An interactive 3D knowledge graph visualization tool that analyzes document relationships using AI-powered semantic analysis. Upload multiple documents and discover connections, shared themes, and conceptual relationships visualized in an immersive 3D graph.

• Force-directed graph layout with optimized node spacing
• Hover interactions - nodes scale up and glow when hovered
• Click to focus - smooth camera animation zooms to selected nodes
• Orbit controls - rotate and zoom around focused nodes
• Auto-rotate mode - automated graph rotation for presentations

• OpenAI GPT-4o-mini integration for intelligent document analysis
• Dynamic theme extraction - no predefined categories, analyzes ANY content
• Semantic similarity using OpenAI embeddings (text-embedding-3-small)
• Relationship detection - automatically finds conceptual connections
• AI Chatbot - ask questions about your document relationships

• Optimized storage system - O(1) lookups with bidirectional indexing
• Theme hierarchies - identifies main themes and subthemes
• Connection strength scoring - quantifies relationship strength
• Shared theme analysis - finds documents with common concepts
• Real-time search - filter nodes by document or theme names

• Dark theme with glassmorphism effects
• Responsive sidebar with custom scrollbars
• Live processing progress indicators
• Detailed node information overlays
• Hover tooltips for quick node identification

• Modern web browser (Chrome, Firefox, Edge, Safari)
• OpenAI API key

1. Clone the repository

   git clone https://github.com/YOUR_USERNAME/document-knowledge-graph.git
   cd document-knowledge-graph

2. Configure your API key

   Edit config.js and add your OpenAI API key:

   window.CONFIG = {
     OPENAI_API_KEY: "your-api-key-here",
   };

3. Open the application

   Simply open enhanced-analyzer.html in your web browser!

   No build process required - this is a pure client-side application.

• Click "Choose Files" button
• Select multiple text files (.txt)
• Files are loaded instantly in the browser

• Click "Analyze Documents" button
• AI analyzes themes, definitions, and relationships
• Progress updates show analysis stages

• Hover over nodes to see quick info
• Click nodes to focus and view details
• Drag to rotate around focused node
• Scroll to zoom in/out
• 🏠 Reset button returns to overview

• Use search bar to find specific documents/themes
• Click search results to focus on nodes

• Ask questions about your documents
• Example: "What themes do these documents share?"
• AI has full context of your graph analysis

Optimized data structure for O(1) lookups:

• Bidirectional Maps for themes ↔ documents
• Efficient relationship queries
• Memory-optimized storage

AI-powered semantic analysis:

• OpenAI embeddings integration
• Cosine similarity calculations
• Conceptual connection detection

Structured LLM prompts for:

• Theme extraction (no predefined categories)
• Definition identification
• Relationship analysis

Main orchestrator that:

• Processes documents through AI pipeline
• Builds optimized graph structure
• Generates analysis reports

React + Three.js frontend:

• 3D graph visualization
• Interactive UI components
• Real-time state management

Edit the config object in enhanced-analyzer.html (lines 476-483):

{
  minThemeConfidence: 0.7,           // Minimum confidence for theme extraction
  maxThemesPerDocument: 8,            // Maximum themes per document
  semanticSimilarityThreshold: 0.7,   // Threshold for semantic connections
  connectionStrengthThreshold: 0.4,   // Minimum connection strength
  enableSemanticAnalysis: true,       // Enable embeddings analysis
  enableHierarchicalThemes: true      // Enable theme hierarchies
}

document-knowledge-graph/
├── enhanced-analyzer.html          # Main application (React + Three.js)
├── config.js                       # API key configuration
├── theme-storage-system.js         # Optimized storage data structure
├── semantic-analyzer.js            # AI semantic analysis engine
├── enhanced-prompts.js             # LLM prompt templates
├── document-connection-analyzer.js # Main analysis orchestrator
├── GRAPH_CONTROLS_GUIDE.md        # User control guide
├── README.md                       # This file
└── test files/                     # Sample documents for testing
    ├── beyblades.txt
    ├── climate_change_and_fish_migration_patterns.txt
    ├── Energy_Analytics_and_Grid_Optimization.txt
    ├── Marine_Biodiversity_and_Coral_Reef_Fish.txt
    ├── Solar_Forecasting_Trends.txt
    ├── sustainable_fisheries_and_ocean_conservation.txt
    └── Wind_Integration_Challenges.txt

• Research Analysis - Find connections between academic papers
• Content Organization - Understand relationships in documentation
• Knowledge Discovery - Uncover hidden themes across documents
• Literature Review - Visualize research landscapes
• Documentation Management - Organize technical documentation
• Creative Writing - Track themes and character relationships

Contributions are welcome! Please feel free to submit a Pull Request.

• Adjust LLM prompt for better responses
  • Improve theme extraction accuracy
  • Better handling of edge cases
  • More nuanced relationship detection
• Fix UI
  • Improve responsive design for mobile
  • Better error message styling
  • Enhanced loading states
  • Accessibility improvements (ARIA labels, keyboard navigation)

This project is licensed under the MIT License - see the LICENSE file for details.

• OpenAI - GPT-4o-mini and text-embedding-3-small APIs
• Three.js - 3D graphics library
• React - UI framework

For questions or issues, please open an issue on GitHub.

Built with ❤️ using React, Three.js, and OpenAI