This repository is meant to serve as a submission to Dev-Datathon conducted jointly by Lambda, Epoch as a part of MILAN 2022.
We are using pipenv virtual environment in this project. We recommend the same. Use the following command to install all the dependencies related to python
pipenv installYou can install all the npm dependencies by following the given commands
cd client
npm installDo you wonder which LA courses should you take so that you would get a good grade with minimal efforts? Perhaps, not just LA. We often want to know how a professor is? are they linient with grading? how is the work load of the course? Who can answer these things better than students themselves. The naive approach is to ask other students. What if you don't know many people who have taken the course?
That's where we come in!
Students can register on the website and post reviews about the course they have taken. In order to provide the users with statistics such as median grade, we require students to upload the html file from view from courses section in the AIMS portal. We summarize all the reviews submitted by our community as a single review and display it separately. The summary is regenerated after every 15 days using a NLP summarizer.
To ensure that our users aren't penalized for being honest, we don't display the names/roll numbers of users along with the review.
This project uses a react front-end, django backend with sqlite3 database and hugging-face transformers library for text-summarization.
The user is required to login to the website which is done by OAuth, the email used to login is limited to organization, only institute email ids will be accepted while logging in.
The user needs to submit their 'view my courses' HTML page from AIMS portal. A backend API extracts all the relevant information from this page using Beautiful Soup. It then populates the data such as the courses taken by this user and the grade obtained in that course to the database.
The reviews are summarized using our summarizer api which runs on Uvicorn server. The requests are handled using the fastApi. The summarizer is a NLP model which uses t5-small model and tensorflow framework.
For testing purposes, the tester is requested to contact the developer to register themselves on the backend. User signup has not been implemented on frontend. Thus this step is required. Currently, reviews have been added only for the course AI3000: Reinforcement Learning. Please click on that to see reviews.
After logging in, go to http://127.0.0.1:3000/courses/ Ignore the signup page on the frontend.