It is important to create unit tests for your functions to make sure they work as you expected, even the experimental code in your Jupyter Notebook. However, it can be difficult to create unit tests in a notebook.
Luckily, ipytest allows you to run pytest inside the notebook environment. To use ipytest, simply add %%ipytest -qq inside the cell you want to run pytest.
Sometimes, it is important to know if your test set contains the same categories in the train set. If you want to check the category mismatch between the train and test set, use Deepchecks’s CategoryMismatchTrainTest.
In the example above, the result shows that there are 2 new categories in the test set. They are ‘d’ and ‘e’.
Deepchecks: Check Category Mismatch Between Train and Test Set Read More »
If you want to test some properties or assumptions, it can be cumbersome to write a wide range of scenarios.
To automatically run your tests against a wide range of scenarios and find edge cases in your code that you would otherwise have missed, use hypothesis.
In the code above, I test if the addition of two floats is commutative. The test fails when either x or y is NaN. Now I can rewrite my code to make it more robust against these edge cases.
Learn more about hypothesis here.
It can be frustrating to wait for a lot of tests to run before knowing the status of the tests. If you want to instantly see the failures and errors with a progress bar, use pytest-sugar.
pytest-sugar is a plugin for pytest. The code above shows how the outputs will look like when running pytest.
pytest-sugar: Show the Failures and Errors Instantly With a Progress Bar Read More »
When testing the outputs of your functions, it can be frustrated to see your tests fail because of something you don’t care too much about such as:
order of items in a list
different ways to specify the same thing such as abbreviation
exact value up to the last decimal point, etc
Is there a way that you can exclude certain parts of the object from the comparison? That is when DeepDiff comes in handy.
DeepDiff Find Deep Differences of Python Objects Read More »
Have you ever wanted to create fake user data for testing? Random User Generator is a free API that generates random user data.
The image above shows how to download and use this data in your code.
Link to previous tips on tools to get new data.
Random User: Generate Random User Data in One Line of Code Read More »
If you want to benchmark your code while testing with pytest, try pytest-benchmark. The code above shows what pytest-benchmark does.
Link to pytest-benchmark.
Link to the source code.
pytest benchmark: A Pytest Fixture to Benchmark your Code Read More »
If you want to output a customized message when the assertion fails, use assert condition, message. The code above is an example of an assertion with an alternative message.
Link to the source code.
Assert in Python: Output a Customized Message When the Assertion Fails Read More »
When working on a data science project, don’t assume that the output will be what you expect. Test your assumption. The easiest way to do that in Python is pytest.
For example, in the code above, I use pytest to test my assumption about the output of the text. When running pytest file_name.py, pytest will run the functions that start with test_.
By running the test, I know where can my function goes wrong.
Check out other examples in my article about pytest.
When testing your function using Pytest, you might want to print only failed tests instead of both failed and successful tests. If so, add –lf to pytest. “lf” stands for “list fails”.
If you are not aware of Pytest, it is the framework that makes it easy to write small tests in Python. You can find more about Pytest from my previous article.