refactor(cache): 重构工具缓存机制并优化LLM请求重试逻辑
将工具缓存的实现从`ToolExecutor`的装饰器模式重构为直接集成。缓存逻辑被移出`cache_manager.py`并整合进`ToolExecutor.execute_tool_call`方法中，简化了代码结构并使其更易于维护。 主要变更： - 从`cache_manager.py`中移除了`wrap_tool_executor`函数。 - 在`tool_use.py`中，`execute_tool_call`现在包含完整的缓存检查和设置逻辑。 - 调整了`llm_models/utils_model.py`中的LLM请求逻辑，为模型生成的空回复或截断响应增加了内部重试机制，增强了稳定性。 - 清理了项目中未使用的导入和过时的文档文件，以保持代码库的整洁。
This commit is contained in:
minecraft1024a
@@ -310,82 +310,94 @@ class LLMRequest:
|
||||
message_builder.add_text_content(processed_prompt)
|
||||
messages = [message_builder.build()]
|
||||
tool_built = self._build_tool_options(tools)
|
||||
|
||||
response = await self._execute_request(
|
||||
api_provider=api_provider,
|
||||
client=client,
|
||||
request_type=RequestType.RESPONSE,
|
||||
model_info=model_info,
|
||||
message_list=messages,
|
||||
tool_options=tool_built,
|
||||
temperature=temperature,
|
||||
max_tokens=max_tokens,
|
||||
)
|
||||
|
||||
content = response.content or ""
|
||||
reasoning_content = response.reasoning_content or ""
|
||||
tool_calls = response.tool_calls
|
||||
# 针对当前模型的空回复/截断重试逻辑
|
||||
empty_retry_count = 0
|
||||
max_empty_retry = api_provider.max_retry
|
||||
empty_retry_interval = api_provider.retry_interval
|
||||
|
||||
if not reasoning_content and content:
|
||||
content, extracted_reasoning = self._extract_reasoning(content)
|
||||
reasoning_content = extracted_reasoning
|
||||
|
||||
is_empty_reply = not tool_calls and (not content or content.strip() == "")
|
||||
is_truncated = False
|
||||
if use_anti_truncation:
|
||||
if content.endswith("[done]"):
|
||||
content = content[:-6].strip()
|
||||
else:
|
||||
is_truncated = True
|
||||
|
||||
if is_empty_reply or is_truncated:
|
||||
# 空回复或截断不进行模型切换，仅记录错误后抛出或返回
|
||||
reason = "空回复" if is_empty_reply else "截断"
|
||||
msg = f"模型 '{model_name}' 生成了{reason}的回复"
|
||||
logger.error(msg)
|
||||
if raise_when_empty:
|
||||
raise RuntimeError(msg)
|
||||
return msg, (reasoning_content, model_name, tool_calls)
|
||||
|
||||
|
||||
# 成功获取响应
|
||||
if usage := response.usage:
|
||||
llm_usage_recorder.record_usage_to_database(
|
||||
model_info=model_info, model_usage=usage, time_cost=time.time() - start_time,
|
||||
user_id="system", request_type=self.request_type, endpoint="/chat/completions",
|
||||
while empty_retry_count <= max_empty_retry:
|
||||
response = await self._execute_request(
|
||||
api_provider=api_provider,
|
||||
client=client,
|
||||
request_type=RequestType.RESPONSE,
|
||||
model_info=model_info,
|
||||
message_list=messages,
|
||||
tool_options=tool_built,
|
||||
temperature=temperature,
|
||||
max_tokens=max_tokens,
|
||||
)
|
||||
|
||||
if not content and not tool_calls:
|
||||
if raise_when_empty:
|
||||
raise RuntimeError("生成空回复")
|
||||
content = "生成的响应为空"
|
||||
|
||||
logger.info(f"模型 '{model_name}' 成功生成回复。")
|
||||
return content, (reasoning_content, model_name, tool_calls)
|
||||
content = response.content or ""
|
||||
reasoning_content = response.reasoning_content or ""
|
||||
tool_calls = response.tool_calls
|
||||
|
||||
if not reasoning_content and content:
|
||||
content, extracted_reasoning = self._extract_reasoning(content)
|
||||
reasoning_content = extracted_reasoning
|
||||
|
||||
is_empty_reply = not tool_calls and (not content or content.strip() == "")
|
||||
is_truncated = False
|
||||
if use_anti_truncation:
|
||||
if content.endswith("[done]"):
|
||||
content = content[:-6].strip()
|
||||
else:
|
||||
is_truncated = True
|
||||
|
||||
if is_empty_reply or is_truncated:
|
||||
empty_retry_count += 1
|
||||
if empty_retry_count <= max_empty_retry:
|
||||
reason = "空回复" if is_empty_reply else "截断"
|
||||
logger.warning(f"模型 '{model_name}' 检测到{reason}，正在进行第 {empty_retry_count}/{max_empty_retry} 次重新生成...")
