/visual-semantic-search

Visual semantic search system | Search across products | Text Query --> Visual Retrieval

Primary LanguagePython

Semantic Search

Setup

  • Step 0 : Make Frontend resources ready | (only needed if the json resources under the assets folder of frontend is missing)
    • pip install -r frontend/requirements.txt
    • cd frontend && python process_data_for_frontend.py
  • Step 1: Build and run services
    • docker-compose up --build -d
    • The above command builds all the services required for the app to load (from redis to qdrant as well as frontend and backend services)
  • Step 2:
    • Head over http://server-ip:8003

Technologies

  • Celery for image encoding task management
  • Redis as both a broker and a backend for celery
  • Qdrant as a vector database
  • Fastapi as the backend framework
  • Streamlit for frontend