Certainly, here's a revised README.md file for your "simple-semantic-search" project:

---

Simple Semantic Search

Overview

This repository contains a project focused on implementing semantic search on PDF documents. It uses the Universal Sentence Encoder model to embed textual data and then performs nearest neighbor search to find the most relevant passages in response to a query.

How to Use

1. Clone the repository to your local machine.
2. Install the necessary Python packages using:

pip install -r requirements.txt

3. Run the main.py script to execute the code.

Project Structure

• main.py: This is the main script where the primary functionality of the app is located.
• utils/helper_functions.py: Contains helper functions required by the main script.
• data/questions.txt: Contains sample questions to query the semantic search.

Dependencies

All necessary dependencies are listed in the requirements.txt file. The core packages include:

• numpy
• openai
• requests
• fitz
• fastapi
• scikit-learn
• tensorflow_hub
• reportlab

... and others.

License

MIT License