Add HGTD information for r21.9 upgrade sample
As said in the comments, please temporarily use https://gitlab.cern.ch/caiyi/training-dataset-dumper-new if you want to use the timing information now.
Hi, this MR is for the code to dump HGTD information, including track time, vertex time and so on.
I would say the current method to get timing information is primary, it is not something should be done by FTAG and I believe that we would be able to do it via Athena in the future. Therefore, those related scripts are stored in an additional folder BTagTrainingPreprocessing/HGTD
rather than BTagTrainingPreprocessing/src
. In particular, the track clustering code ClusteringHelper.h
and the BDT file for vertex time reconstruction TMVA.VBFinv.mu200.Step3p1.8var.weights.xml
are directly obtained from HGTDVertexTimeInterface, while the others are the implement of HGTDVertexTimeInterface and HGTDTrackTimeInterface.
Apart from introducing timing information, there are some more changes:
- Possible cut on tracks is added as
HGTD_delta_maximum
andHGTD_signif_maximum
. To accommodate other configuration files without this option, I usetry {} catch(...) {}
inBTagTrainingPreprocessing/src/SingleBTagConfig.cxx
which might not be a smart approach. - An option
truth_pass_matching
for decorate is added, with this option we can know if jet/event has pass the truth level matching vianEventPassTruthPU/nJetPassTruthPU
. In this casetruth_jet_matching
andtruth_primary_vertex_matching
are recommended to be false. - I want to get SV2 and JF vertex information, but AMVF is unavailable in my samples, so I use
try {} catch(...) {}
again inBTagTrainingPreprocessing/src/TrackVertexDecorator.cxx
. I am not sure if it is okay for you or I should split the optiontrack_sv_info
into two parts. - This study is integrated into the grid submission script, you can directly do
grid-submit upgrade-HGTD
to dump single-top samples with HGTD information.