- Modify BaseParams type to make uiMessages optional.
- Refactor message preparation in HomeWindow and ActionUtils to handle modelMessages and uiMessages separately.
- Ensure compatibility with updated message structures in fetchChatCompletion calls.
- Refactor image generation logic to utilize legacy completions for better support of image editing.
- Introduce uiMessages as a required parameter for image generation endpoints.
- Update related types and middleware configurations to accommodate new message structures.
- Adjust ConversationService and OrchestrateService to handle model and UI messages separately.
- Refined performance optimizations for AI services and improved compatibility with various providers.
- Enhanced error handling and stability in the ModernAiProvider class.
- Updated version in package.json to 1.6.0-beta.6.
* refactor: migrate showWorkspace setting from global settings to notes module
- Move showWorkspace state from settings store to notes store for better module cohesion
- Add useShowWorkspace hook in useNotesSettings for consistent access pattern
- Add smooth animation for workspace panel show/hide transition
- Relocate save to notes action to message toolbar for better accessibility
- Add migration v146 to handle state migration for existing users
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* fix: cli lint error
* feat: add open outside menu
* fix: hooks error
* Update useShowWorkspace.ts
* fix: update icon import in NotesSidebarHeader component
- Replaced FilePlus icon with FilePlus2 in the NotesSidebarHeader for consistency with the latest icon set.
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: suyao <sy20010504@gmail.com>
- Added maxRetries option to buildStreamTextParams for improved error handling.
- Implemented custom fetch logic for 'cherryin' provider to include signature generation in requests.
- Removed hardcoded external dependencies and replaced them with dynamic extraction from package.json.
- Cleaned up the configuration for better maintainability and flexibility in managing dependencies.
- Introduced a new translation key 'invalid_model' in English, Japanese, Russian, Chinese (Simplified and Traditional), Greek, Spanish, French, and Portuguese.
- Updated the SelectModelButton component to display an error tag when no valid provider is found, enhancing user feedback.
- Updated getProviderByModel function to improve provider selection.
- Added fallback logic to return a default or cherryin provider if the specified model provider is not found.
- Ensured that the first provider is returned as a last resort, enhancing robustness in provider retrieval.
* feat: integrate file selection and upload functionality in KnowledgeFiles component
- Added useFiles hook to manage file selection.
- Updated handleAddFile to utilize the new file selection logic, allowing multiple file uploads.
- Improved user experience by handling file uploads asynchronously and logging the results.
* feat: enhance file upload interaction in KnowledgeFiles component
- Wrapped Dragger component in a div to allow for custom click handling.
- Prevented default click behavior to improve user experience when adding files.
- Maintained existing file upload functionality while enhancing the UI interaction.
* refactor(KnowledgeFiles): 提取文件处理逻辑到独立函数
将重复的文件上传和处理逻辑提取到独立的processFiles函数中，提高代码复用性和可维护性
---------
Co-authored-by: icarus <eurfelux@gmail.com>
* refactor(dnd): rename idKey to itemKey for clarity
* refactor: key and id type for draggable lists
* chore: update yarn lock
* fix: type error
* refactor: improve getId fallbacks
- Enhanced web search functionality and optimized knowledge base settings for better performance.
- Added automatic image generation for Gemini 2.5 Flash Image model and improved OCR service compatibility on Linux.
- Refactored AI core architecture and improved message retry mechanisms.
- Fixed various UI component issues and improved overall stability.
- Replaced the switch statement with a source mapping object for improved readability and maintainability.
- Enhanced handling of various web search providers by mapping them to their respective sources.
- Simplified the logic for processing web search results in the onChunk method.
- Updated the 'ai' package version from 5.0.26 to 5.0.29 in package.json and yarn.lock.
- Refactored aiCore's provider schemas to introduce new provider types and improve type safety.
- Enhanced the RuntimeExecutor class to streamline model handling and plugin execution for various AI tasks.
- Updated tests to reflect changes in parameter handling and ensure compatibility with new provider configurations.
- Added support for TypeScript incremental builds by enabling the `incremental` option and specifying the `tsBuildInfoFile` in both `tsconfig.node.json` and `tsconfig.web.json`.
- Updated the `typecheck` script in `package.json` to use `concurrently` for running node and web type checks in parallel.
