File Memoizer

This Python package makes it easy to store function results across executions using cache files. Underlying functionality is provided by cachier, but this package adds a few conveniences, such as being able to ignore parameters that won't serialize.

Installation

Installation is via pip:

pip install file-memoizer

Usage

To cache a function's value, annotate it by calling the memoize function as follows:

import file_memoizer

@file_memoizer.memoize()
def double(n):
    return 2 * n

By default the cached values remain valid for a day. This can be changed with the cache_ttl parameter:

import datetime
import file_memoizer

seven_days = cache_ttl=datetime.timedelta(days=7)
@file_memoizer.memoize(cache_ttl=seven_days)
def triple(n):
    return 3 * n

Cache files are stored in $HOME/.file-memoizer, with one file per combination of input parameters. An alternate location can be specified with the cache_directory parameter:

import datetime
import file_memoizer

custom_path = '/path/to/store/files'
@file_memoizer.memoize(cache_directory=custom_path)
def quadruple(n):
    return 4 * n

Normally all function arguments must be hashable for it to be safely cached. However, there are situations where it's okay to ignore them, such as an object method whose return value doesn't depend on the object's internal state. In these cases, set unhashable_args='ignore' as shown below:

class Arithmetic():

@staticmethod
@file_memoizer.memoize()
def quintuple(n):
     return 5 * n

@file_memoizer.memoize(unhashable_args='ignore')
def multiply(self, x, y):
    return x * y