Refactor statistic analytics to deterministic full-data aggregations

[IanTheMitten]

Motivation

• Replace non-deterministic sampling-based statistic math with deterministic, full-data aggregations driven by transaction timestamps as the single source of truth.
• Provide a single local-day key strategy and reusable aggregation helpers so charts/tables operate consistently on full filtered transaction sets.
• Keep school-only weekday logic explicit by excluding weekends at the aggregation layer.

Description

• Add src/components/statistic/aggregation.ts which centralizes day-key logic and implements filterByDateRange, revenueByWeekday, topDaysByRevenue, revenueByTimePeriod, topProducts, and topCustomers (all operate over full transaction arrays and use the transaction timestamp).
• Rewire src/components/statistic/StatisticPage.tsx to compute explicit date bounds and call filterByDateRange(...), then pass filtered transactions to all statistic UI components; StatisticDateRange now only exposes thisMonth and chosenMonth.
• Update UI/components to consume the new aggregations: WeekdayRevenueBarChart now uses revenueByWeekday with explicit weekend exclusion, timePeriodAnalytics / TimePeriodRevenueBarChart and TimePeriodCumulativeLine use revenueByTimePeriod and cumulative series computed from filtered transactions, and ProductAverageRevenueTable now ranks products using topProducts (inventory adjustment / sampling math removed).
• Remove the sampling subsystem and its statistic-side usage by deleting src/components/statistic/analyticsSampling.ts and stripping sampling controls from DateRangeSelector.tsx.

Testing

• Ran npm run build (Vite production build) and the build completed successfully.

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3138f8b062

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Guard invalid date bounds before filtering transactions

If the month input is cleared (value === ''), parseMonthInputValue produces an invalid Date, and filterByDateRange turns that into NaN bounds. Because the checks only test for null, comparisons like timestamp < startAt and timestamp >= endExclusive both evaluate false, so all transactions are returned instead of an empty/invalid-state result. This is a regression from the prior sampling path (which yielded no selected days for invalid month input) and can silently show incorrect analytics.

Useful? React with 👍 / 👎.