phsl / new-api

- Add new PrefillGroup model with CRUD operations
  * Support for model, tag, and endpoint group types
  * JSON storage for group items with GORM datatypes
  * Automatic database migration support

- Implement backend API endpoints
  * GET /api/prefill_group - List groups by type with admin auth
  * POST /api/prefill_group - Create new groups
  * PUT /api/prefill_group - Update existing groups
  * DELETE /api/prefill_group/:id - Delete groups

- Add comprehensive frontend management interface
  * PrefillGroupManagement component for group listing
  * EditPrefillGroupModal for group creation/editing
  * Integration with EditModelModal for auto-filling
  * Responsive design with CardTable and SideSheet

- Enhance model editing workflow
  * Tag group selection with auto-fill functionality
  * Endpoint group selection with auto-fill functionality
  * Seamless integration with existing model forms

- Create reusable UI components
  * Extract common rendering utilities to models/ui/
  * Shared renderLimitedItems and renderDescription functions
  * Consistent styling across all model-related components

- Improve user experience
  * Empty state illustrations matching existing patterns
  * Fixed column positioning for operation buttons
  * Item content display with +x indicators for overflow
  * Tooltip support for long descriptions

Backend
• model/model_meta.go
  – Added `EnableGroups []string` to Model struct
  – fillModelExtra now populates EnableGroups

• model/model_groups.go
  – New helper `GetModelEnableGroups` (reuses Pricing cache)

• model/pricing_refresh.go
  – Added `RefreshPricing()` to force immediate cache rebuild

• controller/model_meta.go
  – `GetAllModelsMeta` & `SearchModelsMeta` call `model.RefreshPricing()` before querying, ensuring groups / endpoints are up-to-date

Frontend
• ModelsColumnDefs.js
  – Added `renderGroups` util and “可用分组” table column displaying color-coded tags

Result
Admins can now see which user groups can access each model, and any ability/group changes are reflected instantly without the previous 1-minute delay.

Overview
• Re-designed `MissingModelsModal` to align with `ModelTestModal` and deliver a cleaner, paginated experience.
• Improved mobile responsiveness for action buttons in `ModelsActions`.

Details
1. MissingModelsModal.jsx
   • Switched from `List` to `Table` for a more structured view.
   • Added search bar with live keyword filtering and clear icon.
   • Implemented pagination via `MODEL_TABLE_PAGE_SIZE`; auto-resets on search.
   • Dynamic rendering: when no data, show unified Empty state without column header.
   • Enhanced header layout with total-count subtitle and modal corner rounding.
   • Removed unused `Typography.Text` import.

2. ModelsActions.jsx
   • Set “Delete Selected Models” and “Missing Models” buttons to `flex-1 md:flex-initial`, placing them on the same row as “Add Model” on small screens.

Result
The “Missing Models” workflow now offers quicker discovery, a familiar table interface, and full mobile friendliness—without altering API behavior.

Highlights
• Introduced `Typography.Text` link with `IconLink` in `extraText` for the **icon** field, pointing to LobeHub’s full icon catalogue; only “请点击我” is clickable for clarity.
• Added required imports for `Typography` and `IconLink`.
• Removed unnecessary `size="large"` prop from the status `Form.Switch` to align with default form styling.

These tweaks improve user guidance when selecting vendor icons and refine the modal’s visual consistency.

Introduce a generic `renderLimitedItems` helper within `ModelsColumnDefs.js` to eliminate duplicated logic for list-style columns.

Key changes
• Added `renderLimitedItems` to handle item limiting, “+N” indicator, and popover display.
• Migrated `renderTags`, `renderEndpoints`, and `renderBoundChannels` to use the new helper.
• Removed redundant inline implementations, reducing complexity and improving readability.
• Preserved previous UX: first 3 items shown, overflow accessible via popover.

This refactor streamlines code maintenance and ensures consistent behavior across related columns.

fix: correct Gemini channel model retrieval logic

feat: Convert gemini request

refactor: simplify streamResponseGeminiChat2OpenAI by removing hasImage return value and optimizing response text handling

fix: update JSONEditor to default to manual mode for invalid JSON and add error message for invalid data

Revert "feat: add Claude Code channel support with OAuth integration"

1. EditModelModal quality-of-life
   • Added comma parsing to `Form.TagInput`; users can now paste
     `tag1, tag2 , tag3` to bulk-create tags.
   • Updated placeholder copy to reflect the new capability.

