🔐 PKI Expert System with RAG

🎯 Goals

Develop an expert system for Public Key Infrastructure (PKI) using a text-based knowledge base and Retrieval-Augmented Generation (RAG) to provide accurate, context-aware responses.

---

✨ Features

• 🔍 Retrieval-Augmented Generation (RAG) Leverages vector search via ChromaDB and text generation using open models from Hugging Face or OpenRouter.

• 📚 Automated Knowledge Processing Automatically chunks and indexes documents from the knowledge/ folder for efficient and fast retrieval.

• 💬 Interactive Dialog Interface Supports both command-line and Gradio-based user interfaces for interactive expert consultation.

• 🧩 Modular Architecture Components like the LLM, vector store, and prompt logic are modular and easily replaceable.

---

🗂️ Project Structure

.
├── Dockerfile                     # Docker container setup
├── LICENSE                        # License file
├── README.md                      # Project documentation
├── db/
│   ├── chroma_db/                 # ChromaDB local vector database files
│   ├── docker-compose.yml         # Docker configuration for ChromaDB
│   └── otel-collector-config.yaml # OpenTelemetry collector config (if used for observability)
├── docs/
│   └── architecture.png           # System architecture diagram
├── requirements.txt               # Python dependencies
├── src/
│   ├── __pycache__/               # Python cache files (can be ignored)
│   ├── app.py                     # Main application entry point
│   ├── embedding_storage.py       # Embedding storage & retrieval logic
│   ├── media_to_text.py           # (Optional) Convert media (audio/video) to text for indexing
│   ├── rag.py                     # Core RAG implementation
│   └── knowledge/                 # Directory for source documents (.txt, .pdf, etc.)
└── test/
    ├── demo.ipynb                 # Jupyter notebook for testing & exploration
    └── uploads/                   # Folder for uploaded files (used in demos/tests)

---

🧱 System Architecture

---

⚙️ Prerequisites

• Python 3.10.12+
• OpenRouter API key
• Hugging Face API token
• Docker

---

🚀 Getting Started

1. Update Your System

sudo apt update

2. Build and Run the Docker Vector Database

cd db
docker-compose up --build
cd ..

3. Set Environment Variables

export OPENROUTER_API_KEY=your_openrouter_api_key
export HF_TOKEN=your_huggingface_token

4. Create and Activate a Virtual Environment

python3 -m venv .venv
source .venv/bin/activate

5. Install Dependencies

pip install -r requirements.txt

6. Run the app

python3 src/app.py