Why is the "eigvec" in the output of mp.linalg.eig function a set of complex number?

Thanks for your code development! For the Hamiltonian of a one-dimensional TFIM model in MPO form, I don't know why the ground state solution "eigvec" obtained by mp.linalg.eig is a set of obvious complex numbers. Shouldn't it be real number or a complex number with negligible imaginary parts?