fix：通过计时定位LLM异常延时，移除memory模型

src/llm_models/model_client/openai_client.py

@@ -1,6 +1,7 @@
 import asyncio
 import io
 import json
+import time
 import re
 import base64
 from collections.abc import Iterable
@@ -452,6 +453,7 @@ class OpenaiClient(BaseClient):
                 resp, usage_record = await stream_response_handler(req_task.result(), interrupt_flag)
             else:
                 # 发送请求并获取响应
+                # start_time = time.time()
                 req_task = asyncio.create_task(
                     self.client.chat.completions.create(
                         model=model_info.model_identifier,
@@ -469,7 +471,9 @@ class OpenaiClient(BaseClient):
                         # 如果中断量存在且被设置，则取消任务并抛出异常
                         req_task.cancel()
                         raise ReqAbortException("请求被外部信号中断")
-                    await asyncio.sleep(0.5)  # 等待0.5秒后再次检查任务&中断信号量状态
+                    await asyncio.sleep(0.1)  # 等待0.5秒后再次检查任务&中断信号量状态
+                
+                # logger.info(f"OpenAI请求时间: {model_info.model_identifier}  {time.time() - start_time} \n{messages}")
 
                 resp, usage_record = async_response_parser(req_task.result())
         except APIConnectionError as e:

src/llm_models/utils_model.py

@@ -1,6 +1,7 @@
 import re
 import copy
 import asyncio
+import time
 
 from enum import Enum
 from rich.traceback import install
@@ -150,14 +151,22 @@ class LLMRequest:
             (Tuple[str, str, str, Optional[List[ToolCall]]]): 响应内容、推理内容、模型名称、工具调用列表
         """
         # 请求体构建
+        start_time = time.time()
+        
+        
+        
         message_builder = MessageBuilder()
         message_builder.add_text_content(prompt)
         messages = [message_builder.build()]
+        
         tool_built = self._build_tool_options(tools)
+        
         # 模型选择
         model_info, api_provider, client = self._select_model()
-
+        
         # 请求并处理返回值
+        logger.info(f"LLM选择耗时: {model_info.name} {time.time() - start_time}")
+        
         response = await self._execute_request(
             api_provider=api_provider,
             client=client,
@@ -168,6 +177,8 @@ class LLMRequest:
             max_tokens=max_tokens,
             tool_options=tool_built,
         )
+        
+        
         content = response.content
         reasoning_content = response.reasoning_content or ""
         tool_calls = response.tool_calls
@@ -175,6 +186,7 @@ class LLMRequest:
         if not reasoning_content and content:
             content, extracted_reasoning = self._extract_reasoning(content)
             reasoning_content = extracted_reasoning
+            
         if usage := response.usage:
             llm_usage_recorder.record_usage_to_database(
                 model_info=model_info,
@@ -183,6 +195,7 @@ class LLMRequest:
                 request_type=self.request_type,
                 endpoint="/chat/completions",
             )
+        
         if not content:
             if raise_when_empty:
                 logger.warning("生成的响应为空")