Imitating expensive objects

Sometimes, we want to test code that requires an object be supplied that is either expensive or difficult to construct. In some cases, this may mean your API needs rethinking to have a more testable interface (which typically means a more usable interface). But we sometimes find ourselves writing test code that has a ton of boilerplate to set up objects that are only incidentally related to the code under test.

For example, imagine we have some code that keeps track of flight statuses in an external key-value store (such as redis or memcache), such that we can store the timestamp and the most recent status. A basic version of such code might look as follows:

import datetime
import redis


class FlightStatusTracker:
    ALLOWED_STATUSES = {"CANCELLED", "DELAYED", "ON TIME"}

    def __init__(self):
        self.redis = redis.StrictRedis()

    def change_status(self, flight, status):
        status = status.upper()
        if status not in self.ALLOWED_STATUSES:...