288 lines
8.3 KiB
Python
288 lines
8.3 KiB
Python
from collections.abc import AsyncGenerator
|
||
from dataclasses import dataclass
|
||
|
||
from openai import AsyncOpenAI
|
||
from openai.types.chat import ChatCompletion, ChatCompletionChunk
|
||
|
||
|
||
@dataclass
|
||
class ChatMessage:
|
||
"""聊天消息数据类"""
|
||
|
||
role: str
|
||
content: str
|
||
|
||
def to_dict(self) -> dict[str, str]:
|
||
return {"role": self.role, "content": self.content}
|
||
|
||
|
||
class AsyncOpenAIClient:
|
||
"""异步OpenAI客户端，支持流式传输"""
|
||
|
||
def __init__(self, api_key: str, base_url: str | None = None):
|
||
"""
|
||
初始化客户端
|
||
|
||
Args:
|
||
api_key: OpenAI API密钥
|
||
base_url: 可选的API基础URL，用于自定义端点
|
||
"""
|
||
self.client = AsyncOpenAI(
|
||
api_key=api_key,
|
||
base_url=base_url,
|
||
timeout=10.0, # 设置60秒的全局超时
|
||
)
|
||
|
||
async def chat_completion(
|
||
self,
|
||
messages: list[ChatMessage | dict[str, str]],
|
||
model: str = "gpt-3.5-turbo",
|
||
temperature: float = 0.7,
|
||
max_tokens: int | None = None,
|
||
**kwargs,
|
||
) -> ChatCompletion:
|
||
"""
|
||
非流式聊天完成
|
||
|
||
Args:
|
||
messages: 消息列表
|
||
model: 模型名称
|
||
temperature: 温度参数
|
||
max_tokens: 最大token数
|
||
**kwargs: 其他参数
|
||
|
||
Returns:
|
||
完整的聊天回复
|
||
"""
|
||
# 转换消息格式
|
||
formatted_messages = []
|
||
for msg in messages:
|
||
if isinstance(msg, ChatMessage):
|
||
formatted_messages.append(msg.to_dict())
|
||
else:
|
||
formatted_messages.append(msg)
|
||
|
||
extra_body = {}
|
||
if kwargs.get("enable_thinking") is not None:
|
||
extra_body["enable_thinking"] = kwargs.pop("enable_thinking")
|
||
if kwargs.get("thinking_budget") is not None:
|
||
extra_body["thinking_budget"] = kwargs.pop("thinking_budget")
|
||
|
||
response = await self.client.chat.completions.create(
|
||
model=model,
|
||
messages=formatted_messages,
|
||
temperature=temperature,
|
||
max_tokens=max_tokens,
|
||
stream=False,
|
||
extra_body=extra_body if extra_body else None,
|
||
**kwargs,
|
||
)
|
||
|
||
return response
|
||
|
||
async def chat_completion_stream(
|
||
self,
|
||
messages: list[ChatMessage | dict[str, str]],
|
||
model: str = "gpt-3.5-turbo",
|
||
temperature: float = 0.7,
|
||
max_tokens: int | None = None,
|
||
**kwargs,
|
||
) -> AsyncGenerator[ChatCompletionChunk, None]:
|
||
"""
|
||
流式聊天完成
|
||
|
||
Args:
|
||
messages: 消息列表
|
||
model: 模型名称
|
||
temperature: 温度参数
|
||
max_tokens: 最大token数
|
||
**kwargs: 其他参数
|
||
|
||
Yields:
|
||
ChatCompletionChunk: 流式响应块
|
||
"""
|
||
# 转换消息格式
|
||
formatted_messages = []
|
||
for msg in messages:
|
||
if isinstance(msg, ChatMessage):
|
||
formatted_messages.append(msg.to_dict())
|
||
else:
|
||
formatted_messages.append(msg)
|
||
|
||
extra_body = {}
|
||
if kwargs.get("enable_thinking") is not None:
|
||
extra_body["enable_thinking"] = kwargs.pop("enable_thinking")
|
||
if kwargs.get("thinking_budget") is not None:
|
||
extra_body["thinking_budget"] = kwargs.pop("thinking_budget")
|
||
|
||
stream = await self.client.chat.completions.create(
|
||
model=model,
|
||
messages=formatted_messages,
|
||
temperature=temperature,
|
||
max_tokens=max_tokens,
|
||
stream=True,
|
||
extra_body=extra_body if extra_body else None,
|
||
**kwargs,
|
||
)
|
||
|
||
async for chunk in stream:
|
||
yield chunk
|
||
|
||
async def get_stream_content(
|
||
self,
|
||
messages: list[ChatMessage | dict[str, str]],
|
||
model: str = "gpt-3.5-turbo",
|
||
temperature: float = 0.7,
|
||
max_tokens: int | None = None,
|
||
**kwargs,
|
||
) -> AsyncGenerator[str, None]:
|
||
"""
|
||
获取流式内容（只返回文本内容）
|
||
|
||
Args:
|
||
messages: 消息列表
|
||
model: 模型名称
|
||
temperature: 温度参数
