赞
踩
使用 FastAPI 可以帮助我们更简单高效地部署 AI 交互业务。FastAPI 提供了快速构建 API 的能力,开发者可以轻松地定义模型需要的输入和输出格式,并编写好相应的业务逻辑。
FastAPI 的异步高性能架构,可以有效支持大量并发的预测请求,为用户提供流畅的交互体验。此外,FastAPI 还提供了容器化部署能力,开发者可以轻松打包 AI 模型为 Docker 镜像,实现跨环境的部署和扩展。
总之,使用 FastAPI 可以大大提高 AI 应用程序的开发效率和用户体验,为 AI 模型的部署和交互提供全方位的支持。
LangChain基础入门:开源模型应用落地-FastAPI-助力模型交互-WebSocket篇(一),本篇学习如何集成LangChain进行模型交互,并使用工具获取实时信息
FastAPI 是一个用于构建 API 的现代、快速(高性能)的 Python Web 框架。它是基于标准 Python 类型注释的 ASGI (Asynchronous Server Gateway Interface) 框架。
FastAPI 具有以下主要特点:
快速: FastAPI 使用 ASGI 服务器和 Starlette 框架,在性能测试中表现出色。它可以与 Uvicorn 一起使用,提供非常高的性能。
简单: FastAPI 利用 Python 类型注释,使 API 定义变得简单且直观。开发人员只需要定义输入和输出模型,FastAPI 会自动生成 API 文档。
现代: FastAPI 支持 OpenAPI 标准,可以自动生成 API 文档和交互式文档。它还支持 JSON Schema 和数据验证。
全功能: FastAPI 提供了路由、依赖注入、数据验证、安全性、测试等功能,是一个功能齐全的 Web 框架。
可扩展: FastAPI 被设计为可扩展的。开发人员可以轻松地集成其他库和组件,如数据库、身份验证等。
是一种计算机通信协议,它提供了在单个 TCP 连接上进行全双工通信的机制。它是 HTML5 一个重要的组成部分。
WebSocket 协议主要有以下特点:
全双工通信:WebSocket 允许客户端和服务器之间进行双向实时通信,即数据可以同时在两个方向上流动。这与传统的 HTTP 请求-响应模型不同,HTTP 中数据只能单向流动。
持久性连接:WebSocket 连接是一种持久性的连接,一旦建立就会一直保持,直到客户端或服务器主动关闭连接。这与 HTTP 的连接是短暂的不同。
低开销:相比 HTTP 请求-响应模型,WebSocket 在建立连接时需要较少的数据交换,因此网络开销较小。
实时性:由于 WebSocket 连接是持久性的,且数据可以双向流动,因此 WebSocket 非常适用于需要实时、低延迟数据交互的应用场景,如聊天应用、实时游戏、股票行情等。
Tool(工具)是为了增强其语言模型的功能和实用性而设计的一系列辅助手段,用于扩展模型的能力。例如代码解释器(Code Interpreter)和知识检索(Knowledge Retrieval)等都属于其工具。
https://github.com/langchain-ai/langchain/tree/v0.1.16/docs/docs/integrations/tools
基本这些工具能满足大部分需求,具体使用参见:
增加Google Search的依赖包
- conda create -n fastapi_test python=3.10
- conda activate fastapi_test
- pip install fastapi websockets uvicorn
- pip install --quiet langchain-core langchain-community langchain-openai
- pip install google-search-results
1. 输入注册信息
可以使用Google账号登录,但仍要执行下面的认证操作
2. 需要认证邮箱
3. 需要认证手机
4. 认证成功
- # -*- coding: utf-8 -*-
- import os
-
- from langchain_community.utilities.serpapi import SerpAPIWrapper
-
- os.environ["SERPAPI_API_KEY"] = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
- serp = SerpAPIWrapper()
- result = serp.run("广州的实时气温如何?")
