实现像cursor/github copilot那样的ai edit

[gaoconggit]

选中代码出来一个ai对话框,ai只能编辑选中的代码

[gaoconggit]

@claude 实现这个功能，实现好了记得创建pr

[claude]

Claude finished @gaoconggit's task —— View job • claude/issue-52-20251001-1241 • Create PR ➔

---

AI Edit Feature Implementation Complete! ✨

I've successfully implemented the AI code editing feature similar to Cursor/GitHub Copilot! The feature allows you to select code and have AI modify it based on your instructions.

🎯 Features Implemented

1. Selection Widget

• A floating AI Edit button (⭐) appears when you select code in the editor
• Purple gradient design with smooth animations
• Positioned at the bottom-right of your selection

2. Inline Edit Dialog

• Clean, modern dialog that appears near your selection
• Smart positioning that avoids viewport boundaries
• Dark/light theme support

3. AI Integration

• Uses your configured AI model from the chat settings
• Optimized prompts to ensure AI returns only the modified code
• Streams response and applies changes automatically

4. Keyboard Shortcuts

• Ctrl+K - Open AI edit dialog for selected code
• Enter - Submit edit instruction
• Escape - Close dialog

📝 How to Use

1. Select Code: Highlight any code in the Monaco Editor
2. Open Dialog: Click the purple AI button or press Ctrl+K
3. Enter Instruction: Type what you want (e.g., "添加注释", "重构为async/await", "优化性能")
4. Apply: Press Enter or click "编辑" button
5. Review: The modified code replaces your selection and is automatically highlighted

📂 Files Changed

New Files:

• wwwroot/styles/aiEdit.css - Styling for widget and dialog with theme support
• wwwroot/editor/aiEdit.js - Core AI Edit functionality and Monaco integration

Modified Files:

• wwwroot/index.html - Added HTML structure and script imports

🔧 Technical Details

• Integrates seamlessly with existing Monaco Editor
• Reuses the chat API infrastructure (/v1/chat/completions)
• Temperature set to 0.3 for consistent, deterministic edits
• Smart code extraction and markdown cleanup
• Mobile responsive design

✅ All Tasks Completed

• Analyze existing Monaco Editor integration
• Analyze existing AI chat system
• Design the inline edit feature architecture
• Create inline edit dialog HTML structure
• Create inline edit dialog CSS styling
• Implement AI edit JavaScript module
• Integrate with Monaco Editor selection events
• Connect to AI chat API for code editing
• Test the implementation manually
• Commit and push changes
• Create pull request

---

Job Run • Branch: claude/issue-52-20251001-1241