Test Driven Development
So, I started learning Golang recently, and decided that Learn Go with Tests was a good way to go.
However, there was a great ‘buy-one-get-one-free’ with this resource - it also taught me some of the benefits of Test Driven Development (TDD). There are four steps to TDD:
- Write the test
- Write a minimal amount of code for the test to recognise the code (e.g. write the function and just return - the test should still fail)
- Write the function correctly to return the expected output
- Refactor (change the code to better represent the test case)
I’m not using Go that much at work, with Python being the dominant language (for now). This meant I had to learn how to use tests in Python. This blog post represents my learning and notes.
Install Pytest
Pytest does not come with the standard library, so does require an install.
To install pytest, you just run pip install -u pytest.
Write the test
So, my initial test was to just add two numbers together:
from src.add import add
def test_add():
assert add(5, 7) == 12
Here, the code asserts that if I pass 5 and 7 to my add function, the function will return 12. If this returns anything different, then the test will fail.
Directory Structure
This is within a tests directory.
My project file structure is thus:
project/
|
|- src
|
|- tests
The initial problem that needs to be solved is that the test needs to import the add module. To help with this, we need to add __init__.py files to the src directory
The initial problem that needs to be solved is that the test needs to import the add module. To help with this, we need to add __init__.py files to the src and tests directory. So this is how the file structure looks now:
project/
|
|- src
| |- __init__.py
|- tests
|- test_add.py
|- __init__.py
However, the import will still not work. You will need a setup.py file in the top directory to help with that:
from setuptools import setup, find_packages
setup(name="add", packages=find_packages())
This file needs to be at the project level:
project/
|- setup.py
|- src
| |- __init__.py
|- tests
|- test_add.py
|- __init__.py
The important part of the process now is that you run the test to prove that it fails. Run pytest in the project directory. Pytest automatically looks for directories and files that start with test:
================================================= test session starts =================================================
platform win32 -- Python 3.9.6, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: C:\Users\adamc\Documents\test
plugins: anyio-3.3.0
collected 0 items / 1 error
======================================================= ERRORS ========================================================
_________________________________________ ERROR collecting tests/test_add.py __________________________________________
importError while importing test module 'C:\Users\adamc\Documents\test\tests\test_add.py'.
hint: make sure your test modules/packages have valid Python names.
traceback:
..\..\appdata\local\programs\python\python39\lib\importlib\__init__.py:127: in import_module
return _bootstrap._gcd_import(name[level:], package, level)
tests\test_add.py:1: in <module>
from src.add import add
e ModuleNotFoundError: No module named 'src.add'
=============================================== short test summary info ===============================================
eRROR tests/test_add.py
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: 1 error during collection !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
================================================== 1 error in 0.21s ===================================================
This proves that the error is failing.
Then we need to create the file in the src directory to help this pass. Even at the moment, the import is not finding the src directory, which needs to be fixed.
Write the minimal amount of code
We need a file add.py in the src directory. The first iteration of the file we can add is just a function that returns:
def add():
return
This will change the nature of the test - it will run, but will now fail properly:
================================================= test session starts =================================================
platform win32 -- Python 3.9.6, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: C:\Users\adamc\Documents\test
plugins: anyio-3.3.0
collected 1 item
tests\test_add.py F [100%]
====================================================== FAILURES =======================================================
______________________________________________________ test_add _______________________________________________________
def test_add():
> assert add(5, 7) == 12
E TypeError: add() takes 0 positional arguments but 2 were given
tests\test_add.py:4: TypeError
=============================================== short test summary info ===============================================
FAILED tests/test_add.py::test_add - TypeError: add() takes 0 positional arguments but 2 were given
================================================== 1 failed in 0.24s ==================================================
So, pytest is now importing the function, and passing in 5 and 7 parameters, but only getting a return (the function doesn’t accept any values at the moment).
Let’s fix that:
def add(x, y):
return x + y
If we run pytest we now get:
================================================= test session starts =================================================
platform win32 -- Python 3.9.6, pytest-6.2.4, py-1.10.0, pluggy-0.13.1
rootdir: C:\Users\adamc\Documents\test
plugins: anyio-3.3.0
collected 1 item
tests\test_add.py . [100%]
================================================== 1 passed in 0.06s ==================================================
The good news here is the tests\test_add.py . section - specifically the . character. This outlines that the test passed.
Refactor the code
This function is very simple and doesn’t need any refactoring. It might be that my test needs refactoring, or my code needs refactoring. Here this isn’t necessary, but it might be that there is some duplication in my code, or that I need to change a special case (for example, divide by zero).
Summary
This is just an introduction to testing. There are many more ways tests can be ’named’ and then run, but this just aims to get a simple test suite up and running. For more information, consult the pytest documentation.