Repeat-work prevention
A small Python library that tracks which items a recurring job has already processed.
Local library. The caller remains responsible for the processor and its deployment.
The problem
Recurring jobs can send the same items again or compete to process them at the same time.
What I built
I extracted stable item identifiers, a persistent record, atomic reservation, and empty-run checks into a small library. Each run records what it processed.
What you can inspect
The repository includes a CLI demo and tests for concurrent submissions and crash recovery.