Review of cross sections, part I

Review cross sections for backgrounds that are in use and add references for them. Cross sections for signals will be updated in a separate MR later, when the story with the gluon fusion samples is resolved. Fix the bug when the gluon fusion K factors are applied only to the background-only gg→ZZ samples (a part of #33).

@npostiau, the file with DDF stems also mentions some background samples that, I think, were used for the data-driven Drell–Yan estimation at some point but not used anymore (there are no corresponding entries in the lists of datasets in shears). These are all simulated backgrounds in config/samples_2016.yaml that don't have a comment about the cross section. If they are not used indeed, I'll remove them and update this MR.