Pytest fixture allows you to use the same data to test different functions. Every time you use a pytest fixture in a test, a fixture will be executed.
This means that a fixture will be executed twice if used in two different tests.
If a fixture is expensive to execute, you can make the fixture be executed only once per session using scope=session.
If you expect a test to fail, use pytest xfail marker. This will prevent pytest from marking a test as failed when there is an exception.
To be more specific about what exception you expect to see, use the raises argument.
Sometimes, your data might have identical samples with different labels. This might be because the data was mislabeled.
It is good to identify these conflicting labels in your data before using the data to train your ML model. To check conflicting labels in your data, use deepchecks.
In the example above, deepchecks identified that samples 0 and 1 have the same features but different labels.
If you want to skip a test when a condition is not met, use pytest skipif. For example, in the code above, I use skipif to skip a test if the python version is less than 3.9.
Pytest skipif: Skip a Test When a Condition is Not Met Read More »
Have you ever wanted to use the result of one test for another test? That is when pytest_steps comes in handy.
In the code above, I use the result of sum_test as the input of average_2_nums. The argument steps_data allows me to share the data between 2 tests.
When using pytest parametrize, it can be difficult to understand the role of each test case. You can add the ids parameter to pytest parametrize to assign names to test cases.
In the code above, the first test case is shown as neg-neg instead of [-1–2]. This makes it easier for others to understand the roles of your test cases.
Weight and Biases is a tool to track and monitor your ML experiments. deepchecks is a tool that allows you to create test suites for your ML models & data with ease.
The checks in a suite includes:
🔎 model performance
🔎 data integrity
🔎 distribution mismatches
and more.
Now you can track deepchecks suite’s results with Weights & Biases as shown above.
Deepchecks + Weights & Biases: Test and Track Your ML Model and Data Read More »
If you want to test the combinations of two sets of parameters, writing all possible combinations can be time-consuming and is difficult to read.
You can save your time by using pytest.mark.parametrize twice instead. From the output of pytest, we can see that all possible combinations of the given functions and inputs are tested.
pytest parametrize twice: Test All Possible Combinations of Two Sets of Parameters Read More »
It can be time-consuming to create data to test edge cases of your NLP model. If you want to quickly create data to test your NLP models, use Checklist.
In the code above, I use Checklist’s Editor to create multiple examples of negation in one line of code.
