WIP: LB+Lees Edwards fixes

[RudolfWeeber]

• Clean up and fix Lees Edwarsd LB ghostlayer interpolation code
• Tentative fix for particle coupling

Notes:

• The LB grid points are located at (0.5 +i)* agrid in all Cartesian direciotns
• When coupling a particle closer than 1/2 agrid to a periodic boundary, the interpolation contributes to lb nodes across that periodic boundary.
• This is handled by additionally coupling at positions shiftd by +- a box length in al Cartesian directions.
• The exisitn gcode did apply a Lees Edwards position offset, when a shift across the Lees Edwards boudnary was applied, but a corresponding Lees Edwards velocity offset was missing
• This is added here. I'm confused about the sign, but only the present choice resulted in a particle co-moving in sheear flow to be force free (see lb_lees_edwards_particle_coupling.py)