Context Managers & `with` Statements

Advanced Abstractions

Context Managers

`with` statements prevent leaks and keep setup/teardown paired. Build your own context managers for locks, transactions, and temporary overrides.

Implement __enter__/__exit__Use contextlib

Protocol

class temporary_env:
    def __init__(self, key, value):
        self.key = key
        self.value = value

    def __enter__(self):
        self.original = os.getenv(self.key)
        os.environ[self.key] = self.value
        return self.value

    def __exit__(self, exc_type, exc, tb):
        if self.original is None:
            os.environ.pop(self.key, None)
        else:
            os.environ[self.key] = self.original
        return False  # propagate errors

contextlib helpers

• contextlib.contextmanager decorator turns a generator into a context manager.

from contextlib import contextmanager

@contextmanager
def timed():
    start = time.perf_counter()
    yield
    logger.info("elapsed %.3f", time.perf_counter() - start)

• ExitStack manages dynamic numbers of context managers.
• contextlib.AsyncExitStack (future concurrency cluster) supports async resources.

Next up in your learning path

Decorators

Share setup/teardown logic across many functions.

Companion

threading/multiprocessing

Use context managers to manage locks and pools safely.

Concurrency

Error handling

Guarantee cleanup even when exceptions occur.

Resilience