Following instructions are to run the generation of the histograms directly from NanoAOD.
First, log into an LPC node:
ssh -L 9094:localhost:9094 [email protected]
The command will also start forwarding the port 9094 (or whatever number you choose)to be able to use applications like jupyter once on the cluster. Then move into your area on uscms_data:
cd /uscms_data/d?/USERNAME/
where '?' can be [1,2,3]. Fork this repo on github and git clone it:
git clone https://github.com/USERNAME/decaf.git
Then, setup the proper dependences:
source setup_lcg.sh
you will install the necessary packages as user. This is a one-time setup. When you log in next just do:
source env_lcg.sh
by running this script you will also install your grid certificate.
The list of the inputs can be obtained in JSON format per year, packing the files beloning to the same dataset in batches of the desired size. The size of the batch will affect the performances of the following steps. The smaller the batch, the larger number of batches per dataset, the more the jobs you can run in parallel with condor. This needs to be balanced with the number of cores that can be made available through condor. A sweet spot has been found by setting the size of the batch to 75 file. Similarly, packing files in batches of different size will affect the performances once running with Spark. To run with Spark, use a size of 300 files per batch.
To create the JSONS:
cd analysis/metadata
python pack.py -y 2018 -p 300 #(or 75)
The --year option will allow for the generation of the list for a specific year, 2017 in the example. The --pack option will set the size of the batch. The --keep option will save .txt files that list the number of rootfiles per dataset, that will be stored in the metadata/YEAR folder. The JSON files will be stored in the metadata folder and they will include the cross section of the dataset each batch belongs to. A cross section of -1 is assigned to data batches.
The generation of histograms can be launched from the analysis folder. Currently a local generation through python futures is supported together with a condor and a Spark implementation.
Before running, weights from secondary inputs like corrections, ids ecc. need to be compiled and stored in .coffea files. To accomplish this task, simply run the following command:
sh generate_secondary_inputs.sh
the script will run the following python modules:
python secondary_inputs/metfilters.py
python secondary_inputs/corrections.py
python secondary_inputs/triggers.py
python secondary_inputs/ids.py
these modules can also run separately by passing to the bash script the argument corresponding to the name of the module you want to run. For example, to run secondary_inputs/corrections.py, just do:
sh generate_secondary_inputs.sh corrections
Separate .coffea files will be generated, corresponding to the four python modules listed above, and stored in the secondary_inputs folder.
The next step is generating the Coffea processor file corresponding to the analysis you want to run. For example, to generate the Dark Higgs processor, the following command can be used:
sh generate_processor.sh darkhiggs 2018
The script will run the following command, using the first argument to select the corresponding python module in the processors folder and the second argument to set the --year option:
python processors/darkhiggs.py --year 2018
The module will generate a .coffea file, stored in the processors folder, that contains the processor instance to be used in the following steps. The --year option will allow for the generation of the processor corresponding to a specific year.
Phyton futures allows for running on multiple processor on a single node. To run the local histogram generation:
python run.py --year 2018 --dataset MET____0_ --processor darkhiggs2018
In this example, histograms for the 0 batch of the 2018 MET NanoAOD are being generated. The --year option is compulsory and the --dataset is optional. The --lumi is optional. If not provided, it will default to the hard-coded lumi values in run.py. Launching the script without the --dataset option will make the script run over all the batches for all the datasets. If, for example, --dataset TTJets is used, the module will run over all batches and all the datasets that match the TTJets string. The --processor option is compulsory and it will set the processor instance to be used. It corresponds to the name of a specific processor .coffea file that can be generated as described previously.
Condor will allow to parallelize jobs by running across multiple cores:
python submit_condor.py --year 2018 --processor darkhiggs2018 -t
This way jobs to generate the full set of 2018 histograms will be submitted to condor. the -t will allow for tarring the working environment and the necessary dependences to run on condor nodes. The module has a --dataset option that works like described before for run.py. Will allow you to run on a single batch, dataset, or batches/datasets that match the input string.
Taking as example the dark Higgs analysis, run the following command to generate the background model, for example:
python models/darkhiggs.py -y 2018 -f -m 40to120
The models/darkhiggs.py module extracts coffea histgrams from the hists/darkhiggs201?.scaled files and utilize them to generate the different templates that will later be rendered into datacards/workspaces. It also defines transfer factors for the data-driven background models. It produces .model files that are saved into the data folder. More specifically, the models/darkhiggs.py module produces one .model file for each pass/fail and recoil bin.
Running the following command will generate all model files, and the outputs will store in the data directory:
./make_model.sh
In order for the rendering step to be completed successfully and for the resulting datacards and workspaces to be used in later steps, both the python version and the ROOT version should correspond to the ones that are going to be set when running combine. It is therefore suggested to clone the decaf repository inside the src folder of CMSSSW release that will be used to perform the combine fit and run the rendering from there, without running the env_lcg.sh. You should, instead, do the regular cmsenv command.
