[post-processing] Electron SF variations are identical

[mlizzo]

Electron SF variations appear to be identical and I think the issue is due to a small bug in the LeptonSF.py module, in particular by line#455. Indeed, when the isPOGFormat flag is true, the second term under the sqrt is always EleRecoSF_tmp[1]*EleRecoSF_tmp[1], which corresponds to the upwards Electron_RecoSF variation. When the downwards variation is computed, that term should be EleRecoSF_tmp[2]*EleRecoSF_tmp[2] instead.

FYI: @BlancoFS @giorgiopizz @NTrevisani @dittmer