KeyFinder

landing page of KeyFinder

Inspiration

Ever wanted an easy way to review the long article you've just read? KeyFinder was created to help you do just that.

What It Is

KeyFinder pulls keywords from a given text, defines these keywords, and displays them in an easy-to-study manner.

KeyFinder is built in Python Flask. To pull keywords and find definitions, I utilized OpenAI's machine learning model and Google's Knowledge Bank Search API. For the frontend, I used Flickity with HTML and CSS.

Challenges

Originally, I used NLP libraries like rake_nltk to build my program; often I would not get the results I was looking for when pulling keywords, which led me to switch to the more accurate OpenAI API
Navigating the Google API was challenging, as a naive google search would return too many urls for my program to parse through. I solved this by using Google's Knowledge Bank search, which provided a short description for me to use in my program.
Returning data from the backend to display in HTML was a challenge. My strategy to fix this included reading and replacing sections of my HTML template from my python code.
Building the Frontend was a challenge for me as I am not familiar with HTML+CSS. I used Flickity, a CSS library, to help me in this.