How to use memory with VectorDBQAChain

[jteso]

How can i use memory when using VectorDBQAChain?

The QA part works perfect, but I need now to keep track of the conversation.

Thanks

The relevant part of my code below:

// Omitted the part of parsing documents 
const vectorStore = await HNSWLib.fromDocuments(docs, new OpenAIEmbeddings());
const qaChain = VectorDBQAChain.fromLLM(model, vectorStore, { returnSourceDocuments: true });
let question="My question here"; 
const answer = await qaChain.call({ query: question });

[jteso]

I will share my solution in case others are facing a similar problem.
I found that the most effective way to allow for follow-up questions is by implementing something along these lines:

 const llm = new OpenAI({ temperature: 0 });
 const chain = ConversationalRetrievalQAChain.fromLLM(
    llm, 
    vectorStore.asRetriever(), 
  );
  
const question = "Your question here";
let res = await chain.call({ question, chat_history });
chat_history.push(`Question: ${question}. Answer: ${res.text}`);
await chain.call({question: "your followup question", chat_history});

[joaopcm]

Could you share more details with us? I'm facing the same challenge, but I don't think it's enough 👀