Modify to support NanoDST data

[zhenbinwu]

There are two paths stored in the NanoDST PD.

1. The 1.5kHz L1_Zerobias from the L1. One can use them for the L1Menu
   rate estimation. Just run on the NanoDST as the default Zerobias dataset.

2. The L1Accept, with HLT prescaled. This stored all the event fired L1,
   with a certain HLT prescale. To run over them, you need to input the
   SetL1AcceptPS as the HLT prescale of this run. The code will disable
   the Zerobias counting and use nLumi as counting. Also the prescale is
   applied during the run. So the input prescale should consider relative
   prescale.

For testing, you can start with the below option, which should reproduce
the online L1Rate, assume L1AcceptPS is 10:
--UseUnpackTree --nBunches -1 --SetL1AcceptPS 10 --SetNoPrescale

This procedure should be tested and further understood. Given the above
description, it is probabaly not the best option for the L1Menu rate
estimation.

[abrinke1]

Opened a new PR into master; please discuss here:
#7