Swasthya-AI-Placement-GraphQL-Server

GraphQL server for Swasthya AI Placement Assignment

Running Servers on EC2:

http://18.206.209.243:3000/
http://18.206.209.243:4000/graphql

Visit the React Client Repository

Run the MYSQL script dump_2329_22052021.sql on MySQL Workbench
Steps to install and run GraphQL and React Servers:

clone the repository (git clone https://github.com/Rajrahane/Swasthya-AI-Placement-GraphQL-Server.git)
run npm install
change the username,password for GraphQL dbConnection/dbConnection.js
run npm start
view the GraphQL server on http://localhost:4000/graphql
view the React Server on http://localhost:3000/

Approach:

DB Schema:

Create 4 tables: User, Blog, Comment, Friend

User Table: id,first_name,last_name,mobile_number,email_address
Blog Table: id,name
Comment Table: id,message,user_id,blog_id
Friend Table: id,user_id,friend_id
Friend table adds a given friend to the user's friendList.

Assumption: Friends is undirected connection,two entries would be made- for a pair of friends A,B as (A,B) and (B,A).

React UI:
/users Endpoint: User can Register, Search, View Profile
/blogs Endpoint: A new Blog can be created.Search Blog based on ID.
Blog Profile shows blog name and lists the comments.

Summary

Create a DB schema in MYSQL for the tables.
Create a GraphQL Schema to fetch data, create new User and Blog, post a new Comment.
Create getFriends RootQuery in GraphQL to fetch Nth Level Friend of a User
Create a React Frontend for Forms, Search, Profile.

Solution for Nth Level Friend List:
Create a Friends graph and make a BFS recursive call until Nth level.

Consider input as (user_id:int,givenLevel:int) returns list(users)
   Initialize an empty Set of VisitedUsers
   Initialize an empty queue.Push (user_id) into it
   Initialize currentLevel=0
   while(q is not empty and currentLevel lessthan givenLevel){
      init new empty queue q2
      while(q not empty){
        pop each user.
        if(user not visited){
            mark as visited in VisitedUsers
            fetch all friends ids from Friends table in Database
            foreach(friend){
              if(friend not visited){
                push friend to queue q2
              }
            }
         }
      }
      currentLevel++
      push all elements in q2 to q, i.e. q=q2
   }
   if(currentLevel less than givenLevel){ //no nth level depth possible, too few levels
      return empty list
   }
   return to_list(q)

Sample Database Data:
1. 7 users are created
2. 2 blogs are created
3. 1st blog has 2 comments added
4. To mock friend relations, 6 relations are created {1,2,3}, {1,4}, {4,6}, {4,5}

Sample GraphQL Query for getFriends at any level:

{
  getFriends(input:{
    user_id:5
    level:3
  }){
    id
    last_name
  }
}

Expected Output: Users 2 and 3
Explanation: User 5 has 1st level friend- User 4
User 5 has 2nd level friends- User 1,6 via User 4
User 5 has 3rd level friends- User 2,3 via User 1

Sample Query for edge case (level too high or level<=0)

{
  getFriends(input:{
    user_id:5
    level:4
  }){
    id
    last_name
  }
}

Expected Ouput: null list
Explanation: User 5 has friends upto level 3 only

Features of GraphQL Repository

Uses Express NPM Package for Server Setup
Uses GraphQL to create Schema and Query, Mutation operations
Uses Sequelize a Promise based Node.js ORM

Features of ReactJS Repository

Uses ApolloClient to connect to GraphQL Server
Uses React Router for creating Routes Hierarchy
Uses React Bootstrap for Bootstrapping

Improvements:

DFS can be used instead of BFS if some(say 10) and not all users are to be fetched.
Graph DB - Neo4j can be used instead of MySQL, since a friends graph is formed
Pagination can be used to fetch Comments from GraphQL.
GraphQL caching can be used for query caching

Deployment:

GraphQL can be hosted over AWS Lambda + API Gateway

AWS CDK or AWS SAM can be setup for Deployment of Lambda Code from Repository

React JS UI Snaps

EC2 Deployment

Launch EC2 using Amazon AWS console.Open ports 4000(GraphQL),3000(React),22(SSH)
SSH into it. install git `sudo yum install git -y`
install node using nvm
install pm2 Daemon Process Manager. `npm install pm2 -g`
git pull both repositories(React and GraphQL)
Install Docker on EC2
Install docker-compose(linux)
`sudo curl -L "https://github.com/docker/compose/releases/download/1.29.2/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose`
`sudo chmod +x /usr/local/bin/docker-compose`
`docker-compose --version`
Deploy DB using Docker:In /deployment folder: `cd deployment`, `docker-compose up -d` This launches a mysql db and connects a persistent docker volume to it
Deploy GraphQL using pm2:in graphql repository: `pm2 start --name swasthya-ai-graphql-server npm -- start`
Deploy React using pm2:in react repository:change graphql endpoint in src/index.js to http://18.206.209.243:4000/graphql then `pm2 start --name swasthya-ai-react-server npm -- start`

Rajrahane / Swasthya-AI-Placement-GraphQL-Server

Swasthya-AI-Placement-GraphQL-Server

Running Servers on EC2:

Approach:

About

Languages