📚 RAG Pipeline (ChromaDB + SentenceTransformer + Groq)

A simple Retrieval-Augmented Generation (RAG) system that loads local text files, embeds them using SentenceTransformer, stores them in ChromaDB, and generates answers using Groq’s LLaMA 3 model.

---

🚀 Features

• Local embeddings (no OpenAI required)
• Persistent vector DB via ChromaDB
• Query documents with semantic search
• LLaMA 3 (Groq) for answers
• Minimal, single-file implementation (main.py)

---

📁 Project Structure

main.py
news_articles/              # .txt documents
chroma_persistent_storage/  # auto-generated (ignored)
.env                        # contains GROQ_API_KEY
requirements.txt
.gitignore
venv/

---

🔐 Setup

Create a .env file:

GROQ_API_KEY=your_groq_api_key_here

Install dependencies:

pip install -r requirements.txt

---

▶️ Run

python main.py

Edit the question in main.py:

question = "tell me about databricks"

---

🧠 How it Works

1. Load .txt files
2. Chunk them
3. Generate embeddings with all-MiniLM-L6-v2
4. Store in ChromaDB
5. Retrieve relevant chunks
6. Send context → Groq LLaMA 3 → get answer

---