Add optional thinking parameter to the call_ollama function and pass it from the request. Also enhance timezone handling in prompt generation to support configurable timezone preferences.
- Add privacy mode to hide IP and user preferences from AI requests
- Add model thinking levels (low/medium/high) for context analysis depth
- Add PWA support with service worker, manifest, and app icons
- Add SettingsPanel for user preferences (theme, background, language)
- Add i18n translations for en/zh/ja/ko/de/fr
- Add Pinia store for centralized settings management
- Update backend to support user preferences and thinking levels
- Update config to use absolute API URLs
- Add GeoLite2-City.mmdb database for IP lookup
- Create geoip.py module for IP location services
- Extract client IP from requests and log location info
- Pass location context to LLM prompts for enhanced responses
- Deleted `windowDelineation.test.ts` as it is no longer needed.
- Removed `index.ts` and `tokenizer.ts` from the tokenization module due to refactoring.
- Eliminated unused types related to authentication, code citation, context provider API, and core functionalities.
- Cleaned up the `status.ts` file by removing the `StatusKind` type definition.
Remove sensitive environment files from repository tracking and add comprehensive Python ignore patterns including virtual environments, cache files, and environment variables. Also clean up staged __pycache__ binary files.
- Configure Vite dev server to proxy /v1 requests to backend
- Change backend port from 8000 to 8001
- Simplify API URLs to relative paths instead of absolute localhost URLs
- Increase debounce delay from 500ms to 1000ms for better stability
- Update README documentation to reflect all changes
- Implement SHA-256 image hashing to cache OCR results and avoid re-processing identical images
- Add 100MB file size limit for image uploads with user-friendly error messages
- Clear ghost suggestions when uploading new images to prevent interference
- Optimize size limit calculation in copilot plugin to include OCR context
- Remove debug logging from production code
- Add image processing optimization plan document
BREAKING CHANGE: Image upload size limit is now enforced at 100MB (previously unlimited)
- Redesign theme toggle with improved visual effects including gradient overlays, enhanced shadows, and smoother cubic-bezier transitions
- Update toggle dimensions and icon positioning for better visual balance
- Add SVG filter effects for sun/moon icons in dark mode
- Replace English UI text with Chinese localization in MilkdownEditor
- Refactor copilotPlugin by removing unused decoration functions and improving ghost mark text node handling
- Implemented a new composable `useTheme` for managing theme state.
- Added functions to read and write theme preference to local storage.
- Applied theme styles to the DOM based on user preference.
- Introduced a toggle function to switch between light and dark themes.
refactor: enhance copilot plugin functionality
- Improved request handling with sequence and document versioning.
- Refactored ghost text handling to improve clarity and efficiency.
- Updated markdown insertion logic to handle parsed content more robustly.
- Enhanced error handling and logging for better debugging.
style: update global styles for light and dark themes
- Defined CSS variables for light and dark themes to streamline styling.
- Improved overall styling consistency and responsiveness.
- Added transitions for smoother theme changes and interactions.
- Replace HTML comment OCR metadata with inline `<OCR:...>` tags
- Implement serializer-based markdown conversion for prefix/suffix content
- Add extractTextFromOCR utility function for text extraction
- Enable Table, Diagram, and ListCheck features in MilkdownEditor
- Add periodic debug logging for document state analysis
- Add VITE_OCR_URL environment variable with fallback URL construction
- Define IMAGE_NODE_TYPES constant to support 'image', 'image-block', and 'imageBlock' node types
- Add helper functions for safer image attribute access (getImageSrc, isImageNodeWithSrc, getImageLabel)
- Improve OCR error handling with HTTP status checking and error details
- Wrap OCR context in HTML comments to prevent prompt injection issues
- Update MilkdownEditor to use centralized OCR_URL configuration
Add image button with dropdown menu for uploading local images or inserting from URL.
Integrate VLM-based OCR to extract text context from images and include in AI suggestions.
Implement document size limits to disable AI when exceeding threshold.
Refactor copilot plugin with per-view runtime state and OCR context injection.
Add OCR cache utility for managing image metadata.
Add code splitting configuration for optimized bundle size.
