VBF: Update setup to UL by alpakpinar · Pull Request #354 · bu-cms/bucoffea

alpakpinar · 2021-06-19T00:58:29Z

Hey @AndreasAlbert, this PR contains an update of the whole VBF setup to UL settings. I'm currently testing if the code in this branch will reproduce the templates from my local branch, wanted to open the PR for the record and I'll let you know about the outcome of the test. If you have additional comments, please let me know and I can make adjustments, thanks!

… noise estimation

…F weights

…n slots, update the testing file

AndreasAlbert · 2021-06-22T07:05:30Z

bucoffea/limit/legacy_vbf.py

+            'cr_1e_vbf' : re.compile(f'(EW.*|Top_FXFX.*|Diboson.*|QCD_HT.*|DYJetsToLL_M-50_HT_MLM.*|WJetsToLNu.*HT.*).*{year}'),
+            'cr_2m_vbf' : re.compile(f'(EW.*|Top_FXFX.*|Diboson.*|QCD_HT.*|DYJetsToLL_M-50_HT_MLM.*|WJetsToLNu.*HT.*).*{year}'),
+            'cr_2e_vbf' : re.compile(f'(EW.*|Top_FXFX.*|Diboson.*|QCD_HT.*|DYJetsToLL_M-50_HT_MLM.*|WJetsToLNu.*HT.*).*{year}'),
+            'cr_g_vbf' : re.compile(f'(GJets_((HT|DR-0p4)|SM).*|QCD_data.*|WJetsToLNu.*HT.*).*{year}'),


This will match both DR and non-DR, no? OK for now, but have to remember this when the DR samples come out

Yeah good point thanks, we'll need to update the regex once more when we switch to DR again

AndreasAlbert · 2021-06-22T07:09:11Z

bucoffea/monojet/definitions.py

        sf =  np.ones(df.size)
    elif year == 2017:
-        sf = sigmoid(x,0.335,217.91,0.065,0.996) / sigmoid(x,0.244,212.34,0.050,1.000)
+        sf = sigmoid(x,1.140,219.12,0.086,0.996) / sigmoid(x,0.171,207.22,0.092,1.000)


Can you add an if block here to differentiatae UL vs non-UL? I think the way it is written now, it changes the values for non-UL monojet, too

Thanks, this is done in eb0a359

bucoffea/monojet/definitions.py

AndreasAlbert · 2021-06-22T07:17:19Z

bucoffea/vbfhinv/vbfhinvProcessor.py

+        # Mask for 1/5th unlbinding
+        one_fifth_mask = ~pass_all
+        # Only pick each 5 entry in data
+        one_fifth_mask[::5] = True


Just for future reference: This is not entirely replciable because it will depend on the order of the events in the file. Genereally safer, to use the event number: mask = (df['event']%5)==0.

Thanks, done in 517c5e4

AndreasAlbert · 2021-06-22T07:18:36Z

bucoffea/vbfhinv/vbfhinvProcessor.py

+            if cfg.RUN.APPLY_EWK_CORR_TO_SIGNAL:
+                if re.match('VBF_HToInv.*', df['dataset']):
+                    def ewk_correction(a, b):
+                        return 1 + a * df['GenMET_pt'] + b


For future use: Let's use the higgs boson from the gen collection here.
higgs_pt = gen[(gen.pdg==25)&(gen.status==62)].max()

Thanks, done in 517c5e4

alpakpinar · 2021-06-22T23:59:55Z

Hey @AndreasAlbert , I updated the PR, and both monojet and VBF processors work fine for me with this config. Let me know if you want additional updates, thanks!

alpakpinar · 2021-06-23T00:32:08Z

Just pushed new XS from new samples (not the newly measured V+jets ones yet). There is a subtle bug due to Z+jets dataset name right now: The short name for these becomes ZJetsToNuNu_HT-.*-MLM_2017, which won't match the XS entry (mg -> MLM). I'll need to fix this without affecting monojet functionality

AndreasAlbert · 2021-06-23T07:48:57Z

Thanks for the updates. This looks good. Regarding XS I think we have to choices:

We systematically include a tag for UL in the dataset name. For example, DYJets..._2017 -> DYJets..._UL2017. This would ensure that there is never any overlap between EOY and UL. Downside: You have to rename your existing skim files (basically move all folders in your UL skim). Upside: Clear separation.
We separate the XS files for EOY and UL -> xs_eoy.yml, xs_ul.yml. Upside: Easier at the start. Downside: We now have to deal with having separate files, keeping track of which is file is used, etc.

