Backend (controller/ratio_sync.go):
- Add built‑in official upstream to GetSyncableChannels (ID: -100, BaseURL: https://basellm.github.io)
- Support absolute endpoint URLs; otherwise join BaseURL + endpoint (defaults to /api/ratio_config)
- Harden HTTP client:
- IPv4‑first with IPv6 fallback for github.io
- Add ResponseHeaderTimeout
- 3 attempts with exponential backoff (200/400/800ms)
- Validate Content-Type and limit response body to 10MB (safe decode via io.LimitReader)
- Robust parsing: support type1 ratio_config map and type2 pricing list
- Use net.SplitHostPort for host parsing
- Use float tolerance in differences comparison to avoid false mismatches
- Remove unused code (tryDirect) and improve warnings
Frontend:
- UpstreamRatioSync.jsx: auto-assign official endpoint to /llm-metadata/api/newapi/ratio_config-v1-base.json
- ChannelSelectorModal.jsx:
- Pin the official source at the top of the list
- Show a green “官方” tag next to the status
- Refactor status renderer to accept the full record
Notes:
- Backward compatible; no API surface changes
- Official ratio_config reference: https://basellm.github.io/llm-metadata/api/newapi/ratio_config-v1-base.json
- Ran: bun run eslint:fix && bun run lint:fix
- Inserted AGPL license header via eslint-plugin-header
- Enforced no-multiple-empty-lines and other lint rules
- Formatted code using Prettier v3 (@so1ve/prettier-config)
- No functional changes; formatting-only baseline across JS/JSX files
- Rename React components/pages/utilities that contain JSX to `.jsx` across `web/src`
- Update import paths and re-exports to match new `.jsx` extensions
- Fix Vite entry by switching `web/index.html` from `/src/index.js` to `/src/index.jsx`
- Verified remaining `.js` files are plain JS (hooks/helpers/constants) and do not require JSX
- No runtime behavior changes; extension and reference alignment only
Context: Resolves the Vite pre-transform error caused by the stale `/src/index.js` entry after migrating to `.jsx`.
- Switch Collapse from controlled (activeKey) to uncontrolled (defaultActiveKey) so user toggling works
- Add a stable key to reset Collapse state when tab/category changes
- Default all panels to collapsed via defaultActiveKey: []
- Preserve Panel itemKey for consistent behavior
- No linter errors introduced
Scope: web/src/components/table/channels/modals/ModelSelectModal.jsx
- Added unique keys for JSONEditor components to ensure proper re-rendering based on channelId.
- Implemented input reset to clear previous JSON field values when the modal is opened.
- Replace custom dots with Semi Badge types (success/danger/warning); add compact Progress bars
- Remove pie chart and related deps/config; move total key count and mode tags into the modal title
- Rework header using Row/Col; three equal stat cards (enabled/manual-disabled/auto-disabled)
- Integrate toolbar into Table title; wrap content with Card; use Table’s native empty state
- Make “Enable All” conditional (hidden when all keys are enabled), mirroring “Disable All”
- Unify numeric typography (current/total same size) for better readability
- Default page size set to 10; fallback to 10 when backend page_size is absent; page-size options: 10/20/50/100
- Cleanup imports and dead code (remove VChart and pie-spec logic)
- Minor spacing polish (extra bottom margin before table), no footer buttons
- Added support to fetch and render “model prefill groups” in `EditChannelModal.jsx`
- Users can now click a group button to instantly merge that group’s models into the models Select
- Mirrors the prefill-group UX used for tags/endpoints in `EditModelModal.jsx`
Details
- UI/UX:
- Renders one button per model group inside the models field’s extra actions
- Clicking a button merges its items into the selected models (trimmed, deduplicated), updating immediately
- Non-destructive and works alongside existing actions (fill related/all models, fetch upstream, clear, copy)
- API:
- GET `/api/prefill_group?type=model`
- Handles `items` as either an array or a JSON string array for robustness
- If request fails or returns no groups, buttons are simply not shown
- i18n:
- Reuses existing i18n; group names come from backend and are displayed as-is
- Performance:
- Simple set merge; negligible overhead
- Backward compatibility:
- No changes required on the backend or elsewhere; feature is additive
- Testing (manual):
1) Open channel modal (new or edit) and navigate to the Models section
2) Confirm model group buttons render when groups are configured
3) Click a group button → models Select updates with merged models (no duplicates)
4) Verify other actions (fill related/all, fetch upstream, clear, copy) still work
5) Close/reopen modal → state resets as expected
Implementation
- `web/src/components/table/channels/modals/EditChannelModal.jsx`
- Added `modelGroups` state and `fetchModelGroups()` (GET `/api/prefill_group?type=model`)
- Invoked `fetchModelGroups()` when the modal opens
- Rendered group buttons in the models Select `extraText`, merging group items into current selection
Chore
- Verified no new linter errors were introduced.