- Rewrite prompt builder with comprehensive rules for seamless text completion
- Implement Markdown parsing for ghost text with proper mark handling
- Update LLM parameters (temperature 0.7, repeat_penalty, think mode)
- Add CSS styles for formatted ghost text elements
- Add planning documentation for Copilot prompt system analysis
- Implement Markdown parsing for ghost text using Milkdown parser
- Add support for both block and inline content in ghost text
- Refactor prompt system with comprehensive rules and examples
- Adjust LLM parameters: increase temperature to 0.7, add repeat_penalty
- Add CSS styles for formatted ghost text (bold, italic, code, links)
- Add documentation for Copilot prompt system and ghost text rendering
- Replace overlay-based GhostTextOverlay.vue with ProseMirror Mark system
- Add AI toggle button with enable/disable functionality
- Implement new copilotPlugin.ts using copilotGhostMark for inline suggestions
- Fix cursor position offset in prompt.py by moving first suffix char to prefix
- Improve API error handling with abort signal support and debug logging
- Update model configuration from gpt-oss:120b to gpt-oss:20b
- Add button tooltips and improve editor styling
- Remove deprecated inlineSuggestionPlugin.ts
- Update README with new architecture diagram and feature documentation
- Add markdown file upload functionality with upload button
- Fix error handling to throw errors instead of silently returning empty strings
- Fix memory leak by cleaning up debounceTimer in onUnmounted
- Update debounce timing from 150ms to 500ms for stability
- Enhance UI with floating action buttons and extensive style refinements
- Hide toolbar, menu, and line number elements for cleaner interface
Add GhostTextOverlay component for displaying AI-powered completion
suggestions while typing. Includes accept and dismiss actions triggered
by API calls to /v1/completions endpoint with prefix/suffix context.
This commit refactors the LLM integration to use Ollama's native Python client instead of OpenAI-compatible API, while fixing critical template syntax errors and improving project structure.
Key changes:
- Replace openai package with ollama package in backend requirements
- Rewrite llm.py to use ollama.AsyncClient for direct Ollama API calls
- Update main.py to use non-streaming Ollama responses with thinking extraction
- Fix template syntax error in MilkdownEditor.vue (GhostTextOverlay component tags)
- Fix string截取错误 by using slice() instead of substring()
- Add src/utils/api.js and src/utils/config.js for shared configuration
- Add CORS middleware to FastAPI backend
- Update prompt.py with clearer instructions for continuation generation
- Add comprehensive README.md documentation
BREAKING CHANGE: Environment variables OLLAMA_BASE_URL changed to OLLAMA_HOST (remove /v1/ suffix)
- Implemented window delineation tests for indentation-based tokenization.
- Created tokenizer module with various tokenization strategies including TTokenizer and ApproximateTokenizer.
- Added type definitions for authentication parameters and code citation notifications.
- Introduced context provider API for extensions to supply additional context items to Copilot.
- Defined core types and schemas for position and range.
- Established status types for agent status management in IDEs.
Replace the existing contenteditable-based markdown editor with a full-featured WYSIWYG editor using @milkdown/crepe. The new implementation provides:
- True WYSIWYG editing experience with instant Markdown syntax rendering
- Slash command menu support for quick formatting
- Code block highlighting and image paste support
- Built-in export to markdown file functionality
Changes include new MilkdownEditor component, updated App.vue integration, theme styling imports, and optimized Vite configuration for the new dependencies.
- Created a new document for using components in Milkdown.
- Added a guide for using plugins in Milkdown, including toggling plugins programmatically and listing official plugins.
- Introduced a recipe for integrating Milkdown with Angular, including installation steps and component creation.
- Added a recipe for using Milkdown with Next.js, detailing installation and component setup.
- Created a guide for integrating Milkdown with NuxtJS, including installation and component creation.
- Added a comprehensive guide for using Milkdown with React, covering both Crepe and core Milkdown usage.
- Introduced a recipe for SolidJS integration with Milkdown, including installation and component creation.
- Added a guide for using Milkdown with Svelte, detailing installation and component setup.
- Created a comprehensive guide for integrating Milkdown with Vue, covering both Crepe and core Milkdown usage.
- Added a recipe for using Milkdown with Vue2, including installation and component creation.
Add initial project structure including:
- .gitignore and VSCode settings
- Vite configuration and package.json with Vue 3 dependencies
- Basic HTML entry point and README update
- Core source files: App.vue, main.js, style.css
- Markdown editor component with plugin system and related types
- Sample HelloWorld component, router, and Pinia store
- Assets and SVG icons
This commit establishes the foundation for the Vue 3 application.
