Switch to MV2/DL1r and MV2/DL1rmu to include version with IPRNN only

Hi All,

Exo analysis aiming at early 2019 publication (and later) will profit from RNNIP at high pT. Given the issues in calibrating SMT it was discussed to have a version of the high level tagger with RNNIP only.

As discussed at the software meeting 2 weeks ago we have agreed in the following naming convention for the high level discrimimnant:

MV2mu -> MV2r (std+IPRNN)
MV2rnn-> MV2rmu(std+IPRNN+SMT)

and similarly for DL1. The baseline taggers, MV2c10 and DL1, will keep the same name not to confuse the analyzers that already using them.

With jshlomi we generated a new cali file with MC16a training for RNNIP for both MV2 and DL1 following the name convention described above.

this jira is to keep track of this work.

Tagging all FTAG conveners as this is of interested to everybody:

coccaro, mkagan,cpollard,msaimper,dduda,ggilles,cristinz,filthaut,jshlomi

Also Tagging exotics analysis contact that will profit from this in early 2019 publication (feel free to tag more):

nishu,biliu