- Why:
- Avoid introducing `gorm.io/datatypes` for a single field.
- Align with existing pattern (`ChannelInfo`, `Properties`) using `Scanner`/`Valuer`.
- Fix runtime error when drivers return JSON as string.
- What:
- Introduced `JSONValue` (based on `json.RawMessage`) implementing `sql.Scanner` and `driver.Valuer`, with `MarshalJSON`/`UnmarshalJSON` to preserve raw JSON in API.
- Updated `PrefillGroup.Items` to use `JSONValue` with `gorm:"type:json"`.
- Localized comments in `model/prefill_group.go` to Chinese.
- Impact:
- Resolves “unsupported Scan, storing driver.Value type string into type *json.RawMessage”.
- Works with MySQL/Postgres/SQLite whether JSON is returned as `[]byte` or `string`.
- API and DB schema remain unchanged; no `go.mod` changes; lints pass.
Files changed:
- model/prefill_group.go
This commit significantly refactors the `EditModelModal` component to streamline the user interface and enhance usability, aligning it with the interaction patterns found elsewhere in the application.
- **Consolidated Layout:** Merged the "Vendor Information" and "Feature Configuration" sections into a single "Basic Information" card. This simplifies the form, reduces clutter, and makes all settings accessible in one view.
- **Improved Prefill Groups:** Replaced the separate `Select` dropdowns for tag and endpoint groups with a more intuitive button-based system within the `extraText` of the `TagInput` components.
- **Additive Button Logic:** The prefill group buttons now operate in an additive mode. Users can click multiple group buttons to incrementally add tags or endpoints, with duplicates being automatically handled.
- **Clear Functionality:** Added "Clear" buttons for both tags and endpoints, allowing users to easily reset the fields.
- **Code Cleanup:** Removed the unused `endpointOptions` constant and unnecessary icon imports (`Building`, `Settings`) to keep the codebase clean.
Summary
• Introduced standalone `ModelSelectModal.jsx` for selecting channel models
• Fetch-list now opens modal instead of in-place select, keeping EditChannelModal lean
Modal Features
1. Search bar with `IconSearch`, keyboard clear & mobile full-screen support
2. Tab layout (“New Models” / “Existing Models”) displayed next to title, responsive wrapping
3. Models grouped by vendor via `getModelCategories` and rendered inside always-expanded `Collapse` panels
4. Per-category checkbox in panel extra area for bulk select / deselect
5. Footer checkbox for bulk select of all models in current tab, with real-time counter
6. Empty state uses `IllustrationNoResult` / `IllustrationNoResultDark` for visual consistency
7. Accessible header/footer paddings aligned with Semi UI defaults
Fixes & Improvements
• All indeterminate and full-select states handled correctly
• Consistent “selected X / Y” stats synced with active tab, not global list
• All panels now controlled via `activeKey`, ensuring they remain expanded
• Search, vendor grouping, and responsive layout tested across mobile & desktop
These changes modernise the channel model management workflow and prepare the codebase for upcoming upstream-ratio integration.
Summary
• Introduced a unified `selectFilter` helper that matches both `option.value` and `option.label`, ensuring all `<Select>` components support intuitive search (fixes channel “type” dropdown not filtering).
• Replaced all usages of the old `modelSelectFilter` with `selectFilter` in:
• `EditChannelModal.jsx`
• `SettingsPanel.js`
• `EditTokenModal.jsx`
• `EditTagModal.jsx`
• Removed the deprecated `modelSelectFilter` export from `utils.js` (no backward-compat alias).
• Updated documentation comments accordingly.
