A developer has released a new Python library called processes to simplify building reliable periodic routines. The library lets you define a Process composed of Task objects that can have dependencies and run in parallel. If a task fails, non-dependent tasks continue, and you can configure email alerts on failure.

The library uses a task graph approach where the order of definition is handled by the Process. It integrates with uv, ruff, mypy, and pytest, and is now on PyPI. The author built it to address the need for more robust orchestration in automated workflows, like data ETL jobs.

The creator has set up a modern development workflow with GitHub Actions and Trusted Publishers, seeking community feedback on functionality, coding style, and project structure. This tool fills a gap between simple cron jobs and heavier workflow engines, offering a lightweight, dependency-aware solution for Python developers managing recurring tasks.