Problem of running TriggerFloats on 2016 UL samples

[Ang-Li-95]

This assertion sometimes fail for 2016B and 2016C data:
https://github.com/DisplacedVertices/cmssw-usercode/blob/master/MFVNeutralino/plugins/TriggerFloats.cc#L320

The code requires l1 trigger objects to have only one filter but sometimes they have more than one and that's the reason of the problem.
One example of the trigger object that fail the assertion from 2016C is:
obj pt 418.000000 eta -1.348633 phi -0.715088 coll hltCaloStage2Digis:Jet:HLT ids (#=2) -99 -98 is_l1 1 is_ht 0 is_cluster 0 is_photon 0 is_muon 0 is_jet 0