Entanglementplot with complex F symbols

[BramVancraeynest]

Entanglementplot fails for symmetric MPSs constructed from UFCs with complex F symbols. This is because the real singular values are stored in a Vector{ComplexF64}, after which they are sorted, which fails.

using MPSKit,TensorKit,CategoryData,Plots
O = Object{UFC{3,1,2,1,1}} # Z_3
P = Vect[O](2=>1)
V = Vect[O](sector => 2 for sector in (1,2,3))
psi = InfiniteMPS(P,V)
entanglementplot(psi)

[Jutho]

Ok, I guess that's because TensorKit promotes any tensor data to complex if it figures out that the UFC is complex. That's because such tensors would anyway become complex as soon as you do some permutation. Maybe it shouldn't do that for a TensorMap{1,1}. Although that feels rather arbitrary.

I guess adding an explicit real statement in the entanglement plot functionality would not break things and have no overhead in the non-complex cases.