Fix eFermi

This feature allows to set the Fermi energy from input file.

This is needed in some cases to prevent from having unphysical occupations of states.

Case in point: ARPES simulations with a user-defined Kpoint path.

Here the fermi energy is calculated from the MP sampling but, since the path points have zero weight, they are not considered. The occupations of the bands are then chosen according to this fermi energy. In some cases (e.g. graphene), when the path has an higher resolution than the grid, this results in bands with zero occupation along the path and in unphysical holes in the calculated ARPES spectrum.

Without the option:

With the option: