Wrong packing functions in electron-phonon calculations?
Generally, we expect that the unpack2
method should be used to unpack atomic density matrices as it divides the off-diagonal elements by 2 while the unpack
method should be used to reconstruct any Hermitian matrix in packed form.
In elph/electronphonon.py
and elph/supercell.py
the function unpack2
is used on H_sp
objects and based on the naming, I would expect them to be corrections to the Hamiltonian for which unpack
is the appropriate method.