Inspiration
It's difficult to read a text and remember all the important details in it. Using flashcards is a more effective method of studying, however it takes too long to choose all the important bits and write flashcards about them. We want to make it easier and faster.
What it does
This webapp receives text input and analyze it to find all the important keywords. It then asks you which keywords you want to keep, and generate a printable pdf of fill-in-the-blank style flashcards. No more writing flashcards by hand.
How I built it
We use Google Cloud Platform's Natural Language API to analyse the text. We then use the identified words to generate questions based on our custom Python code. The PDF generation is handled by ReportLab. The server is a simple flask server and the HTML/CSS/JS is mostly vanilla, with a little bit of jQuery.
Challenges I ran into
HTML/CSS is actually a lot of hard work and it's challenging to get the format of PDF flashcards output right. We struggled to get a website to POST the way want it to, and the resulting Javascript required to solve it is ugly. Deploying our application on Google Cloud Compute Engine was a challenge as well, since we have no experience working with it.
What's next for Flashcard Generator
We want to add more kinds of questions, such as who/what questions, but this will require significantly more effort, as it requires deep understanding of the text in question. We probably will have to build our all machine learning models for this. Our keyword analysis still chooses a lot of false positives, thus we want to make it more accurate.
Log in or sign up for Devpost to join the conversation.