Small updates, mostly to b-tag studies (+revert ttbb/ttbj split?)

I had this WIP for a while so better merge this now...

• Add an "undef" jet type (so jets can be b/c/light/gluon/quark/undef, where light=gluon+quark+undef) to the b-tag studies.
• Switch back the ttbar sub-components to ttB instead of ttbb and ttbj... we seem to be coming and going with this, but for the unfolding what we will really need is ttB, since the distinction will otherwise be made through the gen-level phase space, which can be a subset of ttbj+ttbb, not just ttbb...