Skip to content

An easy-to-use Python library for lightning-fast persistent function caching.

License

Notifications You must be signed in to change notification settings

umarbutler/persist-cache

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

67 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

persist-cache

persist-cache is an easy-to-use Python library for lightning-fast persistent function caching. It is capable of caching both synchronous and asynchronous functions as well as methods, and is also process-safe and thread-safe.

Features 🎯

  • ⚑ Lightning-fast: a function call can be cached in as little as 145 microseconds and be returned back in as few as 95 microseconds.
  • πŸ’½ Persistent: cached returns persist across sessions and are stored locally.
  • βŒ› Stale-free: cached returns may be given a shelf life, after which they will be automatically flushed out.
  • 🦺 Process- and thread-safe: interprocess file locks prevent processes and threads from writing over each other.
  • ⏱️ Async-compatible: asynchronous functions can be cached with the same decorator as synchronous ones, generators included.
  • πŸ‘¨β€πŸ« Class-compatible: methods can be cached with the same decorator as functions (although the self argument is always ignored).

Installation πŸ§‘β€πŸ”§

persist-cache may be installed with pip:

pip install persist-cache

Usage πŸ‘©β€πŸ’»

The code snippet below demonstrates how both synchronous and asynchronous functions as well as methods may be cached with persist-cache:

from persist_cache import cache

@cache
def my_function(): ...

@cache
async def my_function(): ...

class MyClass:
    @cache
    def my_method(self): ...

    @cache
    async def my_method(self): ...

It is also possible to name caches and specify their shelf life:

from datetime import timedelta

@cache(name='my_shared_cache', expiry=timedelta(months=1))
def my_function(): ...

@cache(name='my_shared_cache', expiry=60 * 60 * 24 * 30)
def my_other_function(): ...

Once created, cached functions may be managed as follows:

# Change cached returns to expire after an hour.
my_function.set_expiry(60 * 60)

# Flush out any expired cached returns.
my_function.flush_cache() or persist_cache.flush(my_function, 60 * 60) or persist_cache.flush('my_shared_cache', 60 * 60)

# Clear out all cached returns.
my_function.clear_cache() or persist_cache.clear(my_function) or persist_cache.clear('my_shared_cache')

# Delete the cache.
my_function.delete_cache() or persist_cache.delete(my_function) or persist_cache.delete('my_shared_cache')

API 🧩

cache()

def cache(
    name: str | Callable | None = None,
    dir: str | None = None,
    expiry: int | float | timedelta | None = None,
) -> None

cache() persistently and locally cache the returns of a function.

The function to be cached must accept and return dillable objects only (with the exception of methods' self argument, which is always ignored). Additionally, for consistent caching across subsequent sessions, arguments and returns should also be hashable.

name represents the name of the cache (or, if cache() is being called as an argument-less decorator (ie, as @cache instead of @cache(...)), the function to be cached). It defaults to the qualified name of the function. If dir is set, name will be ignored.

dir represents the directory in which the cache should be stored. It defaults to a subdirectory named after the hash of the cache's name in a parent folder named '.persist_cache' in the current working directory.

expiry represents how long, in seconds or as a timedelta, function calls should persist in the cache. It defaults to None.

If cache() is called with arguments, a decorator that wraps the function to be cached will be returned, otherwise, the wrapped function itself will be returned.

After being wrapped, the cached function will have the following methods attached to it:

  • set_expiry(value: int | float | timedelta) -> None: Sets the expiry of the cache.
  • flush_cache() -> None: Flushes out any expired cached returns.
  • clear_cache() -> None: Clears out all cached returns.
  • delete_cache() -> None: Deletes the cache.

flush()

def flush(
    function_or_name: str | Callable,
    expiry: int | float | timedelta,
) -> None

flush() flushes out any expired cached returns from a cache.

function_or_name represents the function or the name of the cache to be flushed.

clear()

def clear(function_or_name: str | Callable) -> None

clear() clears out all cached returns from a cache.

function_or_name represents the function or the name of the cache to be cleared.

delete()

def delete(function_or_name: str | Callable) -> None

delete() deletes a cache.

function_or_name represents the function or the name of the cache to be deleted.

Licence πŸ“œ

This library is licensed under the MIT Licence.