Updates for CMSSW_9_4_X - Githubissues

cp3-llbb / Framework

Common framework for all cp3-llbb analyses

https://github.com/cp3-llbb/cp3-llbb.github.io/wiki/CP3-llbb(met):-Getting-Started

0 stars 13 forks source link

Updates for CMSSW_9_4_X #280

Closed pieterdavid closed 6 years ago

pieterdavid commented 6 years ago

I checked out the Framework in CMSSW_9_4_9, removed everything that said 80X or 'Moriond17' etc., and tried to make things run... so there will be differences, and I may have been too enthusiastically removing things in some places - my main aim was to provide a "clean" starting point, but feel free to point out if something doesn't make sense.

Anticipating a few questions:

on-demand scheduling requires a cms.Task to be associated with the process or path for each edm::EDProducer. This required quite a bit of trial-and-error with redoJEC etc., but the tests should run now
I turned off EGM regression and smearing for now, because the way this works changed (new recipes at https://twiki.cern.ch/twiki/bin/view/CMS/EgammaMiniAODV2 and https://twiki.cern.ch/twiki/bin/view/CMS/EgammaPostRecoRecipes) and I haven't implemented the new recipes yet.

cp3-llbb-bot commented 6 years ago

Build triggered. sha1 is merged.

cp3-llbb-bot commented 6 years ago

Build started sha1 is merged.

cp3-llbb-bot commented 6 years ago

Build finished.

cp3-llbb-bot commented 6 years ago

Build triggered. sha1 is merged.

cp3-llbb-bot commented 6 years ago

Build started sha1 is merged.

cp3-llbb-bot commented 6 years ago

Build finished.

pieterdavid commented 6 years ago

I went through the diffs... For unit_tests_data.py the differences are minor: all scale factors gone (I removed them), electron_expectedMissingInnerHits, (fat)jet_pfDeepCSVJetTags_probcc (it seems not to be there in this version of the deepcsv) and muon_isMedium2016. The former and latter will probably be needed still, so I can put them back easily. There are also differences in jet_pfDeepCSVJetTags_probc (maybe the merged probc and probcc?), and in the electron IDs because I removed them... maybe in hindsight I shouldn't have. For unit_test_mc_with_db.py the same, plus the PU weights which I also removed. For unit_test_mc.py there are more differences... I turned off the electron smearing and regression (the recipes changed, see this twiki, I didn't get to implementing that). Also the jet, jet systematic variation and MET branches changed. For the former two I suppose that would be because we pick up newer calibrations/recipes somehow, for the MET I'm reasonably certain (I saw code changes that needed the config to be adjusted). Does that make sense?

OlivierBondu commented 6 years ago

the updateremotes script is just perfect :) thanks for the other fixes !

Do you know why Jenkins fails ? Is it just the outdated ref trees ?

pieterdavid commented 6 years ago

I changed the definition of the target branch, so reopening as a new PR, https://github.com/cp3-llbb/Framework/pull/284