PeerPrep: Collaborative Coding Platform

Overview

PeerPrep is a collaborative coding platform where users can be matched based on question difficulty and topic to solve coding questions together in real-time. It follows a microservices architecture with scalable deployment on Google Cloud.

Features

• Real-Time Collaboration: Two users collaborate live on a shared code editor (CodeMirror).
• Matching Service: Matches users based on question difficulty and topic.
• Question Service: Provides a catalog of coding questions.
• User Management: User registration, authentication, and account management.
• Code Execution: Execute Python code directly on the collaboration page (via Pyodide).
• Code Syntax Highlighting: Syntax highlighting in the shared editor.
• Cloud Deployment: Frontend on Google Cloud Run; Backend on Google Kubernetes Engine.
• Scalability: Horizontal Pod Autoscaler (HPA) support for backend services.
• Secure Communication: API Gateway with SSL/TLS via NGINX Ingress.

Architecture

• Frontend: React.js (deployed to Google Cloud Run)

• Backend Microservices:

  • User Service (Express.js)
  • Question Service (FastAPI)
  • Matching Service (Express + Redis PubSub)
  • Collaboration Service (Socket.io, Express, Redis PubSub)

• Databases:

  • MongoDB Atlas (persistent storage)
  • Redis (message queue and in-memory store)

• Inter-service Communication:

  • REST for frontend-backend communication
  • gRPC for synchronous backend-to-backend communication
  • Redis PubSub for asynchronous messaging

• Containerization: Docker

• Orchestration: Kubernetes (GKE)

• CI/CD: GitHub Actions (multi-architecture matrix builds)

Functional Services

• User Service:
  • User Registration and Authentication
  • JWT-based session management
• Question Service:
  • CRUD for coding questions
  • Rich content support (images, hints, code snippets)
• Matching Service:
  • Match users based on topic and difficulty
  • Publish-Subscribe pattern with Redis
• Collaboration Service:
  • Real-time collaborative code editing
  • WebSocket-backed communication with Redis Adapter

Deployment

• Frontend: Google Cloud Run
• Backend: Google Kubernetes Engine (GKE)
• Database: MongoDB Atlas (Managed Database)
• API Gateway: NGINX Ingress with Let's Encrypt SSL
• Horizontal Pod Autoscaler: Services scale based on memory/CPU usage

Tech Stack

• Frontend: React.js, JavaScript, CodeMirror
• Backend:
  • Express.js (User, Collaboration services)
  • FastAPI + Pydantic (Question Service)
  • TypeScript + Socket.io (Collaboration Service)
• Databases:
  • MongoDB Atlas (User, Question services)
  • Redis (Matching and Collaboration services)
• Containerization: Docker
• Orchestration: Kubernetes
• CI/CD: GitHub Actions
• Hosting: Google Cloud Platform

Contributions

• Ho Qi Yuan, Janald: Frontend development, API integration, Collab page code execution
• Arkar Kyaw Aung: Backend architecture, Kubernetes deployment, Matching and Question service implementation
• Lim Xuan Zheng: Backend services (User, Question, Matching), HPA configuration
• Cyrus Krispin Vijikumar: Kubernetes GKE deployment, SSL setup, Redis integration for Collaboration and Matching
• Austin Huang Deyu: Frontend development for Question, Match, and Collab pages, frontend CI/CD

Nice-to-Haves Implemented

• Code Execution: Sandbox Python execution using Pyodide (WebAssembly)
• Syntax Highlighting: Python syntax highlighting with CodeMirror
• Scalability: Auto-scaling backend with Kubernetes HPA
• API Gateway: NGINX Ingress with SSL/TLS

Further Enhancements

• Comprehensive Integration Testing
• Enhanced Observability (Prometheus + Grafana)
• GitOps Deployment (ArgoCD)
• Helm Charts for Kubernetes resource management
• API Rate Limiting and more advanced API Gateway features

---

Repository: GitHub - PeerPrep