Core: generalize calculation of topological charge

[GPMueller]

Now that #292 has been closed due to the merge of PR #424, the topological charge can be calculated for 2D geometries in the xy-plane. It should be improved in the following ways:

• triangles should be translated into triplets, i.e. indices and translations in order to generalize the implementation for arbitrary orientations of the plane and boundary conditions, using the idx_from_translation or whatever functions which also check atom types etc
• the calculation should be independent of the orientation of the plane
• calculate distribution of charge over atoms (in parallel!) and total charge by reduction
• add checks that the geometry is 2D (ask the Geometry class)