BruceSun
/
MaiBot
mirror of https://github.com/Mai-with-u/MaiBot.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

utils_model

pull/47/head
野原小牛 2025-03-04 15:07:45 +08:00
parent eddc9420e3
commit 6f599aad6d
1 changed files with 31 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
src/plugins/models/utils_model.py
View File

@ -4,17 +4,20 @@ import requests
import time import time
from typing import Tuple, Union from typing import Tuple, Union
from nonebot import get_driver from nonebot import get_driver
from loguru import logger
from ..chat.config import global_config from ..chat.config import global_config
driver = get_driver() driver = get_driver()
config = driver.config config = driver.config
class LLM_request: class LLM_request:
def __init__(self, model ,**kwargs): def __init__(self, model, **kwargs):
# 将大写的配置键转换为小写并从config中获取实际值 # 将大写的配置键转换为小写并从config中获取实际值
try: try:
self.api_key = getattr(config, model["key"]) self.api_key = getattr(config, model["key"])
self.base_url = getattr(config, model["base_url"]) self.base_url = getattr(config, model["base_url"])
except AttributeError as e: except AttributeError as e:
logger.error(f"配置错误：找不到对应的配置项 - {str(e)}")
raise ValueError(f"配置错误：找不到对应的配置项 - {str(e)}") raise ValueError(f"配置错误：找不到对应的配置项 - {str(e)}")
self.model_name = model["name"] self.model_name = model["name"]
self.params = kwargs self.params = kwargs
@ -35,6 +38,7 @@ class LLM_request:
# 发送请求到完整的chat/completions端点 # 发送请求到完整的chat/completions端点
api_url = f"{self.base_url.rstrip('/')}/chat/completions" api_url = f"{self.base_url.rstrip('/')}/chat/completions"
logger.info(f"发送请求到URL: {api_url}") # 记录请求的URL
max_retries = 3 max_retries = 3
base_wait_time = 15 base_wait_time = 15
@ -45,7 +49,7 @@ class LLM_request:
async with session.post(api_url, headers=headers, json=data) as response: async with session.post(api_url, headers=headers, json=data) as response:
if response.status == 429: if response.status == 429:
wait_time = base_wait_time * (2 ** retry) # 指数退避 wait_time = base_wait_time * (2 ** retry) # 指数退避
print(f"遇到请求限制(429)，等待{wait_time}秒后重试...") logger.warning(f"遇到请求限制(429)，等待{wait_time}秒后重试...")
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
continue continue
@ -61,11 +65,13 @@ class LLM_request:
except Exception as e: except Exception as e:
if retry < max_retries - 1: # 如果还有重试机会 if retry < max_retries - 1: # 如果还有重试机会
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"[回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}") logger.error(f"[回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}", exc_info=True)
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
else: else:
logger.critical(f"请求失败: {str(e)}", exc_info=True)
return f"请求失败: {str(e)}", "" return f"请求失败: {str(e)}", ""
logger.error("达到最大重试次数，请求仍然失败")
return "达到最大重试次数，请求仍然失败", "" return "达到最大重试次数，请求仍然失败", ""
async def generate_response_for_image(self, prompt: str, image_base64: str) -> Tuple[str, str]: async def generate_response_for_image(self, prompt: str, image_base64: str) -> Tuple[str, str]:
@ -100,6 +106,7 @@ class LLM_request:
# 发送请求到完整的chat/completions端点 # 发送请求到完整的chat/completions端点
api_url = f"{self.base_url.rstrip('/')}/chat/completions" api_url = f"{self.base_url.rstrip('/')}/chat/completions"
logger.info(f"发送请求到URL: {api_url}") # 记录请求的URL
max_retries = 3 max_retries = 3
base_wait_time = 15 base_wait_time = 15
@ -110,7 +117,7 @@ class LLM_request:
async with session.post(api_url, headers=headers, json=data) as response: async with session.post(api_url, headers=headers, json=data) as response:
if response.status == 429: if response.status == 429:
wait_time = base_wait_time * (2 ** retry) # 指数退避 wait_time = base_wait_time * (2 ** retry) # 指数退避
print(f"遇到请求限制(429)，等待{wait_time}秒后重试...") logger.warning(f"遇到请求限制(429)，等待{wait_time}秒后重试...")
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
continue continue
@ -126,11 +133,13 @@ class LLM_request:
except Exception as e: except Exception as e:
if retry < max_retries - 1: # 如果还有重试机会 if retry < max_retries - 1: # 如果还有重试机会
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"[image回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}") logger.error(f"[image回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}", exc_info=True)
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
else: else:
logger.critical(f"请求失败: {str(e)}", exc_info=True)
return f"请求失败: {str(e)}", "" return f"请求失败: {str(e)}", ""
logger.error("达到最大重试次数，请求仍然失败")
return "达到最大重试次数，请求仍然失败", "" return "达到最大重试次数，请求仍然失败", ""