At the time of writing, the recommended version is CMSSW_10_2_13 (https://cms-analysis.github.io/HiggsAnalysis-CombinedLimit/#cc7-release-cmssw_10_2_x-recommended-version). This version uses python 2.7 and ROOT 6.12.
In addition, you should install the following packages (to be done after cmsenv):
pip install --user flake8
pip install --user cloudpickle
pip install --user https://github.com/nsmith-/rhalphalib/archive/master.zip
Finally, since we are using Python2.7, you should edit the file
$HOME/.local/lib/python2.7/site-packages/rhalphalib/model.py
and change line 94 from
obsmap.insert(ROOT.std.pair('const string, RooDataHist*')(channel.name, obs))
to
obsmap.insert(ROOT.std.pair('const string, RooDataHist*')(channel.name.encode("ascii"), obs))
and also make the following modification:
import ROOT
#add this line
ROOT.v5.TFormula.SetMaxima(5000000)
The models saved in the .model files at the previous step can be rendered into datacards by running the following command:
python render.py -m model_name
the -m or --model options provide in input the name of the model to render, that corresponds to the name of the .model file where the model is stored. The render.py module launches python futures jobs to process in parallel the different control/signal regions that belongs to a single model. Different models can also be rendered in parallel, by using condor. In order to run rendering condor job, the following command should be ran:
python render_condor.py -m <model_name> -c <server_name> -t -x
python render_condor.py -m darkhiggs -c kisti -t -x
this time, all the models stored into .model files whose name contains the sting passed via the -m options are going to be rendered.
The render.py module saves datacards and workspaces into folders that show the following naming:
datacards/model_name
The render_condor.py module returns .tgz tarballs that contain the different datacards/workspaces, and are stored into the datacards folder. To untar them, simply do:
python macros/untar_cards.py -a darkhiggs
Where the -a or --analysis options correspond to the analysis name. The untar_cards.py script will untar all the tarballs that contain the string that is passed through the -a option.
To merge the datacards, run the following script:
python macros/combine_cards.py -a darkhiggs
Where the -a or --analysis options correspond to the analysis name. The combine_datacards.py script will combine all the datacards whose name contains the string that is passed through the -a option. The script will create a folder inside datacards whose name corresponds to the string that is passed through the -a option, will move all the workspaces that correspond to the datacards it combined inside it, and will save in it the combined datacard, whose name will be set to the string that is passed through the -a option.
You can also plot the inputs from a single workspace with the makeWorkspacePlots.C macro. It stacks all the inputs in a single histogram. From the folder you have just created at the previous step, do, for example:
root -b -q ../../makeWorkspacePlots.C\('"zecr2018passrecoil4"',5000,2000\)
where the first argument (zecr2018passrecoil4) is the name of the workspace; the ROOT file name must be the same, but with the .root extension; the second argument (5000) is the limit of the y-axis of the stack; the third argument (2000) is the y-position of an explanatory label; that label will simply be the same as the first argument again.
Make sure you edit your $CMSSW/src/HiggsAnalysis/CombinedLimit/scripts/text2workspace.py module as suggested below:
import ROOT
#add this line
ROOT.v5.TFormula.SetMaxima(5000000)
In the datacards directory, the directory is created based on the <model_name>. We are using the "MultiSignalModel" method to have one datacard includes all signal mass points. Please submint text2workspace job to the condor by using the following command:
python t2w_condor.py -a darkhiggs -c <server name> -t -x
Currently, <server name> is either lpc or kisti. The output will be saved in the same directory with the same name as the datacard.
After you get workspace, you will be able to run fit over condor by doing:
python combine_condor.py -a darkhiggs -c <server name> -m <method> -t -x
python combine_condor.py -a darkhiggs -c kisti -m cminfit -t -x
python macros/dump_templates.py -w datacards/darkhiggs/darkhiggs.root:w --observable fjmass -o plots/darkhiggs2018/model
python macros/hessian.py -w datacards/darkhiggs/higgsCombineTest.FitDiagnostics.mH120.root:w -f datacards/darkhiggs/fitDiagnostics.root:fit_b
Change the path to store outputs in plotConfig.py
python macros/diffNuisances.py -g pulls.root ../fitDiagnostics.root
If you use fast fit command, then the following command will make prefit and postfit histograms
PostFitShapesFromWorkspace -w path/to/darkhiggs.root -d path/to/darkhiggs.txt -f path/to/fitDiagnostics.root:fit_s --postfit --sampling --samples 300 --skip-proc-errs -o outputfile.root
To make postfit stack plots
python macros/postFitShapesFromWorkSpace.py path/to/outputfile.root
This method uses the output from dump_templates.py.
python macros/combine_dump_outputs.py plots/darkhiggs2018/dump_postfit (or dump_prefit)
Then you get combined root files in the plots/darkhiggs2018/dump directory.
python macros/dump_darkhiggs_postfit.py plots/darkhiggs2018/dump/outputs.root
This will make postfit stack plots.