Using repoze.lru

repoze.lru is a LRU (least recently used) cache implementation. Keys and values that are not used frequently will be evicted from the cache faster than keys and values that are used frequently. It works under Python 2.5, Python 2.6, Python 2.7, and Python 3.2.

Using the API programmatically

Creating an LRUCache object:

>>> from repoze.lru import LRUCache
>>> cache = LRUCache(100) # 100 max length

Retrieving from an LRUCache object:

>>> cache.get('nonexisting', 'foo') # return 'foo'
'foo'
>>> cache.get('nonexisting') is None
True

Adding to an LRUCache object:

>>> cache.put('existing', 'value') # add the key 'key' with the value 'value'
>>> cache.get('existing') # return the value for existing
'value'

Clearing an LRUCache:

>>> cache.clear()

Each LRU cache tracks some basic statistics via attributes:

cache.lookups # number of calls to the get method cache.hits # number of times a call to get found an object cache.misses # number of times a call to get did not find an object cahce.evictions # number of times a object was evicted from cache

Decorating an “expensive” function call

repoze.lru provides a class lru_cache, which wrapps another callable, caching the results. All values passed to the decorated function must be hashable. It does not support keyword arguments:

>>> from repoze.lru import lru_cache
>>> @lru_cache(500)
... def expensive_function(*arg): #*
...     pass

Each function decorated with the lru_cache decorator uses its own cache related to that function.

Cleaning cache of decorated function

repoze.lru provides a CacheMaker, which generates decorators. This way, you can later clear your cache if needed.

>>> from repoze.lru import CacheMaker
>>> cache_maker=CacheMaker()
>>> @cache_maker.lrucache(maxsize=300, name="adder")
... def yet_another_exepensive_function(*arg):#*
...     pass

>>> @cache_maker.expiring_lrucache(maxsize=300,timeout=30)
... def another_exepensive_function(*arg):#*
...     pass

This way, when you need it you can choose to either clear all cache:

>>> cache_maker.clear()

or clear a specific cache

>>> cache_maker.clear("adder")