Key Changes:
1. **Removal of Scroll Anchoring Widget**
- Removed the scroll anchoring widget import and its related implementation from `blocksuite/affine/block-root/src/common-specs/widgets.ts`
2. **Enhanced React-Lit Integration**
- Added `ReactWebComponent` type export in `packages/frontend/component/src/lit-react/index.ts`
- Refactored text renderer component to use React integration:
- Added React import and created `LitTextRenderer` component using `createReactComponentFromLit`
- Moved the component declaration to a more appropriate location
3. **AI Feature Flag Integration**
- Added feature flag check for AI functionality in `enableAIExtension`
- Only enables AI extensions if the `enable_ai` flag is true
4. **Component Restructuring**
- Moved several components and utilities to dedicated extension files
- Consolidated Lit adapter implementations
- Removed direct widget imports in favor of extension-based approach
- Reorganized editor component structure for better maintainability
5. **File Reorganization**
- Removed `specs/custom/spec-patchers.ts` and distributed its functionality across multiple extension files
- Created new extension files for various features like:
- Attachment embed views
- Doc mode service
- Doc URL handling
- Edgeless clipboard
- Mobile support
- Note configuration
- Various service patches (notification, peek view, quick search, etc.)
6. **Mobile Support Improvements**
- Refactored mobile extension enablement to be more modular
- Moved mobile-specific widget omissions into a dedicated extension
7. **Type System Improvements**
- Added more specific type imports for editors and components
- Enhanced type safety across the codebase
This PR appears to be a significant refactoring effort focused on:
1. Improving code organization through better separation of concerns
2. Enhancing the integration between React and Lit components
3. Adding feature flag support for AI capabilities
4. Making the codebase more maintainable and modular
5. Improving mobile support
6. Strengthening type safety
The changes suggest a move towards a more extension-based architecture, where functionality is more clearly separated into distinct modules rather than being centralized in larger files.
Let me analyze this diff and provide a clear description of the changes.
This PR introduces several significant changes focused on AI integration and code organization in the AFFiNE codebase:
1. **Enhanced SpecBuilder Functionality** (`blocksuite/affine/shared/src/utils/spec/spec-builder.ts`):
- Added method chaining by returning `this` from `extend`, `omit`, and `replace` methods
- Added new utility methods:
- `hasAll(target: ExtensionType[])`: Checks if all specified extensions exist
- `hasOneOf(target: ExtensionType[])`: Checks if at least one specified extension exists
2. **AI Extensions Modularization**:
- Split the large AI-related code into separate modular files under `packages/frontend/core/src/blocksuite/ai/extensions/`:
- `ai-code.ts`: Code block AI integration
- `ai-edgeless-root.ts`: Edgeless mode AI features
- `ai-image.ts`: Image block AI capabilities
- `ai-page-root.ts`: Page root AI integration
- `ai-paragraph.ts`: Paragraph block AI features
- `enable-ai.ts`: Central AI extension enablement logic
3. **Widget Improvements**:
- Enhanced `AffineAIPanelWidget` and `EdgelessCopilotWidget` with proper widget extensions
- Moved widget-specific extensions into their respective files
- Added proper type definitions and component registrations
4. **Code Organization**:
- Simplified exports in `index.ts`
- Better separation of concerns between different AI-related components
- More modular approach to AI feature integration
5. **AI Integration Architecture**:
- Introduced a new `enableAIExtension` function that handles:
- Replacing standard blocks with AI-enhanced versions
- Conditional enabling of AI features based on the current spec configuration
- Extension of AI chat capabilities
The changes primarily focus on improving code organization, maintainability, and the architecture of AI feature integration in the AFFiNE editor. The modularization will make it easier to maintain and extend AI capabilities across different block types and editor modes.
L-Sun