Visualize projection of third Mandelstam variable in DPD implementation

[redeboer]

There are currently only two 'histograms' in the ampform-dpd.ipynb notebook implemented in #32. This is because we just compute a sum over the meshgrid over both axes. We would have to compute a second Dalitz plot with the third Mandestamm variable ($\sigma_2$) and then we can compute the sum over the Mandelstamm variable that is not $\sigma_2$.