automated-testing-with-pytest-and-fixtures-steven-saporta.md

"Automated Testing with pytest and Fixtures" by Steven Saporta

Features of pytest

• Assertion-based testing pytest tests validity of code through Python's built-in assert mechanisms, whereas unittest has its own conventions pulled from JUnit, the Java testing framework
• Simple test discovery Discovers tests via naming convention of directories, files, and methods
• Run tests by name match Run tests whose names match an expression
• Parallel testing with pytest-xdist Speed up your tests by running many at once. A pre-requisite is to make sure that tests are parallel-safe, i.e. don't manipulate the same data in ways that will make other tests fail since there's no guarantee of run order

Fixtures

• Scope: run once per test, module, or session Fixtures have granular scope control that allow setup steps to be run on a per-test, per-module, or per-run basis
• Parameterize tests Allows running a test or set of tests with different inputs without having to duplicate code. Useful for something akin to fuzz testing or iterating over a sample of real data for sanity checking
• Do useful work like pre-verifying system state Can alter the working directory, create temporary or results directories, etc.
• Can create test records Really useful for performance when many tests will test some features of the same data record