Run only extra variables in production

[watson-ij]

We should only need to run over files in order to get the extra variables (secondary vertices for Vts, and extra jet variables), which should be much quicker than producing NanoAOD. The official NanoAODs will be transferred to our Tier3.

We would need two parts:

1. Be able to read the extra variables into our analyses (similar to how we do HadTruth)
2. New python config to produce only the extra variables