PolyBazar

AI-assisted B2B marketplace for polymer granules and plastic waste trading.

Architecture

flowchart LR
    U[Buyer or Seller] --> FE[Frontend Next.js]
    FE -->|REST| BE[Backend Spring Boot]
    FE -->|WebSocket| BE

    BE --> PG[(PostgreSQL)]
    BE --> MG[(MongoDB)]
    BE --> RD[(Redis)]
    BE -->|HTTP| ML[ML Service FastAPI]

    ML --> VS[(FAISS or Chroma)]

Loading

Repository Layout

polybazar/
├── frontend/          # Next.js app
├── backend/           # Spring Boot API
├── ml-service/        # FastAPI ML service
├── scripts/           # Utility scripts
├── docker-compose.yml
└── .env.example

Prerequisites

• Node.js 18+
• Java 17+
• Python 3.11+
• Docker + Docker Compose (recommended for dependencies)

Environment Setup

cp .env.example .env

Update values in .env as needed. At minimum, configure database credentials and JWT secret for non-demo usage.

Run Locally

Option A: Start Everything with Docker Compose

docker-compose up --build

Expected local endpoints:

• Frontend: http://localhost:3000
• Backend: http://localhost:8080
• ML Service: http://localhost:8000
• Mongo Express: http://localhost:8081
• pgAdmin: http://localhost:5050

Option B: Run Frontend Only

Use this when backend is already running elsewhere.

cd frontend
npm install
NEXT_PUBLIC_API_URL=http://localhost:8080 npm run dev

Frontend runs at http://localhost:3000.

Option C: Run Backend Only

Start required dependencies first (PostgreSQL, MongoDB, Redis), then:

cd backend
./gradlew bootRun

Backend runs at http://localhost:8080.

Important backend env vars:

• POSTGRES_URL, POSTGRES_USER, POSTGRES_PASSWORD
• MONGO_URI
• JWT_SECRET
• ML_SERVICE_URL (optional if ML endpoints are not used)

Option D: Run ML Service Only

cd ml-service
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

ML service runs at http://localhost:8000.

Manual Deployment on Render

This repository does not currently include a committed Render Blueprint file, so use manual service creation.

1. Prepare external services

Create these first:

• PostgreSQL (Render managed)
• Redis (Render managed)
• MongoDB (external provider such as MongoDB Atlas)

Collect connection values for each service.

2. Deploy backend as a Web Service

• Root Directory: backend
• Environment: Docker
• Port: 8080
• Required env vars:
  • SPRING_PROFILES_ACTIVE=prod
  • POSTGRES_URL
  • POSTGRES_USER
  • POSTGRES_PASSWORD
  • MONGO_URI
  • JWT_SECRET
  • ML_SERVICE_URL (if ML service deployed)
  • CORS_ORIGINS (set to frontend Render URL)

3. Deploy ML service as a Web Service

• Root Directory: ml-service
• Environment: Docker
• Port: 8000
• Typical env vars:
  • VECTOR_DB_TYPE
  • EMBEDDING_MODEL
  • OPENAI_API_KEY (if using OpenAI embeddings)

4. Deploy frontend as a Web Service

• Root Directory: frontend
• Environment: Docker
• Port: 3000
• Required env vars:
  • NEXT_PUBLIC_API_URL=https://<your-backend-service>.onrender.com
  • NEXT_PUBLIC_WS_URL=wss://<your-backend-service>.onrender.com/ws

5. Post-deploy checks

• Open frontend URL and verify login/signup pages load.
• Hit backend actuator health endpoint if enabled.
• Verify one end-to-end flow (browse products, view details, call price prediction).

Testing

# Backend
cd backend && ./gradlew test

# ML service
cd ml-service && pytest

# Frontend E2E
cd frontend && npm run cypress:run

Notes

• See project analysis notes in analysis/README.md.
• Contribution guidelines: CONTRIBUTING.md
• License: LICENSE