Fixtures are functions decorated by pytest.fixture() that run before (and maybe after) test functions by pytest to do the setup/teardown. Fixtures can do some setup work and return data to the test functions as well. The name of the fixture(s) are passed to the test/fixtures as parameters and pytest will see that and check for a fixture with that name. We can use the fixture name inside test functions to access returned data or other state(s) that the fixtute set up.
To conclude, fixtures help us separate the “getting ready for” and “cleaning up after” code of test functions. If the fixture has yield statement, then the fixture would run until the yield before yielding to the test function. Once the test function is done, the code following the yield statement would run even if the test function failed.
If we know that some fixtures have to be run always at certain times, we can avoid passing their names to all test functions by using autouse=True parameter in @pytest.fixture which means the fixture will be called always depending on the scope.
We can also rename a fixture using name parameter of @pytest.fixture. This is useful if the fixture name collides with some test functions OR we follow the convention to have the fixture name have fixture suffix/prefix.
Test functions as well as other fixtures can depend on fixtures through their parameter list that would have the same name(s) of the fixture(s).