Lightweight Decorator to cache (memoize) function calls and replay responses for expensive computations and API requests.

Pip installable via pip install functionfaker

Functionfaker offers lightweight and easy to understand function (and method) caching, memoization, similar to some of the functionality of the Joblib package. Function output is recorded once and then replayed from cache from thereon. This facilitates unit testing of applications that involve changing output that needs to be predictable in testing, such as API calls. Testing also often requires API calls to be simulated without actually calling the API. It can also speed up prototyping of computationally expensive code. Functionfaker consists of simple code, and provides a single decorator for your functions and methods, as shown in the following simple example.

from functionfaker import response_player
import os

Add the response_player decorator to an example function called add:

@response_player()
def add(x, y):
    return x + y

This will not affect the function add until you set an environment variable to RECORD:

# enter record mode, to record function responses.
os.environ['RECORD'] = "record"

Call the add function a few times:

# Clear the stored function responses
if os.path.exists('responses.p'):
    os.remove('responses.p')
# call the add function to record some responses.
add(1,2)
add(1,y=3)
add(1,3)
add(2,1)

Recording response function "add"
Recording response function "add"
Recording response function "add"

Set replay mode via the environment variable:

# enter replay mode, so that function will not run, but return stored values instead.
os.environ['RECORD'] = "replay"

Call the add function again, with arguments that it has already seen:

result = add(1,2)
print("The saved result of adding 1 and 2 using function 'add' is %d"%result)

Faking function "add". Response found
The saved result of adding 1 and 2 using function 'add' is 3

The outputs for these inputs (1,2) are now read from cache. Setting the environment variable RECORD to and empty string lets the function add run normally again without functionfaker acting on it (you can leave the decorator on).

Some function arguments might be irrelevant or difficult to serialize. To ignore these arguments, provide the args2ignore argument as a list of integers to the response_player decorator, where the integers represent the index in the argument list.

Default function response storage is in a simple Pickle file responses.p. To use your own storage system, provide a class derived from BaseStore class with an update and get_response method. An object of this class is then provided as the store argument to the response_player decorator.