/ecommerce_semantic_search_app

Primary LanguageJupyter Notebook

Building a Semantic Search Engine

This Semantic Search Engine was built for the purpose of

  • Using Superduperdb to build a more realistic business use case of vector database using MongoDB as the backend
  • Comparing the different between a basic Semantic search engine and a simple Full-Text Search Engine

Here is the infracstructure diagram

image

Basic Components used

  • Backend: MongoDB
  • Embedding: OpenAI
  • Vector search functionality: SuperduperDB
  • Frontend: Streamlit

Here is a Screencast of the App

video