StyleSync - AI Fashion Styling Assistant

An AI-powered fashion styling application that transforms fashion trends into personalized outfit recommendations using knowledge graphs and vector databases.

🎯 What It Does

StyleSync generates personalized outfit recommendations by:

1. Analyzing fashion trends using Neo4j knowledge graphs
2. Finding matching clothes using Weaviate semantic search
3. Creating style boards with AI-generated outfit images

🏗️ Database Architecture

Neo4j - Fashion Knowledge Graph

Use Case: Stores and analyzes fashion trend relationships

• Nodes: Trends, Garments, Vibes
• Relationships: CONSISTS_OF (Trend → Garment), HAS_VIBE (Trend → Vibe)
• Query Example: Find all garments and vibes for "90s Revival" trend

MATCH (t:Trend {name: "90s Revival"})-[:CONSISTS_OF]->(g:Garment),
      (t)-[:HAS_VIBE]->(v:Vibe)
RETURN collect(g.name) as garments, collect(v.name) as vibes

Weaviate - Semantic Wardrobe Search

Use Case: Vector-based semantic matching of clothing items

• Schema: ClothingItem class with vectorized descriptions
• Search: Finds clothes matching trend DNA using semantic similarity
• Filtering: Separates tops and bottoms with precise type filtering
• Scoring: Custom matching algorithm based on garments and vibes

🚀 Quick Start

Prerequisites

• Python 3.8+
• Docker (for databases)
• Neo4j and Weaviate instances

Setup

# Clone and install
git clone <repository>
cd styleSync
pip install -r requirements.txt

# Configure environment
cp .env.example .env
# Edit .env with your database credentials and API keys

# Start databases with Docker
docker run -d --name neo4j -p 7474:7474 -p 7687:7687 \
  -e NEO4J_AUTH=neo4j/your_password neo4j:latest

docker run -d --name weaviate -p 8080:8080 \
  -e AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED=true \
  semitechnologies/weaviate:latest

# Seed databases with sample data
python seed_databases.py

# Run application
python run.py

Visit http://127.0.0.1:5000 to start styling!

🔧 Tech Stack

• Backend: Python Flask
• Knowledge Graph: Neo4j (trend relationships)
• Vector Database: Weaviate (semantic clothing search)
• AI: Google Gemini (image generation)
• Frontend: HTML/CSS/JavaScript

📊 How It Works

The AI Workflow

1. Trend DNA Analysis (Neo4j)

   • User selects a fashion trend
   • System queries Neo4j for trend components (garments + vibes)
   • Example: "90s Revival" → [denim jeans, graphic t-shirt] + [grunge, casual]
   • 

2. Semantic Wardrobe Search (Weaviate)

   • Creates search concept from trend DNA
   • Performs vector similarity search for matching clothes
   • Separates and scores tops vs bottoms
   • Returns best matching items from user's wardrobe

3. Style Board Generation

   • Combines matched clothes with trend aesthetics
   • Generates AI outfit image using Gemini
   • Returns complete style recommendation

Sample Data

Neo4j Trends:

• 90s Revival: denim jeans, graphic t-shirt → grunge, casual, streetwear
• Minimalist Chic: crewneck t-shirt, chinos → clean, simple, professional

Weaviate Wardrobe:

• White cotton t-shirt (casual, basic, minimalist)
• Black oversized hoodie (streetwear, casual, cozy)
• Dark wash denim jeans (casual, classic, streetwear)

🛠️ Project Structure

styleSync/
├── app/
│   ├── agent.py          # AI workflow orchestration
│   ├── db_seeder.py      # Database seeding utilities  
│   └── routes.py         # Flask API endpoints
├── templates/index.html  # Frontend interface
├── static/style.css      # Styling
├── run.py               # Application entry point
└── seed_databases.py    # Database initialization

🧪 Testing

# Test database connections
python test_weaviate.py
python -c "from app.agent import generate_style_board; print('✅ Ready!')"

# Run full workflow test
python test_agent.py

📈 Status

✅ Production Ready - Complete MVP with all core features implemented

• Neo4j knowledge graph with trend relationships
• Weaviate semantic search with custom scoring
• AI agent workflow orchestration
• Modern web interface
• Comprehensive error handling

---

StyleSync - Where fashion trends meet personalized AI styling 🎨✨