Running the Full TBVision Stack

This guide shows how to launch both the backend API and the frontend portal simultaneously so you can run the full TBVision experience locally.

Prerequisites

Python & venv – Python 3.11+ is required. Create/activate the venv in the repo root:

Create virtual env:
```
python -m venv venv
```
Activate virtual env: For Unix/Linux/Mac:
```
source venv/bin/activate
```
For Windows:
```
.\venv\Scripts\activate
```
Dependencies – Install backend and model requirements:
```
pip install -r requirements.txt
```
Node.js – The frontend (under frontend/) expects Node 18+. Install deps when required.
Model weights – Ensure weights/xraytb_net.pth exists or point CHECKPOINT_PATH in .env to your checkpoint.
Vector DB – Start Qdrant (default http://localhost:6333). Create the collection manually if you customized it.
- docker-compose up --build
Load Knowledge base
```
python -m scripts.load_knowledge
```

Environment variables – Copy .env.example to .env with overrides such as:

CHECKPOINT_PATH=weights/xraytb_net.pth
ALLOWED_ORIGINS=http://localhost:3000
GEMINI_API_KEY=...
MISTRAL_API_KEY=...

Start both backend and frontend together via the helper script (it launches the backend, then the frontend dev server with hot reload):
```
./run.sh
```
The backend listens on http://127.0.0.1:8000 by default and exposes /api/predict + /api/rag. The frontend default port is 3000.