Pass prompts in OpenAI's prompt message format while using ConversationalRetrievalQAChain

[whsahhar]

Is there a way to pass prompts and chat history in OpenAI's messages format:

"messages": 
  [
    {"role": "user", "content": "Hello!"},
    {"role": "assistant", "content": "Hi there!"}
  ]

.. while using ConversationalRetrievalQAChain.fromLLM()

The chain is very helpful for my use case, but it only takes a string for chat history which makes managing the chat state challenging (e.g. referring to assistant's previous reply)

Thank you!