ParkPal Backend

This repository contains the backend required to support ParkPal, a mobile application that provides users with real-time parking information in Singapore. It consists of multiple services -- a web api, a data pulling service and a database.

These services can be easily deployed to any server running Docker using Docker swarm and Docker compose after some configuration.

Web api

The web api provides endpoints for users to query information about collated parking information in Singapore.

Data pulling service

The data pulling service consists of a few scripts to pull and normalize carpark information from various providers in Singapore. Upon startup, the Docker container is configured to first retrieve the meta-data of the carparks. Subsequently, cron jobs will be run to pull data about lot availability regularly.

Current support

The backend is currently only able to provide information about public carparks in Singapore from three providers -- URA, LTA and HDB.

The web api currently provides a single end point for users to query nearby carparks based on their current location and a search radius.

Quickstart

The various services can be deployed easily to a server that runs Docker. Simply copy the folders docker, main, scripts along with requirements.txt, docker-compose.yml and docker-compose.staging.yml into your server. This can be done manually or through a deployment pipeline. Afterwards, you can choose to use docker-compose or docker swarm to automatically spin up the respective containers for the various services.

Note that if you choose to use docker swarm, you would have to build the respective Docker images first and use docker-compose.staging.yml as it does not support build contexts in the compose file. If you want your database to be persistent, you can set up a Docker volume in the compose file. More information can be found on Docker's official documentation.

As our services rely on certain third party applications and apis, there a few secrets that have to be injected into the Docker containers through environment variables. Make sure that you have these environment variables set in your server:

• URA_API_ACCESS_KEY - can be obtained through URA's map site
• LTA_API_ACCESS_KEY - can be obtained through LTA's DataMall
• GOOGLE_GEOCODING_API_KEY - your own api key for Google's geocoding services

Google's geocoding services will only be used if the coordinates are missing from a particular carpark's data.

Here are optional variables that can be set based on your preferences:

• CRON_FREQUENCY - the frequency at which to update lot availability (in minutes 0 - 59)
• QUERY_LIMIT - the maximum number of carparks to be stored from each provider.

Technologies

• Python
• Flask
• Docker

Testing

We are currently in the midst of creating more extensive tests.

FAQs

To be updated