agdq-tracker

The various bits and pieces that make up my *gdq tracker.

Here's how it works.

the cron job

This is run by cron every minute. It fetches the viewer count from the twitch API, the donation total from the tracker API and scrapes the schedule from the schedule page.

This script isn't pretty but it's been proven to not die very often over the years. It attempts to recover from any errors and write a null in place of the data it would have written, to ensure there's a record in the JSON for every minute no matter what.

The cron job looks something like this:

* * * * *  cd /home/alligator/dev/agdq-tracker/sgdq19 && /home/alligator/.local/bin/pipenv run python get-data.py

the web interface

This is the code for the actual tracker you see. As before the code is real ugly (and ancient in terms of JS coding standards) but it continues to work with only minor tweaks, so it's here to stay I guess.

---

None of this is generic. Well, it's all generic as in it should work for any GDQ, but I have copy all the files and manually go in and edit all of the names and years for each new marathon. One day I'll get around to automating it but hey, it took me 5 years to put it in version control. One step at a time.

other stuff

misc/comparison contains the code for the comparison chart. This was hacked together quickly and it shows. I've fixed it up a couple of times over the years to make it less dog slow but it's still very rough. It took about 30 minutes to smash together and tends to be more popular than the main tracker come marathon time, so there you go.

misc/original-tracker contains the code for the original tracker , which used the google charts API instead of d3.

misc/gdq-announce.py a very simple python script that looks at the json output from the tracker and posts a message to a discord webhook when a new game is coming up. i point it at a gdq specific channel on a discord server so people can always see what's on now and what's up next.