|
||||
if empty_retry_interval > 0:
|
||||
await asyncio.sleep(empty_retry_interval)
|
||||
continue # 继续使用当前模型重试
|
||||
else:
|
||||
# 当前模型重试次数用尽，跳出内层循环，触发外层循环切换模型
|
||||
reason = "空回复" if is_empty_reply else "截断"
|
||||
logger.error(f"模型 '{model_name}' 经过 {max_empty_retry} 次重试后仍然是{reason}的回复。")
|
||||
raise RuntimeError(f"模型 '{model_name}' 达到最大空回复/截断重试次数")
|
||||
|
||||
# 成功获取响应
|
||||
if usage := response.usage:
|
||||
llm_usage_recorder.record_usage_to_database(
|
||||
model_info=model_info, model_usage=usage, time_cost=time.time() - start_time,
|
||||
user_id="system", request_type=self.request_type, endpoint="/chat/completions",
|
||||
)
|
||||
|
||||
if not content and not tool_calls:
|
||||
if raise_when_empty:
|
||||
raise RuntimeError("生成空回复")
|
||||
content = "生成的响应为空"
|
||||
|
||||
logger.info(f"模型 '{model_name}' 成功生成回复。")
|
||||
return content, (reasoning_content, model_name, tool_calls)
|
||||
|
||||
except RespNotOkException as e:
|
||||
if e.status_code in [401, 403]:
|
||||
logger.error(f"模型 '{model_name}' 遇到认证/权限错误 (Code: {e.status_code})，将尝试下一个模型。")
|
||||
failed_models.add(model_name)
|
||||
last_exception = e
|
||||
continue
|
||||
continue # 切换到下一个模型
|
||||
else:
|
||||
# 对于其他HTTP错误，不切换模型，直接抛出
|
||||
logger.error(f"模型 '{model_name}' 请求失败，HTTP状态码: {e.status_code}")
|
||||
last_exception = e
|
||||
if raise_when_empty:
|
||||
raise
|
||||
break
|
||||
# 对于其他HTTP错误，直接抛出，不再尝试其他模型
|
||||
return f"请求失败: {e}", ("", model_name, None)
|
||||
|
||||
except RuntimeError as e:
|
||||
# 捕获所有重试失败（包括空回复和网络问题）
|
||||
logger.error(f"模型 '{model_name}' 在所有重试后仍然失败: {e}，将尝试下一个模型。")
|
||||
failed_models.add(model_name)
|
||||
last_exception = e
|
||||
continue
|
||||
continue # 切换到下一个模型
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"使用模型 '{model_name}' 时发生未知异常: {e}")
|
||||
failed_models.add(model_name)
|
||||
last_exception = e
|
||||
continue
|
||||
continue # 切换到下一个模型
|
||||
|
||||
# 所有模型都尝试失败
|
||||
logger.error("所有可用模型都已尝试失败。")
|
||||
|
||||
Reference in New Issue
Block a user