Hello, I want to create a chatbot that answers users' questions using a RAG system. Since the users' questions are related to mathematics and involve formulas, they can send their questions as images. I plan to use the Llama Parse service to convert the images into text and then find the answers from a knowledge base. However, every time I test it, the Llama Parse service doesn't convert the images to text well. When I use the UI on the cloud.llama website, it extracts the text correctly, but when I use the Python SDK, I don't know how to configure it to get the best output. I also don't know what the full list of settings for the Llama Parse function is so I can test and see which one works best.