Class-level Setup - setUpClass() and tearDownClass()

diagram pytest fixture scopes, and when each one is torn down mermaid

  flowchart TD
  A["session fixture -- setup"] --> B["module fixture -- setup"]
  B --> C["function fixture -- setup"]
  C --> D["test_one"]
  D --> E["function fixture -- teardown"]
  E --> F["function fixture -- setup"]
  F --> G["test_two"]
  G --> H["function fixture -- teardown"]
  H --> I["module fixture -- teardown"]
  I --> J["session fixture -- teardown"]

A fixture's scope decides how often it is built. Session runs once for the whole test run, module once per file, function once per test -- the default. Teardowns unwind in the reverse order, so a function fixture is always cleaned up before the module fixture it depends on.

Why class-level setup?

Sometimes setup is expensive:

• loading a large fixture file
• initializing test data

Use setUpClass to run once per class.

Example

setupclass_teardownclass.py

import unittest
 
 
class TestExpensiveSetup(unittest.TestCase):
    @classmethod
    def setUpClass(cls):
        cls.data = list(range(1000))
 
    @classmethod
    def tearDownClass(cls):
        cls.data = []
 
    def test_sum(self):
        self.assertEqual(sum(self.data), sum(range(1000)))
 
    def test_len(self):
        self.assertEqual(len(self.data), 1000)

Caution

• Don’t share mutable state between tests unless read-only.
• Tests should not depend on run order.