Retrieval-Augmented Generation: Impact on Information Retrieval Systems for SMEs

This repository contains the source code for the case study exploring the effects of Retrieval-Augmented Generation (RAG) on Information Retrieval Systems tailored for Small and Medium-Sized Enterprises (SMEs) as part of my bachelor thesis.

Getting Started

To begin examining the results yourself, follow these steps to set up the environment:

1. Clone the repository:

   git clone https://github.com/nneubacher/Bachelorarbeit.git

2. Install the necessary dependencies:

   pip install -r requirements.txt

Repository Contents

• data/: Directory containing the Stanford Question Answering Dataset.
• chromaDB/: Directory containing the vector store with the embeddings of SQuAD.
• toChroma.py: Script for embedding and storing data from the data directory into the chromaDB vector store.
• RAG.py: Evaluation script for the RAG-based information retrieval system.
• noRAG.py: Evaluation script for the base model information retrieval system.
• compare.ipynb: Jupyter notebook for interactive analysis and comparison.
• compare.py: Python script for comparing total correct predictions at different thresholds.
• predictions.json: Output from the RAG-based Information Retrieval system.
• predictions_gpt.json: Output from the Information Retrieval system using just the base model.

Usage

After setting up, you can run compare.py or explore compare.ipynb to see how different configurations of the Information Retrieval system perform with respect to your datasets and queries.