README.md

Zarus Smart Mirror

A custom version of a Smart Mirror

Quick Start

Prerequisites

• Raspberry Pi con Raspberry Pi OS (Debian-based)
• Python 3.9+
• Conexión a internet

1. Install system dependencies

# Tkinter (interfaz gráfica), fuentes y dependencias para OpenCV con soporte de ventanas
sudo apt-get update
sudo apt-get install -y python3-tk python3-venv fonts-roboto libgtk2.0-dev pkg-config

2. Clone the repository

git clone https://github.com/<tu-usuario>/smart_mirror.git
cd smart_mirror

3. Create virtual environment and install dependencies

python3 -m venv .venv
source .venv/bin/activate
# Asegúrate de NO tener instalada la versión headless de OpenCV:
pip uninstall -y opencv-python-headless
# Instala la versión normal de OpenCV (con soporte de ventanas):
pip install opencv-python
pip install -r requirements.txt

4. Configure

Copy the configuration template and edit it with your credentials:

cp configuration_template.json configuration.json
nano configuration.json

Fill in:

• mqtt: Your MQTT broker address, user, and password
• homeAssistant: Your Home Assistant URL and long-lived access token
• openWeather: Your coordinates and OpenWeatherMap API key

5. Run

Important: This is a GUI app (tkinter). It must run from a desktop session with a display. It will NOT work over SSH unless you forward the display (see below).

Option A: Run directly on the Pi's desktop

Open a terminal on the Pi (physically or via VNC) and run:

cd ~/smart_mirror
source .venv/bin/activate
python3 smartmirror.py

Option B: Run over SSH (testing only)

# From the Pi's SSH session, set DISPLAY to the local screen:
export DISPLAY=:0
cd ~/smart_mirror
source .venv/bin/activate
python3 smartmirror.py

6. Auto-start on boot (recommended)

The mirror should start automatically when the Pi boots into the desktop.

Enable auto-login to desktop

sudo raspi-config
# Navigate to: System Options → Boot / Auto Login → Desktop Autologin

Install autostart entry

mkdir -p ~/.config/autostart
cp services/smartmirror.desktop ~/.config/autostart/
# Edit the Exec path in the .desktop file if your install path differs from ~/smart_mirror

The mirror will launch fullscreen after the desktop loads on every reboot.

USB Camera RTSP Stream (optional)

If you have a USB camera connected and want to expose it as an RTSP stream (for Frigate or the Smart Mirror face detection), use MediaMTX via Docker.

1. Install Docker

curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER
newgrp docker

# Verify:
docker run hello-world
docker compose version

2. Create the MediaMTX stack

mkdir -p ~/mediamtx
cd ~/mediamtx

Create docker-compose.yml:

# docker-compose.yml
version: "3.8"
services:
  mediamtx:
    image: bluenviron/mediamtx:latest
    container_name: mediamtx
    restart: unless-stopped
    network_mode: host
    volumes:
      - ./mediamtx.yml:/mediamtx.yml

Create mediamtx.yml:

# mediamtx.yml
rtmp: no

paths:
  live:
    source: publisher

3. Start MediaMTX

cd ~/mediamtx
docker compose up -d

# Verify it started correctly:
docker logs -f mediamtx

4. Grant camera access and start the stream

Make sure your user has access to the video device:

ls -l /dev/video0
sudo usermod -aG video $USER
# Log out and back in (or run: newgrp video)

Install ffmpeg if not present:

sudo apt-get install -y ffmpeg

Start publishing the USB camera to MediaMTX:

ffmpeg -f v4l2 -s 640x480 -i /dev/video0 \
  -c:v h264_v4l2m2m -b:v 2M -r 15 -g 30 -an \
  -f rtsp rtsp://localhost:8554/live

The stream will be available at:

• RTSP: rtsp://<PI_IP>:8554/live

5. Use in Frigate

cameras:
  smartmirror_usb:
    ffmpeg:
      inputs:
        - path: rtsp://<PI_IP>:8554/live
          roles:
            - detect

6. Use in Smart Mirror (face detection via stream)

FaceDetectorApp(source="rtsp://localhost:8554/live")

---

Helper services (optional)

These systemd services run background scripts for TV control and temperature monitoring:

# Edit the paths in the .service files to match your install location and username
sudo cp services/checkTv.service /etc/systemd/system/
sudo cp services/tempChecker.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable checkTv.service tempChecker.service
sudo systemctl start checkTv.service tempChecker.service

Auto-start the camera stream (recommended)

To ensure the USB camera RTSP stream is always available after boot, use the provided systemd service:

# Edit the service file if your user is not 'pi':
nano services/usbcam_stream.service
# (Change User=pi to your username if needed)

sudo cp services/usbcam_stream.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable usbcam_stream.service
sudo systemctl start usbcam_stream.service

# Check status:
sudo systemctl status usbcam_stream.service

This will automatically start the ffmpeg stream to MediaMTX on every boot, so the camera is always available for Frigate and the Smart Mirror.

Python dependencies

Package	Purpose
requests	HTTP requests (weather, Home Assistant)
paho-mqtt	MQTT communication
pytz	Timezone handling
ephem	Astronomical calculations (moon, sun)
psutil	System resource monitoring
Pillow	Image processing (PIL)
opencv-python	Camera and face detection (con soporte de ventanas)
RPi.GPIO	Raspberry Pi GPIO pins

Nota importante sobre OpenCV: Si ves errores como "The function is not implemented. Rebuild the library with Windows, GTK+ 2.x or Cocoa support...", asegúrate de:

• Tener instaladas las dependencias del sistema: libgtk2.0-dev pkg-config
• Usar la versión normal de OpenCV (opencv-python), no la versión headless.
• Desinstalar opencv-python-headless si está presente.

Note: tkinter is required but is a system package — installed via python3-tk in step 1.

Image Example

---

Implementations

• Datetime
• Forecast
• Events
• Moon Phase
• Sunset / Sunrise
• Picture Slide
• Face recognition
• Home Assistant Connection
• Mqtt Connection

License

Zarus Smart Mirror is an open source code. All files are licenced under Creative Commons Reconocimiento-NoComercial-CompartirIgual 4.0 Internacional