RAG Document Question Answering API

Overview

This project implements a Retrieval-Augmented Generation (RAG) system using FastAPI.

Users can:

• Upload PDF documents
• Ask questions grounded strictly in document content
• Receive context-aware answers using semantic search + LLM

The system performs:

1. PDF text extraction
2. Text chunking
3. Embedding generation (Sentence Transformers)
4. Vector storage using ChromaDB
5. Similarity retrieval
6. Context-grounded response generation using Groq LLM

The entire system is containerized using Docker for reproducible deployment.

Architecture

User → FastAPI → Chunking → Embeddings → ChromaDB → Retrieve → LLM → Response

Tech Stack

• FastAPI
• Uvicorn
• Sentence Transformers
• ChromaDB(Vector Database)
• Groq LLM API
• Docker

Quick Start

Option 1 — Run with Docker

Install Docker

Download and install Docker Desktop from google:

• Verify installation from the terminsl/cmd:

docker --version

• Build the image:

docker build -t rag-api .

• Run the container:

export GROQ_API_KEY="your_api_key_here" docker run -p 8000:8000 -e GROQ_API_KEY=$GROQ_API_KEY rag-api

• Open in browser:

http://localhost:8000/docs

Option 2 — Run Locally (Without Docker)

• Create virtual environment:

python -m venv .venv source .venv/bin/activate

• Install dependencies:

pip install -r requirements.txt

• Run server:

export GROQ_API_KEY="your_api_key_here" uvicorn app:app --reload

• Open:

http://127.0.0.1:8000/docs

---

API Endpoints

POST /documents
Upload and index a PDF.

POST /chat
Ask questions grounded in the uploaded document.

Environment Variable

This project requires:

GROQ_API_KEY

Do NOT hardcode it inside the source code.

Deployment

The application is fully Dockerized and ready for deployment to:

• Google Cloud Run
• AWS ECS
• Azure Container Apps
• Any container-based platform

Key Concepts Demonstrated

• Retrieval-Augmented Generation (RAG)
• Semantic Search
• Vector Databases
• Embedding-based Similarity
• Backend API Design
• Containerization with Docker