It seems to me that 1 is cleaner, but feel free to let me know what you think.

alpakpinar · 2021-06-23T15:44:36Z

Thanks for the suggestions, I guess method 1 is cleaner indeed. I was thinking since we're going to update V+jets XS as well, maybe separate files could be easier but we can still achieve that by including a UL tag in the dataset names.

alpakpinar · 2021-06-28T00:33:07Z

Hey @AndreasAlbert, do you think instead of changing each of the dataset names in eos (to include UL tag), would it work if we modify the code in 1 to add the UL tag if a parameter like isUL=True is passed to scale_xs_lumi function?

alpakpinar added 6 commits June 18, 2021 19:22

Update XY corrections to UL

dfd2337

Update MET trigger SF to UL, add HF mask scale factors for VBF and HF…

3dd01d6

… noise estimation

Update VBF config to UL: Lepton/photon SF, b-tag. Add ROOT file for H…

a117e73

…F weights

Update VBF processor: Implement HF cuts, fill histograms of HF variables

6ad188c

Fixups, comment out EOY selections as we run out of possible selectio…

f6fc836

…n slots, update the testing file

Point to the new UL skim

3526ff8

alpakpinar added the enhancement New feature or request label Jun 19, 2021

alpakpinar modified the milestones: Consistency check with local branch, UL VBF H(inv): Consistency check with local branch, VBF H(inv): Upgrade to UL + consistency check Jun 19, 2021

alpakpinar added 8 commits June 18, 2021 20:07

Revert: Point to 03Sep20v7 for submission and run_quick

4e9827b

Updates on VBF limit script: Use dipole recoil VBF by default with UL

99e615d

Update photon trigger SF to UL

4575158

Fixup in muon and electron weights

f1cc3c6

Implement GEN checking on electrons and muons

d90248d

Fixup in lepton weights + GEN-check

e0e5a8c

Include QCD W templates in Z CRs

a19d39c

Fixup MET filters

2b1bcd5

AndreasAlbert reviewed Jun 22, 2021

View reviewed changes

bucoffea/monojet/definitions.py Show resolved Hide resolved

AndreasAlbert reviewed Jun 22, 2021

View reviewed changes

alpakpinar added 7 commits June 22, 2021 09:29

UL update in dataset naming

87598aa

Fixup in muon veto weight calculation

758f440

Update run_quick.py: UL testing samples for VBF

eb0a359

Fixup: No GEN check in monojet

c70c774

Cleanup

5e557de

Fixup in muon weights

3978b05

Update 1/5th unblinding mask, EWK corr to signal

517c5e4

Add XS for new samples

dd847a9

alpakpinar added 3 commits June 24, 2021 13:56

Use unsmeared jet pt while calculating b-veto weights

544098f

Fixup: Remove the second loose ID filtering for jets

f6b3b39

VBF: Update EWK correction to latest

8d0bcc2

alpakpinar added 5 commits June 28, 2021 15:57

Electron ID and RECO SF variations

35175f3

Muon ID and ISO SF with variations

dec5672

Update electron trigger SF + variations

0d7c6ea

Cleanup, add ttH to the limit input file SR

ea85d2e

Add ttH XS

6a893d0

Conversation

alpakpinar commented Jun 19, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alpakpinar commented Jun 22, 2021

Uh oh!

alpakpinar commented Jun 23, 2021

Uh oh!

AndreasAlbert commented Jun 23, 2021

Uh oh!

alpakpinar commented Jun 23, 2021

Uh oh!

alpakpinar commented Jun 28, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants