Generate Sophisticated Pytest Fixtures

Acts as an expert in designing sophisticated pytest fixtures for test isolation, performance, and maintainability, covering fixture scoping, dependency injection, parameterization, and complex test-data setup. On scoping strategy it recommends session scope for expensive shared resources like databases and external services, module scope for data shared within one test file, function scope (pytest's default) for per-test isolation and clean state, and class scope for test classes that share setup - paired with guidance to design fixtures as composable, dependency-injected units with proper yield-based teardown and graceful exception handling during cleanup.

It provides ready patterns across several fixture categories: database fixtures (a session-scoped SQLAlchemy engine, a function-scoped session wrapped in a rollback transaction for isolation, and a sample_user fixture layered on top); HTTP client fixtures (a session-scoped test app, a function-scoped test client using test_client()/app_context(), and an authenticated-client fixture that logs in via a POST request); file-system fixtures using tempfile.TemporaryDirectory for a temp directory and a derived sample config file written as JSON; parameterized fixtures using @pytest.fixture(params=[...]) for both structured data (user roles with different permission sets) and simple value sweeps (batch sizes); mock and patch fixtures using unittest.mock.patch for external API calls and for freezing datetime.now()/utcnow() to a fixed value; factory fixtures that yield a callable (user_factory) so tests can create many custom instances and have them auto-cleaned afterward; and conditional fixtures that call pytest.skip() when a dependency like a local Redis server is unavailable, rather than failing the test.

It also covers conftest.py organization across three levels - a root-level conftest.py for global config, a tests/unit/conftest.py using patch.multiple to mock all external dependencies for unit tests, and a tests/integration/conftest.py for a real, module-scoped database connection - plus best-practice guidance on performance (minimizing fixture recreation via correct scoping, lazy-loading expensive resources, pytest-xdist compatibility for parallel runs), error handling (try/finally cleanup, meaningful failure messages, graceful skips), fixture-quality testing (verifying isolation, monitoring setup/teardown cost), and naming conventions (mock_ prefix for mocks, sample_ prefix for test data).

Use this skill when designing or refactoring pytest fixtures for a test suite that needs proper isolation, layered dependencies (database to client to authenticated client), parameterized test cases, or realistic mocked external services - especially when a suite is currently slow, flaky, or has fixtures that leak state between tests. It is not a general pytest tutorial or a guide to writing the test assertions themselves; it is scoped specifically to fixture design, scoping, and conftest.py organization.

Input is a description of the test scenario or fixture need (e.g. "isolated database fixture," "authenticated HTTP client," "parameterized user roles"). Output is pytest fixture code following the scoping, teardown, and naming conventions above, ready to drop into a conftest.py or test module.

Built around the pytest fixture system (@pytest.fixture, yield, params), and commonly paired with SQLAlchemy (create_engine, Session) for database fixtures, Flask-style test_client()/app_context() for HTTP fixtures, Python's tempfile and pathlib.Path for file-system fixtures, unittest.mock.patch/patch.multiple for mocking, and pytest-xdist for parallel-execution-safe fixtures.

Python developers and QA engineers building or maintaining a pytest suite who need composable, well-scoped fixtures for databases, HTTP clients, file systems, mocked services, or parameterized test data, rather than ad hoc setup/teardown code duplicated across tests.