Make audio server more generalised by removing hard-coded installations in favour of adding an Installation Editor panel.

[mitchmindtree]

This way a user could first describe the set of installations that will exist within the exhibition. Then when creating speakers, each speaker could be assigned to one or more installations with unique weights for each (which map to the DBAP weights).

Perhaps the installation editor could also allow for dynamically entering in target IP addresses associated with each installation.

[mitchmindtree]

One issue with going in this route is that it's not clear how to map the input OSC data from a dynamically created installation to the necessary interactive sound generation. E.g. we would probably want some sort of patching node-graph environment to map the incoming OSC packets to audio parameters, however this is a whole large project in itself. @JoshuaBatty are currently in the process of implementing a generic node-graph along these lines in nannou that could be used for this, but it's unclear if we'll get the chance to finish it before the exhibition. As a result it's probably best to leave this as a bonus feature to implement once the rest of the audio_server is working.