🧠 Mini Retrieval-Augmented Generation (Mini-RAG) Prototype

A hands-on project to build a simple RAG system using LangChain, ChromaDB, and Google Gemini embeddings. Designed for learning and demo purposes, no paid OpenAI API needed! 🚀

---

📋 Overview

This repo shows how to:

• Load and split PDF documents into chunks 📄➡️📚
• Generate text embeddings using Google Gemini (or OpenAI if available) 🧩
• Store and query embeddings with ChromaDB (vector DB) 💾
• Build a lightweight Retrieval-Augmented Generation pipeline for search and question answering 🔍🤖
• Use LangChain as an orchestrator for embeddings and retrieval pipelines ⚙️

---

⚙️ Features

• ✅ PDF ingestion with metadata tracking
• ✅ Text splitting with overlap for context preservation
• ✅ Embedding generation via Gemini API
• ✅ Persistent vector store with ChromaDB
• ✅ Query interface with top-k retrieval
• ✅ GitHub Actions for CI testing (mocked embedding generation) 🧪

---

🚀 Getting Started

Prerequisites

• Python 3.10+
• Google Gemini API key (set GOOGLE_API_KEY in your .env)
• (Optional) OpenAI API key if you want to switch embeddings provider

Installation

git clone https://github.com/yourusername/mini-rag.git
cd mini-rag
python -m venv venv
source venv/bin/activate  # macOS/Linux
# venv\Scripts\activate   # Windows
pip install -r requirements.txt

Usage

1. Place your PDF files in the data/pdfs folder.
2. Add your API key to .env file:

GOOGLE_API_KEY=your_google_gemini_api_key_here

3. Run ingestion to build vector store:

python src/ingest.py

4. Query your RAG system (add your own query interface or notebook).

---

🧪 Testing

Run tests locally with:

pytest

GitHub Actions automatically run tests on push and pull requests.

---

💡 Notes

• Gemini embedding API usage is currently limited by quota, so be mindful of your request volume.
• Embeddings are 768-dimensional vectors by default.
• This is a learning/demo project, not production ready.

---