fix: Reduce chunk returns in dify and Hundred Refining Runner to every 8 chunks

pkg/provider/modelmgr/requesters/moonshotchatcmpl.py

@@ -55,6 +55,6 @@ class MoonshotChatCompletions(chatcmpl.OpenAIChatCompletions):
         resp = await self._req(args, extra_body=extra_args)
 
         # 处理请求结果
-        message = await self._make_msg(resp,remove_think)
+        message = await self._make_msg(resp, remove_think)
 
         return message

pkg/provider/runners/dashscopeapi.py

@@ -148,7 +148,7 @@ class DashScopeAPIRunner(runner.RequestRunner):
                     # 将参考资料替换到文本中
                     pending_content = self._replace_references(pending_content, references_dict)
 
-                if idx_chunk % 64 == 0 or is_final:
+                if idx_chunk % 8 == 0 or is_final:
                     yield llm_entities.MessageChunk(
                         role='assistant',
                         content=pending_content,

pkg/provider/runners/difysvapi.py

@@ -115,7 +115,7 @@ class DifyServiceAPIRunner(runner.RequestRunner):
 
         stream_output_pending_chunk = ''
 
-        batch_pending_max_size = 64  # 积累一定量的消息更新消息一次
+        batch_pending_max_size = 8  # 积累一定量的消息更新消息一次
 
         batch_pending_index = 0