Make cache persistent #13
Description
Superopenai currently caches the results of LLM calls only in memory. Is there a way to save the cache on disk for subsequent runs of the program? I don't usually have multiple identical LLM calls within one run of the program, but duplication across multiple program runs is frequent.