FHI-aims band structure alignment
It seems that at least in some cases of FHI-aims band structure calculations, the Fermi energy is reported using the "internal" energy scale, while the band structure energies are reported as already being normalized to some reference state, most likely the Fermi energy. This is breaking the band gap detection in many cases, e.g. https://nomad-lab.eu/prod/rae/gui/entry/id/MF3kN7zlT-q942wC1G33Jw/zoOy4mHCmFI6bpNCqv40h1t9E_S8/archive/section_run/section_single_configuration_calculation/section_k_band
These two values should be given in the same energy scale. If the un-normalized bands are available, they should be used. If only the normalized output is given, then the fermi-level should be adjusted to zero after ensuring that the normalization is based on it.