↑ All examplesPython docs reference

Journey

Reliability

This journey follows the boundaries where programs fail, clean up, split into modules, communicate with the outside world, and run concurrent work.

In this journey

  • Make failure explicit.
  • Control resource and module boundaries.
  • Handle operations that outlive one expression.

Make failure explicit.

Robust Python code distinguishes expected absence, broken assumptions, recoverable errors, and domain-specific failures.

Different failure shapes need explicit signals: assertions, recovery, chained causes, or warnings.
  • Exceptions

    Use this example to signal and recover from errors.

  • Assertions

    Use this example to state internal assumptions.

  • Exception Chaining

    Use this example to preserve the cause while translating an error.

  • Exception Groups

    Use this example to handle multiple failures together.

  • Custom Exceptions

    Use this example to name failures in the language of the problem domain.

  • Warnings

    Use this example to signal soft problems and deprecations.

Control resource and module boundaries.

Cleanup, deletion, imports, and modules define where responsibilities begin and end.

Reliable programs name their boundaries: resources clean up, modules import, environments constrain runtime.
  • Context Managers

    Use this example to pair setup with reliable cleanup.

  • Delete Statements

    Use this example to remove names, attributes, and items intentionally.

  • Modules

    Use this example to split code into importable files.

  • Import Aliases

    Use this example to make imported names clear at use sites.

  • Packages

    Use this example to show package directories, `__init__.py`, and public module boundaries.

  • Virtual Environments

    Use this example to isolate dependencies for a project.

Handle operations that outlive one expression.

I/O, testing, logging, subprocesses, and concurrency require different control boundaries from ordinary expressions.

Async, threaded, test, and logging work cross an operation boundary before evidence comes back.
  • Async Await

    Use this example to await concurrent I/O-shaped work.

  • Async Iteration and Context

    Use this example to consume async streams and cleanup protocols.

  • Logging

    Use this example to record operational events without using `print()`.

  • Testing

    Use this example to write deterministic tests with `unittest` or `pytest`.

  • Subprocesses

    Use this example to run external commands safely.

  • Threads and Processes

    Use this example to contrast concurrency choices beyond `asyncio`.

  • Networking

    Use this example to make HTTP or socket boundaries explicit.