All files pass linting; no runtime changes outside the intended UI updates.

• Removed obsolete `sidebarIconColors` map and `getItemColor` util from
  SiderBar/render; all selected states now use the single CSS variable
  `--semi-color-primary` for both text and icons.
• Simplified `getLucideIcon`:
  – Added `Package` to Lucide imports.
  – Switched “models” case to `<Package />`, avoiding duplication with
    the Layers glyph.
  – Replaced per-key color logic with `iconColor` derived from the new
    uniform highlight color.
• Stripped any unused imports / dead code paths after the refactor.
• Lint passes; sidebar hover/focus behavior unchanged while visual
  consistency is improved.

Highlights
──────────
1. Removed code duplication
   • Introduced `extractItems` helper to safely unwrap API payloads.
   • Simplified `getFormValues` to a single-line fallback expression.
   • Replaced repeated list-extraction code in `loadModels`, `searchModels`,
     and `refreshVendorCounts` with the new helper.

2. Vendor tab accuracy & performance
   • Added `refreshVendorCounts` to recalc counts via a single lightweight
     request; invoked only when必要 (current tab ≠ "all“) to avoid redundancy.
   • `loadModels` still updates counts instantly when viewing "all", ensuring
     accurate numbers on initial load and page changes.

3. Misc clean-ups
   • Streamlined conditional URL building and state updates.
   • Confirmed all async branches include error handling with i18n messages.
   • Ran linter → zero issues.

Result: leaner, easier-to-maintain hook with correct, real-time vendor counts
and no repeated logic.

The update operation for Model previously overwrote `created_time` with zero
because GORM included every struct field in the UPDATE statement.
This commit adjusts `Model.Update()` to:

* Call `Omit("created_time")` so the creation timestamp is never modified.
* Refresh `UpdatedTime` with `common.GetTimestamp()` before persisting.
* Delegate the remainder of the struct to GORM, eliminating the need to
  maintain an explicit allow-list whenever new fields are introduced.

No API contract is changed; existing CRUD endpoints continue to work
normally while data integrity for historical records is now guaranteed.

Backend
• model/model_meta.go
  – Import strconv
  – SearchModels: support numeric vendor ID filter vs. fuzzy name search
  – Explicitly order by `models.id` to avoid “ambiguous column name: id” error

Frontend
• hooks/useModelsData.js
  – Change vendor-filter API to pass vendor ID
  – Automatically reload models when `activeVendorKey` changes
  – Update vendor counts only when viewing “All” to preserve other tab totals
• Add missing effect in EditModelModal to refresh vendor list only when modal visible
• Other minor updates to keep lints clean

Result
Tabs now:
1. Trigger API requests on click
2. Show accurate per-vendor totals
3. Filter models without resetting other counts
Backend search handles both vendor IDs and names without SQL errors.

Backend
• Add `model/model_meta.go` and `model/vendor_meta.go` defining Model & Vendor entities with CRUD helpers, soft-delete and time stamps
• Create corresponding controllers `controller/model_meta.go`, `controller/vendor_meta.go` and register routes in `router/api-router.go`
• Auto-migrate new tables in DB startup logic

Frontend
• Build complete “Model Management” module under `/console/models`
  - New pages, tables, filters, actions, hooks (`useModelsData`) and dynamic vendor tabs
  - Modals `EditModelModal.jsx` & unified `EditVendorModal.jsx`; latter now uses default confirm/cancel footer and mobile-friendly modal sizing (`full-width` / `small`) via `useIsMobile`
• Update sidebar (`SiderBar.js`) and routing (`App.js`) to surface the feature
• Add helper updates (`render.js`) incl. `stringToColor`, dynamic LobeHub icon retrieval, and tag color palettes

Table UX improvements
• Replace separate status column with inline Enable / Disable buttons in operation column (matching channel table style)
• Limit visible tags to max 3; overflow represented as “+x” tag with padded `Popover` showing remaining tags
• Color all tags deterministically using `stringToColor` for consistent theming
• Change vendor column tag color to white for better contrast

Misc
• Minor layout tweaks, compact-mode toggle relocation, lint fixes and TypeScript/ESLint clean-up

These changes collectively deliver end-to-end model & vendor administration while unifying visual language across management tables.

feat: add Claude Code channel support with OAuth integration

fix: handle authorization code format in ExchangeCode function and update placeholder in EditChannelModal

# Conflicts:
#	web/src/components/table/channels/modals/EditChannelModal.jsx