Skip to content

Latest commit

 

History

History
153 lines (101 loc) · 5.48 KB

README.md

File metadata and controls

153 lines (101 loc) · 5.48 KB
Raw

Distributed Online Judge

Tests

Distributed Online Judge is a distributed computing application for testing programs against test cases in programming competitions.

Table of contents:

Final Report

Final Report

Team members

  • Khaled Ismaeel (BS18-SB)
  • Trang Nguyen (BS18-DS-01)
  • Marko Pezer (BS18-SE-01)
  • Kerim Kochekov (BS18-DS-02)

Why yet another online judge?

Hosting a programming competition is a resource-intensive task. In some rounds on Codeforces the judging servers receive an excess of 100,000 solutions to test, each of which runs on around 50 test cases taking a matter of seconds and sometimes minutes to complete. As a result, competitive programming platforms often require high performance servers to accommodate the huge workload.

Obviously, investing in powerful servers is expensive. Furthermore, this approach is error-prone due to its single point of failure. This has already been demonstrated in numerous online round on Codeforces, where the testing servers would be overloaded or DDoS'ed during competitions and the round would eventually be canceled.

This project is an attempt to port the approach of Lichess's computer analysis feature to competitive programming platforms. Lichess employs a volunteer computing paradigm where volunteers perform game analysis on their local machines and report the results back to Lichess servers. In this paradigm the server is not involved in the computationally-intensive task of analyzing game; only in request processing. This has proved to be extremely reliable for Lichess: analysis is almost instant and perpetually available.

Architecture

The application cluster consists of 2 types of network nodes:

  • Server run by the platform administrator.
  • Client run by a single volunteer.

The server hosts the platform interface to contestants along with all units related to solution testing (test cases, checker, validator, interactor, etc). When the server accepts a solution from a user, it forwards this solution along with testing units (which could be cached) to a volunteer. The client performs the testing protocol and reports the results back to the server, which in turn reports the result back to the contestant. All communication channels (contestant-server, server-volunteer) are to be implemented on top of HTTPS.

Structure image

Technology stack

Frontend

  • Pug
  • CSS
  • JavaScript

Backend

  • NodeJS
  • MongoDB

Testing environment (not fully determined)

  • Docker
  • GDB

Testing and documentation

  • Jest
  • Swagger
  • JSDoc

Documentation

Server

Setup

Run locally

To get this project up and running locally on your computer:

  1. Set up a Nodejs development environment.
  2. Once you have node setup, enter the following commands in the root of your clone of this repo: 
    #Go to server directory
cd server

#Install dependencies
npm install

#Start development server 
npm run devstart
  3. Open a browser to http://localhost:3000/ to open the site.

Note: The library uses a default MongoDb database hosted on MongoDB Atlas. You should use a different database for your own code experiments.

Run on Docker
  1. Make sure Docker and Docker Compose installed
  2. Enter the following commands in the root of your clone of this repo: 
    #Go to server directory
cd server

#Run docker compose
docker-compose up
  3. Open a browser to http://localhost/ to open the site.

APIs

Once you have the server run on your machine (locally or using Docker), APIs are documented under /api-docs/ path using Swagger.

API Docs Screenshot

API Test Screenshot

Following is the documentation for our server's frontend.

HTML Verb URL Description
GET /ui/problems Display all available problems
GET /ui/problems/create Show form to make new problem
POST /ui/problems/create Add new problem to database and redirect
GET /ui/problems/:id Show info about one problem
GET /ui/solutions Display all available solutions
GET /ui/solutions/create Show form to make new solution
POST /ui/solutions/create Add new solution to database and redirect
GET /api-docs/ Show API documentation

Client

Run locally

To get this project up and running locally on your computer:

  1. Set up a Nodejs development environment.
  2. Once you have node setup, enter the following commands in the root of your clone of this repo: 
    #Go to server directory
cd client

#Install dependencies
npm install

#Connect to server
npm run devstart

Run on Docker

To be set up

Project

Project progress and planning can be found on GitHub projects page.