Commit Graph

4 Commits

Author SHA1 Message Date
Jason 85334d8dce refine(usage): enhance query robustness and error handling
Backend improvements:
- Add InvalidHttpMethod error enum for better error semantics
- Clamp HTTP timeout to 2-30s to prevent config abuse
- Strict HTTP method validation instead of silent fallback to GET

Frontend improvements:
- Add i18n support for usage query errors (en/zh)
- Improve error handling with type-safe unknown instead of any
- Optimize i18n import (direct import instead of dynamic)
- Disable auto-retry for usage queries to avoid API stampede

Additional changes:
- Apply prettier formatting to affected files

Files changed:
- src-tauri/src/error.rs (+2)
- src-tauri/src/usage_script.rs (+8 -2)
- src/i18n/locales/{en,zh}.json (+4 -1 each)
- src/lib/api/usage.ts (+21 -4)
- src/lib/query/queries.ts (+1)
- style: prettier formatting on 6 other files
2025-11-03 10:24:59 +08:00
Jason 4811aa2dcd refactor(models): migrate to granular model configuration architecture
Upgrade Claude model configuration from dual-key to quad-key system for
better model tier differentiation.

**Breaking Changes:**
- Replace `ANTHROPIC_SMALL_FAST_MODEL` with three granular keys:
  - `ANTHROPIC_DEFAULT_HAIKU_MODEL`
  - `ANTHROPIC_DEFAULT_SONNET_MODEL`
  - `ANTHROPIC_DEFAULT_OPUS_MODEL`

**Backend (Rust):**
- Add `normalize_claude_models_in_value()` for automatic migration
- Implement fallback chain: `DEFAULT_* || SMALL_FAST || MODEL`
- Auto-cleanup: remove legacy `SMALL_FAST` key after normalization
- Apply normalization across 6 critical paths:
  - Add/update provider
  - Read from live config
  - Write to live config
  - Refresh config snapshot

**Frontend (React):**
- Expand UI from 2 to 4 model input fields
- Implement smart fallback in `useModelState` hook
- Update `useKimiModelSelector` for Kimi model picker
- Add i18n keys for Haiku/Sonnet/Opus labels (zh/en)

**Configuration:**
- Update all 7 provider presets to new format
- DeepSeek/Qwen/Moonshot: use same model for all tiers
- Zhipu: preserve tier differentiation (glm-4.5-air for Haiku)

**Backward Compatibility:**
- Old configs auto-upgrade on first read/write
- Fallback chain ensures graceful degradation
- No manual migration required

Closes #[issue-number]
2025-11-02 18:02:22 +08:00
Jason 57552b3159 fix: unify dialog layout and fix content padding issues
- Fix negative margin overflow in all dialog content areas
- Standardize dialog structure with flex-col layout
- Add consistent py-4 spacing to all content areas
- Ensure proper spacing between header, content, and footer

Affected components:
- AddProviderDialog, EditProviderDialog
- McpFormModal, McpPanel
- UsageScriptModal
- SettingsDialog

All dialogs now follow unified layout pattern:
- DialogContent: flex flex-col max-h-[90vh]
- Content area: flex-1 overflow-y-auto px-6 py-4
- No negative margins that cause content overflow
2025-10-18 16:52:02 +08:00
Jason 98c35c7c62 refactor: create modular hooks and integrate API key input
- Create custom hooks for state management:
  - useProviderCategory: manages provider category state
  - useApiKeyState: manages API key input with auto-sync to config
  - useBaseUrlState: manages base URL for Claude and Codex
  - useModelState: manages model selection state

- Integrate API key input into simplified ProviderForm:
  - Add ApiKeyInput component for Claude mode
  - Auto-populate API key into settings config
  - Disable for official providers

- Fix EndpointSpeedTest type errors:
  - Fix import paths to use @ alias
  - Add temporary type definitions
  - Format all TODO comments properly
  - Remove incorrect type assertions
  - Comment out unimplemented window.api checks

All TypeScript type checks now pass.
2025-10-16 17:40:25 +08:00