feat(orderbook): simplify market data and extend analysis

[moscowchill]

Summary

• Replace the animated game with a conventional QRL/USDT spot order book, quote statistics, recent trades, liquidity, grouping and row controls.
• Remove game state, helpers, animations, assets and navigation/metadata terminology. Add real order-book pause/resume and explicit delayed-data states.
• Add 7-day and 30-day fund-flow analysis with readable UTC labels and additive daily-range metadata. Describe taker buy/sell volume and partial retained history explicitly.

Validation

• Frontend: lint passed with 14 existing warnings, TypeScript check passed, 51 Jest suites and 654 tests passed, scoped formatting and API documentation drift checks passed.
• Production build passed. Browser checks passed: 4 order-book scenarios plus 41 navigation, explorer, pending-transaction and faucet regression scenarios.
• Browser coverage includes 320px, 390px and 1440px layouts, dark/light themes, pause/resume, grouping and row controls, long-window charts, keyboard-accessible data and delayed-response isolation. Screenshots were captured and reviewed locally.
• Backend: go test ./..., go vet ./..., go build ./... and focused fund-flow race tests passed.
• Final diff, forbidden-punctuation and public-repository privacy checks passed. Independent source review found no blocking issue.

Runtime and risks

• Deploy both the API and frontend to enable the longer timeframe choices.
• No database migration, trade execution or historical backfill. Retention, indexes, cache policy and query deadline are unchanged.
• Fund-flow history depends on retained recorded executions and may include collection gaps. The UI states this limit and labels incomplete history.
• Live database acceptance remains a rollout gate; local tests use bounded fixtures.

[moscowchill]

Pre-merge review complete: the 23-file diff is scoped to the order book and fund-flow API. Local lint, TypeScript, formatting, 654 unit tests, 45 production-build browser tests, Go tests/race tests/vet/build, and all three CI jobs pass. No blocking review findings remain. Runtime impact is a paired frontend/API rollout; no database migration or retention/index changes. Retained history may be partial and is explicitly labeled. Live read-only canary acceptance is required before deployment.