Management of the L1 and HLT prescales in the NTuples.

The current way of managing the prescales in DAS is to store them in the NTuples. In the normalisation procedure with method 'prescale', it is assumed that the PreL1min and PreL1max prescales are the same. No other situation has been observed for high pile-up datasets. For low pile-up datasets. In fact, preL1Min and preL1Max are the minimum and maximum of the L1 seed used in the HLT trigger. For 2017H dataset, there is also a problem in the prescale that is stored in the dataset after Ntuplisation. The prescales in the dataset does not corresponds to the true prescales.

For the moment, it is advised to always check the prescales stored (e.g.: brilcalc trg --prescale --hltpath HLT_HIPFJet* -r 307055) In case you want to use dedicated prescales, you just need to calculate yourself the luminosities of each triggers and put them into a TriggerLumi.txtfile. Then you could normalise the dataset using applyDataNormalisationwith option lumi.

This part of the code should be generalised to manage properly the prescales in any cases. An important point is to store the proper prescales in the dataset to avoid misleading situations.

Another option is just to not support the prescales anymore and leave the user determine the trigger luminosities himself.