Improve archery no-layout guidance and actionable suggestions

[TaylorHere]

Summary

• change default target diameter from 28mm to 22mm so the initial A4 preset has a feasible strict-symmetry layout
• add diagnoseNoLayoutReason() in shared layout logic to explain why no solution exists
• use the shared diagnosis in both frontend status text and /api/layout failure response
• add suggestNearbyFeasibleLayouts() to search nearby diameters and provide actionable feasible suggestions
• when no layout is found, frontend now shows up to 3 nearby feasible parameter suggestions (with D / n / m / s)
• add automatic correction flow after users edit layout-related fields: if current config is infeasible, the app auto-adjusts other parameters to a nearby feasible solution
  • first lower s_min to feasible upper bound (when applicable)
  • then try nearest feasible D
  • finally fallback to nearest custom page size for fixed D + s_min
• add target appearance options in UI and export:
  • switch between 黑白靶 and 彩色靶
  • toggle 显示环分数
  • keep preview and PDF rendering consistent for both options
• add layout quantity interaction:
  • 自动填满（最大数量） mode (current maximize behavior)
  • 指定数量（尽量接近） mode with user-entered target count
  • solver now supports both modes in shared logic, and API accepts layoutMode + desiredTargets

Why

Users were frequently seeing 无可行排布 without clear guidance. In strict equal-spacing mode, many (page, D) pairs are mathematically unsatisfiable regardless of s_min, so generic advice was misleading.

The new behavior removes trial-and-error: after changing one parameter, users are automatically brought back to a feasible layout by adjusting other parameters.

Also, users requested richer target styles and optional score labels directly on rings.

And users wanted a direct quantity control flow: either fill the page automatically, or set a target count and get the closest feasible strict-symmetry layout.

Validation

• npm run build (TypeScript + Vite build) passed in archery/.