The KPM method for computing conductivity has the advantage that once you have computed all the moments, a simple integration with the Fermi-Dirac distribution should yield results for many temperatures and chemical potentials (see Phys. Rev. Lett. 114. 116602 (2015) equation (1).)

The current implementation supports the input of chemical potentials as an array. However, the variable temperature is taken as one number. This is not very efficient since you need to run the same KPM for many temperatures.