|
||
max_tokens: 最大token数
|
||
**kwargs: 其他参数
|
||
|
||
Yields:
|
||
str: 文本内容片段
|
||
"""
|
||
async for chunk in self.chat_completion_stream(
|
||
messages=messages, model=model, temperature=temperature, max_tokens=max_tokens, **kwargs
|
||
):
|
||
if chunk.choices and chunk.choices[0].delta.content:
|
||
yield chunk.choices[0].delta.content
|
||
|
||
async def collect_stream_response(
|
||
self,
|
||
messages: list[ChatMessage | dict[str, str]],
|
||
model: str = "gpt-3.5-turbo",
|
||
temperature: float = 0.7,
|
||
max_tokens: int | None = None,
|
||
**kwargs,
|
||
) -> str:
|
||
"""
|
||
收集完整的流式响应
|
||
|
||
Args:
|
||
messages: 消息列表
|
||
model: 模型名称
|
||
temperature: 温度参数
|
||
max_tokens: 最大token数
|
||
**kwargs: 其他参数
|
||
|
||
Returns:
|
||
str: 完整的响应文本
|
||
"""
|
||
full_response = ""
|
||
async for content in self.get_stream_content(
|
||
messages=messages, model=model, temperature=temperature, max_tokens=max_tokens, **kwargs
|
||
):
|
||
full_response += content
|
||
|
||
return full_response
|
||
|
||
async def close(self):
|
||
"""关闭客户端"""
|
||
await self.client.close()
|
||
|
||
async def __aenter__(self):
|
||
"""异步上下文管理器入口"""
|
||
return self
|
||
|
||
async def __aexit__(self, exc_type, exc_val, exc_tb):
|
||
"""异步上下文管理器退出"""
|
||
await self.close()
|
||
|
||
|
||
class ConversationManager:
|
||
"""对话管理器，用于管理对话历史"""
|
||
|
||
def __init__(self, client: AsyncOpenAIClient, system_prompt: str | None = None):
|
||
"""
|
||
初始化对话管理器
|
||
|
||
Args:
|
||
client: OpenAI客户端实例
|
||
system_prompt: 系统提示词
|
||
"""
|
||
self.client = client
|
||
self.messages: list[ChatMessage] = []
|
||
|
||
if system_prompt:
|
||
self.messages.append(ChatMessage(role="system", content=system_prompt))
|
||
|
||
def add_user_message(self, content: str):
|
||
"""添加用户消息"""
|
||
self.messages.append(ChatMessage(role="user", content=content))
|
||
|
||
def add_assistant_message(self, content: str):
|
||
"""添加助手消息"""
|
||
self.messages.append(ChatMessage(role="assistant", content=content))
|
||
|
||
async def send_message_stream(
|
||
self, content: str, model: str = "gpt-3.5-turbo", **kwargs
|
||
) -> AsyncGenerator[str, None]:
|
||
"""
|
||
发送消息并获取流式响应
|
||
|
||
Args:
|
||
content: 用户消息内容
|
||
model: 模型名称
|
||
**kwargs: 其他参数
|
||
|
||
Yields:
|
||
str: 响应内容片段
|
||
"""
|
||
self.add_user_message(content)
|
||
|
||
response_content = ""
|
||
async for chunk in self.client.get_stream_content(messages=self.messages, model=model, **kwargs):
|
||
response_content += chunk
|
||
yield chunk
|
||
|
||
self.add_assistant_message(response_content)
|
||
|
||
async def send_message(self, content: str, model: str = "gpt-3.5-turbo", **kwargs) -> str:
|
||
"""
|
||
发送消息并获取完整响应
|
||
|
||
Args:
|
||
content: 用户消息内容
|
||
model: 模型名称
|
||
**kwargs: 其他参数
|
||
|
||
Returns:
|
||
str: 完整响应
|
||
"""
|
||
self.add_user_message(content)
|
||
|
||
response = await self.client.chat_completion(messages=self.messages, model=model, **kwargs)
|
||
|
||
response_content = response.choices[0].message.content
|
||
self.add_assistant_message(response_content)
|
||
|
||
return response_content
|
||
|
||
def clear_history(self, keep_system: bool = True):
|
||
"""
|
||
清除对话历史
|
||
|
||
Args:
|
||
keep_system: 是否保留系统消息
|
||
"""
|
||
if keep_system and self.messages and self.messages[0].role == "system":
|
||
self.messages = [self.messages[0]]
|
||
else:
|
||
self.messages = []
|
||
|
||
def get_message_count(self) -> int:
|
||
"""获取消息数量"""
|
||
return len(self.messages)
|
||
|
||
def get_conversation_history(self) -> list[dict[str, str]]:
|
||
"""获取对话历史"""
|
||
return [msg.to_dict() for msg in self.messages]
|