I wrote this code so that IFTTT and Google Assisstant could control the LEDs on a shelf I built. The server itself is pretty simple (and not very professional) but it works.
When I set all of this up, the only way to enable the Jetson Nano's SPI interface was to manipulate an Excel sheet provided by NVIDIA, run a macro that generated config files, and re-flash the device with those files in a special directory of the SD card. This was not a simple process and it took many tries to get right, but since then it seems Nvidia has made things easier -- see here.
This pinout mapping for the J41 header will also prove useful if you're trying to replicate this project.
You'll need to install Adafruit's dotstar library, but other than that everything is standard Python.
If you have a firewall enabled (such as ufw, make sure to open the HTTP port. Modify the paths and IPs
in the start.sh and start_with_tmux.sh files, chmod +x them, and you should be good to go. Since
I don't have a static IP, I've designed this to SSH into a GCP machine (which does have a static IP) and
forward the ports. You may not want/need to do that.