WordMerge (Backend)

Backend for the WordMerge application

WordMerge is a mobile game where users engagingly try to merge together on a word based upon the clues of their previously uttered words. Behind the scenes, the application builds an ever-persistent graph of words along with their semantic relationships and connections - much like SyntaxNet, but with crowd-sourced data.

Running the Application

To run the backend, provided you have Docker and Docker-Compose installed, fill out your environment variables on the .env file (using the .env_sample file as a guideline), simply run docker-compose build and docker-compose up. Also make sure to install application dependencies using npm install!

Tech Stack

WordMerge's backend is powered by the following technologies

• Docker for environment containerization
• Postgres relational database (for storing relational data)
• Neo4J Graph Database (for storing word relationships as well as being able to query into such relationships)
• NodeJS for the backend application logic
  • Express for HTTP-based routes
  • Socket.IO for WebSocket-based routes

Routes

There are 2 types of routes - HTTP-based routes for synchronous request-response based messages and Websocket-based routes for asynchronous publish-subscribe based messages.

• HTTP Routes - The following routes offer the synchronous request-response based functionality for the backend. All sent and received payloads are encoded in JSON.

  1. POST /auth/login
  • Accepts {username:string, password:string}
  • Returns
    • {status:200, message:string, auth_token:string}
    • {status:400, message:string}
  2. POST /auth/signup
  • Accepts {email:string, username:string, password:string, image_url:string}
  • Returns
    • {status:200, message:string}
    • {status:400, message:string}
  3. POST /room/create
  • Accepts {auth_token:string, game_mode:string}
  • Returns
    • {status:200, message:string, room_id:string}
    • {status:400, message:string}
  4. POST /room/join_specific
  • Accepts {auth_token:string, room_id:string}
  • Returns
    • {status:200, message:string}
    • {status:400, message:string}
  5. POST /room/join_random
  • Accepts {auth_token:string, game_mode:string?}
  • Returns
    • {status:200, message:string, room_id:string}
    • {status:400, message:string}
  6. POST /room/leave
  • Accepts {auth_token:string, room_id:string}
  • Returns
    • {status:200, message:string}
    • {status:400, message:string}

• Websocket Routes - The following websocket flow offer the asynchronous publish-subscribe based functionality for the backend. All sent and received payloads are encoded in JSON.

  1. Client-side driven events (sent by the client, received by the server)
  • {type:'user_joined', auth_token:string, room_id:string}
  • {type:'user_left', auth_token:string}
  • {type:'user_typing', auth_token:string}
  • {type:'say_word', auth_token:string, word:string, parent1:string, parent2:string, duration:int, halt:boolean, created_at:timestamp}
  2. Server-side driven events (received by the client, sent by the server)
  • {type:'room_ready', other_user:{username:string, image:string}}
  • {type:'said_word', other_user:{username:string, image:string}, word:string}
  • {type:'user_typed', other_user:{username:string, image:string}}
  • {type:'game_ended'}

Relational Entities

The following represents the entities as stored on our relational database:

• users: user_id:serial, username:varchar, email:varchar, password:varchar, created_at:timestamp
  • primary_key: user_id
• rooms: room_id:varchar, game_mode:string, created_at:timestamp, destroyed_at:timestamp
  • primary_key: room_id
• room_words: room_id:rooms/room_id, user_id:users/user_id, word:varchar, parent1:varchar, parent2:varchar, duration:int, halt:boolean, created_at:varchar
  • primary_key: [room, user, word]
• room_users: room:rooms/room_id, user:users/user_id, entered_at:timestamp, left_at:timestamp
  • primary_key: [room, user]