import { ICommonObject, INode, INodeData, INodeParams } from '../../../src/Interface' import { initializeAgentExecutorWithOptions, AgentExecutor } from 'langchain/agents' import { getBaseClasses, mapChatHistory } from '../../../src/utils' import { flatten } from 'lodash' import { BaseChatMemory } from 'langchain/memory' import { ConsoleCallbackHandler, CustomChainHandler } from '../../../src/handler' class ConversationalRetrievalAgent_Agents implements INode { label: string name: string version: number description: string type: string icon: string category: string baseClasses: string[] inputs: INodeParams[] constructor() { this.label = 'Conversational Retrieval Agent' this.name = 'conversationalRetrievalAgent' this.version = 1.0 this.type = 'AgentExecutor' this.category = 'Agents' this.icon = 'agent.svg' this.description = `An agent optimized for retrieval during conversation, answering questions based on past dialogue, all using OpenAI's Function Calling` this.baseClasses = [this.type, ...getBaseClasses(AgentExecutor)] this.inputs = [ { label: 'Allowed Tools', name: 'tools', type: 'Tool', list: true }, { label: 'Memory', name: 'memory', type: 'BaseChatMemory' }, { label: 'OpenAI Chat Model', name: 'model', type: 'ChatOpenAI' }, { label: 'System Message', name: 'systemMessage', type: 'string', rows: 4, optional: true, additionalParams: true } ] } async init(nodeData: INodeData): Promise { const model = nodeData.inputs?.model const memory = nodeData.inputs?.memory as BaseChatMemory const systemMessage = nodeData.inputs?.systemMessage as string let tools = nodeData.inputs?.tools tools = flatten(tools) const executor = await initializeAgentExecutorWithOptions(tools, model, { agentType: 'openai-functions', verbose: process.env.DEBUG === 'true' ? true : false, agentArgs: { prefix: systemMessage ?? `You are a helpful AI assistant.` }, returnIntermediateSteps: true }) executor.memory = memory return executor } async run(nodeData: INodeData, input: string, options: ICommonObject): Promise { const executor = nodeData.instance as AgentExecutor if (executor.memory) { ;(executor.memory as any).memoryKey = 'chat_history' ;(executor.memory as any).outputKey = 'output' ;(executor.memory as any).chatHistory = mapChatHistory(options) } const loggerHandler = new ConsoleCallbackHandler(options.logger) if (options.socketIO && options.socketIOClientId) { const handler = new CustomChainHandler(options.socketIO, options.socketIOClientId) const result = await executor.call({ input }, [loggerHandler, handler]) return result?.output } else { const result = await executor.call({ input }, [loggerHandler]) return result?.output } } } module.exports = { nodeClass: ConversationalRetrievalAgent_Agents }