Pytest Advanced Patterns

Overview

Pytest is a powerful testing framework that emphasizes modularity via fixtures and scalability via parametrization. It simplifies setup/teardown logic and enables complex test matrices with minimal code.

When to Use

• Resource Intensive Tests: Using scoped fixtures for database or API connections to avoid redundant setup.
• Input Combinations: Using parametrization to test a single function against dozens of inputs.
• Async Codebases: Testing coroutines directly without manual event loop management.

Decision Tree

1. Is the setup expensive (e.g., creating a DB)?
   • YES: Use a fixture with scope='session' or scope='module'.
2. Do you need to cleanup after a test?
   • YES: Use a yield fixture.
3. Do you have 3+ similar test cases for one function?
   • YES: Use @pytest.mark.parametrize.

Workflows

1. Defining a Scoped Setup/Teardown

1. Create a fixture using the @pytest.fixture(scope="module") decorator for expensive resources like database connections.
2. Use yield to provide the resource to tests.
3. Add cleanup code (e.g., closing the connection) after the yield statement.
4. Request the fixture by name in any test function within the module.

2. Matrix Testing with Stacked Parametrize

1. Apply @pytest.mark.parametrize("user", ["admin", "guest"]) to a test.
2. Apply a second @pytest.mark.parametrize("action", ["read", "write"]) to the same test.
3. Pytest will run the test 4 times, once for each combination (admin-read, admin-write, etc.).

3. Testing Async Code

1. Install pytest-asyncio.
2. Mark the test function with @pytest.mark.asyncio.
3. Define the test as an async def.
4. Use await for the code under test and assertions.

Non-Obvious Insights

• Fixture Caching: Fixtures are executed once per requested scope and the return value is cached for all tests in that scope.
• Reverse Teardown: Teardown logic in yield fixtures is executed in the exact reverse order of the setup sequence.
• Auto-Async Marking: Modern pytest-asyncio can be configured to treat all async def tests as async without needing the explicit decorator on every function.

Evidence

• "pytest won’t execute them again for that test... return values are cached." - Pytest Docs
• "Yield fixtures yield instead of return... Any teardown code for that fixture is placed after the yield." - Pytest Docs
• "provides support for coroutines as test functions. This allows users to await code inside their tests." - pytest-asyncio

Scripts

• scripts/pytest-patterns_tool.py: Examples of fixtures, parametrization, and async tests.
• scripts/pytest-patterns_tool.js: (Comparison) Equivalent logic in Vitest/Jest.

Dependencies

• pytest
• pytest-asyncio

References

• references/README.md