96 lines
3.0 KiB
TypeScript
96 lines
3.0 KiB
TypeScript
import { ICommonObject, IMessage, INode, INodeData, INodeParams } from '../../../src/Interface'
|
|
import { getBaseClasses } from '../../../src/utils'
|
|
import { ConversationalRetrievalQAChain } from 'langchain/chains'
|
|
import { BaseLLM } from 'langchain/llms/base'
|
|
import { BaseRetriever } from 'langchain/schema'
|
|
|
|
class ConversationalRetrievalQAChain_Chains implements INode {
|
|
label: string
|
|
name: string
|
|
type: string
|
|
icon: string
|
|
category: string
|
|
baseClasses: string[]
|
|
description: string
|
|
inputs: INodeParams[]
|
|
|
|
constructor() {
|
|
this.label = 'Conversational Retrieval QA Chain'
|
|
this.name = 'conversationalRetrievalQAChain'
|
|
this.type = 'ConversationalRetrievalQAChain'
|
|
this.icon = 'chain.svg'
|
|
this.category = 'Chains'
|
|
this.description = 'Document QA - built on RetrievalQAChain to provide a chat history component'
|
|
this.baseClasses = [this.type, ...getBaseClasses(ConversationalRetrievalQAChain)]
|
|
this.inputs = [
|
|
{
|
|
label: 'LLM',
|
|
name: 'llm',
|
|
type: 'BaseLLM'
|
|
},
|
|
{
|
|
label: 'Vector Store Retriever',
|
|
name: 'vectorStoreRetriever',
|
|
type: 'BaseRetriever'
|
|
}
|
|
]
|
|
}
|
|
|
|
async init(nodeData: INodeData): Promise<any> {
|
|
const llm = nodeData.inputs?.llm as BaseLLM
|
|
const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever as BaseRetriever
|
|
|
|
const chain = ConversationalRetrievalQAChain.fromLLM(llm, vectorStoreRetriever)
|
|
return chain
|
|
}
|
|
|
|
async run(nodeData: INodeData, input: string, options: ICommonObject): Promise<string> {
|
|
const chain = nodeData.instance as ConversationalRetrievalQAChain
|
|
let chatHistory = ''
|
|
|
|
if (options && options.chatHistory) {
|
|
const histories: IMessage[] = options.chatHistory
|
|
chatHistory = histories
|
|
.map((item) => {
|
|
return item.message
|
|
})
|
|
.join('')
|
|
}
|
|
|
|
const obj = {
|
|
question: input,
|
|
chat_history: chatHistory ? chatHistory : []
|
|
}
|
|
|
|
const res = await chain.call(obj)
|
|
|
|
return res?.text
|
|
}
|
|
|
|
jsCodeImport(): string {
|
|
return `import { ConversationalRetrievalQAChain } from 'langchain/chains'`
|
|
}
|
|
|
|
jsCode(nodeData: INodeData): string {
|
|
const llm = nodeData.inputs?.llm as string
|
|
const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever as string
|
|
|
|
const code = `const input = "<your question>"
|
|
const chatHistory = "<your chat history>"
|
|
const llm = ${llm}
|
|
${vectorStoreRetriever}
|
|
|
|
const chain = await ConversationalRetrievalQAChain.fromLLM(llm, vectorStoreRetriever)
|
|
const result = await chain.call({
|
|
question: input,
|
|
chat_history: chatHistory ? chatHistory : []
|
|
})
|
|
|
|
console.log(result)
|
|
`
|
|
return code
|
|
}
|
|
}
|
|
|
|
module.exports = { nodeClass: ConversationalRetrievalQAChain_Chains }
|