The problem NoteWise solves

All students and professionals have at one point found themselves opening and searching through one pdf after another. It is so easy to just miss the one line you were searching for and be frustrated. Do you find yourself frustrated at your folder structure and files just seem to keep disappearing? We've got you!

NoteWise is the one-stop solution to all your filing problems. With Note Wise, you can upload all your files and in the moment of need, just search for what you want to find. It will go through all your data and find you the spots that mention them. Now you can look at all of your necessary points right infront of you.

Challenges we ran into

Building a seemingly easy site might be pretty easy, or so we thought. We ran into a lot of issues which made us think about dropping the thing altogether multiple times throughout. But we powered through.

Issues we faced?

Bad Stack Choice?

Rust is fast, we wanted fast but didn't understand the time constraints. Being a rustacean (obviously going to brand it out ) I wanted to build it with rust thinking this will be faster than the speed of light. But the development took a toll as documentation was scarce. We(I) admitted defeat and caved in to our friendly neighbour python. FastAPI had everything laid down for us so it was easier. We were stupid but still learned.

Can we really depend on APIs?

Well, we wanted to summarise all the data we were getting from the PDFs, but sadly enough as it seemed the summariser API provider couldn't really handle our requests ( about 50 requests at a time !!! ) so we shifted to putting all the data that we got from pdfs into our indexer. Surprisingly it was fast, boy it was fast so we just winged it. I know it's a perfect solution and we will implement PUB/SUB in order to make this summarisation work, obviously at the cost of performance.

TypeScript is still JavaScript under the hood!

We wanted to show our pdfs, and it was working, we depended on

react-pdf

and it was working fine. It loaded the pdf from a CDN, it showed, it rendered it, the whole deal. But when we ran our integration test, it was just throwing bugs at us, left and right. And for the most part, we couldn't really figure out why.
We debugged and debugged and debugged, and sadly javascript didn't really think we deserved to find the bugs.

But as we already had stored the pdf data in the indexer, we decided why does the user have to even go inside? the data is Right THERE!!!. So we made sure users can see and interact with the data in the search dashboard only, win-win for us!