• Graceful handling of errors (can fall back to cached value)
• Calculate results in background (requires Celery)
• Readable duration strings ('1 day' vs 86400)
• Correct handling of None
• Per-call invalidation

1. pip install django-throttleandcache
2. Set a cache backend in your settings.py file.

from throttleandcache import cache

# Cache the result of my_function for 3 seconds.
@cache('3s')
def my_function():
    return 'whatever'

If you call the function multiple times with the same arguments, the result will be fetched from the cache. In order to invalidate the cache for that call, call my_function.invalidate() with the same arguments:

my_function()
my_function() # Result pulled from cache
my_function.invalidate()
my_function() # Not from cache

If Celery is installed, you can remove the the calculation of new values from the request/response cycle:

@cache('3s', background=True)
def my_function():
    return 'whatever'

Note that, in the case of a cold cache, the value will still be calculated synchronously. Stale values may be used while new ones are being calculated.

Remember that calling the same method on multiple instances means that each invocation will have a different first positional (self) argument:

class A(object):
    @cache('100s')
    def my_function(self):
        print 'The method is being executed!'

instance_1 = A()
instance_2 = A()
instance_1.my_function() # The original method will be invoked
instance_2.my_function() # Different "self" argument, so the method is invoked again.

If you wish to cache the result across all instances, use @cacheforclass.

The first argument to the cache decorator is the timeout and can be given as a number (of seconds) or a string. Since strings contain units, they can make your code much more readable. Some examples are '2s', '3m', '3m 2s', and '3 minutes, 2 seconds'.

The cache decorator also accepts the following (optional) keyword arguments:

• using: specifies which cache to use.
• key_prefix: A string to prefix your cache key with.

• key_func: A function for deriving the cache key. This function will be

  passed the fn, *args, and **kwargs.

• graceful: This argument specifies how errors should be handled. If

  graceful is True and your function raises an error, throttleandcache will log the error and return the cached value. If no cached value exists, the original error is raised.

• background: Specifies that new values should be calculated in the

  background (using Celery).