Using Llama Index for this via the `llama_index.core.base.llms.base.BaseLLM` interface. Using config files to describe the args to different models makes swappi

Using Llama Index for this via the `llama_index.core.base.llms.base.BaseLLM` interface. Using config files to describe the args to different models makes swapping models literally as easy as: chat_model: cls: llama_index.llms.openai.OpenAI kwargs: model: gpt-4 chat_model: cls: llama_index.llms.gemini.Gemini kwargs: model_name: models/gemini-pro