E-Commerce Voice Search Demo

This project demonstrates a voice search interface for an e-commerce product catalog, using speech-to-text and semantic search technology.

Demo

Video Demonstration

GIF showing basic voice search functionality

Installation

Prerequisites

• Python 3.7 or higher

Standard Installation

# Install dependencies (includes NumPy 1.22.0+ required by Whisper)
pip install -r requirements.txt

If you're experiencing dependency issues:

Run the installation helper:

python install_requirements.py

This interactive script will help you resolve common installation issues, including:

• NumPy version conflicts (Whisper requires NumPy 1.22.0+)
• Network/connectivity problems
• Dependency resolution challenges

Known Issues and Solutions

NumPy Version Error

If you see this error:

ImportError: Numba needs NumPy 1.22 or greater. Got NumPy 1.21.

Fix it by:

1. Running python install_requirements.py, or
2. Manually updating NumPy: pip install numpy>=1.22.0

Connection issues with pip

If you see errors like:

WARNING: Retrying... after connection broken by 'NewConnectionError'

Try these solutions:

1. Use the installation helper script: python install_requirements.py
2. Check your network/proxy settings
3. Run pip with trusted hosts:

   pip install -r requirements.txt --trusted-host pypi.org --trusted-host files.pythonhosted.org
   

4. If you're behind a proxy, set environment variables:

   set HTTP_PROXY=http://proxy.example.com:8080
   set HTTPS_PROXY=http://proxy.example.com:8080
   

Running the Application

Once all dependencies are installed:

python app.py

Then open your browser and navigate to: http://127.0.0.1:5000/

Features

• Voice search for products
• Text-based search
• Speech-to-text transcription
• Semantic matching of queries to products

voice-search-demo