Is there a decorator to simply cache function return values?
Starting from Python 3.2 there is a built-in decorator:
@functools.lru_cache(maxsize=100, typed=False)
Decorator to wrap a function with a memoizing callable that saves up to the maxsize most recent calls. It can save time when an expensive or I/O bound function is periodically called with the same arguments.
Example of an LRU cache for computing Fibonacci numbers:
@lru_cache(maxsize=None)def fib(n): if n < 2: return n return fib(n-1) + fib(n-2)>>> print([fib(n) for n in range(16)])[0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610]>>> print(fib.cache_info())CacheInfo(hits=28, misses=16, maxsize=None, currsize=16)
If you are stuck with Python 2.x, here's a list of other compatible memoization libraries:
functools32
| PyPI | Source coderepoze.lru
| PyPI | Source codepylru
| PyPI | Source codebackports.functools_lru_cache
| PyPI | Source code
It sounds like you're not asking for a general-purpose memoization decorator (i.e., you're not interested in the general case where you want to cache return values for different argument values). That is, you'd like to have this:
x = obj.name # expensivey = obj.name # cheap
while a general-purpose memoization decorator would give you this:
x = obj.name() # expensivey = obj.name() # cheap
I submit that the method-call syntax is better style, because it suggests the possibility of expensive computation while the property syntax suggests a quick lookup.
[Update: The class-based memoization decorator I had linked to and quoted here previously doesn't work for methods. I've replaced it with a decorator function.] If you're willing to use a general-purpose memoization decorator, here's a simple one:
def memoize(function): memo = {} def wrapper(*args): if args in memo: return memo[args] else: rv = function(*args) memo[args] = rv return rv return wrapper
Example usage:
@memoizedef fibonacci(n): if n < 2: return n return fibonacci(n - 1) + fibonacci(n - 2)
Another memoization decorator with a limit on the cache size can be found here.
Python 3.8 functools.cached_property
decorator
https://docs.python.org/dev/library/functools.html#functools.cached_property
cached_property
from Werkzeug was mentioned at: https://stackoverflow.com/a/5295190/895245 but a supposedly derived version will be merged into 3.8, which is awesome.
This decorator can be seen as caching @property
, or as a cleaner @functools.lru_cache
for when you don't have any arguments.
The docs say:
@functools.cached_property(func)
Transform a method of a class into a property whose value is computed once and then cached as a normal attribute for the life of the instance. Similar to property(), with the addition of caching. Useful for expensive computed properties of instances that are otherwise effectively immutable.
Example:
class DataSet: def __init__(self, sequence_of_numbers): self._data = sequence_of_numbers @cached_property def stdev(self): return statistics.stdev(self._data) @cached_property def variance(self): return statistics.variance(self._data)
New in version 3.8.
Note This decorator requires that the dict attribute on each instance be a mutable mapping. This means it will not work with some types, such as metaclasses (since the dict attributes on type instances are read-only proxies for the class namespace), and those that specify slots without including dict as one of the defined slots (as such classes don’t provide a dict attribute at all).