How to use monkeypatch in a "setup" method for unit tests using pytest?

Question

I am attempting to mock out a utility class (In this case the python logger utility) in a unit test.

While I know how to do it using monkeypatch on a per test level, I was hoping I could simply do it as part of the setup/globally in some way.

Here is what I am hoping I can do (but I am getting errors):

import logging
...
def setup(self, monkeypatch):
  class fake_logger(level):
    def __init__(self, val):
      pass
    def setLevel(self, level):
      # Do something
  def mock_logger(level):
    return fake_logger(level)
  monkeypatch.setattr(logging, 'getLogger', mock_logger)

What is the right way to do this?

EDIT: Example error

name = 'setup'
def call_optional(obj, name):
 method = getattr(obj, name, None)
 isfixture = hasattr(method, "_pytestfixturefunction")
 if method is not None and not isfixture and py.builtin.callable(method):
 # If there's any problems allow the exception to raise rather than
 # silently ignoring them
> method()
E TypeError: setup() missing 1 required positional argument: 'monkeypatch'

• Tadhg McDonald-Jensen: Please include the errors you are getting, it really helps us understand what is going wrong.

Answer 1

monkeypatch works as a normal pytest fixture. If you want to use it, you need to make your method as a fixture as well.

import logging
import pytest
@pytest.fixture
def setup(monkeypatch):
 class fake_logger(object):
 def __init__(self, val):
 pass
 def setLevel(self, level):
 # Do something
 pass
 def mock_logger(level):
 return fake_logger(level)
 monkeypatch.setattr(logging, 'getLogger', mock_logger)
def test_fake_logger(setup):
 # test steps

and if you check the type of logging.getLogger('any level') in test, it will be fake_logger you defined.

• The Compiler: You can also do @pytest.fixture(autouse=True) so it's automatically applied for all tests without having to have a setup argument.