Skip to main content

Memoiz is a memoization decorator that makes reasonable assumptions about how and if to cache the return value of a function or method based on the arguments passed to it. The decorator can be used on both free and bound functions.

Project description

Memoiz

A thread-safe memoization decorator for functions and methods.

Introduction

Memoiz is a memoization decorator that makes reasonable assumptions about how and if to cache the return value of a function or method based on the arguments passed to it. The decorator can be used on both free and bound functions.

Features

  • A thread-safe cache
  • Use the Memoiz decorator on free and bound functions
  • Support for parameter and return type hints
  • Cache invalidation

Table of Contents

Installation

pip install memoiz

Usage

Bound Functions (methods)

In this example you will use Memoiz to memoize the return value of the greeter.greet method and print the greeting.

from memoiz.cache import Cache

cache = Cache()


class Greeter:

    def __init__(self):
        self.adv = "Very"

    @cache
    def greet(self, adj: str) -> str:
        return f"Hello, {self.adv} {adj} World!"


greeter = Greeter()

print("1:", cache._cache)

greeting = greeter.greet("Happy")

print("2:", greeting)
1: {}
2: Hello, Very Happy World!

As a continuation of the example, you will selectively invalidate cached articles using the cache.invalidate method.

greeter = Greeter()

print("1:", cache._cache)

greeting = greeter.greet("Happy")

print("2:", greeting)

greeting = greeter.greet("Cautious")

print("3:", greeting)

# The cache has memoized the call using the `Happy` argument and the call using the `Cautious` argument.
print("4:", cache._cache)

# Invalidate the call to `greeter.greet` with the `Happy` argument. Provide a reference to the method, the object, and the argument.
cache.invalidate(greeter.greet, greeter, "Happy")

print("5:", cache._cache)

# Invalidate the call to `greeter.greet` with the `Cautious` argument.
cache.invalidate(greeter.greet, greeter, "Cautious")

# The cache is empty.
print("6:", cache._cache)
1: {}
2: Hello, Very Happy World!
3: Hello, Very Cautious World!
4: {<bound method Greeter.greet of <__main__.Greeter object at 0x7fa5e7f837f0>>: {((<__main__.Greeter object at 0x7fa5e7f837f0>, 'Happy'), ()): 'Hello, Very Happy World!', ((<__main__.Greeter object at 0x7fa5e7f837f0>, 'Cautious'), ()): 'Hello, Very Cautious World!'}}
5: {<bound method Greeter.greet of <__main__.Greeter object at 0x7fa5e7f837f0>>: {((<__main__.Greeter object at 0x7fa5e7f837f0>, 'Cautious'), ()): 'Hello, Very Cautious World!'}}
6: {}

Free Functions

In this example you will use Memoiz to memoize the return value of the greet function and print the greeting.

from memoiz.cache import Cache

cache = Cache()


@cache
def greet(adj: str) -> str:
    return f"Hello, {adj} World!"


print("1:", cache._cache)

greeting = greet("Happy")

print("2:", greeting)
1: {}
2: Hello, Happy World!

As a continuation of the example, you will selectively invalidate cached articles using the cache.invalidate method.

print("1:", cache._cache)

greeting = greet("Happy")

print("2:", greeting)

greeting = greet("Cautious")

print("3:", greeting)

print("4:", cache._cache)

cache.invalidate(greet, "Happy")

print("5:", cache._cache)

cache.invalidate(greet, "Cautious")

print("6:", cache._cache)
1: {}
2: Hello, Happy World!
3: Hello, Cautious World!
4: {<function greet at 0x7fa5cefb8430>: {(('Happy',), ()): 'Hello, Happy World!', (('Cautious',), ()): 'Hello, Cautious World!'}}
5: {<function greet at 0x7fa5cefb8430>: {(('Cautious',), ()): 'Hello, Cautious World!'}}
6: {}

Limitations

Memoization relies on the behavior of pure functions; given the same input the function will produce the same output. It isn't safe to assume that a callable is pure in Python; hence, you must use discretion.

Memoiz uses a Python dictionary in order to cache callable inputs and outputs. Memoiz will attempt to transform a callable's arguments into a hashable representation. If it succeeds, the hashable representation of the callable's arguments is used as the dictionary key in order to store and look up the cached ouput. If it fails, Memoiz will call the decorated function or method and return the result.

Memoiz employs a few strategies to produce a hashable lookup key. Memoiz will iteratively recurse through dict, list, set, and tuple type arguments, transforming these mutable objects into hashable representations. See the Type Transformation table for type tranformations. When a primitive is encountered (e.g., int, float, complex, bool, str, None), it is left as is. If allow_hash is set to True (the default), Memoiz will additionally attempt to discern if an object is hashable using Python's hash function.

Effectively what this all means is that if you are using common Python iterables and primitives as arguments to your callable, and if your callable doesn't have side effects, Memoiz should be able to accurately cache your input arguments and output for subsequent function calls.

Type Transformation

Type Hashable Type
dict tuple of tuples
list tuple
tuple tuple
set tuple

API

The Cache Class

memoiz.Cache(immutables, allow_hash, deep_copy)

  • immutables Tuple[type, ...] An optional list of objects that are assumed to be immutable. Default: (int, float, complex, bool, str, type(None))
  • allow_hash bool An optional flag that indicates if an object's hash is sufficient for indexing the callable's arguments. Default: True
  • deep_copy bool Optionally return the cached return value using Python's copy.deepcopy. This can help prevent mutations of the cached return value. Default: True.

cache.__call__(callable)

  • callable typing.Callable The function or method that you want to feature memoization.

A Cache instance (see above) is a callable. This is the @cache decorator that is used in order to add memoization to a callable. Please see the above usage for how to use this decorator.

cache.invalidate(callable, *args, **kwargs)

  • callable typing.Callable The callable.
  • args Any The arguments passed to the callable.
  • kwargs Any The keyword arguments passed to the callable.

Invalidates the cache for the specified callable and arguments. See the usage for for how to invalidate the cache.

NB The first argument of a method (i.e., a bound function) is the object instance e.g., the self in the method definition.

cache.invalidate_all()

Resets the cache making the old cache eligible for garbage collection.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

memoiz-1.1.0.tar.gz (6.8 kB view details)

Uploaded Source

Built Distribution

memoiz-1.1.0-py3-none-any.whl (5.9 kB view details)

Uploaded Python 3

File details

Details for the file memoiz-1.1.0.tar.gz.

File metadata

  • Download URL: memoiz-1.1.0.tar.gz
  • Upload date:
  • Size: 6.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: python-httpx/0.27.2

File hashes

Hashes for memoiz-1.1.0.tar.gz
Algorithm Hash digest
SHA256 19626848bf513a2acb402dd73b0deea5cb64d5aa904ef857bf1ddea047e360f1
MD5 e1827868771ed66ee8bf44188908a53b
BLAKE2b-256 1a91acb19569955b60569d17bc558f400763f3a7a5ef965a908be28e85c8c3f1

See more details on using hashes here.

File details

Details for the file memoiz-1.1.0-py3-none-any.whl.

File metadata

  • Download URL: memoiz-1.1.0-py3-none-any.whl
  • Upload date:
  • Size: 5.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: python-httpx/0.27.2

File hashes

Hashes for memoiz-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 e75cff319b5179fc2e6be8f6001fc889501a3da2cd1290bf11e51a2da271ddf2
MD5 75cd77bd434f2e75ff6ff45226e8a7dc
BLAKE2b-256 f844fac568437d7001cd0bfcb6b1bfbf5dbc9a1ba42cb1cf00f2f487634b6459

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page