There was a beautiful idea of building an app that would show the upcoming weather. The developers wrote a nice backend and a frontend following the latest principles and - to be honest - bells and whistles. However, the developers did not remember to add any information about the infrastructure or even setup instructions in the source code.
Luckily we now have docker compose saving us from installing the tools on our computer, and making sure the app looks (and is) the same in development and in production. All we need is someone to add the few missing files!
- An openweathermap API key.
- Docker and docker compose installed.
- Make a copy of this repository in your own github account (do not fork unless you really want to be public).
- Create a personal repository in github.
- Make changes, commit them, and push them in your own repository.
- Send us the url where to find the code.
- Clone this repository.
- Make changes and commit them.
- Create a .tgz -package including the .git-directory, but excluding the node_modules-directories.
- Send us the archive.
There are a few things you must do to get the app up and running. After that there are a few things you can do to make it better.
-
Get yourself an API key to make queries in the openweathermap.
-
Either run the app locally (using
npm i && npm start) or move to the next step. -
Add Dockerfile's in the frontend and the backend directories to run them virtually on any environment having docker installed. It should work by saying e.g.
docker build -t weatherapp_backend . && docker run --rm -i -p 9000:9000 --name weatherapp_backend -t weatherapp_backend. If it doesn't, remember to check your api key first. -
Add a docker-compose.yml -file connecting the frontend and the backend, enabling running the app in a connected set of containers.
-
The application now only reports the current weather. It should probably report the forecast e.g. a few hours from now. (tip: openweathermap api)
-
The developers are still keen to run the app and its pipeline on their own computers. Share the development files for the container by using volumes, and make sure the containers are started with a command enabling hot reload.
-
There are eslint errors. Sloppy coding it seems. Please help.
-
The app currently reports the weather only for location defined in the backend. Shouldn't it check the browser location and use that as the reference for making a forecast? (tip: geolocation)
-
There are no tests. Where are the tests? (tip: mocha or robot framework) Disclaimer: this is not an easy task. If you really want to try writing robot tests, start by creating a third container that gives expected weather data, and direct the backend queries there by redefining the MAP_ENDPOINT.
-
Set up the weather service in a free cloud hosting service, e.g. AWS or Google Cloud.
-
Write ansible playbooks for installing docker and the app itself.