What would be the best way to add memory to the code below?

[Elindo586]

Given the code below, what would be the best way to add memory, or to apply a new code to include a prompt, memory, and keep the same functionality as this code:

import { TextLoader } from "langchain/document_loaders/fs/text";
import { CharacterTextSplitter } from "langchain/text_splitter";
import { OpenAIEmbeddings } from "langchain/embeddings/openai";
import { FaissStore } from "langchain/vectorstores/faiss";

import { OpenAI } from "langchain/llms/openai";
import { PromptTemplate } from "langchain/prompts";
import {
  SimpleSequentialChain,
  LLMChain,
  RetrievalQAChain,
  loadQAStuffChain,
} from "langchain/chains";

let loader;
let docs;
let splitter;
let documents;
let embeddings1;
let vectorstore;

let model1;

let embeddings;
let vectoreStore;
let vectorStoreRetriever;
let vectorChain;
// let overallChain;

const useStore = async (req, res) => {
  if (req.method === "POST") {
    loader = new TextLoader("./documents/restaurant.txt");
    docs = await loader.load();
    splitter = new CharacterTextSplitter({
      chunkSize: 200,
      chunkOverlap: 50,
    });
    documents = await splitter.splitDocuments(docs);
    embeddings1 = new OpenAIEmbeddings();
    vectorstore = await FaissStore.fromDocuments(documents, embeddings1);
    await vectorstore.save("./");

    const message = req.body.newText;
    const firstOne = req.body.firstMsg;

    console.log(message);
    if (firstOne) {
      model1 = new OpenAI({
        openAIApiKey: process.env.OPENAI_API_KEY,
        temperature: 0,
      });

      embeddings = new OpenAIEmbeddings();
      vectoreStore = await FaissStore.load("./", embeddings);
      vectorStoreRetriever = vectoreStore.asRetriever();

      vectorChain = RetrievalQAChain.fromLLM(model1, vectorStoreRetriever);
    }

    const result = await vectorChain.call({
      query: message,
    });

    console.log(result);

    return res.status(200).json({ output: result });
  } else {
    res.status(405).json({ say: "Only POST is allowed" });
  }
};
export default useStore;

[dosubot[bot]]

🤖

To add memory to your code, you can use the PromptGenerator class and add a new property memory to it. You can then create methods to add and retrieve memory. Here's an example of how you can do it:

export class PromptGenerator {
  // ...
  memory: string[];

  constructor() {
    // ...
    this.memory = [];
  }

  /**
   * Adds a memory to the list of memories.
   * @param memory The memory to add.
   * @returns void
   */
  add_memory(memory: string): void {
    this.memory.push(memory);
  }

  /**
   * Generates a memory string that includes the memories.
   * @returns A string representing the memory.
   */
  generate_memory_string(): string {
    return this._generate_numbered_list(this.memory);
  }

  /**
   * Generates a prompt string that includes the constraints, commands,
   * resources, performance evaluations, memory, and response format.
   * @returns A string representing the prompt.
   */
  generate_prompt_string(): string {
    // ...
    const prompt_string =
      `Constraints:\n${this._generate_numbered_list(this.constraints)}\n\n` +
      `Commands:\n${this._generate_numbered_list(
        this.commands,
        "command"
      )}\n\n` +
      `Resources:\n${this._generate_numbered_list(this.resources)}\n\n` +
      `Memory:\n${this.generate_memory_string()}\n\n` + // Add memory here
      `Performance Evaluation:\n${this._generate_numbered_list(
        this.performance_evaluation
      )}\n\n` +
      `You should only respond in JSON format as described below ` +
      `\nResponse Format: \n${formatted_response_format} ` +
      `\nEnsure the response can be parsed by Python json.loads`;

    return prompt_string;
  }
}

You can then add memory to the PromptGenerator in your function:

export function getPrompt(tools: ObjectTool[]): string {
  const prompt_generator = new PromptGenerator();

  // ...

  prompt_generator.add_memory(
    "Previous task details and outcomes."
  );
  prompt_generator.add_memory(
    "Important information saved from previous tasks."
  );

  const prompt_string = prompt_generator.generate_prompt_string();

  return prompt_string;
}

You can find more details about this in the langchain/src/experimental/autogpt/prompt_generator.ts and langchain/src/experimental/generative_agents/generative_agent_memory.ts files in the langchainjs repository.

As for the OpenAI class, it is a JavaScript class that provides an interface to interact with the OpenAI API. It is designed to handle the generation of text using OpenAI's language models. You can find more details about this in the langchain/src/llms/openai.ts file in the langchainjs repository.

I hope this helps! Let me know if you have any other questions.

Sources

• How can we add Memory to the dynamictool
• langchain/src/experimental/autogpt/prompt_generator.ts
• langchain/src/experimental/generative_agents/generative_agent_memory.ts
• langchain/src/llms/openai.ts
• examples/src/models/chat/integration_azure_openai.ts

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.