Python：数组中使用的模拟补丁模块

Question

Here is my method:

def _eda_workflow(self, workflow_id, account_id):
    uuids = [str(uuid.uuid4()) for x in range(2)]

    return {
        'id': workflow_id,
        'root': uuids[0],
        'accountId': account_id,
        'steps': [
            {
                'this': uuids[0],
                'next': [uuids[1]],
                'type': 'on-contact-signed-up',
                'displayName': 'Sign Up',
                'constraints': {},
            },
            {
                'this': uuids[1],
                'prev': uuids[0],
                'next': [],
                'type': 'on-time-elapsed',
                'displayName': 'Send the email after delay or not!',
                'delay': 'PT0M',
                'sideEffect': {
                    'task': 'sendEmail',
                    'constraints': {},
                    'mailing_data': {},
                    'email': {}
                }
            }
        ]
    }

The problem I am having is when I write this

def generate_uuids():
    return ["10e6e848-dc77-4057-890e-5acd4ed9aeb3", "d8d501a7-f7e7-4423-921c-e0f39b7e1301"]

@mock.patch('path.to.uuid')
def test_translte_workflow(self, uuid_mock):
   uuid_mock.uuid4.side_effect = generate_uuids

The return method is returning a list of values instead of what I am expecting. Not sure how to write this test correctly

Answer 1

Side effect is just for specifying something additional to happen whenever the mock is called. I think this is more what you were looking for:

def fake_uuid4():
    yield "10e6e848-dc77-4057-890e-5acd4ed9aeb3"
    yield "d8d501a7-f7e7-4423-921c-e0f39b7e1301"

@mock.patch('path.to.uuid.uuid4', fake_uuid4().next)
def test_translte_workflow(self):
   ...

Answer 2

Your generate_uuids function is returning a list of UUIDs, so that's what you get.

If you want to create a function that returns one UUID at a time, from a pool of exactly two UUIDs, create an iterator from a list of two UUIDs and return iterator.next() from that function. If you also want to make assertions on those UUIDs, store them separately from the iterator:

import mock
import unittest
import uuid


TEST_UUIDS = ["10e6e848-dc77-4057-890e-5acd4ed9aeb3",
              "d8d501a7-f7e7-4423-921c-e0f39b7e1301"]


uuid_pool = iter(TEST_UUIDS)


def generate_uuid():
    return uuid_pool.next()


def func():
    uuid_1 = uuid.uuid4()
    uuid_2 = uuid.uuid4()
    return [uuid_1, uuid_2]


class TestUUIDs(unittest.TestCase):

    @mock.patch('uuid.uuid4', generate_uuid)
    def test_uuid_mock(self):
        actual = func()
        expected = TEST_UUIDS
        self.assertEquals(expected, actual)


unittest.main()

Once you exhaust that iterator (after fetching two elements), it will raise StopIteration .

If you want to create a function that always returns just one an UUID from an endless stream of UUIDs, you can use itertools.cycle :

from itertools import cycle

uuid_pool = cycle(["10e6e848-dc77-4057-890e-5acd4ed9aeb3",
                   "d8d501a7-f7e7-4423-921c-e0f39b7e1301"])

Answer 3

I needed to solve something similar and, after some experimenting, preferred using an iterable with @patch and side_effect . In my case, instead of populating an array, I was populating the return value of a function called a layer deeper than the function in my test.

TEST_UUIDS = ["10e6e848-dc77-4057-890e-5acd4ed9aeb3",
              "d8d501a7-f7e7-4423-921c-e0f39b7e1301"]
TEST_UUID_POOL = iter(TEST_UUIDS)

@patch('path.to.function1', side_effect=TEST_UUID_POOL)
def test_translate_workflow(self, uuid_mock):
    # function 1 is called twice in function2
    function2()