Inspiration

Having had a test a few days ago, we were distraught to realise that we only had 2 past papers to practice with. Even worse, these past papers contained material we hadn't covered yet, as the previous years had done it later. With this experience fresh in mind, we had the idea to make a question finder app, before deciding to include a note generating function.

What it does

The user can enter a prompt, exercise amount, and grade level. The Perplexity API will be used to find the main topic and subtopics from the prompt, before generating JSON data. This JSON data is converted to Markdown, which is then converted to a PDF. Users can also view and download previously generated PDFs which other users generated.

How we built it

We used Flask as the backend. We used a frontend template online. We used pypandoc to convert the Markdown to PDF with LaTeX formatted equations.

Challenges we ran into

We originally had a different idea, but at around 12:00am we realised that it wasn't feasible, at least not with the time constraints and our skill levels. We originally decided to make a holiday planner, but unfortunately the API was unable to find accurate data.

Accomplishments that we're proud of

We managed to actually generate the PDFs, after a long tug-of-war with trying to format LaTeX properly, we were able to sucessfully do so.

What we learned

LaTeX is very hard to parse. We also learned that we should plan better, as we changed ideas.

What's next for WonderBot

Vector embeddings to search through previously generated PDFs more effectively, as well as better quality PDFs (possibly with images).

Built With

Share this project:

Updates