🕵️‍♂️ MythInformation Bringing Lore and Order to the Multiverse. An AI-powered "Story Detective" that transforms chaotic wiki articles and scripts into interactive 3D Galactic Knowledge Graphs. Extract characters and relationships with Gemini, predict relationship types with ML, and analyze patterns in Databricks. Try it Here: MythInformation Site IMPORTANT: Backend is hosted on free-tier Render, so it times out after a while. Give a couple of minutes for the backend to start. 🏆 Event 🛠️ Technologies

Layer	Technologies
The Brain
The ML Engine
The Face
The Graph

• User Authentication: Secure JWT-based register/login system with password hashing via bcrypt.
• Persistent Analysis Storage: Save, load, and manage multiple lore analyses per user in PostgreSQL.
• Multiversal Clustering: Characters are automatically grouped into "Solar Systems" based on their source work (e.g., FNAF 1 vs. FNAF 2).
• Star Magnitude: Characters with more connections glow brighter and appear larger using Degree Centrality metrics.
• Bridge Detection: Characters appearing in multiple works are rendered with Multi-Segment Segmented Coronas, visually showing their multiversal history.
• AI Case Files: Click a node to decrypt a character's dossier, including an AI-generated biography and mission history, rendered with rich Markdown support.
• 🤖 ML Relationship Predictor: AI-powered relationship type prediction using Random Forest classifier (87.5% accuracy on 20 relationship types). When creating new connections, the system suggests the most likely relationship type with confidence scoring.
• 📊 Analytics Dashboard: Databricks SQL dashboards showing KPI metrics, character connectivity analysis, relationship type distribution, work comparisons, user activity timelines, and centrality distributions.
• Fluid Physics Engine: Real-time collision detection and spring-forces create a "liquid" movement feel when adding new lore.
• Procedural Starfield: A native 20,000-point 3D background for a cinematic immersive experience.

• Python 3.10+
• Node.js 18+
• Docker & Docker Compose
• Google Gemini API key

1. Clone & Setup Backend:

   cd backend
   python -m venv venv
   venv\Scripts\activate  # Windows
   pip install -r requirements.txt

2. Setup Environment Variables:

   # backend/.env
   GEMINI_API_KEY=your_gemini_key_here
   
   # Optional: Databricks Analytics
   DATABRICKS_HOST=your_workspace_url
   DATABRICKS_HTTP_PATH=your_cluster_path
   DATABRICKS_TOKEN=your_pat_token
   DATABRICKS_CATALOG=lore_graph
   DATABRICKS_SCHEMA=analytics
   DATABASE_URL=postgresql://loreuser:lorepass@localhost:5432/loredb
   SECRET_KEY=your-secret-key-change-in-production

3. Start PostgreSQL Docker Container:

   docker-compose up -d

4. Run Backend Server:

   cd backend
   uvicorn main:app --reload

5. Setup Frontend:

   cd frontend
   npm install
   npm run dev

6. Access Application:

   • Frontend: http://localhost:5173
   • API Docs: http://127.0.0.1:8000/docs

1. While editing: Click "Add Connection" under a character's dossier
2. Select target: Choose another character from the dropdown
3. View prediction: The AI suggests the relationship type with a confidence percentage
4. Modify if needed: Accept the suggestion or choose a different relationship type
5. Confirm: Click "Create Connection" to add it to your graph

1. Go to Databricks Workspace: Access your Databricks SQL Warehouse
2. View SQL Queries: 6 pre-built queries available in the workspace
3. Track Metrics: Monitor character connectivity, analysis trends, and graph growth over timeect Gutenberg by ID
4. View the 3D Graph as characters and relationships materialize
5. Click nodes to view character dossiers
6. Delete systems by clicking the ✕ next to their name

1. Click "💾 Save Analysis" after building your graph
2. Name your analysis and add notes
3. Click "📂 Load Analysis" to browse your saved work
4. Load an analysis to restore it and continue exploring

• id - UUID primary key
• username - Unique username

• lore_analyses - Analysis metadata synced from PostgreSQL (user_id, names, total_characters, total_relationships)
• lore_characters - Character data with graph metrics (analysis_id, name, degree_centrality, work_source, description)
• lore_relationships - Relationship data for analytics (analysis_id, source_id, target_id, relationship_type, work_source)
• email - Unique email address
• hashed_password - Bcrypt hash
• created_at - Timestamp

• id - UUID primary key
• user_id - Foreign key to users
• name - Analysis title
• description - User notes
• nodes - JSON array of graph nodes
• links - JSON array of graph edges
• work_meta - JSON object with system metadata (colors, positions)
• created_at, updated_at - Timestamps

and custom scrollbars

• ✅ NEW: ML Relationship Predictor with 87.5% accuracy on 20 relationship types
• ✅ NEW: Databricks Analytics with Delta Lake table syncing and 6 SQL dashboard queries
• ✅ NEW: Async/Non-blocking Backend - Databricks writes happen in background threads, API never hangs
• ✅ Full 3D force-directed graph visualization with 20k-point starfield
• ✅ Character extraction & relationship mapping via Google Gemini 2.5 Flash
• ✅ Interactive node selection with smooth camera transitions
• ✅ Rich character dossiers (showing ML predictions in action)
• 🌐 Deployment (Backend to Render/Railway, Frontend to Vercel)
• 📱 Mobile responsiveness (lower priority)
• 📊 Advanced analytics queries and visualization customization

Developed Solo in 36 Hours for Hacklytics 2026 - Entertainment Track *Featuring AI-powered lore extraction, graph visualization, and ML relationship prediction.

• ✅ Cyberpunk UI with Tailwind theming
• ✅ Comprehensive error handling and user feedback

• 🎬 Demo video walkthrough
• 🌐 Deployment (Backend to Render/Railway, Frontend to Vercel)
• 📱 Mobile responsiveness (lower priority)
• 📊 Optional: Databricks integration for advanced analytics

Developed Solo in 36 Hours for Hacklytics 2026.