Building a Multimodal RAG Example with Base64 Images for GPT-4 Vision

i'm building a multimodal rag example so i'm getting base64 images as part of the example to feed to gpt-4 vision