mocker fixture usage - assistance please

[kevindawson]

Trying to come to terms with mocking AND pytest
the following is ok,

test code from 2020 youtube -> @nikhilkumarsingh tut9

from unittest import mock
import pytest
from tut9.myapp.sample import guess_number


@pytest.mark.parametrize("_input,expected", [(3, "You won!"), (4, "You lost!")])
@mock.patch("tut9.myapp.sample.roll_dice")
def test_guess_number(mock_roll_dice, _input, expected):
    mock_roll_dice.return_value = 3
    assert guess_number(_input) == expected
    mock_roll_dice.assert_called_once()

but when I try to code it to use mocker

import pytest
from tut9.myapp.sample import guess_number


@pytest.mark.parametrize("in_put,out_put", [(3, "You won!"), (4, "You lost!")])
def test_guess_num(mocker, in_put, out_put):
    mocker.patch("tut9.myapp.sample.roll_dice", return_value=3)
    assert out_put == guess_number(in_put)
    #tut9.myapp.sample.roll_dice.assert_called_once()

the above partial conversion has taken me hours
Why
the lack of/no examples with plan English text

i have been using the IDE intellisense

I have no idea as to how to implement .assert_called_once()

all I can see are redirects to unittest.mock API which is of NO help at all as there examples do not translate

I might be old and senile now

what about the teen pythonites

[kevindawson]

https://docs.pytest.org/en/stable/search.html?q=mocker

<div id="search-results">
<h2>Search Results</h2>
<p class="search-summary">Your search did not match any documents. Please make sure that all words are spelled correctly and that you've selected enough categories.</p>
<ul class="search"></ul>
</div>

[nicoddemus]

Hi @kevindawson,

mocker is a fixture from the pytest-mock plugin, and not a part of pytest, which explains why you didn't find any results in the pytest's documentation.

The mocker fixture is a thing wrapper around the mocker functionality, this is your original code using mocker:

import pytest
from tut9.myapp.sample import guess_number


@pytest.mark.parametrize("_input,expected", [(3, "You won!"), (4, "You lost!")])
def test_guess_number(mocker, _input, expected):
    mock_roll_dice = mock.patch("tut9.myapp.sample.roll_dice")
    mock_roll_dice.return_value = 3
    assert guess_number(_input) == expected
    mock_roll_dice.assert_called_once()

Particularly I find using the decorator facilities of unittest.mock a bit confusing when used with pytest, that's why pytest-mock exists, see the rationale here.

[kevindawson]

@nicoddemus
did you mean

mock_roll_dice = mock.patch("tut9.myapp.sample.roll_dice")
mock_roll_dice.return_value=3

OR

mock_roll_dice = mocker.patch("tut9.myapp.sample.roll_dice", return_value=3)

Also in your Why bother with a plugin? you seam to be saying what not to do, rather than you could do this with an example

[nicoddemus]

did you mean
OR
...

My bad. Both will work though (updated the original post).

Also in your Why bother with a plugin? you seam to be saying what not to do rather that you could do this with an example

There I try to exemplify how one would use unittest.mock directly and why I believe it is not a great fit with pytest, hence the plugin.

[kevindawson]

@nicoddemus
1, are you sure about that update?
2, I know python unittest.mock is not a great FULL STOP
3, I could not find any examples exemplifying good practice - hence over to you ;)

4, I found some documentation: Data Pipelines with Apache Airflow 8.1.4   pytest project structure. Cool
5, discussions will be more popular in time than #pylib is currently

[nicoddemus]

1, are you sure about that update?

Why do you ask? If you see something strange, please point it out, no need to make me review it entirely.

3, I could not find any examples exemplifying good practice - hence over to you ;)

Hence what over to me?

[kevindawson]

1, I think you should have

Assign mock to a variable, enables capture behaviour

mock_roll_dice = mock.patch("tut9.myapp.sample.roll_dice")

2, @nicoddemus you are the author are you not. you have written a cool plugin because you are a cool dude, but you have forgotten to turn the light on for all the new users, especially the teen pythonites

3, I will ask @nikhilkumarsingh about using his code so you can have some examples for us mere mortals when we first come across something NEW

[nicoddemus]

Assign mock to a variable, enables capture behaviour

Ahh thanks. Fixed.

@kevindawson can you mark my first post as answering your question if that's the case? This will help others find the answer quickly. Thanks. 👍

[kevindawson]

@nicoddemus I have pondered over this for some time now:

1, the use of _input to avoid stomping ove the python built-in function(input)

• I am lead to believe -> it should have been input_ both of which I do not like
• as mocking is akin to black boxes -> in_put and conversely out_put make more sense

2, Assign mock to a variable, enables capture behaviour

• but why O why do we have to prefix it with mock_roll_dice as per unittest
• as pytest-mock let's us define the variable name we could be more descriptive loaded_dice in this case

3, the use of the helper method assert_called_once()

• issues as per engineroom & Threat or Menace
• plus you can observe the variable call_count and see it incrementing via the python debugger (pyCharm community ed.)

All of the above goes into the following:

@pytest.mark.parametrize("in_put,out_put", [(3, "You won!"), (4, "You lost!")])
def test_guess_number(mocker, in_put, out_put):
    loaded_dice = mocker.patch("tut9.myapp.sample.roll_dice", return_value=3)
    assert guess_number(in_put) == out_put
    assert loaded_dice.call_count == 1