Natural Language Search Interface

For demonstration purposes, embeddings are locally generated. In production, OpenAI embeddings can be used

This project implements a Natural Language Search system using:

• PostgreSQL
• pgvector
• OpenAI LLM
• Streamlit

Features

• Convert English queries to SQL
• Secure SQL validation
• Hybrid semantic + SQL search
• Simple Streamlit UI

Example Queries

• Show employees in Engineering department
• Top 5 expensive products
• Orders handled by Sales employees
• Products similar to laptop

Setup

1. Create PostgreSQL database
2. Run schema.sql and sample_data.sql
3. Install dependencies
4. Run embeddings.py
5. psql -U postgres nl_search_db -f db\schema.sql
6. psql -U postgres nl_search_db -f db\sample_data.sql
7. psql -U postgres nl_search_db (verify tables exist \dt \q)
8. Start Streamlit app

python -m streamlit run app.py