A Retrieval-Augmented Generation (RAG) assistant for interactive document-based Question Answering.
Upload PDFs, DOCX, TXT, Markdown, or JSON files, and interact with them via Streamlit UI, CLI, or FastAPI API.

---

🚀 Features

• 📂 Multi-format ingestion: PDF, DOCX, TXT, MD, JSON
• ✂️ Smart text splitting with overlapping chunks
• 🧠 Vector embeddings using SentenceTransformers
• 🗂 ChromaDB persistence for long-term storage
• 🤖 LLM providers: Groq & OpenAI
• 🖥 Multiple interfaces: Streamlit UI, CLI, FastAPI
• 📝 Config-driven via config.yaml
• 🔎 Source citation for transparency

---

⚡ Quick Start (5 Minutes)

1. Clone and Install

git clone https://github.com/ak-rahul/RAG-Assistant.git
cd RAG-Assistant
pip install -r requirements.txt

2. Set API Key

echo "GROQ_API_KEY=your_key_here" > .env

💡 Get free key at: https://console.groq.com/

3. Run the App

python cli.py web

4. Try the Sample

• We included data/sample.txt to get started
• Ask: "What is RAG?" or "List the benefits"
• Upload your own documents and explore!

📚 More examples in examples/

---

📸 Screenshots

Upload and process documents

Get answers with cited sources

---

📂 Project Structure

rag-assistant/
│
├── app.py                 # Streamlit UI
├── cli.py                 # CLI entrypoint
├── config.yaml            # Config file
├── requirements.txt       # Dependencies
├── scripts/               # Helper scripts
│   ├── rag.sh 
│   └── rag.bat
├── src/
│   ├── config.py
│   ├── logger.py
│   ├── server.py          # FastAPI app
│   ├── pipeline/
│   │   └── rag_pipeline.py
│   ├── db/
│   │   └── chroma_handler.py
│   ├── ingestion/
│   │   └── ingest.py
│   └── utils/
│       ├── file_loader.py
│       └── text_splitter.py
│
├── data/                  # Uploaded docs
├── logs/                  # Logs
└── README.md

---

⚙️ Configuration

config.yaml controls everything:

data:
  source_dir: "./data"
  allowed_ext: [".pdf", ".docx", ".txt", ".md", ".json"]

vector_store:
  persist_directory: "./.chroma"
  embedding_model: "sentence-transformers/all-MiniLM-L6-v2"
  top_k: 4

ingestion:
  chunk_size: 1200
  chunk_overlap: 150

llm:
  provider: groq
  model: llama-3.1-8b-instant 
  temperature: 0.2
  max_tokens: 512

🔑 Environment overrides via .env:

GROQ_API_KEY=...
OPENAI_API_KEY=...

---

📋 Prerequisites

• Python 3.9 or higher [attached_file:1][attached_file:2]
• Groq API key (free tier available at console.groq.com)
• 2GB disk space for vector embeddings
• (Optional) OpenAI API key for alternative LLM

---

⚙️ Setup

1. Clone the Repository

git clone https://github.com/ak-rahul/rag-assistant.git
cd rag-assistant

2. Create Virtual Environment

python -m venv .venv
source .venv/bin/activate   # (Linux/Mac)
.venv\Scripts\activate      # (Windows)

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment

Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key
OPENAI_API_KEY=your_openai_api_key   # optional

---

🖥 Usage

Web UI

Launch the Streamlit UI:

python cli.py web

• Upload documents in the sidebar

• Ask questions in the chat interface

• Inspect DB stats or clear DB

CLI

Ingest Documents

python cli.py ingest

Query

python cli.py query "What is Kali Linux?"

Run API Server

python cli.py serve

Show DB Stats

python cli.py stats

Clear DB

python cli.py clear

---

🧩 Architecture

flowchart TD
    A[User Uploads Files] --> B[File Loader]
    B --> C[Text Splitter]
    C --> D[ChromaDB Vector Store]
    D --> E[Retriever]
    E --> F[LLM via Groq/OpenAI]
    F --> G[Answer + Sources]

Loading

---

🧾 Example Workflow

• Upload a PDF in the Streamlit sidebar
• The file is chunked and ingested into ChromaDB
• Ask a question like:
  • "What is covered in Chapter 2 of the Kali Linux PDF?"
• The system retrieves relevant chunks → sends them to the LLM → returns an answer with sources

---

📊 Vector Store Management

• All documents are persisted in ChromaDB inside ./.chroma
• You can check stats (total docs, embeddings, metadata)
• Use Clear DB to reset your database

---

❓ Troubleshooting

GROQ_API_KEY not set : Create .env file with your API key

echo "GROQ_API_KEY=gsk_your_key" > .env

ChromaDB errors

python cli.py clear # Clear database
python cli.py ingest # Re-ingest documents

No documents found

• Ensure files are in ./data folder
• Run python cli.py ingest

Slow responses

• Reduce top_k in config.yaml (4 → 2)

Need help? Open an issue

---

🛠 Tech Stack

• Python 3.9+
• Streamlit – Web UI
• LangChain – RAG pipeline
• ChromaDB – Vector store
• Groq LLM – LLM provider
• OpenAI (optional) – Alternative LLM provider

---

📝 License

This project is licensed under the MIT License.