Rationalize stellar/dust prior/grid weight calcuations

[karllark]

The dust prior weights are implemented as a class. The stellar dust/grid weights are implemented as functions. Would be good if they were done uniformly one way or the other. Likely the MegaBEAST development will inform which way is best.

To be fully consistent, the dust grid weights need to be computed and used. Currently it is assumed that the dust grid is always uniformly spaced. So the dust grid weights are not computed at all. Would be good to get them computed and used so that we can use non-uniform spaced dust grids in the future.

[karllark]

The grid weights should be implemented like the new prior weights class (#649). The generic grid weight class should handle non-uniformly spaced grids including ones like the current variable width mass that is variable with age. Need to be able to do the same with age spacing variable with mass (evolutionary tracks). Not sure if other variables need such flexibility.