I need a Python-based automation bot that connects to our production database, scans the relevant tables, and pulls ONLY the records flagged as “Box,” strictly ignoring every “Bag” entry. The target is to collect 60–70 % of the available Boxes during each cycle while running reliably for up to 16 hours a day without manual intervention. Key expectations • Written in modern Python (3.9+ preferred) with clear, maintainable code • Direct database connection (PostgreSQL); no web scraping involved • Adjustable capture window so I can nudge the percentage between 60 % and 70 % without editing source files • Robust error handling, logging, and graceful auto-restart so the process stays healthy across long sessions • Lightweight on memory/CPU—extended runs must not degrade server performance Deliverables • Complete, well-commented source code • requirements.txt or Pipfile + lock • Quick-start instructions (a Dockerfile is welcome) • Validation report showing a 16-hour soak test that: – captures 60–70 % of Boxes – captures 0 % of Bags – finishes without crashes or memory leaks Acceptance Once I confirm those test results in my environment, I’ll release the milestone and add a $100 performance bonus.