[FEATURE] Caching

[RichardZhu2]

Use Case
Deterministic tasks that perform the same computation again and again would benefit from caching mechanisms.

Caching behaviour could be configured via an extra task parameter. For example:

from pyper import task, CachePolicy

def extremely_expensive_calculation(x: int):
    return x + 1

pipeline = task(
    extremely_expensive_calculation,
    multiprocess=True,
    cache_policy=CachePolicy(maxsize=100, ...)
)

Alternatives
Currently, functools.cache and functools.lru_cache can be used for normal synchronous functions. However, this approach won't straightforwardly extend to asynchronous tasks and multiprocessed tasks. There is a need for a unified caching API across all modes of execution.