Light Assistant

Light Assistant is a lightweight home automation tool focused on easy and intuitive control of lights. The name is a pun on both its simplicity in usage and its main purpose.

The project is provided as-is; it is a project I wrote for my own use at home, but if it can help you out as well, I am glad 😊

My motivation

This project was started out from my own need for a "home automation" system but not finding the available solutions viable for my needs – for various reasons. In brief, my needs (and thus reasons for starting this project) were these:

• It should be simple to set up and use. I found Home Assistant too complicated, and it still could not quite do what I wanted it to (without resorting ugly yaml, or C# code with hard-coded device names).
• It should be able to dim the lights very slowly. Off-the-shelf product can do this (with Zigbee lights) in the "bright" spectrum of light intensities. However, at low light conditions they fail (due to 8 bit limitations in Zigbee).
• Most lights (all the lights I have encountered) have a gamma curve response to light intensities. That is; at low values, there is a lot of light intensity change with each increment, while at high values, the intensity change is very small. I wanted some software that would correct for this gamma curve.

There you have it. On the other hand, I have not any need (yet) for the ocean of different devices out there, nor any user / password management. So I skipped that.

Screenshots

A few screenshots just to give you an impression of the frontend side:

Main panel (desktop)

LED – General tab

LED – Schedule tab

Smart Knob routing

Main panel (mobile)

Launch debug session

This is a quick start for getting a feeling for the project:

• Open the project with VS Code, opening it with the Dev Container option.
• Start the offline emulation of the backend server: Press F5 to start debugging it.
• Download npm dependencies: Press ctrl-shift-b (to launch task) and choose Install npm dependencies.
• Start the webgui frontend: Press ctrl-shift-b and choose Run web server.
• Point your web browser at https://localhost:5173

Warning

There is currently no user/password protection implemented. Only run this on trusted home networks!

Build for deployment

In order to set this system up, you need the following services running:

• A message broker (MQTT). I use Mosquitto.
• The Zigbee2Mqtt service.
• The Light Assistant C# backend.
• A webserver (I use Lighttpd) hosting the Light Assistant WebGui frontend.
• Optional: The PiPwm service (for PWM-controlling LEDs with a GPIO).

To make this easier, I have prepared the following:

• A Dockerfile building the frontend and backend, and packaging it together in a Docker image with Lighttpd serving the webgui.
  • Build it with docker build . -t light-assistant:latest (in this directory).
• A Dockerfile for PiPwm that encapsulates what this service need.
  • Build it with docker build . -t pipwm:latest (in the pipwm directory).
• A set of systemd files, using Docker to download and run the Mqtt-related services.

To build the main project and encapsulate into a docker image for deployment, run:

Child Project: pipwm

pipwm is a PWM controller for Raspberry Pi, publishing PWM controller data on an MQTT bus.

License

Licensed under the BSD 3-Clause License.