notebook/notebook-demo_README.md

TSMPD Search Demo Notebook

This directory contains a working Jupyter notebook that demonstrates semantic product search on the TSMPD-US-Public v1.0 dataset.

What's Inside

• tsmpd_search_demo.ipynb: A step-by-step demo of embedding-based search using faiss and sentence-transformers.

Features

• Load SBERT embeddings from parquet chunks
• Build and query a FAISS index for nearest neighbor search
• Explore how consumers might search across 355k+ small U.S. stores

Setup

To run this notebook, install the following:

pip install faiss-cpu polars sentence-transformers

Licensing

Code in this notebook is licensed under the MIT License.