laundry

GMU Laundry

A site that scrapes http://gmu.esuds.net and creates a pretty graph of the current machine usage data for each hall.

This used to be part of my personalsite project (which is a django site), but I decided to make this it's own website.

Getting Started

• Git clone the repo: git clone https://github.com/thallada/laundry
• Set-up a database. Create a secrets.py file in the laundry/ folder with the following variables set:
  • SECRET_KEY
  • DATABASE_NAME
  • DATABASE_USER
  • DATABASE_PASSWORD
  • DATABASE_HOST
• Modify the ADMINS variable in settings.py so you get emails when the site errors (happens when eSuds changes something on their site).
• Run syncdb: python manage.py syncdb
• Import the GMU Laundry Hall data: python manage.py loaddata laundry.json
• Run the server: python manage.py runserver

Errors

Whoever is in the ADMINS variable in settings.py will get an e-mail whenever an error on the site occurs. This most likely happens when eSuds changes something on their site that breaks the webscraper (which is in laundry.py).

The most often change is that machines are added or removed from halls on the eSuds website. To fix this issue you will need to manually modify the database entry for the hall in the django shell. For example, adding a washer to Whitetop Hall:

python manage.py shell

>>> from laundry_app.models import Hall, LaundryMachine
>>> whitetop = Hall.objects.get(name="Whitetop")
>>> LaundryMachine.objects.create(number=10, type=LaundryMachine.WASHER, hall=whitetop)

The number is from the eSuds site in the chart for the hall (under the "Washer #" column).

Sometimes errors happen when eSuds goes down too. Nothing we can really do about that, but perhaps we can improve our downtime by caching previous results and showing those in the intrim.

Todo

Move away from Django to Javascript. There's no real reason to have a backend for this site. It can just query esuds and make a chart in D3 or something. Django is a lot of overhead for such a small site.

Add weekly usage stats. I was in the process of writing this but never finished. It's a lot more complicated than the current usage stats because a record of historical queries needs to be kept somehow.

Host this on a SRCT server. Right now it's somewhere that I probably only have control over and with a database at MIT that I might loose access to eventually.

Make the site more reliable to eSuds changing, like machines getting added or removed from halls.