spliteasy

This is an educational project (submitted as minor project for a Bachelor Of Technology Degree program). It utilises a collection of frontend technologies (web including ReactJS, JS) and backend technologies (Python3 including flask, json, requests) to create a simple application where the following purpose is served -

A group of people can create a list (called listing) and then add details of their mutual payments in it over time. Finally, they can see a simple settlement recommendation for the payments.

Details about Frontend

The frontend is a static website written in Javascript.

The user-interface is written using ReactJSand communicates with the backend over http using fetchAPI. JSON format is used for request & response.

Main frontend author - @prathamagre (for ReactJS)
Secondary frontend author - @SmartyGrasshopper (for backend interaction and data logic)

Details about Backend

The backend is an API written in Python3 programming language using the flask framework. It communicates with the frontend using json over http. Suitable response codes, especially 400, 404 and 200 are provided with the return response.

Storage

Plain json files are used for storage of the data. The decision to use json instead of some database was reached as the expected scope and scale of the project is very very limited. These files are stored with pretty-printing of indentation 2 so that they are human-readable.

Important Before using the app, add the following folder structure at the root of the project -

    data/
      listOfListing.json
      payments.json

Then, add the following content to listOfListings.json -

{
    "lastListingID": 0,
    "listings": [
    ]
}

And the following content to payments.json -

{
  
}

The original data folder was added to .gitignore because it was causing unnecessary tasks while working with the git.

Dependencies

The package dependencies for the backend are listed in sebackend/requirements.txt file. gunicorn is required only if one wants to serve using a production WSGI server on a linux based operating system. Apart from that, the python interpreter is a dependency (obviously).

Serving the backend app

Currently, the backend is designed to be hosted only on localhost.

To start a production server with gunicorn on a linux based os - source startProdServer.sh on the terminal from root folder.

To start a development server with flask on linux based os - source testing/startDevServer.sh on the terminal from root folder.

To start a development server with flask on windows - testing/startDevServer.bat on the command prompt from the root folder.

Known errors and problems

• Don't know why serving with 1 worker on gunicorn hangs with /payment/settlement/ while not with 2 or more workers or the flask-development server. Maybe using requests.get inside settlement route causes issue or maybe any timeouts since the settlement route does bit more processing compared to the other routes. Whether requests.get is the issue can be easily tested by removing the requests and using the actual code of the requested route instead (so there is no requests.get).

Main backend and overall technical-design author - @SmartyGrasshopper