Re: [Kwant] gauge invariance

Dear Zhang Bing, The gauge that is used for each lead _must_ be such that the hoppings are translation invariant for that lead. If you have leads with different directions, this means that you need to use different adopted gauges. It is possible to interpolate between the different gauges using a gauge transformation such that you have the same constant magnetic field vector in all of your system. Please see footnote 4 and Reference 21 of version 1 of the Kwant paper: http://arxiv.org/abs/1309.2926v1 Best, Christoph