chat.js

import { ConversationalRetrievalQAChain } from "langchain/chains"
import { PineconeClient } from "@pinecone-database/pinecone"  
import { PineconeStore } from "langchain/vectorstores/pinecone" 
import { OpenAIEmbeddings } from "langchain/embeddings/openai"
import { OpenAI } from "langchain/llms/openai"
import dotenv from 'dotenv'
dotenv.config()

//initialize the OpenAI model to use to answer the question
const model = new OpenAI({
    temperature:0
})

//initialize vector library
const pineconeClient = new PineconeClient()
await pineconeClient.init({
    apiKey: process.env.PINECONE_API_KEY,
    environment: process.env.PINECONE_ENVIRONMENT,
})
const pineconeIndex = pineconeClient.Index(process.env.PINECONE_INDEX)
const pineconeStore = await PineconeStore.fromExistingIndex(new OpenAIEmbeddings(),{
    pineconeIndex,
    textKey:"text",
    namespace:"vue3-document"
}) 

//create the chain using the OpenAI model
const chain = ConversationalRetrievalQAChain.fromLLM(model,pineconeStore.asRetriever(),{
    returnSourceDocuments:true  
})

//ask question
const res = await chain.call({
    question:"我是vue新手，给我一些学习建议",
    chat_history:[]
})

//ask question with previous response as chat history.
const secondRes = await chain.call({
    question:"是否可以提供更多的建议",
    chat_history:["我是vue新手，给我一些学习建议",res.text]
})

console.log(secondRes)