Stockroom: read an unfamiliar stock service, fix three reports, add low-stock alerts, then make reservations fast

Context

You have just joined the checkout team, and the stock service in stockroom/ is yours now. Nobody who wrote it is around. Read README.md and the code first: the reports below assume you know which file does what.

Work through the three phases. Every check is graded on its own, so each fix counts even if you do not finish.

Phase 1: debug

STK-1: Lapsed holds block stock for one second too long

Report: A cart reserved the last scarf at 12:00:00 and never paid. At 12:15:00 exactly, available still says 0.

• A reservation made at t holds nothing from t + 900 on.

STK-2: Two shoppers reserved the last unit

Report: One lamp on hand. Two carts reserved it a few seconds apart, and both got a reservation id.

• reserve refuses when fewer units are available than asked for, counting what other carts hold.

STK-3: Confirming an order takes its units off twice

Report: Six cups on hand, a cart reserves two and confirms. available says 2, not 4. Confirming the same reservation again also succeeds.

• A confirmed reservation holds nothing, and confirming it again is False.

Phase 2: extend

STK-4: Low-stock alerts

Add low_stock(threshold, now) to Stockroom: every SKU ever received with fewer than threshold units available at now, A to Z. Build it on the service's existing pieces.

• low_stock returns the SKUs as described.

Phase 3: optimise

STK-5: Reservations get slower all day

Report: On sale days the stock service crawls: by evening each reservation takes longer than the last, though only a few hundred carts are open at a time. Find out why and fix it without changing what any method returns.

• A day of 18,000 reservations runs within 1 second, the time limit of the speed checks.

Notes

• Run pytest tests/visible. Three tests fail on the current code; not every report has a failing test.
• In the reported round candidates may use an AI assistant. Whatever you use, be ready to explain every change you make.
• Practice task in the format reported for Google's code comprehension interview round, piloted in 2026. Original StackYard code, not an actual Google question; not affiliated with Google.