Unify and simplify the imposition of boundary conditions on the charged particle distribution function and collision operator calculation.

[mrhardman]

The charged particle boundary conditions are used in #149 to impose numerical conservation on the density, upar, and pressure. The boundary conditions are imposed on the distribution function in initial_conditions.jl, but the boundary conditions on the collision operator are imposed via a function in fokker_planck_calculus.jl.

The reason for this behaviour is that the @loop macro that is naturally used is different in each case. This is related to issue #140.