Consider setting negative PMT gains to zero

[ramirezdiego]

A follow-up of https://github.com/XENONnT/fuse/pull/231could be to set negative gains (convoluted with SPE acceptance) to zero, to make sure we do not negatively bias larger areas, as we already import the gains-corrected pattern maps before this correction kicks in.