def generate_response_for_image_sync(self, prompt: str, image_base64: str) -> Tuple[str, str]: def generate_response_for_image_sync(self, prompt: str, image_base64: str) -> Tuple[str, str]:
@ -165,6 +174,7 @@ class LLM_request:
# 发送请求到完整的chat/completions端点 # 发送请求到完整的chat/completions端点
api_url = f"{self.base_url.rstrip('/')}/chat/completions" api_url = f"{self.base_url.rstrip('/')}/chat/completions"
logger.info(f"发送请求到URL: {api_url}") # 记录请求的URL
max_retries = 2 max_retries = 2
base_wait_time = 6 base_wait_time = 6
@ -174,8 +184,8 @@ class LLM_request:
response = requests.post(api_url, headers=headers, json=data, timeout=30) response = requests.post(api_url, headers=headers, json=data, timeout=30)
if response.status_code == 429: if response.status_code == 429:
wait_time = base_wait_time * (2 ** retry) # 指数退避 wait_time = base_wait_time * (2 ** retry)
print(f"遇到请求限制(429)，等待{wait_time}秒后重试...") logger.warning(f"遇到请求限制(429)，等待{wait_time}秒后重试...")
time.sleep(wait_time) time.sleep(wait_time)
continue continue
@ -191,11 +201,13 @@ class LLM_request:
except Exception as e: except Exception as e:
if retry < max_retries - 1: # 如果还有重试机会 if retry < max_retries - 1: # 如果还有重试机会
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"[image_sync回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}") logger.error(f"[image_sync回复]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}", exc_info=True)
time.sleep(wait_time) time.sleep(wait_time)
else: else:
logger.critical(f"请求失败: {str(e)}", exc_info=True)
return f"请求失败: {str(e)}", "" return f"请求失败: {str(e)}", ""
logger.error("达到最大重试次数，请求仍然失败")
return "达到最大重试次数，请求仍然失败", "" return "达到最大重试次数，请求仍然失败", ""
def get_embedding_sync(self, text: str, model: str = "BAAI/bge-m3") -> Union[list, None]: def get_embedding_sync(self, text: str, model: str = "BAAI/bge-m3") -> Union[list, None]:
@ -220,6 +232,7 @@ class LLM_request:
} }
api_url = f"{self.base_url.rstrip('/')}/embeddings" api_url = f"{self.base_url.rstrip('/')}/embeddings"
logger.info(f"发送请求到URL: {api_url}") # 记录请求的URL
max_retries = 2 max_retries = 2
base_wait_time = 6 base_wait_time = 6
@ -230,7 +243,7 @@ class LLM_request:
if response.status_code == 429: if response.status_code == 429:
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"遇到请求限制(429)，等待{wait_time}秒后重试...") logger.warning(f"遇到请求限制(429)，等待{wait_time}秒后重试...")
time.sleep(wait_time) time.sleep(wait_time)
continue continue
@ -244,13 +257,13 @@ class LLM_request:
except Exception as e: except Exception as e:
if retry < max_retries - 1: if retry < max_retries - 1:
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"[embedding_sync]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}") logger.error(f"[embedding_sync]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}", exc_info=True)
time.sleep(wait_time) time.sleep(wait_time)
else: else:
print(f"embedding请求失败: {str(e)}") logger.critical(f"embedding请求失败: {str(e)}", exc_info=True)
return None return None
print("达到最大重试次数，embedding请求仍然失败") logger.error("达到最大重试次数，embedding请求仍然失败")
return None return None
async def get_embedding(self, text: str, model: str = "BAAI/bge-m3") -> Union[list, None]: async def get_embedding(self, text: str, model: str = "BAAI/bge-m3") -> Union[list, None]:
@ -275,6 +288,7 @@ class LLM_request:
} }
api_url = f"{self.base_url.rstrip('/')}/embeddings" api_url = f"{self.base_url.rstrip('/')}/embeddings"
logger.info(f"发送请求到URL: {api_url}") # 记录请求的URL
max_retries = 3 max_retries = 3
base_wait_time = 15 base_wait_time = 15
@ -285,7 +299,7 @@ class LLM_request:
async with session.post(api_url, headers=headers, json=data) as response: async with session.post(api_url, headers=headers, json=data) as response:
if response.status == 429: if response.status == 429:
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"遇到请求限制(429)，等待{wait_time}秒后重试...") logger.warning(f"遇到请求限制(429)，等待{wait_time}秒后重试...")
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
continue continue
@ -299,11 +313,11 @@ class LLM_request:
except Exception as e: except Exception as e:
if retry < max_retries - 1: if retry < max_retries - 1:
wait_time = base_wait_time * (2 ** retry) wait_time = base_wait_time * (2 ** retry)
print(f"[embedding]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}") logger.error(f"[embedding]请求失败，等待{wait_time}秒后重试... 错误: {str(e)}", exc_info=True)
await asyncio.sleep(wait_time) await asyncio.sleep(wait_time)
else: else:
print(f"embedding请求失败: {str(e)}") logger.critical(f"embedding请求失败: {str(e)}", exc_info=True)
return None return None
print("达到最大重试次数，embedding请求仍然失败") logger.error("达到最大重试次数，embedding请求仍然失败")
return None return None