BMI Calculator

This is a Node.js based standalone backend application that calculates BMI of a large set people and categorizes them into following :

• underweight
• normal
• overweight
• moderately obese
• everely obese
• very severely obese

Focus of the application is on calculating BMI as efficiently as possible for a large dataset. A tally of people falling into each category is returned to any client requesting this applications endpoint.

Dataset

The Dataset consists of an array of JSON objects with following properties :

{"Gender": "Male", "HeightCm": 171, "WeightKg": 40}

For simplicity we create dataset as follows :

• Create 6 JSON objects as shown above, each object belongs to one of the BMI category mentioned.
• Repeat these 6 objects 20,000 times and store in bigData.json file.
• bigData.json file has 1,20,000 json objects.

Specification

The application takes a positive integer as input(N) from client and reads 'N' bigData.json files parallely to calculate their bmi and categories.

Constraints :

• 1 <= N <= 100
• N is a positive integer

This input can be given in following way :

• https://ak-bmi-app.herokuapp.com/?size=1

Use above URL to access the application, here N is taken input as size=N

According to above constraints if

• N = 1, The application reads 1.2 Lakh json objects from one file
• N = 100, The application reads 12 Million json objects from 100 files parallely

To provide best performance files are read asynchronously.

Output

• A JSON object with total tally of people in each BMI category is given to the client.

{"underweight":20000,"normal":20000,"overweight":20000,"moderately obese":20000,"severely obese":20000,"very severely obese":20000,"Execution Time in Milliseconds":95}

• This output will be shown when size=1 as one bigData.json file will have 20,000 records of people belonging to each of 6 BMI categories. (Total 1,20,000 people records in bigData.json file)
• Notice the last property of the object "Execution Time in Milliseconds". This is total execution time taken by the request in milliseconds, this is shown in object to verify performance.

Implementation

Follwoing tech stack is used to implement the application :

• Node.js
• Express
• Jest
• Travis CI
• Heroku

Node.js handles the core logic of the application. Express provides a thin layer of fundamental web application features, without obscuring Node.js features.

Core Logic :

• According to input given by client(N), we read bigData.json 'N' times asynchronously (to focus on logic, same file is used to read multiples times!)
• A promise is returned by each async file read, hence 'N' promises are tracked in an array.
• On fulfillment of each promise, the statistics JSON object is returned to client along with time of execution.

Deployment

• Heroku is used for deployment of the application, it can be accessed with following URL
  • https://ak-bmi-app.herokuapp.com/?size=1
• Continous Delivery(CD) feature is utilized from Heroku platform by linking it with github repo, each git push will deploy to Heroku automatically.

Testing

• Jest framework is used to write test for Node.js application
• Travis CI is used to run test suite by linking it with github repo, each push will run the test suite and deployed if the build passes.
• Tests are written to check basic functionality and specification of the application.