Why
The old filter only inspected `option.value`, causing searches to fail when `label` carried the meaningful text (e.g., numeric IDs for channel types). The new helper searches both fields, covering all scenarios and unifying the API across the codebase.
Notes
No functional regressions expected; all components have been migrated.
Previously, the "Force Format" switch was displayed for every channel type
although it only applies to OpenAI (type === 1).
This change wraps the switch in a conditional so it renders exclusively when
the selected channel type is OpenAI.
Why:
- Prevents user confusion when configuring non-OpenAI channels
- Keeps the UI clean and context-relevant
Scope:
- web/src/components/table/channels/modals/EditChannelModal.jsx
No backend logic affected.
- Extract channel extra settings into a dedicated Card component for better visual hierarchy
- Replace custom gray background container with consistent Form component styling
- Simplify layout structure by removing complex Row/Col grid layout in favor of native Form component layout
- Unify help text styling by using extraText prop consistently across all form fields
- Move "Settings Documentation" link to card header subtitle for better accessibility
- Improve visual consistency with other setting cards by using matching design patterns
The channel extra settings (force format, thinking content conversion, pass-through body, proxy address, and system prompt) now follow the same design language as other configuration sections, providing a more cohesive user experience.
Affected settings:
- Force Format (OpenAI channels only)
- Thinking Content Conversion
- Pass-through Body
- Proxy Address
- System Prompt
The VolcEngine Ark/Doubao channel now has a hard-coded base URL inside the backend, so it no longer requires any API-address settings on the front-end side.
Previously, the input field was hidden but the surrounding “API Config” card still rendered, leaving a blank, confusing section.
Changes made
• Added `showApiConfigCard` flag (true when `inputs.type !== 45`) right after the state declarations.
• Wrapped the entire “API Config” card in a conditional render driven by this flag.
• Removed the duplicate declaration of `showApiConfigCard` further down in the component to avoid shadowing and improve readability.
Scope verification
• Checked all other channel types: every remaining type either displays a dedicated API-related input/banner (3, 8, 22, 36, 37, 40, …) or falls back to the generic “custom API address” field.
• Therefore, only type 45 requires the card to be fully hidden.
Result
The “Edit Channel” modal now shows no empty card for the VolcEngine Ark/Doubao channel, leading to a cleaner and more intuitive UI while preserving behaviour for all other channels.
- Add left-right pagination layout for desktop (total info on left, controls on right)
- Keep mobile layout centered with pagination controls only
- Implement proper i18n support for pagination text using react-i18next
- Add pagination translations for Chinese and English
- Standardize t function usage across all table components to use xxxData.t pattern
- Update CardPro footer layout to support justify-between on desktop
- Use CSS variable --semi-color-text-2 for consistent text styling
- Disable built-in Pagination showTotal to avoid duplication
Components updated:
- CardPro: Enhanced footer layout with responsive design
- createCardProPagination: Added i18n support and custom total text
- All table components: Unified t function usage pattern
- i18n files: Added pagination-related translations
The pagination now displays "Showing X to Y of Z items" on desktop
and maintains existing centered layout on mobile devices.
Fix "Rendered fewer hooks than expected" error caused by conditional hook calls
in createCardProPagination function. The issue occurred when paginationArea was
commented out, breaking React's hooks rules.
**Problem:**
- createCardProPagination() internally called useIsMobile() hook
- When paginationArea was disabled, the hook was not called
- This violated React's rule that hooks must be called in the same order on every render
**Solution:**
- Refactor createCardProPagination to accept isMobile as a parameter
- Move useIsMobile() hook calls to component level
- Ensure consistent hook call order regardless of pagination usage
**Changes:**
- Update createCardProPagination function to accept isMobile parameter
- Add useIsMobile hook calls to all table components
- Pass isMobile parameter to createCardProPagination in all usage locations
**Files modified:**
- web/src/helpers/utils.js
- web/src/components/table/channels/index.jsx
- web/src/components/table/redemptions/index.jsx
- web/src/components/table/usage-logs/index.jsx
- web/src/components/table/tokens/index.jsx
- web/src/components/table/users/index.jsx
- web/src/components/table/mj-logs/index.jsx
- web/src/components/table/task-logs/index.jsx
Fixes critical runtime error and ensures stable pagination behavior across all table components.
CaIon