nicecache

A cache that persists data between runs. Trust me, you need this.

@functools.cache is nice, but there's a slight problem -- each program starts its own cache fresh. In comes @nicecache.nicecache. Use it just like you'd use @functools.cache, but rest assured that you really only have to compute this once and they'll persist across runs (by saving to disk!).

Uses diskcache to manage things behind the scenes.

@nicecache decorator

Got some slow data-loading or preprocessing that you'd rather just compute once? Frustrated that functools.cache spawns a new cache each time you re-run your program? You're in the right place.

Here's a really simple example:

from time import time
from nicecache import nicecache
@nicecache(tmp_cache_dir)
def slowtimes(a, b):
    print('saving to cache:', (a, b))
    time.sleep(5)
    return a * b

def test():
    t1 = time()
    print(slowtimes(5, 3))
    t2 = time()
    print(f'Func 1 time: {t2-t1}s')
    print(slowtimes(5, 3))
    t3 = time()
    print(f'Func 2 time: {t3-t2}s')
    print(slowtimes(4, 2))
    t4 = time()
    print(f'Func 3 time: {t4-t3}s')
    print(slowtimes(4, 2))
    t5 = time()
    print(f'Func 4 time: {t5-t4}s')
saving to cache: (5, 3)
15
Func 1 time: 5.0264012813568115
15
Func 2 time: 0.0017404556274414062
saving to cache: (4, 2)
8
Func 3 time: 5.011389970779419
8
Func 4 time: 0.0015897750854492188

Specifying a cache path

By default values are cached to ~/.nicecache. Specify a different folder path by passing an argument to the decorator. For example, @nicecache('/tmp/cache/path').

NiceCache