RAG Framework

RAG-Framework

A Python framework to implement Retrieval-Augmented Generation (RAG) using a Vector-store Database. This framework utilizes the Vector-Store repository to store and retrieve embeddings for effective RAG implementation.

Structure

• main.py - The main RAG pipeline framework, processing input and orchestrating the RAG flow.
• llm.py - A module that interfaces with the Google Gemini API, simulating a Large Language Model (LLM) for generating responses.
• get_embedding.py - A script for generating embeddings from .txt files using the SentenceTransformer model.
• comm-Vectorstore.py - A communication module that uploads embeddings to the Vector Store for storage and retrieval.

Usage

1. Clone the Vector-Store repository and set it up according to the instructions.
2. Configure the main.py and get_embedding.py to use your Vector-store Database.
3. Run main.py to execute the RAG pipeline with any input text.

Requirements

• Python 3.x
• Sentence-Transformers - For generating text embeddings
• nltk - For sentence tokenization
• Google Gemini API credentials for llm.py (or any compatible LLM API for inference)

Installation

pip install -r requirements.txt