Separation of concerns: speaker and protocol

[dge8]

Hi all, thanks for your work on this project, it's been really useful!

I'm working on a project that runs on a Raspberry Pi and allows controlling Kef speakers via HDMI CEC control, basically emulating HDMI ARC on sound bars (and in the newer LS50W IIs and LS60s). Speakers turn on with the TV, changing TV volume changes speaker volume, turn off again when the TV turns off, etc. The actual audio path is of course still via optical.

I would like to use the Kef protocol in this library in a more direct way - without the tenacity retry module, or volume scale, or helpers for splitting e.g. volume and mute. I'd rather implement that stuff higher up in my own project.

Would you be open to separating the module into a submodule focussed entirely on modelling the protocol ("proto"), and another submodule focussed on abstracting the speaker functions ("speaker")? I've submitted another issue covering a third possible submodule "discovery".

I'd be happy to attempt this in a branch and bring it for review although I will only be able to test on my LSX.

Thanks and regards, Dan.