- Added `.tsbuildinfo` to `.gitignore` to prevent build info files from being tracked.
chore: update electron.vite.config.ts and yarn.lock for dependency management
- Added additional external dependencies in electron.vite.config.ts to improve build configuration.
- Updated multiple package versions in yarn.lock, including @napi-rs/wasm-runtime, @oxc-project/runtime, and @rolldown packages to their latest beta versions for better compatibility and performance.
- Adjusted fdir and picomatch versions to ensure alignment with the latest features and fixes.
* refactor(ocr): streamline OCR service registration and improve image preprocessing
- Simplified the registration of the system OCR service by removing the conditional check for Linux.
- Updated SystemOcrService to directly import necessary modules, enhancing clarity.
- Refactored image preprocessing to use a static import of the 'sharp' library for better performance.
* add patch for system-ocr
* add patch
* add patch again
* add patch
* delete setting
* delete i18n
* lint error
* add isLinux
* Revert "delete i18n"
This reverts commit 173e65bbd0.
* Revert "delete setting"
This reverts commit de39c76f83.
* fix: add system check for error message
---------
Co-authored-by: icarus <eurfelux@gmail.com>
* feat(chat/input): allow sending during streaming; remove loading guard
- Inputbar: drop loading check; keep Send clickable; right-click to pause
- Message/MessageMenubar: render menubar even while streaming
- Minor cleanup of unused imports and lints
* feat(chat/menubar): one-click regenerate (remove confirm); add empty-context fallback
- MessageMenubar: remove Popconfirm; click to regenerate
- ApiService: ensure last user message is included if filters empty (avoid messages=[])
- Minor cleanup
* feat(chat/ui): decouple generation from sending; enable actions during streaming
- MessageGroupMenuBar: add “Retry all” (ReloadOutlined) to retry errored/empty/no-block replies; tooltip via i18n
- Context fallback: always include last user message in both messageThunk and ApiService
- i18n: add message.group.retry_failed (zh-CN, en-US, ja-JP, ru-RU, zh-TW)
- Cleanup: remove unused imports; fix lints
* feat(chat/settings): Add confirmation settings for message actions
- Add "Confirm before deleting message" and "Confirm before regenerating message" options to the settings page, allowing users to customize action confirmations.
- Update internationalization files to support multi-language prompt messages.
- Modify the message menu bar to integrate the confirmation logic, enhancing the user experience.
* fix(chat/ui): Apply regeneration confirmation to user messages
Previously, the "Regenerate" action on user messages would trigger immediately, bypassing the `confirmRegenerateMessage` setting. This behavior was inconsistent with the regeneration logic for assistant messages, which correctly showed a confirmation dialog.
This commit wraps the user message's regenerate button in a `Popconfirm` component, conditioned on the `confirmRegenerateMessage` setting. This aligns its behavior with the existing logic for assistant messages.
Now, all regeneration actions are uniformly governed by the user's confirmation preference, creating a more consistent and predictable user experience.
* fix(ui): Only show 'Retry failed' button when errors exist
fix(ui): Conditionally render the 'Retry failed' button
The 'Retry failed messages' button in the message group menu bar was previously always visible, even when no messages had failed.
- The 'Retry failed' button is now conditionally rendered and will only appear if one or more messages in the group meet the failure criteria.
* feat(chat/ui): Add dedicated button to pause message generation
Replaced the undiscoverable right-click-to-pause functionality on the send button with a dedicated, visible "Pause" button. This new button only appears during message generation, making the action intuitive and accessible.
- Removed `onPause` and context menu logic from `SendMessageButton`.
- Added a conditional `CirclePause` button to the `Inputbar` when loading.
* feat(settings/migrate): initialize confirm message flags for legacy users
- Add migration 138 to default confirmDeleteMessage/confirmRegenerateMessage
- No behavior change for fresh installs (uses initialState)
fix(format): correct indentation in MessageMenubar
fix(settings/migrate): fix persistedReducer verison
* fix(ui): resolve React Hook dependency warnings
- Remove unnecessary `topic.prompt` dependencies from Message components
- Remove `loading` dependency from Inputbar useCallback
Resolves ESLint exhaustive-deps warnings and conflicts
---------
Co-authored-by: n2yt584v2t4nh7y <117180266+n2yt584v2t4nh7y@users.noreply.github.com>
