Metadata-Version: 2.4
Name: ragas_testing
Version: 0.1.0
Summary: RAG pipeline using Azure + Chroma
Author: Mannat Sharma
Description-Content-Type: text/markdown
Requires-Dist: langchain
Requires-Dist: langchain-openai
Requires-Dist: langchain-chroma
Dynamic: author
Dynamic: description
Dynamic: description-content-type
Dynamic: requires-dist
Dynamic: summary

# RAGAS Demo 🚀

A lightweight **Retrieval-Augmented Generation (RAG)** library built with **Azure OpenAI** and **ChromaDB**.

This package allows you to query your own vector database using natural language and generate accurate, context-aware responses.

---

## ✨ Features

* 🔍 Semantic search using vector embeddings
* ⚡ Fast retrieval with ChromaDB
* 🤖 Answer generation via Azure OpenAI
* 🔐 User-controlled API keys (no key storage)
* 📦 Easy to install via pip

---

## 📦 Installation

```bash
pip install ragas-demo-mannat
```

---

## 🚀 Quick Start

```python
from ragas_demo import RAG

rag = RAG(
    api_key="YOUR_AZURE_API_KEY",
    endpoint="YOUR_AZURE_ENDPOINT",
    chat_deployment="YOUR_CHAT_DEPLOYMENT",
    embedding_deployment="YOUR_EMBEDDING_DEPLOYMENT",
    db_path="./chroma_db"   # Path to your vector DB
)

response = rag.ask("What is RAG?")
print(response)
```

---

## 🧠 How It Works

1. User question is converted into embeddings
2. Relevant documents are retrieved from ChromaDB
3. Context is passed to Azure OpenAI
4. LLM generates a grounded response

---

## 📁 Requirements

* Python 3.8+
* Azure OpenAI account
* Pre-built ChromaDB vector store

---

## ⚠️ Important Notes

* The embedding model used for queries **must match** the one used to build the ChromaDB
* This package does **not store API keys**
* All requests are processed using user-provided credentials
* Ensure your `chroma_db` directory exists before running

---

## 🔐 Security

* No API keys are stored internally
* No external data is logged
* Users retain full control over their data and credentials

---

## 🛠️ Future Improvements

* CLI support
* Auto vector DB creation
* Remote vector DB integration
* Caching for faster responses

---

## 👨‍💻 Author

Mannat Sharma

---


