Unit testing

Unit testing#

Now lets use the Test-driven development methodology with pytest to create a efficient Unit Testing.

The Problem#

For this example lets considering three functions:

even_odd: function that returns Even or Odd for a given number.
get_largest: function that returns the largest value between three inputs.
dice_rolling: function that that works as a dice-rolling simulator, i.e., that randomly returns a value between 1 and 6.

Which tests should we create for each function?

Project Structure#

The first step is to create the proper structure for creating the main code and test units.

some-functions
├── src
│   └── __init__.py
│   └── functions.py
└── tests
│   ├── __init__.py
│   └── test_dice.py
│   └── test_evenodd.py
│   └── test_largest.py
└── environment.yml

Setting the environment#

For this task we are going to write some functions in python and then, test them with pytest. Therefore we need a python environment with these packages. We also want to use black to linting the code, so the environment.yml should look like:

name: testing
dependencies:
  - python
  - black
  - pytest

To install/activate the environment:

$ conda env create -f environment.yml
$ conda activate testing

Even or Odd?#

Before jump to the function itself lets think about what tests we should do for an even/odd function. Some ideas are:

Is the output a string?
Does the function work for a small even number?
Does the function work for a large even number?
Does the function work for a small odd number?
Does the function work for a large odd number?
Does the function work for negative number (in ever situation above)?
Does the function work for zero?
What the function should return for a non-integer value? >> define here >> message: “Please use an integer value”

Even or odd test unit#

Now inside the file test_evenodd.py write a test function for every item listed above. Something like:

import src.functions as fc

def test_output_type():
    assert isinstance(fc.even_odd(5), str)

def test_even_small():
    assert fc.even_odd(4) == "Even"

def test_even_large():
    assert fc.even_odd(178418932) == "Even"

def test_odd_small():
    assert fc.even_odd(178418931) == "Odd"

def test_odd_large():
    assert fc.even_odd(178418931) == "Odd"

def test_zero():
    assert fc.even_odd(0) == "Even"

def test_neven_small():
    assert fc.even_odd(-4) == "Even"

def test_neven_large():
    assert fc.even_odd(-178418932) == "Even"

def test_nodd_small():
    assert fc.even_odd(-178418931) == "Odd"

def test_nodd_large():
    assert fc.even_odd(-178418931) == "Odd"

def test_no_integer():
    assert fc.even_odd(1.5) == "Please use an integer value"

The `even_odd` function#

With the test unit in hand, you can now create the even_odd function. To do this, open the file functions.py and type your function.

def even_odd(number):
    if isinstance(number, int):
        if number % 2 == 0:
            return "Even"
        else:
            return "Odd"
    else:
        return "Please use an integer value"

Testing the `even_odd` function#

To test if your function is working in scenarios that you planned before, you just need to use the pytest command.

$pytest
============================= test session starts =============================
platform linux -- Python 3.11.0, pytest-7.1.2, pluggy-1.0.0
rootdir: /localhome/home/earpte/Documents/RSE-support/FLUDS/testing
plugins: lazy-fixture-0.6.3
collected 11 items                                                                                                                                                                               

tests/test_evenodd.py ...........                                         [100%]

============================== 11 passed in 0.02s ==============================

The largest and the dice problems#

Now, do the same with the other two functions.

First list all scenarios and the desired outputs.
Write a test unit that covers all scenarios
Write the function
Test the function using pytest

For the Dice problem some additional knowledge may be helpful.

You can define variables to be used by pytest via @pytest.fixture
You may have problems in verify if two float numbers are equal, try the pytest.approx function to set the number of decimal places to be considered.