Rating Service

A web-server (with HTTP APIs) that handles all tasks required for a rating service.

---

Steps to run:

1. Install Node.js version 8.9.1
2. Install MongoDB Community Server edition version 3.3.5.
3. On command line type mongo and hit enter. (Will run Mongo on its default port 27017)
4. On command line type redis-server and hit enter. (Will run Redis on its default port 6379)
5. Go to root directory of rating-service project.
6. On command line type npm install and hit enter.
7. On command line type npm start and hit enter.

• Yay! 🎉 our server is up and running on default port 3000.
• Go through below docs, to understand the Entities and supported API's.
• Optional: You can also run test cases by npm test.

---

Architecture

Assumptions

• We have assumed that the response for getting a rating of any product would be count of all the ratings (i.e onestar, twostar .... fivestar).
• This is a similar example of any other rating service such as in google play etc.
• No auth service is developed, it is assumed that user is logged in.
• To know which user is hitting the API we are adding userId for the post request in ProductRating API.
• Every user can rate a unique product only once.
• Every user has rented each and every product.

Explanation

• As per the, rating service it is should be read-heavy than compared to writes. At peak times, if a product a trending the writes can be heavy.
• We will focus on retrieval of ratings. Hence, as it is read heavy. We are using Redis for caching the product ratings into the redisStore.
• The cache strategy can be multiple ways like cache aside, write through, read through. But we have chose, CRON jobs to minimize writes and reads, but we traded consistency for the period of time between CRON job. We achieve eventual consistency as CRON jobs will update the redis in the near future.
• Also among all the NOSQL DB's we have chosen MongoDB document data store, because it provides atomicity at the document level & is strongly consistent (providing CP of the CAP theorem).

---

Entities:

• User, Product, Ratings are the 3 collections formed. We have separated out Ratings Schema by which the document size of product and user wont exceed the max limit 16MB if number of ratings increase.
• The productId and UserId are String Type Because in JS (as most of our web clients would be in JS) the Number cannot go beyond 9007199254740992 ie. 2^53. So we don't want to limit the number of records in our table by mere JS language's data type limitation! That's why all id's are choosen to be of String data type.

---

Schema

• Used https://github.com/arb/celebrate for schema validation.

• Ratings Schema

  userId: String. Field is required.
  productId: String. Field is required.
  rating: String, Field is required. Enum data  [
              Constants.RATING.TYPE.ONESTAR,
              Constants.RATING.TYPE.TWOSTAR,
              Constants.RATING.TYPE.FOURSTAR,
              Constants.RATING.TYPE.THREESTAR,
              Constants.RATING.TYPE.FIVESTAR
          ]
  

---

API

API Documentation link:

https://documenter.getpostman.com/view/4946631/SWE57z5z?version=latest

User API

1. Add User API

• API: localhost:3000/user

• HTTP method: POST

• Adds new user in the DB

  Example using cURL
  
  curl --location --request POST "{{url}}/user" \
  --header "Content-Type: application/json" \
  --data "{
  \"userId\": \"abdc\",
  \"userName\": \"Piyush Nalawade\"
  }"  
  
  

  Success response
  Status code: 200 OK
  
  {
      "success": true,
      "message": "Success"
  }   
  

  Success response
  Status code: 409 Conflict
  
  {
      "success": false
  }   
  

2. Add Product API

• API: localhost:3000/product

• HTTP method: POST

• Adds new Product into the DB

  Example using cURL
  
  curl --location --request POST "http://127.0.0.1:3000/product" \
  --header "Content-Type: application/json" \
  --data "{
  \"productId\": \"abc\",
  \"productName\": \"Sofa\"
  }"
  
  

  Success response
  Status code: 200 OK
  
  {
      "success": true,
      "message": "Success"
  }   
  

  Success response
  Status code: 409 Conflict
  
  {
      "success": false
  }   
  

3. Adds ProductRating API

• API: localhost:3000/product/:productId/ratings

• HTTP method: POST

• Adds new Product rating into the DB

  Example using cURL
  
  curl --location --request POST "{{url}}/product/2/ratings" \
  --header "Content-Type: application/json" \
  --data "{
  \"userId\": \"1\",
  \"rating\": \"FIVESTAR\"
  }"
  
  

  Success response
  Status code: 200 OK
  
  {
      "success": true,
      "message": "Success"
  }   
  

  Success response
  Status code: 409 Conflict
  
  {
      "success": false
  }   
  

4. Get ProductRating API

• API: localhost:3000/product/:productId/ratings

• HTTP method: GET

• Gets Product rating of a product 99% from Redis, just uses MongoDb when keys not present into the redis.

  Example using cURL
  
  curl --location --request GET "{{url}}/product/1/ratings"
  
  

  Success response
  Status code: 200 OK
  
  {
      "success": true,
      "message": "Success",
      "rating": {
          "fiveStar": 9,
          "fourStar": 0,
          "threeStar": 0,
          "twoStar": 0,
          "oneStar": 1,
          "averageRating": 4.6
      }
  }