phsl / new-api

- Add comprehensive i18n support to TwoFASetting.js component
- Add all required English translations to en.json for 2FA settings
- Update component to accept t function as prop and use translation keys
- Fix prop passing in PersonalSetting.js to provide t function
- Maintain all existing UI improvements and functionality
- Support both Chinese and English interfaces for:
  * Main 2FA settings card with status indicators
  * Setup modal with guided steps (QR scan, backup codes, verification)
  * Disable 2FA modal with impact warnings and confirmation
  * Regenerate backup codes modal with success states
  * All buttons, placeholders, messages, and notifications
- Follow project i18n conventions using t('key') pattern
- Ensure seamless language switching for enhanced user experience

This enables the 2FA settings to be fully localized while preserving
the modern UI design and improved user workflow from previous updates.

- Replace lucide-react icons with Semi UI icons for consistency
- Implement Steps component for guided 2FA setup modal flow
- Redesign disable and regenerate backup codes modals to match setup modal style
- Extract duplicate backup codes display logic into reusable BackupCodesDisplay component
- Move modal navigation buttons to proper footer parameter following Semi UI standards
- Replace custom styled dots with Badge components (warning/danger/success types)
- Use Banner and Divider components for better visual hierarchy
- Remove redundant modal step titles and download functionality
- Apply consistent rounded corners, spacing, and color scheme across all modals
- Improve responsive design with maxWidth constraints

This unifies the 2FA settings visual design with other settings pages and
enhances user experience through better component usage and layout structure.

Feats: use the types of gjson,the error expection,the invert and the key missing process  of the condition

feats: the flexable params override and compatible format

feat: use proxy HTTP client fetch models

feat(relay): add OpenAI gpt-4.1 o3 o4 gpt-image-1 models

Fix reasoning adaptor for openrouter

fix: prevent OpenRouter cache calculation with custom model ratios

手机客户端下拉菜单采用导航栏的样式

refactor: Introduce pre-consume quota and unify relay handlers