Skip to content

QUICKSTART_ENHANCED.md

Latest commit

 

History

History
300 lines (221 loc) · 5.82 KB

QUICKSTART_ENHANCED.md

File metadata and controls

300 lines (221 loc) · 5.82 KB

🚀 Quick Start Guide - Enhanced Artistry AI

Prerequisites

  • Python 3.10+
  • Node.js 18+
  • Docker & Docker Compose (optional)
  • 8GB+ RAM (16GB recommended)
  • CUDA GPU (optional, for faster inference)

Option 1: Docker Setup (Recommended)

1. Start Backend Services

cd artistry-backend
docker-compose up --build

This starts all 6 services:

2. Start Frontend

cd frontend
npm install
npm run dev

Access: http://localhost:5173/enhanced-workflow

Option 2: Manual Setup

Backend Services

Each service runs independently:

1. Detect Service (Port 8001)

cd artistry-backend/detect
pip install -r requirements.txt
uvicorn app.main:app --port 8001 --reload

2. Segment Service (Port 8002)

cd artistry-backend/segment
pip install -r requirements.txt
uvicorn app.main:app --port 8002 --reload

3. Advise Service (Port 8003)

cd artistry-backend/advise
pip install -r requirements.txt
uvicorn app.main:app --port 8003 --reload

Note: First run will download LLaVA model (~13GB). This may take time.

4. Generate Service (Port 8004)

cd artistry-backend/generate
pip install -r requirements.txt
uvicorn app.main:app --port 8004 --reload

Note: First run will download Stable Diffusion models (~5GB).

5. Commerce Service (Port 8005)

cd artistry-backend/commerce
pip install -r requirements.txt
uvicorn app.main:app --port 8005 --reload

6. Gateway Service (Port 8000)

cd artistry-backend/gateway
pip install -r requirements.txt
uvicorn app.main:app --port 8000 --reload

Frontend

cd frontend
npm install
npm run dev

Testing the System

Test Individual Services

1. Test Detection

curl -X POST http://localhost:8001/detect/ \
  -H "Content-Type: application/json" \
  -d '{"image_b64": "YOUR_BASE64_IMAGE"}'

2. Test Commerce

curl http://localhost:8005/commerce/products/bed

3. Test Health

curl http://localhost:8000/health
curl http://localhost:8001/health
curl http://localhost:8002/health
curl http://localhost:8003/health
curl http://localhost:8004/health
curl http://localhost:8005/health

Test Complete Workflow

  1. Visit: http://localhost:5173/enhanced-workflow
  2. Upload a room image
  3. Select budget (Low/Medium/High)
  4. Enter design preferences
  5. Select items to replace
  6. Click "Generate Design"
  7. View results + shopping recommendations

Environment Variables

Backend (.env in artistry-backend/)

# MongoDB (optional)
MONGO_URI=mongodb://localhost:27017

# Service URLs (for Docker)
DETECT_URL=http://detect:8001
SEGMENT_URL=http://segment:8002
ADVISE_URL=http://advise:8003
GENERATE_URL=http://generate:8004
COMMERCE_URL=http://commerce:8005

# Service URLs (for local development)
# DETECT_URL=http://localhost:8001
# SEGMENT_URL=http://localhost:8002
# ADVISE_URL=http://localhost:8003
# GENERATE_URL=http://localhost:8004
# COMMERCE_URL=http://localhost:8005

Frontend (.env in frontend/)

VITE_API_BASE=http://localhost:8000

Troubleshooting

Issue: Services won't start

Solution:

# Check if ports are in use
netstat -ano | findstr :8000
netstat -ano | findstr :8001

# Kill conflicting processes or change ports

Issue: Model download fails

Solution:

# Check internet connection
# Ensure sufficient disk space (20GB+ free)
# Try manual download:
cd artistry-backend/advise/app
# Download LLaVA model manually from HuggingFace

Issue: Out of memory

Solution:

  • Use Docker with memory limits
  • Reduce batch sizes in generation
  • Close other applications
  • Use CPU-only mode (slower but less memory)

Issue: Frontend can't connect to backend

Solution:

# Check CORS settings in backend services
# Verify VITE_API_BASE in .env
# Check if gateway is running on port 8000

Performance Tips

For CPU-only Systems

  1. Reduce inference steps:

    • Detection: Use YOLOv8n (nano) - already set
    • Generation: Reduce to 20 steps (default: 30)

  2. Use smaller models:

    • Already using optimized variants (YOLOv8n, MobileSAM)

  3. Enable attention slicing:

    • Already enabled in Generate service

For GPU Systems

  1. Enable CUDA:

    • Models automatically detect and use GPU
    • Verify: torch.cuda.is_available() returns True

  2. Use FP16:

    • Already enabled for GPU inference

  3. Increase batch sizes (if multi-image processing)

What's Next?

Try Different Workflows

  1. Budget Comparison:

    • Same room, different budgets
    • See material quality differences

  2. Style Exploration:

    • "Modern minimalist"
    • "Cozy bohemian"
    • "Luxury hotel"
    • "Scandinavian bright"

  3. Partial Redesign:

    • Replace only specific items
    • Keep furniture, change colors

  4. Shopping Integration:

    • Browse matched products
    • Compare vendors
    • Track prices

Next Features to Build

  1. Save/Load designs
  2. User accounts
  3. Design history
  4. Social sharing
  5. Real product API integration
  6. AR preview (mobile)

API Documentation

Full API documentation available at:

Support

Issues? Open a GitHub issue with:

  • Error logs
  • System specs (CPU/GPU, RAM)
  • Steps to reproduce

Questions? Check ENHANCED_ARCHITECTURE.md for detailed documentation.

Happy Designing! 🎨✨