feat: polish spatial view and extract public deck proofs

[FHoffarth]

Summary

• Spatial Visual Polish: Implemented zoom-adaptive stateroom density rendering (thin outlines at low zoom, progressive label reveal at zoom $\ge 1.75$), rich semantic color coding (Stateroom: white, Public Venue: amber, Infrastructure: slate), and glowing amber selection rings.
• Multi-Deck Forensic Extraction: Extracted versioned forensic geometry proof artifacts from ART-0001 Page 3 for public decks:
  • Deck 5 (geometry/proofs/bellissima/deck05/deck05.proof.json — 30 objects)
  • Deck 6 (geometry/proofs/bellissima/deck06/deck06.proof.json — 26 objects)
  • Deck 7 (geometry/proofs/bellissima/deck07/deck07.proof.json — 15 objects)
• Fail-Closed Dual-Gate Governance: All newly extracted public deck proof objects retain DRAFT / UNKNOWN / PUBLISH_BLOCKED and are strictly excluded from passenger UI exposure until owner review adjudication.
• Passenger Visibility Bounds: Only Deck 14 is presented to passengers with 244 admitted spatial objects (243 cabins + 1 Vertical Core Area).

Tests & CI Coverage

• Frontend Vitest Suite: 76/76 tests passing (including 20 spatial visual polish & boundary tests).
• Frontend Build: Clean Vite production build with zero TypeScript errors.
• Backend Test Suite: 892/892 pytest tests passing across 72 test suites.
• Diff Hygiene: git diff --check is clean with zero whitespace warnings.