Is it possible to use multi-modal messages with ContextChatEngines? I found the MultiModal Refactor epic and the task for consolidating ChatMessages is completed, but I'm still running into Pydantic typing issues when I try to send content that is a list (GPT4V style).
