Redis 支持的聊天记忆

为了在聊天会话之间实现更长期的持久性，您可以将支持聊天记忆类（如 BufferMemory）的默认内存中 chatHistory 替换为 Redis 实例。

设置

您需要在项目中安装 node-redis

提示

查看本节了解安装集成包的通用说明。

npm

npm install @langchain/openai redis @langchain/community

Yarn

yarn add @langchain/openai redis @langchain/community

pnpm

pnpm add @langchain/openai redis @langchain/community

您还需要一个 Redis 实例来连接。有关在本地运行服务器的说明，请查看Redis 官方网站。

使用

存储在 Redis 中的每个聊天历史会话都必须具有唯一的 ID。您可以提供可选的 sessionTTL，以使会话在给定秒数后过期。config 参数直接传递给 node-redis 的 createClient 方法，并接受相同的参数。

import { BufferMemory } from "langchain/memory";
import { RedisChatMessageHistory } from "@langchain/redis";
import { ChatOpenAI } from "@langchain/openai";
import { ConversationChain } from "langchain/chains";

const memory = new BufferMemory({
  chatHistory: new RedisChatMessageHistory({
    sessionId: new Date().toISOString(), // Or some other unique identifier for the conversation
    sessionTTL: 300, // 5 minutes, omit this parameter to make sessions never expire
  }),
});

const model = new ChatOpenAI({
  model: "gpt-3.5-turbo",
  temperature: 0,
});

const chain = new ConversationChain({ llm: model, memory });

const res1 = await chain.invoke({ input: "Hi! I'm Jim." });
console.log({ res1 });
/*
{
  res1: {
    text: "Hello Jim! It's nice to meet you. My name is AI. How may I assist you today?"
  }
}
*/

const res2 = await chain.invoke({ input: "What did I just say my name was?" });
console.log({ res2 });

/*
{
  res1: {
    text: "You said your name was Jim."
  }
}
*/

API 参考

• BufferMemory 来自 langchain/memory
• RedisChatMessageHistory 来自 @langchain/redis
• ChatOpenAI 来自 @langchain/openai
• ConversationChain 来自 langchain/chains

高级使用

您还可以直接传入之前创建的 node-redis 客户端实例

import { Redis } from "ioredis";
import { BufferMemory } from "langchain/memory";
import { RedisChatMessageHistory } from "@langchain/community/stores/message/ioredis";
import { ChatOpenAI } from "@langchain/openai";
import { ConversationChain } from "langchain/chains";

const client = new Redis("redis://localhost:6379");

const memory = new BufferMemory({
  chatHistory: new RedisChatMessageHistory({
    sessionId: new Date().toISOString(),
    sessionTTL: 300,
    client,
  }),
});

const model = new ChatOpenAI({
  model: "gpt-3.5-turbo",
  temperature: 0,
});

const chain = new ConversationChain({ llm: model, memory });

const res1 = await chain.invoke({ input: "Hi! I'm Jim." });
console.log({ res1 });
/*
{
  res1: {
    text: "Hello Jim! It's nice to meet you. My name is AI. How may I assist you today?"
  }
}
*/

const res2 = await chain.invoke({ input: "What did I just say my name was?" });
console.log({ res2 });

/*
{
  res1: {
    text: "You said your name was Jim."
  }
}
*/

API 参考

• BufferMemory 来自 langchain/memory
• RedisChatMessageHistory 来自 @langchain/community/stores/message/ioredis
• ChatOpenAI 来自 @langchain/openai
• ConversationChain 来自 langchain/chains

Redis Sentinel 支持

您可以使用 ioredis 启用 Redis Sentinel 支持的缓存

这需要在项目中安装 ioredis。

npm

npm install ioredis

Yarn

yarn add ioredis

pnpm

pnpm add ioredis

import { Redis } from "ioredis";
import { BufferMemory } from "langchain/memory";
import { RedisChatMessageHistory } from "@langchain/community/stores/message/ioredis";
import { ChatOpenAI } from "@langchain/openai";
import { ConversationChain } from "langchain/chains";

// Uses ioredis to facilitate Sentinel Connections see their docs for details on setting up more complex Sentinels: https://github.com/redis/ioredis#sentinel
const client = new Redis({
  sentinels: [
    { host: "localhost", port: 26379 },
    { host: "localhost", port: 26380 },
  ],
  name: "mymaster",
});

const memory = new BufferMemory({
  chatHistory: new RedisChatMessageHistory({
    sessionId: new Date().toISOString(),
    sessionTTL: 300,
    client,
  }),
});

const model = new ChatOpenAI({ temperature: 0.5 });

const chain = new ConversationChain({ llm: model, memory });

const res1 = await chain.invoke({ input: "Hi! I'm Jim." });
console.log({ res1 });
/*
{
  res1: {
    text: "Hello Jim! It's nice to meet you. My name is AI. How may I assist you today?"
  }
}
*/

const res2 = await chain.invoke({ input: "What did I just say my name was?" });
console.log({ res2 });

/*
{
  res1: {
    text: "You said your name was Jim."
  }
}
*/

API 参考

• BufferMemory 来自 langchain/memory
• RedisChatMessageHistory 来自 @langchain/community/stores/message/ioredis
• ChatOpenAI 来自 @langchain/openai
• ConversationChain 来自 langchain/chains