Currently the backend is a bunch of nodejs express stuff that should upload the robot.py but also acts as a bridge to start some commands on the robot. I think this should be solved differently:
- The upload part could probably be just a PUT server command in eg nginx.
- All the commands on the robot could/should probably be run through Xterm.js. This aligns with the ideology that the web interface is a layer around the terminal interface.
- The tty for xtermjs could be served via https://github.com/tsl0922/ttyd
- Currently we are not using the bonjour/mDNS in the web inteface.
- For extra robot statusinfo (eg battery of MM, robot name, etc) we could just serve a changing file via GET on nginx
Currently the backend is a bunch of nodejs express stuff that should upload the robot.py but also acts as a bridge to start some commands on the robot. I think this should be solved differently: