python-lru-cache

Saw this:

@lru_cache(maxsize=1)
@pyrolog.log_duration(slogger, "preparing go_terms")
def get_go_term_data() -> pd.DataFrame:
    # ...

learn about lru_cache
how does maxsize=1 help?
when would this help in general?
when would this NOT be a good idea in general?
I think functools.cache may be recent syntax sugar for functools.lru_cache(maxsize=None)…when to choose which?
When to use this instead of diskcache (which requires extra calls)? When is this good enough?
- Is this good when just caching in memory for current session?
- While diskcache is for caching indefinitely?