Хабр Курсы для всех
РЕКЛАМА
Практикум, Хекслет, SkyPro, авторские курсы — собрали всех и попросили скидки. Осталось выбрать!
The linked source code just seems to be the web assets of the front-end.
It connects to trueratr-backend.herokuapp.com. Reading the about there points to the real source code which is hosted here:
github.com/MathieuCliche/Sarcasm_detector
EDIT: Which is weird though since it's created by someone not listed among those involved with this project while their posted source seems to get it's actual results from Mathieu Cliche's sarcasm detector.
EDIT2: Looking through it some more it seems that the open sourced part is just a material design front-end. The back-end seems to be a modification of the above linked sarcasm detector with two additional routes which allow searching on iTunes app names and iTunes app IDs instead of inputting direct strings.
I'm really quite curious as to what exactly the team actually did? Is it just a nice looking front-end and does it still use the algorithm for sarcasm detection of the original source? Or are there real modifications to the back-end beyond linking it with iTunes and will those modifications be open sourced?
В Корнельском университете создали самообучающуся систему по определению сарказма в Сети