This post is about my setup of the popular MagicMirror2 application. I show the steps needed to set it up on a Raspberry Pi Zero W and connect it to a TV set. As a bonus, I share my thoughts on the Chuvash localization work.
MagicMirror2 is a DIY project and an open-source application, voted to number one of the best Raspberry Pi Projects. In essence, it shows information of your choice (weather, calendar, news) on a screen that is embedded in a mirror.
I was introduced to MagicMirror2 by my colleague, who uses it in another way: rather than having it in a mirror, he has it on his smart tv. The customisability is the beauty of the whole DIY and the Raspberry Pi.
I also decided to use it on my tv, through the built-in web browser. My raspberry pi zero w required some special steps due to its processor architecture. Fortunately, I found a guide for MagicMirror2 and Raspberry Pi Zero W which I used as a starting point.
After trying it, I found that I needed those steps for my server-only magic:
Every time I work with software I try to think: can I help to translate/localize it to Chuvash. Chuvash is a “little” language, it is only spoken by 1.5 million people. Almost in all cases, it is hard to even to register the Chuvash as such (like in Windows, or macOS), and it is even harder to localize applications, because they tend to have thousands or sometimes millions of strings to translate (and maintain!).
With MagicMirror I experienced how work from the past can help today, how small pieces can become connected parts of a bigger picture.
That was nice! I changed language to “cv” in the config/config.js and the most of the interface turned into Chuvash. (cv – is the iso code for Chuvash language). What a feeling of joy!
Why that excitement? Well, almost everything in the Chuvash IT (UX, localization, keyboard layouts, speech recognition, machine translation etc) is driven and sponsored by volunteers and a community. That’s why it is a special joy to see pieces come together, even though it is a small DIY project.
The MagicMirror2 itself does a minimal set of strings to translate. I translated it and submitted a Pull Request, which has been already merged to the develop branch.
So first of all, why should we test? The best answer is actually: Life is to short for manual testing (it was actually the slogan at the Google London Test Automation Conference 2007.
To be able to run nodeunit tests on server we have to install node.js. Then we have to install nodeunit with npm globally:
npm install nodeunit -g
Then create a test file: TestCustomerModel.js:
Btw, here is my solution:
Next, run the test:
Allright, something is wrong:
Aha, we forgot the space. Let’s fix it:
And the test result is much better 😀
Nodeunit in web browser
nodeunit can be run in a web browser as well. I’ll test the same CustomerModel in this lab. Very simple stuff so far 😀
I know, node.js has been present for a while. But I actually had no time to try it. I was surprised that now it is very straight forward to start with node.js. Allright, everything begins with Hello world. Eventhough it is available for allmost all combinations of operating systems and servers, the easiest way to test it was actually Ubuntu. To install just run:
sudo apt-get install nodejs
Then make a new directory and create the hello.js: