Extracting schedule information from timeedit

At liu we use timeedit to track our schedules. Recently they updated their interface and improved some parts. It’s now possible to create a prenumeration of a collection of courses exported as csv which can then be imported to other calendar apps. But they also started to obfuscate their urls.

This is the schedule for a single course TGTU49 in the interval “now - 2015-01-17”:


This is for the courses TGTU49 and TATA49 in the same interval:


There seem to be some logic behind the urls, but first, let’s look at how they handle search.

Using firebug we can intercept the json requests when we click the search button at


We find a GET request to


which has has some search results formatted as html:

types=219 means we’re searching for a course and if can change it to types=205 we can search for student groups instead. If we change objects.html to objects.json we even get a json response:

json looks much better and it’s a lot easier to parse, but unfortunately if we instead search for “TGTU” we only get one match but using html we get six. What’s worse, there’s a lot less information given with json. data-id for example seems important as it’s the unique id signifying the course.

As I’m writing this I tried .txt as well and we get a much larger json response which contains data-id, among other things.

Investigating the page source in ta schedule page we find

Which seems to suggest that we can send a request directly to https://se.timeedit.net/web/liu/db1/schema/ri.html with url parameters.

For example


Is a search for TGTU49 (data-id: 363972.219) between 2014-07-03 and 2014-12-31 and


searches for TGTU49 (data-id: 363972.219) and TATA49 (data-id: 363741.219).

sid is an important flag here which specifies what information you’re interested in.

sid               Information given
3                 Date Course Local Type Teacher Group
4                 Date Type Teacher Local
others            Date

There’s a bunch of other parameters which we can simply ignore. This time if we replace .html to .json we get a nice json request:

We can also get information formatted as .txt and .csv. For parsing purposes the best one should be json, as we can use rust’s json deriving, but then we need to create a bunch of structs. .csv is actually very easy to parse and it looks like this:

TimeEdit, Linköpings universitet
"TGTU49, Teknikhistoria, 1, TGTU49 1445-1503 Valla", 2014-07-03 - 2014-12-31 ,
Startdatum, Starttid, Slutdatum, Sluttid, Kurs, Lokal, Undervisningstyp, Lärare, Studentgrupp, Fria grupper, Information till student, Studentförening, URL
2014-11-04, 13:15, 2014-11-04, 15:00, TGTU49, KEY1, Föreläsning, Dick Magnusson, "KA3, EMM3, EL3, PRO2", , , ,

Which is what I did. For a more stable library I should probably switch to json parsing, but there it is.

I made a rust crate which uses the outlined approach to retrieve scheduling information from timeedit. There’s a bunch of things I’m not supporting, but it’s enough for my needs. I made a simple cli and integrated a .schema command into my irc bot.