- print("实时搜索结果:", result)
调用结果:
本章代码将开源模型应用落地-FastAPI-助力模型交互-WebSocket篇(三)基础上进行拓展
服务端:
- import uvicorn
- import os
-
- from typing import Annotated
- from fastapi import (
- Depends,
- FastAPI,
- WebSocket,
- WebSocketException,
- WebSocketDisconnect,
- status,
- )
- from langchain.agents import create_structured_chat_agent, AgentExecutor
- from langchain_community.utilities import SerpAPIWrapper
-
- from langchain_core.prompts import ChatPromptTemplate, SystemMessagePromptTemplate, HumanMessagePromptTemplate
- from langchain_core.tools import tool
- from langchain_openai import ChatOpenAI
-
- os.environ["OPENAI_API_KEY"] = 'sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx' #你的Open AI Key
- os.environ["SERPAPI_API_KEY"] = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
-
-
- class ConnectionManager:
- def __init__(self):
- self.active_connections: list[WebSocket] = []
-
- async def connect(self, websocket: WebSocket):
- await websocket.accept()
- self.active_connections.append(websocket)
-
- def disconnect(self, websocket: WebSocket):
- self.active_connections.remove(websocket)
-
- async def send_personal_message(self, message: str, websocket: WebSocket):
- await websocket.send_text(message)
-
- async def broadcast(self, message: str):
- for connection in self.active_connections:
- await connection.send_text(message)
-
- manager = ConnectionManager()
-
- app = FastAPI()
-
- async def authenticate(
- websocket: WebSocket,
- userid: str,
- secret: str,
- ):
- if userid is None or secret is None:
- raise WebSocketException(code=status.WS_1008_POLICY_VIOLATION)
-
- print(f'userid: {userid},secret: {secret}')
- if '12345' == userid and 'xxxxxxxxxxxxxxxxxxxxxxxxxx' == secret:
- return 'pass'
- else:
- return 'fail'
-
- @tool
- def search(query:str):
- """只有需要了解实时信息或不知道的事情的时候才会使用这个工具,需要传入要搜索的内容。"""
- serp = SerpAPIWrapper()
- result = serp.run(query)
- print("实时搜索结果:", result)
- return result
-
-
- def get_prompt():
- template='''
- Respond to the human as helpfully and accurately as possible. You have access to the following tools:
-
- {tools}
-
- Use a json blob to specify a tool by providing an action key (tool name) and an action_input key (tool input).
-
- Valid "action" values: "Final Answer" or {tool_names}
-
- Provide only ONE action per $JSON_BLOB, as shown:
-
- ```
-
- {{
-
- "action": $TOOL_NAME,
-
- "action_input": $INPUT
-
- }}
-
- ```
-
- Follow this format:
-
- Question: input question to answer
-
- Thought: consider previous and subsequent steps
-
- Action:
-
- ```
-
- $JSON_BLOB
-
- ```
-
- Observation: action result
-
- ... (repeat Thought/Action/Observation N times)
-
- Thought: I know what to respond
-
- Action:
-
- ```
-
- {{
-
- "action": "Final Answer",
-
- "action_input": "Final response to human"
-
- }}
-
- Begin! Reminder to ALWAYS respond with a valid json blob of a single action. Use tools if necessary. Respond directly if appropriate. Format is Action:```$JSON_BLOB```then Observation
- '''
- system_message_prompt = SystemMessagePromptTemplate.from_template(template)
- human_template='''
- {input}
-
- {agent_scratchpad}
-
- (reminder to respond in a JSON blob no matter what)
- '''
- human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
- prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
-
- return prompt
-
- async def chat(query):
- global llm,tools
- agent = create_structured_chat_agent(
- llm, tools, get_prompt()
- )
-
- agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True, handle_parsing_errors=True)
-
- result = agent_executor.invoke({"input": query})
- print(result['output'])
- yield result['output']
-
- @app.websocket("/ws")
