flake8-mock-spec 1.1.0

A linter that checks mocks are constructed with the spec argument

flake8-mock-spec

Do you use mocks and are concerned you are calling methods or accessing attributes the mocked objects don't have? If not, you should be as that is a sure way to inject bugs into your code and still have your tests pass. The flake8-mock-spec linter enforces the use of the spec argument on mocks ensuring that your use of mocks is compliant with the interface of the object being mocked.

Getting Started

python -m venv venv
source ./venv/bin/activate
pip install flake8-mock-spec
flake8 test_source.py

On the following code:

# test_source.py
from unittest import mock

def test_foo():
    mocked_foo = mock.Mock()

This will produce warnings such as:

flake8 test_source.py
test_source.py:5:22: TMS001 unittest.mock.Mock instances should be constructed with the spec or spec_set argument, more information: https://github.com/jdkandersson/flake8-mock-spec#fix-tms001

This can be resolved by changing the code to:

# test_source.py
from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.Mock(spec=Foo)

Rules

A few rules have been defined to allow for selective suppression:

• TMS001: checks that unittest.mock.Mock instances are constructed with the spec or spec_set argument.
• TMS002: checks that unittest.mock.MagicMock instances are constructed with the spec or spec_set argument.
• TMS003: checks that unittest.mock.NonCallableMock instances are constructed with the spec or spec_set argument.
• TMS004: checks that unittest.mock.AsyncMock instances are constructed with the spec or spec_set argument.

Fix TMS001

This linting rule is triggered by creating a unittest.mock.Mock instance without the spec or spec_set argument. For example:

from unittest import mock

def test_foo():
    mocked_foo = mock.Mock()

This example can be fixed by using the spec or spec_set argument in the constructor:

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.Mock(spec=Foo)

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.Mock(spec_set=Foo)

Fix TMS002

This linting rule is triggered by creating a unittest.mock.MagicMock instance without the spec or spec_set argument. For example:

from unittest import mock

def test_foo():
    mocked_foo = mock.MagicMock()

This example can be fixed by using the spec or spec_set argument in the constructor:

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.MagicMock(spec=Foo)

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.MagicMock(spec_set=Foo)

Fix TMS003

This linting rule is triggered by creating a unittest.mock.NonCallableMock instance without the spec or spec_set argument. For example:

from unittest import mock

def test_foo():
    mocked_foo = mock.NonCallableMock()

This example can be fixed by using the spec or spec_set argument in the constructor:

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.NonCallableMock(spec=Foo)

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.NonCallableMock(spec_set=Foo)

Fix TMS004

This linting rule is triggered by creating a unittest.mock.AsyncMock instance without the spec or spec_set argument. For example:

from unittest import mock

def test_foo():
    mocked_foo = mock.AsyncMock()

This example can be fixed by using the spec or spec_set argument in the constructor:

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.AsyncMock(spec=Foo)

from unittest import mock

from foo import Foo

def test_foo():
    mocked_foo = mock.AsyncMock(spec_set=Foo)