LangChain

在 Python 和 JavaScript/TypeScript 的 LangChain 中将邮箱验证集成为自定义工具。

Python

安装

pip install langchain langchain-openai requests

基础工具定义

from langchain.tools import tool
import requests
import os

@tool
def verify_email(email: str) -> dict:
    """验证电子邮箱地址是否有效且可送达。

    参数:
        email: 要验证的电子邮箱地址

    返回:
        验证结果,包括状态、可送达性和风险标志
    """
    response = requests.post(
        'https://api.emailverify.ai/v1/verify',
        headers={
            'Authorization': f'Bearer {os.environ["EMAILVERIFY_API_KEY"]}',
            'Content-Type': 'application/json',
        },
        json={'email': email}
    )
    return response.json()

与代理一起使用

from langchain.tools import tool
from langchain_openai import ChatOpenAI
from langchain.agents import AgentExecutor, create_openai_functions_agent
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
import requests
import os

@tool
def verify_email(email: str) -> dict:
    """验证电子邮箱地址是否有效且可送达。

    参数:
        email: 要验证的电子邮箱地址

    返回:
        验证结果,包括状态、可送达性和风险标志
    """
    response = requests.post(
        'https://api.emailverify.ai/v1/verify',
        headers={
            'Authorization': f'Bearer {os.environ["EMAILVERIFY_API_KEY"]}',
            'Content-Type': 'application/json',
        },
        json={'email': email}
    )
    return response.json()

# 创建代理
llm = ChatOpenAI(model="gpt-4-turbo")
tools = [verify_email]

prompt = ChatPromptTemplate.from_messages([
    ("system", "你是一个能够验证电子邮箱地址的有用助手。"),
    MessagesPlaceholder("chat_history", optional=True),
    ("human", "{input}"),
    MessagesPlaceholder("agent_scratchpad"),
])

agent = create_openai_functions_agent(llm, tools, prompt)
agent_executor = AgentExecutor(agent=agent, tools=tools)

# 使用代理
result = agent_executor.invoke({
    "input": "john@google.com 是有效的邮箱吗?"
})
print(result["output"])

批量验证工具

@tool
def verify_emails_bulk(emails: list[str]) -> dict:
    """一次性验证多个电子邮箱地址。

    参数:
        emails: 要验证的电子邮箱地址列表(最多 100 个)

    返回:
        批量验证结果
    """
    response = requests.post(
        'https://api.emailverify.ai/v1/verify/bulk',
        headers={
            'Authorization': f'Bearer {os.environ["EMAILVERIFY_API_KEY"]}',
            'Content-Type': 'application/json',
        },
        json={'emails': emails[:100]}
    )
    return response.json()

错误处理

@tool
def verify_email(email: str) -> dict:
    """验证电子邮箱地址是否有效且可送达。

    参数:
        email: 要验证的电子邮箱地址

    返回:
        验证结果或错误消息
    """
    try:
        response = requests.post(
            'https://api.emailverify.ai/v1/verify',
            headers={
                'Authorization': f'Bearer {os.environ["EMAILVERIFY_API_KEY"]}',
                'Content-Type': 'application/json',
            },
            json={'email': email},
            timeout=30
        )

        if response.status_code == 429:
            return {'error': '超出频率限制。请稍后重试。'}

        response.raise_for_status()
        return response.json()

    except requests.exceptions.Timeout:
        return {'error': '请求超时。请重试。'}
    except requests.exceptions.RequestException as e:
        return {'error': f'请求失败: {str(e)}'}

JavaScript/TypeScript

安装

npm install @langchain/core @langchain/openai langchain zod

基础工具定义

import { DynamicStructuredTool } from '@langchain/core/tools';
import { z } from 'zod';

const verifyEmailTool = new DynamicStructuredTool({
  name: 'verify_email',
  description: '验证电子邮箱地址是否有效且可送达',
  schema: z.object({
    email: z.string().email().describe('要验证的电子邮箱地址'),
  }),
  func: async ({ email }) => {
    const response = await fetch('https://api.emailverify.ai/v1/verify', {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${process.env.EMAILVERIFY_API_KEY}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify({ email }),
    });
    return JSON.stringify(await response.json());
  },
});

与代理一起使用

import { DynamicStructuredTool } from '@langchain/core/tools';
import { ChatOpenAI } from '@langchain/openai';
import { AgentExecutor, createOpenAIFunctionsAgent } from 'langchain/agents';
import { ChatPromptTemplate, MessagesPlaceholder } from '@langchain/core/prompts';
import { z } from 'zod';

const verifyEmailTool = new DynamicStructuredTool({
  name: 'verify_email',
  description: '验证电子邮箱地址是否有效且可送达',
  schema: z.object({
    email: z.string().email().describe('要验证的电子邮箱地址'),
  }),
  func: async ({ email }) => {
    const response = await fetch('https://api.emailverify.ai/v1/verify', {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${process.env.EMAILVERIFY_API_KEY}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify({ email }),
    });
    return JSON.stringify(await response.json());
  },
});

const llm = new ChatOpenAI({ modelName: 'gpt-4-turbo' });
const tools = [verifyEmailTool];

const prompt = ChatPromptTemplate.fromMessages([
  ['system', '你是一个能够验证电子邮箱地址的有用助手。'],
  new MessagesPlaceholder('chat_history'),
  ['human', '{input}'],
  new MessagesPlaceholder('agent_scratchpad'),
]);

const agent = await createOpenAIFunctionsAgent({ llm, tools, prompt });
const agentExecutor = new AgentExecutor({ agent, tools });

const result = await agentExecutor.invoke({
  input: '检查 test@example.com 是否有效',
  chat_history: [],
});

console.log(result.output);

流式传输

import { AgentExecutor, createOpenAIFunctionsAgent } from 'langchain/agents';

// ... 设置代码 ...

const stream = await agentExecutor.stream({
  input: '为我验证 john@google.com',
  chat_history: [],
});

for await (const chunk of stream) {
  if (chunk.output) {
    console.log(chunk.output);
  }
}

下一步

OpenAI 函数调用

直接 OpenAI 集成

Anthropic 工具

Claude 工具使用

API 参考

完整 API 文档

LangChain

OpenAI 函数调用

Anthropic 工具

API 参考

On this page