A simplified implementation of an AI-powered wine blending assistant built on Val Town.

---

The Terroir Harmony AI Blend Designer is a custom AI application designed for boutique winemakers. It leverages AI to propose wine blend compositions, predict sensory profiles, and assess market suitability.

• Blend Composition Generator: Input desired wine characteristics and get AI-generated blend ratios
• Sensory Profile Prediction: Receive detailed tasting notes and aroma descriptors for proposed blends
• Market Suitability Scoring: Get predicted consumer appeal scores
• Chemical Analysis Integration: Store and analyze vineyard block data
• Historical Blend Tracking: Maintain records of successful blends

This Val Town implementation uses:

• Backend: Hono.js API framework
• Database: SQLite for data persistence
• AI: OpenAI GPT-4 for blend generation and sensory prediction
• Frontend: React with TailwindCSS
• Storage: Val Town Blob storage for file uploads

├── backend/
│   ├── index.ts              # Main API entry point
│   ├── database/
│   │   ├── migrations.ts     # Database schema
│   │   └── queries.ts        # Database operations
│   └── routes/
│       ├── blends.ts         # Blend generation endpoints
│       ├── data.ts           # Data management endpoints
│       └── static.ts         # Static file serving
├── frontend/
│   ├── index.html            # Main HTML template
│   ├── index.tsx             # React app entry point
│   └── components/
│       ├── App.tsx           # Main app component
│       ├── BlendGenerator.tsx
│       ├── DataUpload.tsx
│       └── BlendHistory.tsx
├── shared/
│   └── types.ts              # Shared TypeScript types
└── data/
    └── sample_data.json      # Sample wine data

🚀 Live Application: The Terroir Harmony AI Blend Designer is now running at: https://spooky--ed5633146df111f09a7d0224a6c84d84.web.val.run

1. Explore Vineyard Blocks: View the pre-loaded sample vineyard blocks (4 blocks with different varietals)
2. Generate Your First Blend:
   • Go to the "Blend Generator" tab
   • Select a target wine style (e.g., "Bordeaux-style Red Blend")
   • Choose 2-3 vineyard blocks
   • Click "Generate Blend" to get AI recommendations
3. Review Blend History: See all generated blends and add feedback on actual results
4. Manage Data: Upload your own chemical analysis data or add new vineyard blocks

The application comes pre-loaded with:

• 4 sample vineyard blocks (Cabernet Sauvignon, Merlot, Cabernet Franc, Petit Verdot)
• Chemical analysis data for each block
• Ready-to-use blend generation capabilities

This Val Town implementation successfully delivers:

✅ Complete Backend API - Fully functional REST API with all endpoints
✅ AI Blend Generation - GPT-4 powered wine blend recommendations
✅ Database Integration - SQLite with comprehensive wine data schema
✅ React Frontend - Modern, responsive user interface
✅ Data Management - File upload and processing capabilities
✅ Sample Data - Pre-loaded vineyard blocks for immediate testing
✅ Blend History - Track and evaluate blend performance
✅ Market Analysis - AI-powered market positioning insights

🍷 Generate Wine Blends: Select vineyard blocks and get AI-optimized blend ratios
📊 View Sensory Predictions: Detailed aroma, taste, and mouthfeel forecasts
💰 Market Intelligence: Consumer appeal scores and pricing recommendations
📈 Performance Tracking: Record actual results vs. predictions
🍇 Vineyard Management: Add and manage vineyard block data
📁 Data Upload: Process chemical analysis files

• GET / - Main application interface
• POST /api/blends/generate - Generate new blend compositions
• GET /api/blends - List historical blends
• POST /api/data/upload - Upload chemical analysis data
• GET /api/data/vineyard-blocks - Get vineyard block data

• Input: Target wine style + available vineyard blocks + preferences
• Output: Optimized blend ratios + predicted sensory profile + market analysis
• AI Model: GPT-4 with specialized wine knowledge prompts

• Vineyard Blocks: Store chemical analysis data for each vineyard block
• Upload System: Process CSV files with chemical analysis data
• Sample Data: Pre-loaded with realistic vineyard data for immediate testing

• Track Results: Record actual tasting results vs. AI predictions
• Performance Metrics: Rate market performance and sensory accuracy
• Continuous Learning: Feedback helps improve future recommendations

• Target Market Analysis: AI predicts consumer appeal and positioning
• Price Point Recommendations: Suggests appropriate pricing strategy
• Consumer Profiling: Identifies ideal customer segments

• Backend: Hono.js, SQLite, OpenAI GPT-4
• Frontend: React, TailwindCSS
• Platform: Val Town serverless environment
• Database: SQLite with comprehensive wine data schema
• AI Integration: OpenAI API for blend generation and sensory prediction