Allow custom dust models

[rosteen]

A few years ago, there was some discussion of allowing custom dust models is Specutils, which never went anywhere. These requests were recorded elsewhere, this issue is just to record the idea here for future reference if it ever becomes a priority or someone wants to work on it.

[karllark]

If someone want's to work on this, I'd be happy to support. I would suggest that using the dust_extinction package would be a good way. Note there is the G23 R(V) dependent extinction relationship model that would provide the best model - based on spectroscopic data from the FUV (912 A) to the MIR (32 micron).