SouL Search

Introducing SouLSearch, the ultimate document search engine! It's not just about finding files; it's about finding what you need faster and smarter.

Comment

SouLSearch

Inspiration

Tired of sifting through endless search results in Windows or relying on "Spotlight" in Mac to find files? We've been there too, and we wanted a better way – to search for file contents, not just file names. Imagine searching for something like "Photo of the day when I wore a red jacket" or "1960 War pictures" and getting instant, accurate results. That's the inspiration behind SouLSearch.

What it does

SouLSearch is a powerful document search engine that enables you to find files based on their content. Whether it's text, images, audio, videos, pdf, or docs, SouLSearch can locate your files quickly and effortlessly. We've harnessed the capabilities of Elastic Search, Python, and Fast API on the backend, Angular, and Iconic native for the frontend, and incorporated machine learning models for image captions(Vision Transformer GPT v2), audio-video converters, PDF readers for document extraction, and text vector embeddings ((MpNet v2) to make your searches lightning-fast and precise. Say goodbye to cumbersome folder navigation and hello to instant, intelligent results.

How we built it

We built SouLSearch by combining cutting-edge technologies and components. Here's a brief overview of our tech stack:

  • Backend: Elastic Search, Python and Fast API
  • Frontend: Angular, Iconic Native
  • Machine Learning Models: For generating image captions and enhancing search accuracy
  • Audio-Video Converters: To handle multimedia content
  • PDF Readers: For extracting content from PDF documents
  • Text Vector Embeddings: To enhance search precision
  • AngularJS: Powering our frontend user interface

Challenges we ran into

Throughout the development of SouLSearch, we encountered several challenges. Some of the key obstacles included:

  • Integrating multiple technologies seamlessly
  • Optimizing search speed and accuracy
  • Handling a variety of document types and formats
  • Implementing a user-friendly interface for a versatile search engine
  • Finding appropriate metrics to judge our ML system
  • Overall performance measurement

Accomplishments that we're proud of

We've achieved several milestones with SouLSearch that we're truly proud of:

  • We now use SouLSearch ourselves, and it feels like we have our personal search engine, akin to Google.
  • Our team worked cohesively to complete this project on time.
  • We've created a powerful and versatile tool that we believe can transform how people search for files.

What we learned

During the development of SouLSearch, we gained valuable insights into:

  • Integrating diverse technologies
  • Leveraging machine learning for enhanced search capabilities
  • Building a user-friendly and efficient frontend interface

What's next for Team SouL

Our journey with SouLSearch is far from over. Here's what's on the horizon:

  • Explainability - We want to highlight searched reason in file. And explain why that result occurred.
  • Product Expansion: We aim to take SouLSearch from a project to a product that everyone can use and benefit from.
  • Productionization: We plan to refine and enhance the system to make it production-ready.
  • Cloud Deployment: Our goal is to deploy SouLSearch on the cloud, making it accessible for a wide range of applications beyond just personal use.

Stay tuned for exciting developments from Team SouL as we continue to revolutionize the way you search for and manage your documents.

Share this project:

Updates

Submission history