Inspiration

We are or have been polytechnic students, and we tend to have too many documents on our laptops/PCs, which makes it a pain to refer to them. For example, I want to know about a single concept that I learnt in a year-long module. If I don't remember where it is, I would have to sift through a years worth of teaching materials.

What it does

This is why we developed our application, which makes searching details from our school notes much easier. By uploading our notes to a RAG chain, we hope to be able to ask the document questions, which with the assistance of an LLM (Large Language Model), will provide answers to our questions by referencing the documents (notes) we have uploaded. However, the process of uploading notes, will put your notes to a cloud database, allowing you to go off saved notes, and immediately get to question asking. This eliminates the start-up and preprocessing time for RAG processes. We also added helpful features with the info

How we built it

We first developed the idea first, thinking about what we or polytechnic students would want to make their studying much easier. Then we created a simple prototype UI with some features we would like. Then we set out to create a working prototype.

Challenges we ran into

Due to the short period of time, we seriously had trouble putting together our ambitious project together, especially the working prototype, especially since we didn't really have experience doing this before. As a result, our UI is a placeholder. We also originally intended to utilize cloud, but due to a lack of time, we had to stick to a locally-hosted app. The utilization of a cloud database was also not able to be achieved in time.

Accomplishments that we're proud of

We like the base idea we have, and what we have implemented so far.

What we learned

We learnt about how we can utilize AI to solve our problems, especially with RAG utilizations.

What's next for Notes+++

Future development of this project would include completing all planned features for the project, hosting the app securely on a website, and a mobile app version

Credit

https://github.com/slvg01/10d_RAG_OnTheFly This github gave us a base to work off, credit to them for this template that we would not have been able to create in time.

Built With

  • beautifulsoup4
  • faiss
  • figma
  • glob2
  • langchain
  • langsmith
  • openai
  • pypdf2
  • rag
  • streamlit
  • tiktoken
Share this project:

Updates