest:
                     time.sleep(10)  # 等待1秒后重试
 
     def get_wiki_list_by_msgid(self,msg_id):
-        if msg_id is None or pd.isna(msg_id):
+        try:
+            if msg_id is None or pd.isna(msg_id):
+                return ""
+            msg_debug_info = self.exporter.dify_tool.get_message_debug_info_by_id(msg_id)
+            if not msg_debug_info:
+                return ""
+            wiki_list = self.exporter.get_wiki_list(msg_debug_info)
+            if len(wiki_list) == 0:
+                return ""
+            else:
+                return "\n".join(list(set(wiki_list)))
+        except Exception as e:
+            logging.error(f"获取词条列表失败: {e}")
             return ""
-        msg_debug_info = self.exporter.dify_tool.get_message_debug_info_by_id(msg_id)
-        if not msg_debug_info:
-            return ""
-        wiki_list = self.exporter.get_wiki_list(msg_debug_info)
-        if len(wiki_list) == 0:
-            return ""
-        else:
-            return "\n".join(list(set(wiki_list)))
 
     def process_single_row(self, index, row):
         """处理单行数据的方法"""
         try:
             query = row["提问"]
+            current_software = row["当前软件"]
+            if pd.isna(query) or len(query) == 0 or pd.isna(current_software) or len(current_software) == 0:
+                result_row = row.copy()
+                result_row["message_id"] = ''
+                result_row["本次回答"] = ''
+                result_row["回答对比"] = ''
+                result_row["检索到的词条"] = ''
+                return index, result_row
+            
             if "回答" in row:
                 old_answer = row["回答"]
             else:
                 old_answer = ""
-            current_software = row["当前软件"]
             
             inputs = {
                 "current_softname": current_software,
@@ -247,14 +259,14 @@ if __name__ == "__main__":
         
         # 处理第一个文件
         excel_files = [
-            ("data/excel/300专业提问.xlsx", "data/excel/300专业提问_问答测试.xlsx"),
+            ("data/excel/第一轮的专业问题.xlsx", "data/excel/第一轮的专业问题_dify.xlsx"),
             # ("data/excel/有知识的.xlsx", "data/excel/有知识的_问答测试.xlsx")
         ]
         
         for excel_path, save_path in excel_files:
             logging.info(f"开始处理文件: {excel_path}")
             try:
-                dify_compare_test.run(excel_path=excel_path, save_path=save_path, max_workers=10)
+                dify_compare_test.run(excel_path=excel_path, save_path=save_path, max_workers=5)
                 logging.info(f"文件处理完成: {excel_path}")
             except Exception as e:
                 logging.error(f"处理文件 {excel_path} 时出错: {e}")