faster-async-lru 2.0.5

A faster fork of async-lru: Simple LRU cache for asyncio. Written in Python and compiled to C using mypyc.

info:

Simple lru cache for asyncio

Installation

pip install async-lru

Usage

This package is a port of Python’s built-in functools.lru_cache function for asyncio. To better handle async behaviour, it also ensures multiple concurrent calls will only result in 1 call to the wrapped function, with all awaits receiving the result of that call when it completes.

import asyncio

import aiohttp
from faster_async_lru import alru_cache


@alru_cache(maxsize=32)
async def get_pep(num):
    resource = 'http://www.python.org/dev/peps/pep-%04d/' % num
    async with aiohttp.ClientSession() as session:
        try:
            async with session.get(resource) as s:
                return await s.read()
        except aiohttp.ClientError:
            return 'Not Found'


async def main():
    for n in 8, 290, 308, 320, 8, 218, 320, 279, 289, 320, 9991:
        pep = await get_pep(n)
        print(n, len(pep))

    print(get_pep.cache_info())
    # CacheInfo(hits=3, misses=8, maxsize=32, currsize=8)

    # closing is optional, but highly recommended
    await get_pep.cache_close()


asyncio.run(main())

TTL (time-to-live in seconds, expiration on timeout) is supported by accepting ttl configuration parameter (off by default):

@alru_cache(ttl=5)
async def func(arg):
    return arg * 2

The library supports explicit invalidation for specific function call by cache_invalidate():

@alru_cache(ttl=5)
async def func(arg1, arg2):
    return arg1 + arg2

func.cache_invalidate(1, arg2=2)

The method returns True if corresponding arguments set was cached already, False otherwise.

Benchmarks

async-lru uses CodSpeed for performance regression testing.

To run the benchmarks locally:

pip install -r requirements-dev.txt
pytest --codspeed benchmark.py

The benchmark suite covers both bounded (with maxsize) and unbounded (no maxsize) cache configurations. Scenarios include:

• Cache hit

• Cache miss

• Cache fill/eviction (cycling through more keys than maxsize)

• Cache clear

• TTL expiry

• Cache invalidation

• Cache info retrieval

• Concurrent cache hits

• Baseline (uncached async function)

On CI, benchmarks are run automatically via GitHub Actions on Python 3.13, and results are uploaded to CodSpeed (if a CODSPEED_TOKEN is configured). You can view performance history and detect regressions on the CodSpeed dashboard.

Thanks

The library was donated by Ocean S.A.

Thanks to the company for contribution.