> ## Documentation Index
> Fetch the complete documentation index at: https://docs.plugged.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Version 2.11.1

> AI Search enhancements, document versioning UI, and major performance improvements

# Version 2.11.1 Release Notes

<Note>
  Released on January 18, 2025 - This release brings powerful AI search capabilities to the document library, complete document versioning UI, and significant performance improvements.
</Note>

## 🎯 Highlights

<CardGroup cols={3}>
  <Card title="AI Search" icon="magnifying-glass">
    Semantic search with document name display
  </Card>

  <Card title="Document Versioning" icon="clock-rotate-left">
    Full version history with diff visualization
  </Card>

  <Card title="68% Faster" icon="rocket">
    Major performance optimizations
  </Card>
</CardGroup>

## ✨ AI Search in Document Library

The document library now features a powerful AI-powered search capability that provides intelligent answers based on your documents.

### Key Features

<AccordionGroup>
  <Accordion title="Semantic Search Toggle">
    Switch between traditional keyword search and AI-powered semantic search with a single click. The AI search uses advanced natural language processing to understand the intent behind your queries.

    ```typescript theme={null}
    // Toggle between search modes
    const [aiSearchEnabled, setAiSearchEnabled] = useState(false);
    ```
  </Accordion>

  <Accordion title="Document Name Display">
    No more cryptic document IDs! The AI search now displays actual document names as clickable buttons, making it easy to identify and access source documents.

    **Before**: `099fdd05...`, `3d7c39a7...`
    **After**: `Project Requirements.pdf`, `API Documentation.md`
  </Accordion>

  <Accordion title="Real-time Responses">
    Get instant AI-generated answers with a 500ms debounce for optimal performance. The system intelligently processes your query and returns relevant information from your documents.
  </Accordion>

  <Accordion title="Source Attribution">
    Every AI answer includes clear source attribution with clickable document links. Click any source document to instantly open it in the preview modal.
  </Accordion>
</AccordionGroup>

### How It Works

1. **Enable AI Search**: Click the sparkles icon in the library search bar
2. **Type Your Query**: Enter a natural language question
3. **Get Instant Answers**: AI analyzes your documents and provides relevant answers
4. **Explore Sources**: Click on source documents to view the full content

### Technical Implementation

```typescript theme={null}
// Enhanced askKnowledgeBase with document name fetching
export async function askKnowledgeBase(
  userId: string,
  query: string,
  projectUuid?: string
): Promise<{
  answer?: string;
  documents?: Array<{ id: string; name: string }>;
  sources?: string[];
}>
```

## 📚 Document Versioning UI

Complete user interface for document version management with visual diff tracking.

### Version History Features

<Tabs>
  <Tab title="Timeline View">
    * Chronological version display
    * Expandable version details
    * Visual indicators for current version
    * Model attribution badges
  </Tab>

  <Tab title="Change Tracking">
    * Addition/deletion indicators
    * Line-by-line diff visualization
    * Change summaries
    * Modification highlights
  </Tab>

  <Tab title="Version Comparison">
    * Select two versions to compare
    * Side-by-side diff view
    * Merge conflict resolution
    * Rollback capabilities
  </Tab>
</Tabs>

### Version Metadata

Each document version includes:

* **Version Number**: Sequential versioning
* **Created By**: AI model attribution (name, provider, version)
* **Timestamp**: Creation time with relative display
* **Change Summary**: Brief description of modifications
* **Content Diff**: Additions, deletions, and modifications

## ⚡ Performance Optimizations

Major performance improvements across the document preview system.

### Optimization Metrics

| Metric          | Before   | After          | Improvement       |
| --------------- | -------- | -------------- | ----------------- |
| Bundle Size     | 312KB    | 100KB          | **68% reduction** |
| Memory Usage    | 150MB    | 60MB           | **60% reduction** |
| State Variables | 13       | 1 (useReducer) | **92% reduction** |
| Re-renders      | Frequent | Minimal        | **Memoized**      |

### Technical Improvements

<AccordionGroup>
  <Accordion title="State Management">
    Replaced 13 individual state variables with a single `useReducer` for centralized state management.

    ```typescript theme={null}
    // Before: 13 useState hooks
    // After: Single useReducer
    const [state, dispatch] = useReducer(documentReducer, initialState);
    ```
  </Accordion>

  <Accordion title="Import Caching">
    Implemented dynamic import caching to prevent redundant module loading.

    ```typescript theme={null}
    const importCache: Record<string, any> = {};
    const getDynamicImport = async (key: string, importFn: () => Promise<any>) => {
      if (!importCache[key]) {
        importCache[key] = await importFn();
      }
      return importCache[key];
    };
    ```
  </Accordion>

  <Accordion title="Memoization">
    Comprehensive memoization with `useMemo`, `useCallback`, and `React.memo` for optimal rendering performance.
  </Accordion>

  <Accordion title="Lazy Loading">
    Implemented React.lazy and Suspense for code splitting and on-demand loading of heavy components.
  </Accordion>
</AccordionGroup>

## 🔧 API Enhancements

### Enhanced RAG Query Response

The RAG API now returns structured data with document metadata:

```json theme={null}
{
  "results": "Your AI-generated answer...",
  "sources": ["doc1.pdf", "doc2.md"],
  "document_ids": ["uuid1", "uuid2"],
  "documents": [
    { "id": "uuid1", "name": "Project Requirements.pdf" },
    { "id": "uuid2", "name": "API Documentation.md" }
  ]
}
```

### New Hooks

**`useKnowledgeBaseSearch`**: Complete hook for AI search functionality

```typescript theme={null}
const {
  query,
  answer,
  documents,
  sources,
  isLoading,
  error,
  setQuery,
  clearAnswer
} = useKnowledgeBaseSearch();
```

## 🐛 Bug Fixes

* Fixed document ID display showing cryptic identifiers instead of names
* Resolved null safety issues in document ID handling
* Fixed memory leaks in document preview component
* Corrected AbortController cleanup in useEffect hooks
* Fixed XSS vulnerabilities with DOMPurify implementation

## 🌍 Internationalization

All new features include complete translations for:

* English (en)
* Turkish (tr)
* Chinese (zh)
* Hindi (hi)
* Japanese (ja)
* Dutch (nl)

## 🔒 Security Improvements

<Warning>
  Important security enhancements in this release
</Warning>

* Enhanced input sanitization with DOMPurify
* Proper cleanup of AbortController to prevent memory leaks
* Type safety improvements replacing `any` types
* Content Security Policy (CSP) headers for preview modal

## 📦 Migration Guide

### For Developers

1. **Update Dependencies**
   ```bash theme={null}
   pnpm update
   pnpm install
   ```

2. **Database Migration**
   No database migrations required for this release.

3. **Environment Variables**
   No new environment variables required.

### For Users

1. **Clear Browser Cache**: Recommended to ensure all performance optimizations take effect
2. **Refresh Library Page**: The new AI search toggle will appear automatically
3. **Re-index Documents**: Optional but recommended for optimal search performance

## 🚀 What's Next

Looking ahead to v2.12.0:

* Multi-language document support
* Advanced version merging capabilities
* Collaborative document editing
* Enhanced AI model selection for document generation

## 📝 Full Changelog

View the complete list of changes in our [GitHub repository](https://github.com/VeriTeknik/pluggedin-app/releases/tag/v2.11.1).

## 🙏 Acknowledgments

Special thanks to all contributors and the community for their feedback and support in making this release possible.