- async def websocket_endpoint(*,websocket: WebSocket,userid: str,permission: Annotated[str, Depends(authenticate)],):
- await manager.connect(websocket)
- try:
- while True:
- text = await websocket.receive_text()
-
- if 'fail' == permission:
- await manager.send_personal_message(
- f"authentication failed", websocket
- )
- else:
- if text is not None and len(text) > 0:
- async for msg in chat(text):
- await manager.send_personal_message(msg, websocket)
-
- except WebSocketDisconnect:
- manager.disconnect(websocket)
- print(f"Client #{userid} left the chat")
- await manager.broadcast(f"Client #{userid} left the chat")
-
- if __name__ == '__main__':
- tools = [search]
- llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0, max_tokens=512)
- uvicorn.run(app, host='0.0.0.0',port=7777)
客户端:
- <!DOCTYPE html>
- <html>
- <head>
- <title>Chat</title>
- </head>
- <body>
- <h1>WebSocket Chat</h1>
- <form action="" onsubmit="sendMessage(event)">
- <label>USERID: <input type="text" id="userid" autocomplete="off" value="12345"/></label>
- <label>SECRET: <input type="text" id="secret" autocomplete="off" value="xxxxxxxxxxxxxxxxxxxxxxxxxx"/></label>
- <br/>
- <button onclick="connect(event)">Connect</button>
- <hr>
- <label>Message: <input type="text" id="messageText" autocomplete="off"/></label>
- <button>Send</button>
- </form>
- <ul id='messages'>
- </ul>
- <script>
- var ws = null;
- function connect(event) {
- var userid = document.getElementById("userid")
- var secret = document.getElementById("secret")
- ws = new WebSocket("ws://localhost:7777/ws?userid="+userid.value+"&secret=" + secret.value);
- ws.onmessage = function(event) {
- var messages = document.getElementById('messages')
- var message = document.createElement('li')
- var content = document.createTextNode(event.data)
- message.appendChild(content)
- messages.appendChild(message)
- };
- event.preventDefault()
- }
- function sendMessage(event) {
- var input = document.getElementById("messageText")
- ws.send(input.value)
- input.value = ''
- event.preventDefault()
- }
- </script>
- </body>
- </html>
调用结果:
用户输入:你好
不需要触发工具调用
模型输出:你好!有什么我可以帮忙的吗?
用户输入:广州现在天气如何?
需要调用工具
模型输出:The current weather in Guangzhou is partly cloudy with a temperature of 95°F, 66% chance of precipitation, 58% humidity, and wind speed of 16 mph. This information was last updated on Monday at 1:00 PM.
PS:
1. 在AI交互中,LangChain框架并不是必须引入,此处引用仅用于简化Openai的交互流程。
2. 页面输出的样式可以根据实际需要进行调整,此处仅用于演示效果。
3. 目前还遗留两个问题,一是如何实现流式输出,二是如何更好维护prompt模版,篇幅有限,下回分解
在提示词模版加入:Remember to answer in Chinese. 暗示模型一定要以中文进行回复。
修改后的提示语为:
- Respond to the human as helpfully and accurately as possible. You have access to the following tools:
-
- {tools}
-
- Use a json blob to specify a tool by providing an action key (tool name) and an action_input key (tool input).
-
- Valid "action" values: "Final Answer" or {tool_names}
-
- Provide only ONE action per $JSON_BLOB, as shown:
-
- ```
-
- {{
-
- "action": $TOOL_NAME,
-
- "action_input": $INPUT
-
- }}
-
- ```
-
- Follow this format:
-
- Question: input question to answer
-
- Thought: consider previous and subsequent steps
-
- Action:
-
- ```
-
- $JSON_BLOB
-
- ```
-
- Observation: action result
-
- ... (repeat Thought/Action/Observation N times)
-
- Thought: I know what to respond
-
- Action:
-
- ```
-
- {{
-
- "action": "Final Answer",
-
- "action_input": "Final response to human"
-
- }}
-
- Begin! Reminder to ALWAYS respond with a valid json blob of a single action. Use tools if necessary. Respond directly if appropriate. Remember to answer in Chinese.Format is Action:```$JSON_BLOB```then Observation
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。